I

a complete repertoire for black

1an
in this most dynamic of openings

by Joe Gallagher

EVERYMAN CHESS
Gloucester Publishers pic www.everymanchess.com
First published in 2004 by Gloucester Publishers pic (formerly. Everyman Publishers
pic), Northburgh House, 10 Northburgh Street, London EC1V OAT

Copyright© 2004 Joe Gallagher

The right of Joe Gallagher to be identified as the author of this work has been as­
serted in accordance with the Copyrights, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means, electronic, electrostatic, magnetic
tape, photocopying, recording or otherwise, without prior permission of the publisher.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British library.

ISBN 1 85744 324 1

Distributed in North America by The Globe Pequot Press, P.O Box 480,
246 Goose Lane, Guilford, CT 06437-0480.

All other sales enquiries should be directed to Everyman Chess, Northburgh House,
10 Northburgh Street, London EC1V OAT
tel: 020 7253 7887 fax: 020 7490 3708
email: [email protected]
website: www.everymanchess.com

Everyman is the registered trade mark of Random House Inc. and is used in this
work under license from Random House Inc.

EVERYMAN CHESS SERIES (formerly Cadogan Chess)

Chief advisor: Garry Kasparov
Commissioning editor: Byron Jacobs

Typeset and edited by First Rank Publishing, Brighton.

Cover design by Horatio Monteverde.
Production by Navigator Guides.
Printed and bound in the US by Versa Press.
 CONTENTS I

Bibliography 4
Introduction 5

Part One: The Classical Variation

1 d4 tt'lf6 2 c4 g6 3 tt'lc3 i..g7 4 e4 d6 5 tt'lf3 0-0 6 i..e2 e5

1 The Classical Variation: Modern Main Line 7

2 The Classical Variation: 9 lbe1 liJd7 without 10 i.e3 26
3 The Classical Variation: The Bayonet Attack 9 b4 40
4 The Classical Variation: White's 9th Move Alternatives 57
5 The Classical Variation: 7... lba6 69
6 The Classical Variation: White's 7th Move Alternatives 83

Part Two: Other Lines

1 d4 tt'lf6 2 c4 g6 3 tt'lc3 i..g7 4 e4 d6

7 The Samisch 94
8 The Fianchetto Variation 120
9 The Four Pawns Attack 144
10 White Plays an early h3 164
11 The Averbakh Variation 178
12 Other Systems 191

Index of Variations 205

Index of Complete Games 207

 BIBLIOGRAPHY I

Books
101 Attacking Ideas in Chess, Gallagher (Gambit 2000)
Beating the Anti King's Indians,Gallagher (Batsford 1996)
Beating the King's Indian and Benoni, Vaisser (Batsford 1997)
ECO Volume A and Volume E
King's Indian and Griinfeld: Fianchetto Lines, Janjgava
(Gambit 2003)
King's Indianfor the Attacking Plqyer,Burgess (Batsford 1993)
NCO, Nunn, Burgess, Emms, Gallagher (Everyman/Gambit 1999)
Plqy the King's Indian Defence, Marovic (Pergamon 1984)
Siimisch King's Indian, Gallagher (Batsford 1995)
Starting Out: The King's Indian, Gallagher (Everyman 2002)
The Complete King's Indian, Keene and Jacobs (Batsford 1992)
The New Classical King's Indian, Nunn & Burgess (Batsford 1997)
Victor Korchnoi: My Best games Volume 1: Games with White, Korchnoi (Olms 2000)

Chess Software
MegaBase 2003
TWIC
Fritz

Internet Sources
ChessPublishing Website: King's Indian section by Martin, Gallagher and Mikhalevski

Magazines and Periodicals

b!fimnators 1- 89
New in Chess:_yearbooks and maga�nes
British Chess Maga�ne
jtN7RODUCTIONI

\X'hat? Another Gallagher book on the King's Indian, I hear you say. It seems like only yesterday
when his last one was out. Yes, at times it feels a bit like that to me as well. But it is in fact well
over two years since Starting Out: The King's Indian (Everyman 2002 and henceforth referred to as
SOKID) was released and this book, which has been in the pipeline for sometime, was always
intended as a sequel to that one. Let me quickly recap what SOKID was about and then highlight
the similarities and differences to the current offering.
SOKID was a book primarily aimed at the inexperienced player or the more experienced
player who was new to the King's Indian. It covered all the major variations in the King's Indian
from an objective, middle of the road, viewpoint. It concentrated on explaining the first dozen or
so moves of each line and it was heavy on verbal explanations and light on theoretical variations.
This book is aimed at all players who play the King's Indian, or who want to play it, and it as­
sumes a slightly higher level of chess understanding than SOKID. Inexperienced players can still
derive plenty of benefit from this book but I strongly recommend that they start with SOKID.
This book also deals with all the major variations of the King's Indian but strictly from a Black
point of view. It is largely based on my own personal King's Indian repertoire and the variations
that have served me well over the years. Some readers may be disappointed that their favourite
variations are not included but in order to cover the chosen variations to the required depth I
have had to be selective. Here is a brief overview of the repertoire this book provides against the
main lines:
1) In the Classical both the main lines with 7...l"iJc6 and the modem 7...tba6 are covered.
2) In the Samisch I recommend that Black plays 6... c5.
3) In the Fianchetto Variation I recommend the classical 6...tbbd7, concentrating on the Gal-
lagher variation and a closely related system based on 8...a6.
4) In the Four Pawns Black plays with 6...c5 and 9.....i.g4.
5) In the Averbakh and early h3 systems Black plays, in general, for ...e7-e5.
Within the chosen repertoire, however, I do not restrict myself to one particular line of play
and there is plenty of choice for the black player if one or two of the recommended lines run into
trouble.
\X'hereas SOKID concentrated on the early stages of the opening this book devotes most of
its firepower to the late opening/ early middlegame stage of the game. There is a lot more analysis
and theoretical variations than in SOKID but also plenty of verbal explanation. I have tried to
reach a happy medium between the two. A myth I often hear spouted (in reviews, for example) is
that books with plenty of prose are for the weak to average player whereas books that resemble

5
Play the King's Indian

the telephone directory are for the expert players. I strongly disagree and my personal opinion is
that, especially in this database age, books that resemble the telephone directory are boring, ir­
relevant and for nobody.
I have used a complete games format to cover the material. On occasion this is merely win­
dow dressing but not here. In a book entitled PI<!] the King's Indian I felt it was important to anno­
tate the games fully throughout, although I have usually drawn the line at placing long endgames
under the microscope.
One of the main dilemmas I have faced in writing this book is what to do about the overlap
with SOKID. The fact is that in certain cases I have already said what I wanted to say about a
position in SOKID. I have, on the whole, just decided to repeat what I said in SOKID here. The
alternative was to refer the reader to SOKID for these parts, and whilst I have done this for the
more basic explanations I didn't want to do this regularly as this is a stand-alone King's Indian
book. The other option was to change things round a bit but this seemed like a waste of time
(and a bit false) when I have already worked hard on a variation or on drafting a piece of text.
Apart from SOKID, I also wrote two other King's Indian books in the 1990's, Beating the Anti
King's Indians (Batsford 1996) and The Samisch King's Indian (Batsford 1995). Where appropriate, I
have drawn on material from these works as well.
Another important source has been the ChessPublishing website. For some of 2002 and all of
2003 I produced a monthly article on the latest King's Indian games for this highly recom­
mended openings site and the annotations to a number of games in this book have been adapted
from what first appeared there.
It goes without stating, however, that everything from whatever source, has been updated and
these are my King's Indian views of mid-2004.
This introduction has dealt mainly with technical issues but I would just like to finish on a
more emotional note. During my chess career many openings have come and gone but only one
has been ever present - the King's Indian. I started playing it at age 11 and have not stopped for
nearly 30 years. It has brought many highs and some lows - it is not always easy playing Black no
matter what opening you choose - but rest assured that if you master its secrets it will bring far
more pleasure than misery.

Joe Gallagher,
Neuchattel,
September 2004

6
 CHAPTER ONE I
The Classical Variation:
Modern Main Line

1 d4 lt:lf6 2 c4 g6 3 lt:lc3 i.g7 4 e4 d6 5 which is simply described as me Main Line

i.e2 0-0 6 lt:lf3 e5 7 0-0 lt:lc6 8 d5 ltle7 9 King's Indian. By playing 7...lLlc6 Black invites
ttJe 1 lt:ld7 1 0 i.e3 White to close me centre wim gain of tempo.
The Classical is by far me most imponant He is even willing to have his knight driven to
variation in me King's Indian. You will proba­ me inferior e7-square to achieve rhis central
bly have to face it as often as all me omer varia­ clarification. The reason Black wants me centre
tions put togemer. Consequently, it takes up closed is mat he wants to attack on me kingside
about half me book. White's plan is simple and and rhis is not possible wim a fluid centre.
unassuming. He just develops his kingside sen­ There are quite a few alternatives as me
sibly and rapidly and awaits developmenrs. Let sharp meoretical battles mat arise after 7...lLlc6
us now consider me above sequence in detail. are not too everyone's taste. 7...ll:la6 is me most
1 d4 lt:lf6 2 c4 g6 3 lt:lc3 i.g7 4 e4 d6 5 imponant alternative and it is covered exten­
i.e2 0-0 6 lt:lf3 e5! sively in Chapter 5.
The best move and me only one we shall be 8 d5
looking at in me book. Alternatives such as White usually accepts Black's invitation to
6...c5 and 6...i.g4 lead to completely different close me centre as mere are no promising alter­
types of games where it is easier, in my opinion, natives. The one move mat does occur from
for �'hire to maintain an edge. In me past time to time is 8 i.e3 (8 dxeS dxeS is com­
6...lLlbd7 was played sometimes as a means of pletely harmless):
preparing eS but we can't do that unless we play
the variation 6...e5 7 0-0 lLlbd7 (which we do
not as Black is worse after 8 i.e3�. The one
plausible alternative to me text is 6 ... lLla6. This
can be adopted by mose of you who play me
7 ...ll:la6 variation but wish to avoid, for exam­
ple, me exchange variation in its purest form.
7 0-0
This is by far the most popular and flexible
of \X'hite's 7th moves. The first five chapters
are devoted to it whilst Chapter 6 deals wim the
alternatives, 7 dS, 7 .i.e3 and 7 dxeS.
7 . lt:lc6
. .

Now we are getting close to me territory

7
Play the King's Indian

a) It would be played more often if it wasn't We also have some pretty major alternatives
for the reply 8...l:e8! when the natural 9 dS here:
leads to complete equality after 9.)od4! 10 a) 9 b4 is the famous Bayonet Attack and
lLlxd4 exd4 11 i.xd4 lLlxe4 12 lLlxe4 :Xe4 13 that is the subject of Chapter 3.
i.xg7 �g7. Therefore White's only try is the b) 9lLld2 is not quite as popular as it used to
almost equally tedious 9 dxeS dxeS with the be but is still an important line. That is covered
following possibilities: in Chapter 4.
at ) 10 ..xd8 lLlxd8 (more reliable than c) The interesting 9 i.gS is also covered in
to...:Xd8) 11 lLlbs lLle6 12 lLlgs l:e7! Chapter 4 along with other 9th move alterna­
(12...lLlxgS?! 13 i.xgS l:e7 14 l:adl is unpleas- tives:
ant for Black) 13 l:fdl b6 14 a4 c6 1StLlxe6 (lS
lLlc3 lLld4� 1S.....txe6 16 lLlc3 l:b8 (or
16...l:b7) 17 b4 l:d7 18 £3 i.£8 19 :Xd7tLlxd7
20 l:bt aS with a level game, Hjartarson-Nunn,
Szirak 1987.
a2) 10 cS i.g4 11 i.bS ..c8 12 h3 i.hS! is
an idea of Smirin which diffused a White idea
that was beginning to gain a supporter or two.
The point is that it will now be difficult for
White to break the pin (13 g4? lLlxg4). After 13
..txc6 bxc6, both 14 i.gS l:b8 tS l:b1 ..tx£3 16
..x£3 lLld7 17 b4 aS, 011-Smirin, Rostov 1993,
and 14 ..e2 h6 lS �h2 gS 16 ..c4 l:b8 17 b3
aS 18lLld2 g4 19 £3 gxh3 20 gxh3 l:b4, Van der
Sterren-Smirin, Tilburg 1992, were quite satis­ 9...tLld7
factory for Black. Black gets ready to play ...f7-f5. As to why I
a3) 10 h3 i.e6 11 cS and now Black can prefer this move to 9...lLle8 I once again refer
choose between ll.....e7, 11...a6 and 1t ...lLlhS you to SOKID.
all of which are thought to offer a roughly level Now:
game 1) 10 i.e3 is the subject of the remainder of
b) Despite the fact that these lines offer this chapter. Our coverage starts after the addi­
Black easy equality most King's Indian players tional and almost universal moves 10...£5 11 £3
prefer to go fishing in murky water with f4 12 i.£2 gS.
8...tLlg4 9 i.gS f6 and now: 2) 10lLld3 and 10 £3 are seen in Chapter 2.
b1) 10 i.h4 is not played so often when
White has already castled as his options are The Modem Main Line:
severely reduced (in comparison to 7 i.e3lLlg4 7... lLic6 8 d5lD87 9 lbe1 tLld7 10 .b3
8 i.gS f6 9 ..th4, the Gligoric Variation, that
is). The immediate 10...lLlh6 is one possibility 1 d4 tLlf6 2 c4 g6 3 lbc3 ..tg7 4 e4 d6 5
whilst tO...gS 11 i.g3 lLlh6 12 dxeS fxeS is lLif3 0-0 6 ..te2 e5 7 0-0 lbc6 8 d5 lbe7 9
thought to be fine for Black, e.g. 13 cS g4 14 tLle1 tLld7 10 ..te3 f5 11 f3 f4 12 ..tf2 g5
lLld2 dxcS tStLlb3 b6 16 ..dS+ 1i'xdS 17tLlxdS This is one of the sharpest and most theo­
l:f7 18 i.c4 �h8 when White has compensa­ retical lines in the whole King's Indian. It also
tion for the pawn but no more. represents the King's Indian in its purest form.
b2) 10 ..tel is more common and it trans­ White will attempt to tear Black limb from limb
poses to the Gligoric System (see the introduc­ on the queenside while Black will endeavour to
tion to that line in Chapter 6). hang, draw and quarter White on the kingside.
8 . .lD87
. 9 tLle1 Sometimes both players succeed in their aim
For a lengthy explanation of this prophylac­ and then Black wins. That is the advantage of
tic retreat I refer you to SOKID. attacking the king.

8
 The Classical Varia tion: Mode rn Main Line

taring. The manoeuvre ...l:f6-h6, in order to

gain a quick attack on the h-tile, is also an im­
portant device. If White defends well then
Black normally reverts to the strategy of playing
for ... g4. Apart from in Game 1 where White's
move order presented Black with an opportu­
nity to block up the queenside, Black usually
leaves this side of the board alone. Even if
White attacks a rook in the comer, for example,
then Black may ignore it in favour of continu­
ing his kingside attack. In such a position every
tempo is worth its weight in gold.
From the diagram position White has six
reasonable moves and they are covered during
The great champion of this line with the the next five games. Game 1 deals with 13 a4
white pieces is the legendary Grandmaster Vic­ and 13l"bb5, Game 2 with 13 b4, Games 3 and
tor Korchnoi. I have been fortunate enough to 4 with 13 l:ct and Game 5 with 13 lbd3 and 13
play in the same team as him on many occa­ g4. At the moment the most fashionable lines
sions. We have had many lively discussions on for White are 13 a4 and 13 l:ct but this can
the merits of the King's Indian. I am afraid he easily change as theory does not stand still in
does not hold this opening in the same high this sharp line.
esteem as I do. For example, when discussing The coverage of these lines is denser than in
Kasparov in his recent games collection the rest of the book but this is unavoidable due
Korchnoi lets slip, 'the World Champion ha­ to the tactical nature of the position. They are
bitually takes his life in his hands by depending also very entertaining if you like sacrificial
on his miserable King's Indian'. Korchnoi is chess.
particularly scathing about 7...ttk6 in the Clas­
sical Variation and 10 i..e 3 is his weapon of
Gamel
choice.
Koutsin-Frolov
I, on the other hand, used to be of the opin­
ion that 10 i..e3 just lost by force. Admittedly I
Kiev199 5
was rather young when I held that view but I
simply couldn't believe that anyone could be so 1 d4 lLlf6 2 c4 g6 3 lLlc3 .i.g7 4 e4 d6 5
generous as to let me play f5-f4 with gain of lLlf3 0-0 6 .i.e2 e5 7 0-0 lLlc6 8 d5 lLle7 9
tempo. Of course, I now realise that things are lLle1 lLld7 10 .i.e3 f5 11 f3 f4 12 .i.f2 g5
not so simple. The bishop on £2 is perfectly 13 a4
placed to assist in the queenside attack while in
other lines it becomes the white king's most
valuable protector. However, it is also a target
for a black pawn landing on g3. Black's main
strategy in this line is based around playing g5-
g4-g3. He doesn't care if it costs him a pawn as
in return he will get open lines and outposts for
his pieces. Perhaps the most important black
piece in this line is the light-squared bishop.
This is because if White meets ...g3 with h3
then Black may be able to sacrifice his bishop
to rip open the white king position. You will
find countless examples of ...i.xh3 sacrifices
throughout this chapter. They are usually devas-

9
Play the King's Indian

This move is an invention of Korchnoi. To went back to the drawing board and discovered
understand why White wants to start his queen­ that if only he could meet 13...b6 with a2-aS
side attack with this particular move we must Black would not be able to kick the knight with
first look at another of Korchnoi's inventions, a6 as White just replies axb6. Unfortunately,
13 lLlbS. This was introduced by Korchnoi pawns are not allowed to move this far so
back in 1987 when he was having a few prob­ Korchnoi's idea was to start with 13 a2-a4, in
lems with the more conventional lines. The order to 'threaten' 14 lbbS' (not everyone
basic idea is to play lba7 and lbxc8 to remove agrees it is a threat) when 14...b6 can now be
the dangerous bishop on c8. Of course Black is met by 15 aS.
not going to let this happen as then he will have Such is the way chess theory develops. It is
nothing to sacrifice on h3. The most obvious often very hard to understand why a particular
try for Black is 13...a6 14 lba7 L7 1S i.xa7 move is fashionable without understanding the
b6 as the bishop on a7 is in deep trouble. history of a whole variation.
Korchnoi's idea is to play 16 b4 followed by cS 13... a51?
with every chance of extricating the bishop or,
failing that, to at least ensure it is sold quite
dearly. Black's results in this line were very poor
but only, according to John Nunn, because he
tried to win the bishop with moves such as
.i.b7 and lbc8. If instead he just got on with his
kingside attack (16 ...lbf6, 17...g4) then the situa­
tion is not so clear.
Another line Black has tried is 13...lbf6, just
sacrificing the a-pawn to give his bishop a
square. The critical test is 14 lbxa7 i.d7 1S cS
and Black's results there do not inspire.
The third try is what put Korchnoi off 13 lbbS:
13...b6 14 b4 a6 1S lbc3. The position is the
same as after 13 b4 (see next game) except that Black decides to try and punish White for
Black has two extra moves on the queenside. his move order. By blocking the queenside with
Korchnoi must have hoped that they would ...aS and ...b6 he makes it impossible for White
weaken his position but in practice they turned to achieve his most dangerous queenside ad­
out to not be so damaging. Black can play vance c4-cS (at least for a very long time). For a
15...lbg6 or 1S...lLlf6 but there is another inter- while 13...a5 was thought to be inaccurate be­
esting idea, namely 15...l:tf6. The rather crude cause White is still able to develop his attack on
idea is to play ...:116, ...11i'e8, ...11i'h5 and mate the queenside by playing b4. The time that
White down the h-file. This is possible in other Black has spent on moves such as ...aS and ...b6
lines as well (See Game 3) but it is particularly could have been put to better use on the king­
dangerous here as Black has been given the side, so the argument went. However, due to a
very useful move ...a6 for free. The point is that lack of success with his other 13th moves Black
he doesn't have to worry about White playing started to examine 13...aS again and this time
lbbS (remember the queen is intending to leave found it much more to his liking.
d8). Kiltti-Maki, Tampere 1998 continued (after The main line after 13 a4 was considered to
1S...l:tf6) 16 lbd3 l:th6 (16...l:tg6!?) 17 cS?! (17 be 13...lbg6 14 aS l:tf7 but Black has been un­
.i.e1 in order to defend with lLlf2 and h3 is der pressure there on two fronts. The first nasty
more to the point) 17...'ife8 18 �h1 bxcS 19 line is Kozul's pawn sacrifice 1S cS (the main
bxcS lLlf6! 20 11i'd2 11i'hs 21 i.g1 g4 22 lba4 g3 line runs lS...lLlxcS 16 i.xcS dxcS 17 i.c4 �h8
23 cxd6 (23 h3 i.xh3) 23...11i'gS! 24 dxe7 18 a6) while the second is Korchnoi's 15 b4
l:txh2+l 0-1. .i.fB 16 cS lLif6 17 c6!. Often it is possible to
So, discouraged with 13 lL!bS Korchnoi patch a line up against just one aggressor but

10
 7 he Classical Varia ti on: (Vturunn IVIBin Line

\\·hen there are two it is usually time to move his attack underway while White is taking it) 19
on. I also thought about recommending bxaS bxaS 20 itJbS i.£8 21 .i.c3 l:.g7 22 l:.a2
t3...l:.f6 (I have played it myself) with similar lbh8!? 23 1i'e1itJfl 24 .ixaS lbh6 2S cS g4 26
ides to the previous note and Game 3. How­ fxg4 hxg4 27 lbxg4 .ixg4 28 hxg4 lbhxg4 29
ever, it is probably a slightly inferior version and .ixg4 lbxg4 30 lt£3 dxcS Vz-11> Ikonnikov­
1 also want to look at a broader range of posi­ Hassan, Egypt 2001.
tions. White was much higher rated than his op­
14 tZ'ld3 b6 ponent so obviously wasn't too keen on his
position.
c2) 17 ...'�h8 18 ltlbS lbeg8 (the ...lbe7-g8-
h6 is another typical manoeuvre to try and
force through ...g5-g4) 19 b4 ltf7 20 bxaS bxaS
21 cS .i.£8 22 cxd6 cxd6 23 l:.ct lbh6 24 l:.c4,
P.Nielsen-Kottonias, Hastings 2003/04, and
now it is probably best for Black to play the
immediate 24 ...g4 with unclear play.
15 ... axb4 16 tZ'lb5
After 16lbxb4 Black has a choice:
a) 16...lbcS!? 17 aS bxaS 18 .ixcS dxcS 19
itJd3 lbc6! is a nice trick which has occurred a
couple of times. The knight is heading for d4
and after 20 dxc6 1i'd# Black regains the piece
The consistent follow up. There is no point with a good position. 17 aS looks premature.
pla)ing ...aS if Black just allows White to play cS b) Black may prefer to bring his knight to
afterwards. the kingside. Yusupov-Kasparov, Yerevan
15 b4 Olympiad 1996 continued 16...itJf6 17 lta3 (17
White must play b4 if he is going to open itJbS transposes to the next note while on 17
the queenside but he has several ways to im­ lbc6 Black should probably avoid 17...1i'd7 on
plement this plan: account of 18 aS and play 17...lbxc6 18 dxc6
a) He can play 1S b4 and after 1S...axb4 just 1i'e8 with unclear play) 17 ....id7 18 lbbS �h8
recaprure with the knight (see notes to move (18...g4 19 fxg4 lbxe4 was a serious alternative)
16). The knight then has access to the hole on 19 .ie1l:.g8 20 g4! (otherwise White will come
c6 but plays no part in the defence of the king­ under a heavy attack) 20...fxg3 21 hxg3 g4 with
side. equal chances according to theory. Given the
b) He can play 1S b4 and meet 1S...axb4 choice I would take Black.
with 16itJbS followed by .ie1xb4. This was his 16.. lZ'lf6
.

choice in the main game.

c) There is a third way. White starts with 1S
i.e1 and follows up with ltJ£2 and h3 in order
to slow down Black's kingside attack. Only then
does he play for b4. Certain sources have sug­
gested that White can gain the advantage in this
fashion but I don't really believe it. Let's have a
look: IS .iet hS 16lb£2lbf6 17 h3 and now:
cl) 17 ...l:.f7 (a typical manoeuvre which
combines defence and attack) 18 b4 lbg6 (on
18...axb4 19 itJbS White will play i.xb4 fol­
lowed by aS so Black prefers to keep a pawn on
aS to prevent this; he understands that the
pawn is likely to drop off but will hope to get

11
 Play tne King's Indian

17 i.e171 knight on bS is irrelevant) 3S l:te2 'iV£3+ 36 l:tg2

17 lilib4 g4 18 .i.h4 g3 19 h3 (as usual a line 'iVfl+ 37 �h2 e2 0-1 Nikitin-Solovjov, Alushta
like 19 hxg3 fxg3 20 .txg3 .i.h6 would be 2002.
&aught with danger for White; we will see many 18...g3 19 h3 i.xh31
examples of similar positions) is given by Nwm
as better for White. His assessment is based on
the game P.Nielsen-Harestad, Gausdal 1996
which continued 19...i0g6 20 lL\c6 'iVd7 21
.txf6 :Xf6 22i0ba7 but Black may be able to
do better than this.
Firstly, the immediate sacrifice on h3 doesn't
work. After 19....i.xh3 20 gxh3 'iVd7 21 'it>g2
i0g6 22 .txf6 :Xf6 23 l:ht l0h4+ 24 'it>gt
Black can make no further progress on the
kingside. The problem is that White's bishop
has managed to get rid of an important knight,
a knight that Black usually needs for the follow­
up sacrifice on e4.
Instead of 19....txh3 Black should play the This bishop was born to lay down its life on
quiet 19...i.d7, covering c6 and threatening h3. If Black delays then White will shore up his
'iVc8 followed by .txh3. A possible continua­ defences with l:tet and .tft. There is no reason
tion: 20 l:tet (so that he can defend with .i.ft) to delay!
20 ...�h8 21 l:ta3 l0g6 22 .i.xf6 :Xf6 23 'iVat 20 gxh3 1Wd7 21 1Wc2
l:taS! The only variation that Black needed to cal­
There is also an argument for Black delaying culate before sacrificing was 21 'it>g2? lOg6 22
g4 for a while. 17....i.d7 would lead to very l:thtl0h4+ 23 �gl .!ilie4! 24 fxe4 £3 and White
similar play to the Yusupov-Kasparov game gets blown away. He didn't need to look at
given above. anything else because if White can't play 'it>g2
17 ...g4 18 i.xb4 then Black will pick up the crucial pawn on h3.
It was because the knight on bS no longer That's what I wrote in SOKID two years
protects e4 (as it did from c3) that Black was ago. Now we have an exameJe that illustrates
able to play g4 without further preparation. perfectly the danger of ignoring my King's
After 18 fxg4 l0xe4 Black's central pawns are Indian books. Pogorelov-Matamoros, Dos
potentially deadly. After 19 .txb4 we have a Hermanas 2003 actually reached the position
couple of examples: after 24...£3 and the game concluded: 2S i.d2
a) 19....td7 20 'iVc2iOgS 21 h4l0e4 22 .tel (2S .i.x£3 :X£3 is just hopeless for White)
i0g6 23 hSi0g3 24 hxg6 1i'h4 2Si0£2 .txbS 26 2S... £2+ 26 lOx£2 (26 'it>ft g2 mate would be a
gxh7+ �h8 27 cxbS e4 28 l:tdt £3 29 gx£3 prettier way to go) 26...:Xf2 27 l:[h2 gxh2+ 28
.!ilie2+ 30 'iVxe2 ex£3 0-1 Chabanon-Degraeve, �£2 'iVxh3 29 'iVht l:tfBt 30 'it>et 'iVg2! 31
French Ch. 1999 .i.£3 .!ili£3+ 0-1.
b) 19...l0g6 20 aS tOes 21 lilieS bxcS 22 The game Krivoshey-Xie Jun, Linares 1997
.i.c3 e4 23 .i.xg7 'it>xg7 (Fritz likes White here deserves a mention. There White played 21
but I don't really care what it says; give me 'iVd2 and now 2t...'iVxh3 22 .i.dti0g6 23 'iVg2
Black's centre any kingside attacking chances would have transposed to the main game.
any day) 24 'iVc2 'iVe7 2S 'iVc3+ (2S �hl is However, Xie Jun preferred to manoeuvre her
Fritz's choice which is certainly a better move; pieces into more threatening positions before
Black should probably play 2S....i.a6 and l:tae8) capturing on h3. After 2t...i0g6 22 .i.dt .i.h6
25...l0e5 26 l:tael 'it>g8 27 .i.dt e3 28 g3 .i.xg4 23 iOet �h8 24 l:[a2 l:tg8 25 'iVd3? (25 'ifg2
29 .i.xg4 .!ilig4 30 :Xf4 l:txf4 31 gxf4 'iVe4 32 li:'lh4 26 'ifh1 is not so clear) 25...'ifxh3 26 l:[g2
h3 'iVxf4! 33 hxg4 'ifg3+ 34 'it>ht .:e8 (the ltJh4 27 'ife2l:tg5! my database says that a draw

12
 ,,� ..,,t�:.:.tr.an vanauon: Mo ctern
Main Line

was agreed. I suspect this is a mistake and that '5'xh4 loses immediately to 27..�5
and after
White actually resigned. Black just plays ...l:hS 27 �4 l:h5 28 fxe4l:l.xh4 29 '5'f3l:l.h2
Black
and delivers mate on the h-file. also has a winning attack.
21 ...1i'xh3 22 ..i.d1 lljg6 23 1i'g2 1i'h61 24 27 ... f3 28 ..txf3
1i'h1 At first glance 28 '5'xh4 loses immediately to
After 24 ttlxc7 �h4 25 '5'e2 ttlxe4! 26 '5'xe4 28...fxg2 but White has the absolutely amazing
(both 26 fxe4 and 26 �8 are hopeless) the move 29 l:l.f4! Black still wins the rook but he
beautiful 26...lbg2! gives Black a winning attack. doesn't get a new queen as well. After 29...llxf4
These tactics may look amazingly complicated 30 'irxh6 ..i.xh6 Black is better but White can
but it is no surprise they work. The unifying fight on.
theme is ...ttlxe4 as this brings to life the other 28...<Llxf3+ 29 l:lxf3 1i'xh1+ 30 �xh1 l:lxf3
Black pieces (and pawns)
24 ..lDM 25 ltJ81
.

If White can play just one more move, 26

.'Dg2, then he should be able to beat off the
attack. It's time for another sacrifice to keep up
the momentum of the attack.
25 ...<Llxe4!

The main wave of complications have ended

and Black has emerged with the slight material
advantage of rook and 2 pawns against two
minor pieces (in an endgame anyway). Black's
passed pawns on the kingside are potentially a
decisive factor but for the moment it is the
weakness of the respective queenside pawns
For those of you who have been checking that is the most relevant.
the variations in the notes this can come as no 31 �g1!
surprise. Not falling for 31 ttlxc7l:l.xa4! 32l:l.xa4 l:l.fl
26 <Llg2! mate.
26 fxe4 £3 is completely hopeless for White 31 " .l:lb3?
26...l:lf5?! Black needed to play one careful defensive
Black had another fascinating possibility: move to keep the advantage. After 31....i.f8 32
26...�xf3+! 27 ..i.x£3 'irxht+ 28 �xh1 �£2+ 29 ttlxc7 l:l.c8 it is important that the d6 pawn is
l:.xl2 (29 q;,g1 e4! and despite the fact that defended. White could try 32 aS but the posi­
queens have been exchanged and White has tion after 32...bxa5 33 l:l.xaS l:l.xaS 34 .i.xaS
two extra pieces he is just lost) 29...gxl2 30l:l.fl l:l.bt+ 35 �e1l:l.ct is good for Black.
(there is nothing better, e.g. 30 ..i.e4 l:l.xa4!) 32 <Llxc7! l:lf8 33 ..te1?I
30...l:l.xa4 31 ..i.a3 l:l.xc4 32 l:l.x£2 and the rook White probably didn't like the look of 33
and five pawns should beat the three minor ..i.xd6!l:l.£2 but 34 l:l.fl holds the balance. Now
pieces. This looks more convincing than Black's Black is pressing again but the odds are still on
choice in the game. White holding the draw. The remaining moves
27 fxe4 were:
White has to take one of the pieces; 27 33.....i.f61 34 lLlb5 ..i.e7 35 l:le3 l:lxa3 36

13
Piay the King's Indian

t!Jxa3 :t3 37 t!Jc2 .tg5 38 .tb4 h5 39 bishop. For example 1S l:.cl l:.f7 16 a4 �£8 17
t!Jce1 :t6 40 a5 bxa5 41 .txa5 :t2 42 cxd6 �xd6 18 tt'ld3 b6, preventing tt'lcS, is
.tc7 .te7 43 tlJ83 :e2 44 t!J1g2 lla2 45 c5 supposed to be comfortable for Black.
:a1+ 46 t!Jf1 h4 47 t!Jxh4 .txh4 48 .txd6 I would like to look at another continuation
:e1 49 c6 :xe4 50 c7 :c4 51 .txe5 �7 for White, 1S a4. If Black now continues as in
52 t!Jxg3 .txg3 % ·% the main game with 1S... l:.f7 16 aS �£8 we have
transposed into a position which is normally
Game2 reached via the 13 a4 move order. This position
is mentioned briefly in the 13th move notes to
Piket-Kasparov
the previous game. There it was pointed out
Tilburg,1989 that Black has been suffering after the move 17
c6. The reason is that whilst 1S...l:[f7 is a good
1 d4 t!Jf6 2 t!Jf3 g6 3 c4 .tg7 4 t!Jc3 0-0 5 reply to 1S l:.cl (it defends c7 in anticipation of
e4 d6 6 .te2 e5 7 0-0 tlJc6 8 d5 t!Je7 9 White's planned invasion there) it is a little pas­
tlJ81 ttJd7 10 .te3 f5 11 f3 f4 12 .i.f2 g5 sive in reply to 1S a4. It is probably better for
13 b4 Black to play 1S...hS with the following possi­
bilities:

In some ways this is White's most obvious

move. Now there is nothing Black can do to a) 16 aS?! (too slow) 16...g4 17 cxd6 cxd6 18
stop White playing cS but at least a tempo has tt'lbS g3! (this is the thematic breakthrough;
been spent to prepare it (there are many lines Black gives up a pawn to open lines on the
where White gets to play cS without b4). 13 b4 kingside) 19 hxg3 (White must accept as 19
used to be considered as the main line but after �xa7? loses by force in this position; Black
what happened to Piket in this game its popu­ plays 19... tt'lh7!, threatening ... 'itb4, and meet­
larity rapidly diminished. ing 20 �h 1 with 20...:Xa7! 2 1 tt'lxa7 1ih4 22
13...t!Jf6 h3 �xh3) 19... fxg3 20 �xg3 a6 2 1 tt'la3 h4 22
Sometimes the knight stays on d7 to prevent �£2 tt'lhS 23 f4 (this is the only way for White
cS but after b4 this advance has become un­ to try and defend the kingside) 23... tt'lhxf4 24
stoppable. The knight may as well go and join tt'lc4 'ifgS with a clear advantage to Black,
the kingside attack. From f6 it obviously has a Ftacnik-Smirin, Biel Interzonal 1993. Only one
key role in supporting the advance g4. side has an attack in this position.
14 c5 t!Jg6 15 cxd6 b) 16 h3 l:.f7. The inclusion of the moves h3
Wben White plays the line with 13 b4 his and ...hS changes the dynamics of the position
usual plan is to play cS, cxd6, l:.cl and tt'lbS. It as now White is going to have to take on g4.
is also customary for him to play cxd6 before This means that the h-file will be opened so
Black can play �£8 as this gives Black the seri­ that is where Black's rook is heading. 17 aS �£8
ous option of recapturing on d6 with the 18 c6 l:.h7 (if Black takes on c6 then White will

14
 7 ne C.;Jasstcat vanartan: tll'laaern 1111atn Ltne

ha" e the possibility of llJdS as well as danger­ stronger with the pawn still at home. For ex­
ou s queenside pawns) and now: ample, after White accepts the sacrifice on g3
bl) 19 bS b6! 20 axb6 cxb6 is very interest­ Black may be able to hit the bishop immediately
ing strategically. Black has allowed White a with ...lLlhS (instead of ...h5-h4 and then ...lbhs
massive passed pawn on c6 but in return the when White has a tempo) while a pawn on h4
q ueenside is totally blocked. Of course an end­ can just get in the way of Black's queen. So
game won't be fun for Black but before the ideally, Black would like to force through ...g4
en dgame there is the middlegame! But make without playing ...hS but this is not always pos­
sure you don't consider playing in such a posi­ sible, especially if White leaves his knight on c3
cion until White has played the move b4-b5. If to defend e4. Let's have a quick look at the old
the pawn were back on b4 White would have move 17...h5 as this is also not bad and there
use of the bS- square, especially for his knight, are some interesting thematic lines. White re­
and things would not be so easy for Black. plies 18 aS and now:
After 20...cxb6 the game Hausner­
Dolm atov, Bundesliga 1993 continued 21 l:.a3
'iic7 22llJci3 g4 23 fxg4 hxg4 24 hxg4 1i'g7! 25
.ixb6 'i'h6 with a strong attack for Black.
b2) 19 cxb7! (an improvement) 19...i.xb7 20
bS (20 a6 .ic8 21 lbbS is too slow on account
of 21...g4� 20...i.c8 21 b6 (White sacrifices a
pawn to open the queenside) 21...axb6 22 axb6
l:.xal 23 1i'xa1 cxb6 24 1i'a8 g4 (24...i.e7!?
would prevent White's next move) 25 .ia6
lL'le7 26 .ih4 l:.h6 27 i.xf6 .:Xf6 28 .ixc8
lL'lxc8 29 lL!bS l:.£7 30 fxg4 hxg4 31 hxg4 1i'd7
32 'i'c6 'i'xg4 33 lL!f3 and White had some
compensation for the pawn (I think the posi­
cion is level), Stanec-Arakhamia, Vienna 1996. a) 18...g4 t 9lL!bS g3 20 .ixa7!lL!h7 21 �ht
c) 16 c6!? .ih6 17 bS b6 18 aS g4 19 axb6 .:Xa7 22 l:.xc8! 1i'xc8 23 lbxa7 1i'd8 24 h3
cxb6 20 fxg4!? hxg4 21 g3 with an unclear should be compared with line 'a' in the notes to
game, Korchnoi-Xie Jun, Marbella 1999. White's 15th move. There, with his rook on at ,
Korchnoi's sense of danger did not desert him. White was unable to return the exchange with
He realised that with the queenside closed dras­ .:XeS and just got mated. This position after 24
tic action was required on the kingside. Black, h3 used to be considered good for White but in
too, can now have a protected passed pawn on the latest game Black played 24...lL!gS followed
the 6th rank but her attack would also be by a sacrifice on h3 (White didn't accept) and
blocked. Xie Jun preferred to hold back on ...f3 the game ended in a draw.
and eventually lost a complicated game. b) 18....id7 19 lLlbS i.xbS (one of the rare
15 .. cxd6 16 l:tc1 l:tf7
. occasions Black is willing to give up his bishop)
Black must defend his c7-square. Now 17 20 .ixbS g4 21 �ht g3 22 .igt gxh2 23 i.£2 is
lObS can just be met by 17...a6. the main line of 17...h5. This is a well-known
17 a4 .itS!? defensive tactic from White - he uses Black's
Kasparov's new move. Previously Black had pawn on h2 to shield his king. Black will now
played 17...h5 to suppon g4 but Kasparov real­ try and attack with ...h4, ...lbh5-g3 and very
Ised that if White plays lbbS, as he wants to do, often he will throw in the move ...h3 to open
then Black does not actually need ... hS. He can the h-file. White will obviously try and break­
just play ...g4 and if White plays fxg4 then through on the queenside. The game is very
...lt'lxe4 destroys the white centre. Not only is unclear and Black is supposed to stan with
Kasparov hoping to save rime by omitting ...hS 23...a6 in order to prevent White from advanc­
but sometimes the attack can actually be ing his own pawn to a6.

15
Play the King's Indian

18 aS .i.d71 25 �7 �7 26 0.xa8 gxh2 27 i.f2 1fxf2! with

Kasparov is trying to tempt White into play­ the point 28 l:.xf2 �3+ 29 �xh2 llh6t 30
ing ll:)b5. Piket obliges and becomes yet an­ �g1 llh1 mate. However, White could have
other victim of the champ's legendary prepara­ defended with 26 �3 as then after 26...gxh2
tion. 27 i.f2 the queen sacrifice no longer works and
19 lL!b57! on 26...llh6 he can play 27 1t'e1. This is aU very
Afterwards White tried to rehabilitate this frightening for White but I'm not sure that he
line with 19 �h1. This is a useful defensive can't defend. Perhaps Kasparov was thinking of
move and it makes sense to play it before initi­ the proverb 'better a bird in the hand than two
ating complications on the queenside. One idea in the bush' as he recommended 23...gxh2.
for Black is to play 19...1t'e8 to stop ll:)b5 (note Kasparov also gave 20 fxg4 ll:)xe4 21 �7
that 19...a6 is bad as this concedes the b6- i.a4 22 1fxa4l:.xc7 as White's best chance but
square) but 19...llg7 is more direct. Now: there have been no takers for this variation.
a) 20 ll:)b5 g4 21 0.xa7 g3 22 i.b6 1t'e8! (to 20...g31
prevent i.b5) 23 llc7 gxh2 24l:.xb7 ll:)h5 was I hope this move doesn't come as a surprise
very unpleasant for White, Burgess-Badea, to you. We have seen it many times already.
Manchester 1990. Burgess confinns this in the 21 lL!xaB?
book he co-authored (with John Nunn) on the This loses by force. Kasparov produced
Classical King's Indian. realms of analysis to show that Black is better
b) 20 i.b5 g4 21 i.xd7 1fxd7 22 fxg4 �4 after 21 hxg3 fxg3 22 i.xg3 i.h6!. We just have
23 ll:)£3, D.Gurevich-Griinberg, New York room for his main line. It runs 23 0.xa8 ll:)h5
1991, was more solid but Black is still doing 24 i.f2 ll:)gf4 25 �3! llg7! 26 0.xf4 i.xf4 27
quite weU after 23...i.e7. He plans to play g4! i.xc1 28 'it'xc1 ll:)f4 29 'it'e3 h5! with a clear
...ll:)h4 in order to exchange some of White's advantage for Black.
defensive pieces. 21 ••. lL!h5!
19...g4!

Piket had probably been expecting 21...gxf2+

20 lL!c77! 22 llxf2 'it'xa8 when his rooks might become
After 20 ll:)xa7 g3 Kasparov gives the varia­ very active on the queenside. Kasparov's pow­
tion 21 i.b6 1t'e7 22 �h1 ll:)h5 23 ll:)b5 gxh2 erful continuation leads to the same position
24 i.f2 i.xb5 25 i.xb5 ll:)g3+ 26 i.xg3 fxg3 as except with his knight on g3 instead of f6. We
clearly better for Black. In effect he is a king up! shaU see what a difference this makes.
This didn't satisfy Black in the game Friesen­ 22 �h1
Lomineishvili, Rotterdam 1998. He decided to To avoid getting checkmated White must be
improve upon Kasparov's analysis and suc­ in a position to meet "ifh4 with -'.gt. He can
ceeded in spectacular fashion! Instead of never play h3, of course, as Black just sacrifices
23... gxh2 the game went 23...1Vh4 24 i.gt l:tf6 the bishop on d7.

16
 The Classical Varia tion: M odern
Main Line

22 ... gxf2 23 l:l.xf2 �g3+1 24 �g1 'it'xa8 However, there is an interesting pawn sacrifci e
Of course if W'hite had taken the knight on that enables 13 l:tcl to stand alone as an inde­
his last move he would have been mated at pendent line. After 13 ...li)g6 White can play 14
once. There is still no time to take it as after 25 cS (Kozul's idea again) 14...li)xc5 (14...dxc5 is
bxg3 fxg3 the queen just returns to d8 and worse) 15 b4li)a6 and now with either 16 ttxl3
Black has a winning attack. or 16 li)bS he obtains reasonable compensation
25 i.c4 a6! for the pawn (mainly because the knight is mis­
placed on a6). lbis line has become very theo­
retical and also leads to the sort of position we
don't really want.

A beautiful little move which enables the

black queen to occupy the most important
diagonal on the board. The respective pawn
chains mean that it is almost impossible for Therefore, I suggest that it is time to indulge
\V'hite to contest this diagonal. in a spot of rook swinging. The possibility of
26 '*d3?! 'it'a7 27 b5 playing ...l:tf6 has already been mentioned sev­
\V'hite overlooks the threat but after eral times in this chapter. Now it is time for this
27...'ifd2 Black would have won by 27...i.e7 idea to take centre stage. It can be played
and i.h4. against virtually all of W'hite's 13th moves but
27 ... axb5 28 ..ixb5 �h1!! 0-1 most theoretical sources agree that this is one
If people liked 25...a6 then they were in rap­ of the best moments to play 13...l:tf6. That is
tures over this. It was wonderful how Kasparov because 13 :ct puts less immediate pressure
co-ordinated his queen and knight from the on the black queenside than some of the other
21st move onwards. Piket resigned, as he will moves and unlike 13l[)d3, it does not contrib­
now be a whole piece down. ute defensively either.
I have already touched upon the reasoning
Game3 behind 13...:f6 in the analysis of 13 li)bS in
Game 1. A quick recap. The basic idea is to play
Kaganskiy-Golod ...:h6 and ...1Ve8 and if W'hite fails to take the
Israe/1999 necessary defensive precautions then Black
plays ...'ifhs and mates him. Take a look at the
1 d4 �f6 2 �f3 g6 3 c4 ..ig7 4 �c3 0-0 5 foUowing examples that demonstrate some of
e4 d6 6 ..ie2 e5 7 0-0 �c6 8 d5 �e7 9 the tactical possibilities available to Black. In
ti:Je1 �d7 10 ..ie3 f5 11 f3 f4 12 ..if2 g5 the first two W'hite was careless while the third
13 l:tc1 l:tf6!? features a classic panic reaction.
The most common reply, at least in the past, They all start after 13 l:tcl l:tf6:
to 13 l:tcl is 13...li)g6 when play can easily a) 14l[)d3 l:th6 15 b4 1Ve8 16 cS? 'ifhS 17
transpose into one of the other lines. For ex­ h3li)xc5 (17...li)f6 also does the trick) 18 �xeS
ample, 14 b4 li)f6 15 cS takes us into Game 2. �xh3 19 lf:)f2 �xg2 20 �g4 1Vh1+ 21 �f2

17
Play the King's Indian

'ifh4+ 22 �gl llh5 0-1 Borges Mateos­ 14 b4

Pecorelli, Cali 2000. White is no mug but a
member of the strong Cuban national team.
b) 14 b4 l:th6 15 c5 a6 16 tLla4 ...e8 17 �hl
�h8 18 cxd6 (the prophylactic ..igl should
have been played at some point) 18...cxd6 19
l:tc7 b5 20 tLlb2 tLlf6 21 a4 'ifh5 22 ..igt
tLlfxd5! 23 l:txc8+ (23 exd5 tLl£5 24 h3 tLlg3t 25
�h2 ..ixh3 wins) 23...l:txc8 24 exd5 tLl£5 25 h3
e4 26 �h2 e3 27 tLled3 tLlh4 0-1 Opalic-Socko,
Passau 1999.
c) 14 t0d3 llh6 15 c5 ...e8 16 g4 fxg3 17
..ixg3 tLlf6 18 cxd6 cxd6 19 �£2 tLlg6 20 �e3
tLlf4 21 �d2 t06h5 22 �c2 ..id7 23 �bt l:tc8
24 ..i£2 tLlxe2 25 ...xe2 tLlf4 26 tLlxf4 exf4 27
..ixa7 llh3 28 ..igt g4 29 fxg4 llhxc3 30 bxc3 White prepares to play c5 and as I men­
..ib5 31 �2 ...xe4+ 0-1 Hoensch-Kiundt, tioned above playing it at once as a pawn sacri­
Bundesliga 1991. fice is possible and that is covered in Game 4.
How sad to have made it all the way to the The position after 14 t0d3 llh6 featured in a
queenside just to get mated there. couple of our quick wins for Black. After 15 c5
There are four basic defensive strategies ...e8 White should not play 16 g4 or 16 b4 (see
against Black's attack: . above), nor 16 cxd6 which can be met by
1) White can play �h1 in order to meet 16...1i'h5!. Neither is there time for 16 ..iet but
...'ifh5 with ..igt. If Black can then land with a White could, conceivably, try 16 tLlb5 intending
knight on g3 that will be very nice but more to meet 16...'ifh5 with 17 h4. See the note to
often he ends up playing for g5-g4. White's 16th move for analysis to a similar
2) White can play t0d3, ..iet, and tLl£2 so position. I don't know why White players avoid
that he can meet ...'ifh5 with h3. Assuming that 16 �hl in this position as 16...'ifh5 17 ..igt
White arrives in time (i.e. he is not going to get tLlxc5 18 tLlxc5 dxc5 19 tLlb5 is good for White.
blown away by a sacrifice on h3) then Black Black should prefer 16...tLlf6 with similar play
should switch strategies and play ...l:tg6 (with­ to the main game.
out ...'ifh5) ...h5 and ...g4. White's queenside Incidentally, the no-nonsense American
attack will be much slower as his knight and Grandmaster Yermolinsky played 15 ..iet (after
bishop are less actively placed. 14 t0d3 llh6) so that he can defend with tLl£2
3) White can play g4. In practice this tends and h3 if need be. Yermolinsky-Gurcan, Istan­
to occur out of desperation when it is already bul Olympiad 2000 continued 15...a6 16 b4
too late, but if White played it on move 14, for tLlf6 17 tLl£2 l:tg6 18 c5 h5 19 h3 and now I like
example, it would be quite acceptable. the idea of 19...�h7 followed by tLleg8-h6 and
4) Once again White can sacrifice a pawn g4.
with 14 c5. You may not think of this as a de­ Before moving on it should be pointed out
fensive strategy but it does create a diversion. that Piket, one of l:tct's main supponers, has
Black must deal with concrete problems on the played 14 g4 here. After 14...h5 play will be very
queenside before he can contemplate an attack similar to the 13 g4 line we are going to exam­
against the white king. I don't believe the sacri­ ine within the notes to Game 5.
fice is so dangerous for Black in this particular 14...:116 15 c5
position. See Game 4 for the details. I have switched the move order of the game
The three illustrative games are examples of so that we can look at some interesting stuff
White failing in his defensive effons. Below we after 15 c5. The actual move order was 15 t0d3
shall see what happens when White plays bet­ tLlf6!? 16 c5 'ife8 17 �h1. Instead of 16 c5
ter. White could have played 16 ..iet in order to

18
 The CIBSSICBI Vana tton : M O Oern MBtn L ine

meet 16 ...'W'e8 17 tD£2 1i'h5 with 18 h3. Black

should play 17 ...l:tg6 18 h3 h5 with a fairly
standard King's Indian position.
1 5 . 'ife817
. .

a) 17 ...tDg6 is an attempt to keep the h-file

open. Now 1 8 tDxc7 loses to 18 ... �4 (threat­
ening ...tDx£3+ or ...tDxg2 or tD anywhere), e.g.
19 i.xh4 'W'xh4 20 tDxas 1i'h2+ 21 �£2 'W'g3+
This is actually a new move, or at least a new 22 �g1 .:th2! and there is nothing White can do
move order that I have conjured up for the about ...1i'h4 and ....:th1 mate. But White does
readers of this book. Everyone else gives 1 5 ... a6 have a defence: 18 g4! fxg3 19 i.xg3. Now
in order to prevent tDb5. However, there are Black has to be careful. For example,
drawbacks to playing ...a6. Firstly, it wastes a 19 ...gxh4?? 20 f4! costs him his queen while
tempo and secondly, although it covers the b5- after 19 ... �4 20 tDxc7 I can't see anyway
square, it weakens the b6-square. In Korchnoi­ through for Black. The best move is 19 ... tDf4
Nataf, Cannes 1 998, the old war-horse tried to and after 20 tDxc7 l:tg6. Black now threatens
exploit this at once with 1 6 tDa4!?. Now if the 21...gxh4 so White has to choose between 21
black knight moves from d7 White exchanges �h1 and 21 i.xf4. I have analysed masses of
on d6 and invades on b6. And if the knight variations and I still don't know what's happen­
can't move it means that the whole of the black ing. Let's just say the position is unclear. Per­
queenside is tied down. In the second of our haps that is unsatisfactory but the editor would
three quick wins above we saw that Black has have cut all the variations out anyway.
tactical ideas here as well but they only worked b) 1 7 ... gxh4 1 8 tDxc7 h3 1 9 g4 fxg3 20 i.xg3
because White failed to take prophylactic action (taking the rook is also possible; perhaps Black
with i.gt . The continuation of Korchnoi-Nataf starts with 20...'W'g5!? as 21 i.xg3 'W'xg3+ �h1
was (after 16 tDa4) 16 ...'W'e8 17 �h1 1i'h5 l:tg6 is winning; I think White must play 21
( 1 7 . ..�h8!?) 18 i.g1 1i'h4 (threatening lbf6-h5) �h1 when 21 ...h2 is fun and maybe good for
1 9 tDd3 b5 (19 ...tDf6 20 'W'e1) 20 cxb6 axb6 21 Black but there are no guarantees) 20... h2+!?
'i'e 1 with a slightly better game for White. (20 ...l:tg6 21 tDxa8 1i'h4! wins for Black but I
1 6 �h 1 don't see anything good against 21 �h1!) 21
So what happens if White takes up the chal­ �h 1 1i'h3 and now:
lenge and plays 1 6 tDb5!?. Well, the answer is I b1) 22 cxd6 'W'xg3 23 dxe7 tDf6 24 tDxa8
don't really know but I'll share some of my l:tg6! 25 l:txc8+- �t7 and White gets mated after
thoughts with you. After 16 ...1i'h5! 17 h4 (this he runs out of checks.
suicidal-looking move is actually forced as 1 7 b2) 22 tDxa8 'i1Vxg3 23 tDd3 l:tg6 24 .l:t£2
h3 loses t o 1 7. . .tDf6 followed by . . .i.xh3) Black tDxc5!! (threatening i.h3) 25 l:txh2 tDxd3 26
can play either 1 7 ...tDg6 or 17 ... gxh4. When I i.xd3 i.h3 and with ...i.h6 to follow Black has
staned my analysis I assumed that Black could the advantage.
crush him with either move but White's posi­ b3) 22 i.£2! is best and after a few hours of
tion turned out to amazingly resilient: unsuccessfully sacrificing my rook I finally set-

19
Pla y th e King 's Indian

tied for 22...l:tb8. The position is totally unclear. game after the text. Black has managed to pre­
Black's queenside has collapsed but so has vent lLlbS without having to reson to ...a6. His
White's kingside. There were many lines I plan now is to break through on the kingside
looked at where White seemed to be doing with ...g4. Note that 17...lt:\h5 would be met by
quite well but then he suddenly got mated. I'll 181i'et.
just point out that Black has no need to fear 23
cxd6lbg6 24 i.xa7 as he has the reply 24.)Llf4
and if White plays 25 ..i.xb8 then he is mated
after 2S...liJhS!
To conclude, if you don't trust these varia­
tions then play 15...a6, but in my opinion it will
take an extremely brave man to play 16lLlbS.
1 6 . . .llJf6!
White has defended against the threat of
"iVhS which he can now meet with 17 i.gt.
Black needs to feed some more pieces to the
kingside if he's going to attack successfully
there. It was possible to play 16...a6 but it is still
unnecessary.
1 8 b5!?
Another game also reached this position and
went 18 lLlb2 1i'g6 19 ltJc4 lLlhS 20 1i'e1 but
Black could have played 19...g4!
1 8 . . .b6 1 9 cxd6 cxd6 20 a4 �hB 21 '1Ve1
White intends to play g4 to try and relieve
some of the pressure on the kingside. Black
could play 21...g4 himself but he only wants to
play this when he has more forces available. As
we shall see from the course of the game there
is no need to stop White playing g4.
21 . . .'ii'g 6 22 g4 fxg3 23 ..ixg3 llJh5 24 .l:l.g1
.l:l.fB 25 llJd1 g4

1 7 llJd3
This time 17 lLlbS looks much more plausi­
ble but it is in fact even worse than on the pre­
vious move. After 17lLlb5 'ii'hs 18 ..i.g1 Black
has the delightful move 18...1ih4!! Now if
White plays 19 lLlxc7 then after 19...lLlhS! he
gets checkmated. The only way to prevent
...ltJhS is with the sad 19 g3 and after 19... fxg3
White must try 20 lt:\xc7. The main line of my
analysis then runs 20...lt:\g6! 21 lt:\g2 (21 lLlxa8?
1i'xh2+! 22 i.xh2 :.Xh2+ 23 �g1 lLlf4 leads to
mate) 2t...'iih3 22 l:.e1 (22 lLlxa8 lLlf4 wins)
22...lt:\h5 23 ..i.ft gxh2 24 ..i.t2 g4! with a pow­
erful attack for Black (I think it wins by force). 26 .l:l.c7?
1 7 .....id7 White makes his first aggressive move for a
After my move order trip to unchaned terri­ long time and it's a mistake. The siruation was
tories we arrive back in the Kaganskiy-Golod less clear after 26 i.h4!lLlf6 27 l:.g2.

20
 The Classi cal Vari a ti o n : Mo dern Main Line

26 ... gxf3 27 ..tf1 "ifeB 28 lba7 ..tf6 29 Rotstein-Isonzo, Arco 2003 conrinued 17..ll . g6
'ii'e3 tLlxg3+ 30 lbg3 'ih15 31 "ifg1 .tea 32 (18 ixa6 bxa6 19 'ifc2 tt:lc8 20 tt:\c6 1i'e8 is
ltl3f2? 11gB 33 llxg8+ tLlxgB unclear) 18 a4 hS 19 h3 ih6 20 llc3 �h8 21
Now we see what was wrong with White's tt:lbS tt:\g8 22 'ifc2 tt:lf6 23 tt:lxc7 tt:lxc7 24
32nd move. It boxed in his own queen and the llxc7 g4! 25 fxg4 hxg4 26 hxg4 .i.xa4 27 1i'c4
threat of 34...llg8 now forces the knight to .i.bS! 28 'ifxbS 'ifxc7 with a decisive advantage
retrace its steps. for Black.
34 tLld3 llg6 35 "ife3 ..th3 36 "ife1 ..txf1 1 6 ...llh6
0-1 Black just gets on with his attack.
1 7 a4 "ife8 1 8 'it>h1 ..td7!
Game 4 Up until here the two players had been fol­
lowing a game they played the previous year in
Vera-Nataf
Cuba. That went 18...'ifhs 19 .i.g1 .i.d7 20
Montrea/2003 tt:la2llc8 21 tt:lf21i'e8 22 tt:\g4llg6 23 bS tt:\cs
24 ixcS dxcS 25 :XeS with advantage to
1 d4 tLlf6 2 c4 g6 3 tLlc3 ..tg7 4 e4 d6 5 White, Vera-Nataf, Havana 2002.
.ie2 0-0 6 tLlf3 e5 7 0-0 tLlc6 8 d5 tLle7 9 1 9 'ifb3 �h8 20 tLlb2
ltle1 tLld7 10 ..te3 f5 1 1 f3 f4 1 2 ..tf2 g5
13 llc1 llf6 1 4 c5!?
The lines of the previous game are not a lot
of fun for White so in the last year or two he
has introduced a familiar pawn sacrifice in a
new guise. It has to be said, though, that Black
players seem to have come to terms with this
particular version of the c4-c5 sacrifice quite
quickly.
14...tLlxc5
14... dxc5 15 lbci3 b6 16 b4! cxb4 17 tt:lbS cS
18 dxc6 tt:lxc6 19 llxc6! llxc6 20 tt:lxb4 1-0
Kallio-Nordenbaek, Copenhagen 2003 is not
the way to go.
15 b4 lDa6 Natafs improvement is quite clever. By de­
laying .. .'ifhS for a couple of moves he has kept
pressure on the a4 pawn and prevented White
from repearing the strong plan he found in their
previous game. But this is a good moment to
launch the attack as the knight on d3, which
often retreats to f2 to help in the defence of the
king, has disappeared to the queenside
20 . . .'ih15 21 ..tg1 g4 22 fxg4
White must capture on g4 as otherwise
Black will win with ...g3, ...1i'g5, ...llxh2+ and
...'ifh4.
22 . . ...txg4 23 ..txg4 "ifxg4 24 tLlb5 ..tf6 25
tLlxc7?
25 'iff3 looks better when Black can reply
16 tiJd3 25...1i'd7.
White can regain his pawn with 16 tt:lbS 25 ...ttJxc7 26 llxc7 "ife2! 27 lld1 "ifxe4
.i.d7 17 tt:lxa7 but it is far from clear if that is a 28 tLlc4
better continuation. For example, the game He probably has to play 28 'ifc4 but

21
Pla y the King 's Indian

28. .'�Jf5 29 .i.f2 'ii'xc4 30 lbxc4 e4 is no fun

. i s not well placed here. It has few active possi­
for White. bilities and just tends to get in the way of the
28 ...lLlf5 29 �f2 �h4! 30 �xh4 l:tg8!? other pieces.
The simple 3 0. . .lbxh4 31 'ii'h3 llg8 looks
crushing as 32 l:r.c8 is met by 32...liJf5.
31 'iff3 'ifxf3 32 gxf3 l:txh4

There is one other 1 3th move that has to be

considered and that is 1 3 g4. The motivation
behind such a move is to try and stop Black's
The attack is even stronger after the ex­ kingside attack before it has even started. This
change of queens now that the g-file is open move crops up from time to time in various
33 l:td2 variations of the King's Indian and is mainly
Or 33 l:r.xb7 lbg3t 34 'iti>g2 lbe2t! 35 'it>h 1 played by players who are nervous about their
(35 �fl l:r.xh2) 35 ...l:r.h6! and Black has a mat­ king coming under fire. Immediately Black has
ing attack an important decision to make. Should he cap­
33 ...e4! 34 l:tf7 lLlg3+ 35 �g1 lD82+ 36 ture en passant or not? Let's have a look at
�1 l:tg1 + 0-1 both options.
Sweet revenge for Nataf after his earlier de­
feat. The ball is now in White's court in this
variation.

Game S
Korchnoi-Xie Jun
Amsterdam 2001

1 d4 lLlf6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5

�e2 0-0 6 lLlf3 e5 7 0-0 lLlc6 8 d5 lDe7 9
lLle1 lLld7 1 0 �e3 f5 1 1 f3 f4 1 2 �f2 g5
1 3 lLld3
This move has a bad theoretical reputation
based on some old games where Black domi­ a) 1 3. .. fxg3 14 hxg3. The character of the
nated on the kingside. This reputation seems to position has now changed dramatically. With
be more or less justified but Korchnoi, at peri­ the disappearance of the pawn on f4 White
odic intervals, has taken it upon himself to try now has more space for his pieces on the king­
and rehabilitate this line. Even he hasn't had side. His king may look more exposed but there
great success. Although lbd3 ensures that is much less long-term danger to it than before.
White can force through c5 without wasting a In practice White scores heavily after Black
tempo on a pawn move such as b4, the knight exchanges on g3 but the move is not necessarily

22
 The Classical Varia tio n : Mo dern Main Line

bad. After 14...lt:Jg6 1S ii'd2 hS 16lbg2 aS 17 white kingside is about to cave in. Black threat­
lDa4 b6 18 a3 an interesting moment has arisen. ens 23...hxg4 24 hxg4 llh2+ so White prevents
Black has made some useful prophylactic this with 23 ii'g1 but after 23...ii'g6! he is totally
moves on the queenside but now he has to embarrassed as his rooks can no longer protect
decide whether or not to play 18...h4. This has each other. After 24 bS hxg4 White played the
the advantage of gaining control over f4 (White amazing 2S hxg4 just giving up his queen for a
must reply g4) but at the same rime concedes rook and resigned a few moves later. 2S fxg4
the f5-square to White. It also totally blocks up offers slightly more resistance when 2S...lt:lxe4
the kingside so Black can forget about any di­ is probably stronger than 2S...lt:lxg4 (though
rect attack against the white king. In both should win). This game may have given
Savchenko-Fedorov, Nikolaev 1993 Black you the impression that Black wins by force
played 18...h4 and after 19 g4lLlf4 20 'it>h2lLlf6 against 13 g4 but this is certainly not the case.
21 lt:Je3 �d7 22 lt:Jc3 a4 23 llfb1 he was very At least you know what you are aiming for
solid but had no active ideas at all. White's posi­ now. This material should be studied in con­
tion must be slightly better. junction with C.arne 9 where g4 is played in a
In his notes to this game Savchenko sug­ similar position.
gested that 18...lt:JcS is better. He gave the fol­ 1 3 ...lt:lf6
lowing line: 19 lLlxcS bxcS 20 b4 axb4 21 axb4 This is not thought to be the best moment
:Xa1 22 llxa1 cxb4 23 ii'xb4 g4 (this is one of to play the ...llf6 plan, e.g. 13..Jlf6 14 cS llh6
the reasons for holding back with ...h4 - Black 1S cxd6 cxd6 16 lLlbS is considered good for
has more options on the kingside) 24 fxg4 White.
.ixg4 2S �xg4 hxg4 26lLle3 which he assessed 1 4 c5ltjg6 1 5 a4
as unclear. Burgess then chipped in with
26...lLlf4 27 lLlfS lLlh3+ 28 ..tg2 lLlxf2 29 ..txf2
and said that White was better. 28...lt:lxf2 is a
sill y move. Black should play 28...ii'd7 and then
sacrifice the exchange on f5 with what appears
to be adequate compensation.
b) 13...hS is the alternative. Black allows
White to keep his blockading pawn on g4 and
plans to attack by transferring some major
pieces to the h-file. He is in no rush to play
hxg4 as this will just lead to exchanges on the
h-file. He will only take on g4 when he has a
decisive idea in mind. Black's attack requires a
lot of build-up but White's queenside play is
also more laboured than usual as he has to Korchnoi's attempts to give this line a new
make sure that the g4 pawn has sufficient cover lease of life have centred around the advance of
- a piece sacrifice there is likely to be devastat­ the a-pawn. While nearly everyone else concen­
ing. Let's have a look at a game that went like a trates on cxd6 Korchnoi often prefers cS-c6.
dream for Black. Jurga-Baecker, Germany 199S First though, he likes to get a pawn to aS, hav­
continued 14 h3 llf6 1S �g2 llh6 16 llh1 ing been scarred by his game with Kasparov
'itf7!? (Black clears the way for his queen to the (see below).
h-file; another typical plan in such positions is The alternative is 1S llcl ll£7 and now:
to play ...lbg6 followed by ...�f8-e7) 17lt:ld3 a6 a) 16 llc2 was the choice in Taimanov­
(Black does not want ...ii'h8 to be met by lLlbS) Najdorf, from the famous 19S3 Candidates
18 b4lLlf6 19 cS �d7 20 a4 ii'h8 21lt:lc1? (this tournament in Zurich (made famous by Bron­
move is just ridiculous; White had to play 21 bS stein's excellent book on the tournament). After
with an unclear game) 21...ii'h7! (now Black 16...�f8 17 cxd6 cxd6 18 '1Vd2 g4 19 llfcl g3
just triples on the h-file) 22 lt:ld3 llh8. The 20 hxg3 fxg3 21 .llxg3 lLlhS Black had a dan-

23
Pla y t h e King 's Indian

gerous attack on the kingside. lbxg4! 29 fxg4 £3 he is not going anywhere)

b) 16 cxd6 cxd6 17 lbbS is the more direct 27....i.xh4 28 gS ...xgS 29 l:e2 lbg4 30 l:b1
approach but is weU met by 17...g4! Now 18 .i.g3 31 ...d3 �4 0-1. One of Korchnoi's
fxg4lbxe4 is better for Black, 18 .i.xa7 loses to worst defeats in this line. He usually gets at least
..id7, 18lbxa7 to 18...g3 and the rather desper­ a bit of counterplay.
ate 18 g3 fxg3 19 hxg3 gx£3 20 .i.x£3 .i.h3 21 1 6 . . .g4
l:e1 .i.h6 also led to trouble for White in Mi­
ladinovic-Georgiev, Belgrade 1988. Other
moves allow Black his customary attack on the
kingside. Here is a recent example: 18 'itb3 g3
19 hxg3 fxg3 20 .i.xg3 lbhS 21 .i.h2 .i.h6 22
l:c2 a6 23 lbc3 .i.e3t 24 lb£2 lOg£4 2S lbcd1
..id4 26 lbe3 ...gS 27 lbeg4 .i.xg4 28 fxg4
lbxe2+ 29 :Xe2 ...xg4 30 l:e3 .l:a£8 0-1 Saif-
Timoshenko, Dubai 2000.
1 5 ... h5
Another idea is to prepare for ...g4 with
1S...�h8 16 aS l:g8. Korchnoi-Lmka, Linz
1997 now continued 17 cxd6 cxd6 18 lbbS g4
19 fxg4 lbxe4 20 .i.xa7 .i.d7 21 .i.b6 ...e7 22
lbc7 .l:a£8 with an unclear game. Sometimes Black has to work hard to
achieve ...g4 - not this time. Perhaps Korchnoi
just wanted to show that the ...g3 pawn sacrifice
is not so dangerous for White if he just wants
to make a draw.
1 7 c6
Korchnoi is going to accept the pawn on g3.
This is better than 17 cxd6 cxd6 18lbbS g3! 19
ha7lbh7! 20 h3 (or 20 �h1 :Xa7!- we have
already seen this tactic before) 20...�4 and
with ... .i.xh3 to come Black had a decisive ad­
vantage in U!rsen-Torre, Bauang 1973.
1 7 ...g3 1 8 hxg3 fxg3 1 9 �xg3 h4 20 �h2
lt:lh5
It is much worse for Black to exchange on
1 6 a5 c6 as then White gets use of the dS-square. For
In Korchnoi-Kasparov, Amsterdam 1991 example, if Black plays 20...bxc6 21 dxc6 and
White played 16 h3 l:f7 17 c6 and was sur­ then continues as in the game with 2t...lbhS 22
prised by the reply 17...aS!. This game was such f4 lbhxf4 23lbxf4 lbxf4 24 .i.xf4 exf4 White
a bad experience for Korchnoi that never again now has 2S ...dS+.
did he allow c6 to be met by ...aS. Play contin­ 21 cxb7 �xb7 22 f4
ued 18 cxb7 .i.xb7 19 b4 .i.c8! 20 bxaS ..ih6 21 Korchnoi has great experience in this line.
lbb4 (Kasparov considered 21 a6 to be unclear) He knows when it is time to defend on the
21...g4 22 lbc6 (111 the King's Indian the c6- kingside and when it is time to bail out. With­
square sometimes looks a more dangerous out this move he would come under heavy
outpost than it really is; for example, here the pressure and if he doesn't play it now he might
knight attacks nothing) 22......£8 23 fxg4 hxg4 not get the chance again. There now foUows a
24 hxg4 ..igS 2S ..i£3 �6 26 l:e1 lbh4 27 series of exchanges after which we suddenly
..ixh4 (White could have tried to evacuate his find ourselves in an endgame.
king with 27 �ft but after 27...lbxf3 28 gx£3

24
 Th e Classi cal Vari a ti o n : Mo dern Main Line

1) Black's main aim is to play ...g5-g4 as

quickly as possible (though there is another idea
where he plays .. .l:tf6-h6) and, assuming White
doesn't take on g4, then ...g4-g3. As time is of
the essence this advance usually has to be
played as a pawn sacrifice. If White takes the
pawn, and often he has no choice, then Black
will follow up with moves such as ...h5-h4 (if
the h-pawn has moved), ...tbh5-f4 and ...i.h6.
Basically he takes over the dark squares on the
kingside and gets a strong attack. Experienced
White players may try and give the pawn back
with f3-f4 to avoid suffocating on the kingside.
The ...g3 pawn sacrifice usually persuades White
22... iL!hxf4 23 .!LJxf4 iL!xf4 24 .bf4 exf4 25 to forget about his queenside attack and forces
i.g4 •gs 26 .te6+ �h8 27 l:lf3 .tea 28 him to devote his energies to kingside survival.
:h3 .txe6 29 dxe6 l:lae8 30 ••1 l:lxe6 31 2) If White meets ...g3 with h3 then Black
•xh4+ •xh4 32 l:lxh4+ �g8 will be looking to sacrifice his bishop on that
square. If White just ignores the black pawn on
g3 Black will usually try and force him into
playing h3 (e.g. by threatening mate on h2) so
that he can sacrifice his bishop. Usually Black is
in no rush to play the move ... g3xh2 as White
may be able to use Black's pawn as a shield for
his own king. Sometimes an opponent's pawn
can be better protection than one of your own
as a piece cannot be sacrificed to remove it.
3) Black must guard his light-squared bishop
most preciously as without it will be much
harder to conduct a successful attack.
4) If White attacks any other piece on the
queenside then Black's first reaction should be
In the end Xie Jun had to work a little to to see if he can ignore the threat and just get on

hold the draw but at this particular moment with the kingside attack. Many a game has been
Black is very comfortable. The remaining lost by White wasting precious tempi capturing
moves were: 33 l:tg4 c6 34 l:tcl f3 35 gx£3 l:tx£3 a rook on aS.
36 ltk2 l:tff6 37 l:txc6 l:tg6 38 l:txg6 l:txg6+ 39 5) Apart from against 13 a4, where I am
�£2 i.xb2 40 tbf4 i.d4+ 41 �£3 l:tf6 42 �g4 suggesting Black plays 13...a5, Black should
i.e5 43 tbd3 l:t£7 44 a6 i.h2 45 l:tc8+- �g7 46 leave the queenside alone and just concentrate
l:tb8 �f6 47 l:tb7 �e6 48 �h3 i.g1 49 �g3 on his kingside attack. Sometimes it is worth
i.e3 'l2-'l2 playing ... a6 to prevent White playing tbb5, but
Black should take care that this doesn't allow an
Summary even more serious invasion on b6.
This chapter includes some of the sharpest and 6) Black's knight on d7 makes it harder for
most complex chess in the King's Indian. It is White to play c5. Although Black is itching to
very heavily theoretical but there are some gen­ move this piece to f6 he should try and stay
eral rules and strategies that Black can follow to there until White has wasted a move preparing
help him through the maze. c5 (tbd3 or b4 for example).

25
 CHAPTER TWO I
The Classical Variation :
9 ltJe 1 ltJd7 without 1 0 ..te3

1 d4 .!Df6 2 c4 g6 3 .!Dc3 �g7 4 e4 d6 5 an d gives White another possible entry point

.b2 0-0 6 .!Df3 e5 7 0-0 .!Dc6 8 d5 .!D&7 9 into the black camp. Very often Black wiU just
.!D&1 .!Dd7 let a7 go if it means gaining a tempo or rwo for
The old main line of the King's Indian is a the kingside attack (see Game 6). He has to be
mere shadow of its former sel£ Nowadays it is more careful about letting c7 go as this is not
rarely seen and has been surpassed in popularity such a peripheral square. But Black's objective
by all the other lines in the Classical during the is not to hold the queenside for ever, as that
last rwenty years. I think it just became too theo­ would be almost impossible, but to just hold it
retical for its own good and reached a sort of together long enough for his attack on the king­
saruration point where everyone just decided to side to become a potent force.
move on to pastures new. I can remember one As in the previous chapter, Black's main idea
frustrating experience as a young player where I is to break through on the kingside with ...g5-g4
had a club game sent for adjudication after 36 and in the long run White wiU not be able to
moves and we still weren't out of theory! Still, prevent this advance (unless he plays g2-g4
even if the lines are no longer so topical they are himself). Don't forget Black can always play it
of great instructional value as they illustrate weU as a sacrifice the moment things start to go
the attacking and defensive ideas for each side. wrong on the queenside. Also, bear in mind
White's main plan in these lines is to pene­ that the black attack can afford to be a few
trate into the heart of the black position via the moves behind the white attack as his prize is
c7-square. In games 6-9 he starts with tO lbd3 much the greater. A queenside attack may win
in order to force through the advance c4-c5 and material but a kingside attack can deliver mate!
open the c-file for his major pieces. Once cS In Games 9 and 10 we examine a radical
has been achieved the knight is poorly placed idea where White breaks the rules by advancing
on d3 so it usually retreats to f2 for defensive g2-g4 in front of his own king. The thinking
duty and White piles up his forces against c7. behind this is explained below.
Black can usually prevent an invasion here with
the help of moves like ...ltJf6-e8 and ...l:.f8-f7 Game 6
and sometimes the interesting manoeuvre
Miles-Sax
...�f8-e7-d8 or ...�f6-d8 (see Game 7). With
c7 so weU protected White often turns his at­ London 1980
tention to the other vulnerable point in the
black camp, a7. He wiU try and persuade Black 1 d4 .!Df6 2 c4 g6 3 .!Dc3 �g7 4 e4 d6 5
to play ...a7-a6 as this weakens the b6-square �e2 0-0 6 .!Df3 e5 7 0-0 .!Dc6 8 d5 .!De7 9

26
 The Cla ssical Varia tio n : 9 £De 1 lDd 7 with o u t 10 i.e3

ltJe1 �7 1 0 �3
For 10 f3 f5 11 g4 see Game 10.
10 ...f5 1 1 �d2

So the stage is set. Black closes the centre in

anticipation of an attack against the white king.
This is the real starting point of the variation
Unless White is planning to block the king­ and in a tournament game this position would
side with g4 he normally waits for Black to play usually be reached after just a couple of minutes
it:lf6 before playing £3. This gives Black the play.
chance to make the positional mistake 11 ...f4 1 3 c5 g5 1 4 cxd6
which is strongly met by 12 .i.g4!. Also after 11 White has little to gain in delaying this ex­
f3 f4 Black may be able to derive some benefit change on d6. There is a speculative sacrifice,
by leaving his knight on d7 for a while in order 14 .l:l.ct lt:'lg6 15 lt:'lbS a6 (Black should take the
to hold up cS. It probably doesn't make all that piece as 1S....l:l.f7 16 .i.aS is awkward) 16 cxd6
much difference but 11 .i.d2 is considered the axbS 17 dxc7, although the current view is that
more accurate move order. it is doubtful White has enough compensation
Very rarely White plays 11 ex£5. This leads after 17...'iVe8.
to equal play after both ll...gx£5 12 f4lbg6 13 1 4. . .cxd6 1 5 llc1
i.e3 exf4! t40Jxf40Jxf4 15 .i.xf4 a6 (prevent­ 15 lZ'lf2 is equally popular and this is the sub­
ing ...0JbS-d4) 16 'iVd2 1i'f6 17 .l:l.ael lDeS and ject of Game 8.
tt...0Jxf5 12 f3 0Jf6 13 0Jf2 c6 14 0Jfe4 lt:'lxe4 1 5 ...lL\g6 1 6 lLib5
1Slt:'lxe41i'b6+ 16 �h1 .i.d7. White plans to invade on c7. Black must
However, in general in the King's Indian prevent this.
Black meets ex£5 with ...gx£5 in order to keep 1 6 ...llf7
control of the e4-square.
1 1 ...lLif6
In the 1990's it become quite fashionable for
Black to play an early ...�h8, either here or on
the next move. The idea (apart from freeing g8
for the knight on e7) is to adopt a more flexible
set-up, not rushing with ...f4 and reserving the
option of challenging White on the queenside
with c6 or exchanging in the centre with ...fxe4.
Unfortunately we don't have the space to cover
this as weU as the main line and I consider it
more important to look at the main line, if only
to develop your King's Indian culture!
12 f3 f4
You will see this move time and time again

27
Pla y t h e King 's Indian

throughout our coverage of the Classical King's will be open) and create threats against the
Indian. The rook now both defends along the black king. Remember it has no pawn cover
second rank and can move to g7 or h7 where it left. Therefore, I suggest that after 20 h3 Black
can play an important role in the assault on the plays 20.....i.d7!? and after White's most likely
white kingside. reply 21 'ifb3 we have transposed into Roeder­
1 7 1fc2 tbe8 Hebden (the next game).
It may seem like a concession for Black to Black can use the move order of this game if
retreat the knight but that is not the way to look he wants to give White the chance to play 20
at. In fact it's only because this knight on eS tLlxa7. Otherwise just play 19....i.d7.
does such a wonderful job of holding up the 20 .....td7!?
White queenside attack that Black can play in This leads t o a full-blooded struggle where
this fashion at all. It is possible for Black to Black will hope to make White pay the price on
attack at once with 17...g4 but the position after the kingside for his queenside greed.
1S l"Dc7 gxf3 19 gxf3 ..i.h3 20 tfr6! is consid­ The alternative 20...:c7 leads by force to a
ered favourable for White. drawish endgame, e.g. 21 ..i.a5 l:xc2 22 ..i.xdS
1 8 a4 l:xe2 23 lLlxcS :Xa4 24 tLld3 g4 25 l:£2 :C3 26
Note that 1SlDxa7 fails to 1S.. 'ifb6+. White
. tfr1 :as 27 l:fc2 l:b3 2S 'it£2 l:a2 29 l:b1
needs to play a4 so that he can meet 1S...a6 'itt7 30 �e2 ..i.e7 31 tLlxe7 tLlxe7 32 tLld3 l:aS
with 19 tLla3 and not have to worry about Black 33 ..i.xe7 'itxe7 and Novikov-Giek, Lvov 19S5
playing ...b5. soon ended peacefully.
18 ... h5 1 9 ltlf2 .i.f8 21 ltlb5 l:tg7 22 h3 lbh4 23 1fb3 �h8!
19.....i.d7!? is the subject of the next game.
20 ltlxa7

Black's plan is, of course, to play ...g4. He

could have done so at once but he first decides
This little tactical trick should be familiar to to remove his king from the same diagonal as
the King's Indian player. In this position it is no White's queen. This is a very sensible precau­
longer considered to be very dangerous for tion. Why, you may ask, is it so important to get
Black and, as we shall see, he even has two the king off the same diagonal as White's
satisfactory replies. queen? After all, isn't it just blocked by a pawn
The alternative is to play 20 h3 when tradi­ on d5? Yes, true, but when Black commits
tionally Black has played 20...:g7 21 '1Vb3 iLlh4 himself to the kingside attack with ...g4 he will
and tried to force through ...g4 as quickly as need to support this action by returning his
possible. The theoretical verdict here is that knight from eS to f6. This will allow White to
White has slightly the better chances. One of play l"Dc7 and subsequently tfr6. A knight on
the problems is that if Black plays ...g4 too soon e6 is hard for Black to live with and he will
the white major pieces can swing along the probably have to take it. White will recapture
third rank (after lots of exchanges on g4 this with dxe6 and with the black king on gS he

28
 The Classical Varia tion: 9 lbe 1 0Jd 7 with o u t 1 0 � e 3

might be threatening a nasty discovered check traumatic.

by e6-e7. The tempo Black will have to spend 37 . . . .l:l.xg2+ 38 .l:l.xg2 We3+ 39 ..tf2 .l:l.xg2+
to prevent this could cost him the game in such 40 1Vxg2 fxg2 0-1
a sharp situation. The tempo is much better
spent now before the hand to hand fighting Game l
starts and when White has only relatively quiet
Roeder-Hebden
moves at his disposal.
24 a5 Bern 1992
Black would have played ...g4 against virtu­
ally any White move, e.g. 24 l:tc3 g4 25 fxg4 1 d4 lt:lf6 2 c4 g6 3 lt:\c3 ..tg7 4 e4 d6 5
lt:lf6 26 'Wd1 hxg4 27 hxg4 lt:lh5! gave Black a ..te2 0-0 6 lt:lf3 e5 7 0-0 &6 8 d5 lbe7 9
sttong attack in Frias-Wilder, New York 1 984. lt:\e1 lbd7 1 0 lbd3 f5 1 1 ..td2 lt:lf6 1 2 f3 f4
24. . .g4! 25 fxg4 1 3 c5 g5 1 4 cxd6 cxd6 1 5 .l:l.c1 lt:\g6 1 6
Or 25 hxg4 hxg4 26 lLlxg4 lLlf6! with the lt:lb5 .l:l.f7 1 7 'iVc2 lbe8 1 8 e4 h 5 1 9 lt:lf2
usual attacking chances. ..td7!?
25 ...hxg4 26 hxg4 lt:lf6 27 lt:\c7 lt:\xg4 28
.bg4
After 2B lLlxaB lLle3 29 �xe3 l:txg2t 30 �h 1
'i'g5 Black has a decisive attack.
28...�xg4 29 lt:\xg4 .l:l.xg4

The previous game saw all out attack by the

two sides on their respective wings. The text is
a slightly more cautious move. Black still has
every intention of crashing through on the
kingside but this time he is also willing to take
30 .l:l.f2? defensive measures on the queenside if he be­
The key line is 30 �e1 ! t1 when Black is los­ lieves he has time for them.
ing after 30 .. .I:txg2t 31 �h1 but 30 ... t1! 31 l:txt1 20 "ifb3 ..tf8
'i'g5 leads to very unclear play. Perhaps it is a Considering that in the main game Hebden
draw, e.g. 32 l:lg3 'Wxcl 33 l:txg4 'Wxet+ 34 adopted a plan with ... 'ii'dB-bB, ... �e7 and
�h2 l:tcB 35 lLle6 l:tct 36 lLlxfB 'Wh 1+ 37 �g3 ... �dB it is certainly worth considering
'i'e 1+ 3B �h2 'Wh 1+ etc. 20...�f6!?, which has the intention of transfer­
30. . .1Vg5 31 "ifh3 .l:l.g3 32 'Wh1 ring the bishop &om g7 to dB in two moves
After 32 'Wxg3 1i'xg3 33 lLlxa8 t1 34 l:tcB instead of three. Still, it's not totally clear to me
�g7 35 l:tc3 'it'g4! Black should be winning. that this is an improvement as although
32 ... .1:1.c8! 33 ..te1 20 ... �f6 gains a tempo it is less flexible. A typi­
33 lLle6 loses to 33 ...1:txc1+ 34 �xcl 'it'g4! cal continuation now would be 21 l:tc2 'iVbB 22
35 lLlxfB l:txg2t 36 l:txg2 'Wd1+ 37 'ith2 lLlGt. l:tfcl �dB and I refer you to the notes in the
33 . . ...th6 34 a6 bxa6 35 .l:l.c6 .l:l.g8 36 .l:l.xd6 main game for a discussion of such positions.
f3 37 .l:l.xa6 In Neverov-Vokac, Bled Olympiad. 2002
There goes the black queenside. Not too White tried instead 21 �b4 but Black was able

29
Pla y th e King 's Indian

to solve the position tactically: 21....i.e7 22 l:lc2 21 'ifb8!?

...

a6 23lba3 g4! 24. fxg4 hxg4 25 .i.xg4 .i.xg4 26

lbxg4 'Wb6t 27 �h1 aS 28 lLlh6t �g7 29lbxf7
'W'xb4 30 'W'h3 �xf7 31 'W'f5+ �g7 32 g3 f:xg3
(32...li:lf6 is a safer continuation and after 33
gxf4 'W'xe4+ 34 'W'xe4 lLlxe4 35 fxe5 dxe5 36
l:lc7 l:ld8 a draw is the most likely outcome) 33
1i'f7+? (better was 33 .l:lg1! as the text leaves e4
unprotected) 33...¢>h6 34 lbc4 lbg7! (now
White can't avoid the disastrous simplification
that arises when Black plays l:l£8) 35 11'£3 l:l£8
36 1i'd3 l:lxfl+ 37 1i'xfl lbh5 with a decisive
advantage for Black who went on to win the
game.
21 h3
White usually needs to play this move at White's idea is to double rooks on the c-file
some point but it wasn't forced just yet. He and play lbc7. Black can stop this by chasing
could have started by doubling his rooks on the the knight with a6 but he is reluctant to do this
c-file. Normally he does this with 21 l:lc3 or 21 too soon because the manoeuvre lba3-c4-b6
l:lc2, but in the latest top class game to reach may hurt. He can also remove the knight with
this position, Kozul-Radjabov, Sarajevo 2003, .i.xb5 but as you already know Black hates
White opted for the strange 21 l:lc4!?. Play parting with this bishop. There is a much more
continued 21...a6 (2t....l:lg7!?) 22 lba3 (normally imaginative solution to the problem of White
the knight quickly re-emerges on c4 after it is invading on c7. Black has played 'W'b8 to enable
kicked back to a3, but with his rook there Ko­ the manoeuvre .i.e7-d8. After this White can
zul must have other ideas) 22...l:lg7 (of course forget about entering the black position via c7.
Black is getting ready to play ...g4; all he needs In fact it becomes quite difficult for White to
to do is to hold his queenside together long make any son of progress at all on the queen­
enough for his own attack to get out of the side. The black bishop is extremely well placed
starting blocks - then White can have it all) 23 on d8. On the kingside it had no active play at
a5li:lf6 24 'Wb6 1i'e8 25 h3 (25 1i'xb7 .i.b5 26 all but now it is just a short distance away from
'Wb6 is not just the exchange but a queen after the diagonal of its dreams (g1-a7). Once Black
26...l:lb8) 25...g4 26 f:xg4 hxg4 27 hxg4 lLlh4 28 takes control of this diagonal the white king will
l:lc7 1i'g6 29 .i.e1 (with the idea of playing start to feel very uncomfortable. White nor­
.i.xh4 in certain variations; also, it would be mally does everything in his power to prevent
very risky for White to start taking queenside this happening.
pawns here, no matter what your computer A word about the black queen. It may look
programme may say) 29...l:lh7! (Radjabov horribly passive on b8 but it too has visions of
switches the attack to the h-ffie now that the activity on the g1-a7 diagonal. The sequence
move .i.e1 has hemmed in the white king) 30 ...a7-a6 followed by 1i'a7 is not unusual in this
'Wb3! (the queen rushes back to defend along position. The queen is not as effective as the
the third rank as Black wins after 30 1i'xb7 bishop on this diagonal as White may be able to
1i'h6! 31 lbh3 lbxg2) 30...lbxe4 (Radjabov arrange .i.e1 and .i.£2 to kick it off. Still, that
settles for the important central pawn) 31 1i'd3 takes time. Have a look at Game 2 (Piket­
lbg3! 32 l:lxb7 and now instead of 32...e4?, Kasparov) to see how dangerous this idea can
Black should have played 32....i.f5! when 33 be.
l:lxh7 (33 gx£5 lLlxe2+ 34 1i'xe2 1i'xg2 mate) 22 a5
33....i.xd3 (33...¢>xh7) 34 .i.xd3 �xh7 35 White understands Black's plan and starts by
..i.xg6t lbxg6! gives him a much better end- taking control of the b6-square. It is less good
game. for him to double rooks at once. For example,

30
 Th e Classical Varia tio n : 9 liJe 1 liJd 7 with o u t 1 0 �e3

after 22 l:.c2 i..e7 23 l:.fc1 i..d8 24 aS (to stop would be easy to play something like 2S ...tbh4.
.ib6) 24...a6 2S tba3 bS! (Oicayoz-Grivas, The problem is, then what? It is much better to
Mangalia 1 992) White, if he wants to see move reposition the knight to h6 so that Black can
30, dare not capture en passant. But if he can't force through ...g4. It is also interesting that
do this that means his queenside play has come Hebden refrained from 2S...a6 26 tba3 bS. Per­
to an abrupt halt. Black is better as he will even­ haps because in this position White can play 27
tually force through ...g4. Grivas prepared this tbc2 foUowed by tbb4-c6, whereas in the simi­
advance in the same way as Hebden, i.e. ...l:.g7 lar position in the notes to move 22 he was
foUowed by ...tbh8-f7-h6. nowhere near ready to play this manoeuvre.
22 ...:tg7 26lt:'la3 a6!
As there is no immediate need to cover c7 Preventing i..bS -we already know this rule.
Black takes a time out to improve his rook. 27 lt:'lc4 .i.b5 28 �1 lt:'lf7 29 .i.e1 liJh6 30
Remember that ...gS-g4 is still his long-term lt:'lb6 .i.xb6 31 axb6 .i.xe2+ 32 �xe2 1Wd8
ambition. In Krush-Fedorowicz, Somerset 1998 33lt:'ld3 g4
Black did play 22...i..e7 and after 23 tDa3 i..d8?! It wasn't such a bad idea for White to evacu­
24 i..bS White had achieved the positionally ate his king to the centre but Black's kingside
desirable exchange of light-squared bishops. play still gives him the initiative.
The basic rule here is that when White plays 34 hxg4 hxg4 35 .i.f2 lt:'lf7
lba3 Black replies with a6 to prevent this ex­ The knight is redundant on h6 once g4 has
change. White will then try to play tbc4-b6 and been achieved but as we shall see there is plenty
Black should be able to arrange his pieces so of work for it elsewhere.
that he can meet this with i..xb6 and White has 36 l:h1 gl 37 .i.e1 1Wg5 38 .b5
to take back with the pawn. For example after White chooses this square for his bishop as
23 tDa3 a6 24 tbc4 Black plays 24...i..bS! to pin he's planning an exchange sacrifice with l:.c7.
the knight on c4 and next move he plays 38 . . .1Wg6 39 :tac1 liJg5
2S....id8. Black is now threatening to sacrifice a piece
23 :tel?! on e4.
Perhaps not the best square for the rook as 40 :tc4 :th7 41 :txh7 1Wxh7 42 :tc7
in interferes with the defence of the pawn on
aS. Instead 23 l:.c2 i..e7 24 l:.fc1 i..d8 2S tba3
(Georgiev-Sahovic, Lvov 1984) 2S...a6! 26 tbc4
.ibS is fine for Black.
23....b7 24 :ta1 .i.d8 25 :tcc1
Black was threatening 2S...a6 26 tba3 i..xaS.
25 ...lt:'lh8!?

White was probably looking forward to a

dangerous passed pawn on the 7th rank after
Black took the rook but Hebden had prepared
a beautiful riposte ...
42 ...lt:'lxf3! !
Now 43 �xf3 'W'hs mate and 43 gx£3 'iiVh2+
foUowed by ...g2 are out of the question so
If one was playing without a plan then it White takes the queen.

31
Pia y t h e King 's Indian

43 .l:lxh7 liJd4+ 44 �e1 ltJxb3

Black wins a piece as he attacks both the
rook and bishop. The remaining moves were:
45 :Xb7 ltJxa5 46 .l:la7 .l:lb8 47 b4 lbc4 48
b7 �7 49 lbc5 dxc5 50 bxc5 �e7 51 c6
ltJc7 52 �e2 �d6 0-1

GameS
Andruet-Spassov
Sofia 1990

1 d4 ltJf6 2 c4 g6 3 ltJc3 .i.g7 4 e4 d6 5

ltJf3 0-0 6 .i.e2 e5 7 0-0 lbc6 8 d5 ltJe7 9
lOa 1 ltJd7 1 0 ltJd3 f5 1 1 .i.d2 ltJf6 1 2 f3 f4 a) 19.....if8 20 aS. Now White is threatening
1 3 c5 g5 1 4 cxd6 cxd6 1 5 ltJf2 lDa4-b6 so Black must react before it's too late.
In the previous games we saw White playing 20... bS 21 axb6 'it'xb6 looks tempting to get the
1 S l:.ac1 followed by lDbS. This time he is not queen on the dangerous diagonal but after 22
in such a rush to play lDbS but prefers a slower �4 'it'a7 23 ..iaS! it is White who takes con­
set-up with 'it'c2 and l:lfcl. He will then bring trol. Instead Black took to playing 20 ...g4 21
the queen's rook into play by a4 and l:r.a3. If fxg4 hxg4 22 hxg4 and only then 22...bS. How­
possible he will then continue with lDbS and ever, a large number of games established that
l:.ac3. If Black doesn't take action before this 23 axb6 'it'xb6 24 lDa4 'it'a7 2S ..iaS! l:lb8 26
happens he will find himself in serious trouble. gS! was in White's favour. This is why Black is
1 5 ...h5 1 6 h3 lOg& 1 7 "Wc2 .l:lf7 sacrificing a pawn with ...g4 immediately - to
The main drawback to White's system is that steer clear of this position.
the knight on f6 doesn't have to retreat to e8. b) But there is an intriguing new develop­
This means that Black can play a quick ...g4 and ment in this line. In Kekelidze-Baklan, Barurni
he should do so even if it means sacrificing a 2002 Black preferred 19 ... lDh4 20 aS g4 and
pawn. The text is a useful move as the rook is after 21 fxg4 (maybe 21 hxg4 hxg4 22 fxg4 is
more active on the second rank but it is also better) he didn't take back on g4 as players had
possible to play ...g4 at once. For example done unsuccessfully in the past, but sacrificed
1 7 ...g4 1 8 fxg4 hxg4 1 9 hxg4 �8 (Black plans another pawn with 21 ...£3!. Now:
to activate his bishop with .....if6-h4) and now:
a) 20 l:lfct ..if6 21 ..ie1 ..ih4 22 lDbS ..id7
23 1i'b3 ..ig3 24 lDh 1 'it'b6+- 2S ..if2 ..ixf2+ 26
lDxf2 f3 27 gx£3 lDf4 28 �fl lDxc2 0-1 Re­
sende-Van Riemsdijk, San Jose 1 99S.
b) 20 a4 ..if6 21 lDh3 ..ih4 22lbd1 l:lt7 23
lbdf2 l:lh7 (23.. .l:lg7!?) 24 l:la3 lDf6 2S aS ..id7
26 'it'd1 , Moehring-Uhlmann, Halle 1 981, and
now 26 ...�8 is supposed to be unclear.
18 .l:lfc1
Sometimes White plays 1 8 a4 and Black usu­
ally replies 1 8.....if8. However, I don't see why
he can't play 1 8 ...g4 1 9 hxg4 hxg4 20 fxg4 �8
with similar play to the main line.
1 8 ...g4 b1) According to Kostakiev (annotating for
Formerly Black used to play 1 8 ... a6 to pre­ ChessRase) the critical position is reached after
vent lDbS. After 19 a4 most games continued: 22 gxf3 lDh7! 23 'it'd3 lDgS 24 ..ixgS 'W'xgS 2S

32
 Th e Classical Varia tio n : 9 ltJe 1 ljjd 7 with o u t 1 0 i&.e3

l:tab1 .i.d7 26 �h1 which he assesses as slightly

better for White. The one thing that is certain is
that this is not the critical position! How could
it be with that move 25 l:tab1 included. It is,
perhaps, the most pointless move of all time.
But even after White improves on this (the
immediate 25 �h1 or 25 tDa4 are given by
Andrew Martin) his position remains highly
suspicious. His kingside and dark squares are
riddled with weaknesses and any King's Indian
player worth his salt should be happy to play
such a position in return for a couple of weak
backward pawns.
b2) 22 .i.gS!? is considered unclear by
Baklan. Black's plan is the same as in the notes to
b3) The game continued 22 .i.xf3lbxg4!? (a move 17, i.e. to play ....i.f6-h4 and hopefully
fascinating piece sacrifice but 22...�£3+ 23 develop a strong attack against the white king.
gx£3 ljjh7 also deserves consideration) 23 hxg4 21 a4
(maybe White should decline the offer, e.g. The logical way to develop the queen's rook.
after 23 .i.xg4!? hxg4 24 ljjxg4 il.xg4 25 hxg4 On the third rank it will be well placed for both
'ireS 26 i&.gS 1i'xg4 27 .i.xh4 1i'xh4 Black has defence and attack.
good compensation for the pawn but perhaps 21 i&.f6 22 l:la3 i&.h4 23 �d1
...

no more) 23...tLlxf3+ 24 gx£3 1i'h4 when it was White meets the immediate threat of
not easy to see how White can defend his king­ 23....i.xf2+ and plans to use both his knights to
side defend his king.
b31) The game finished: 25 lLlcd1 1i'g3+ 26 23 i&.g3
...

�fl .i.d7 27 .i.e3 l:tx£3 28 l:ta3 .i.b5+- 29 �e1 Preeing the h4-square for the queen.
hxg4 (Black has fantastic compensation for his
piece: in fact the white position is already be­
yond redemption) 30 l:tb3 l:taffi (30...1i'gt+ 31
�d2 .i.h6! wins in more spectacular fashion) 31
l:txbS l:txe3+ 32 ljjxe3 1i'xe3+ 33 �fl g3 34
:l.b3l:txf2+ 35 �g11i'h6 361i'c8+�h7 0-1
b32) 25 ljjh1 l:tx£3 26 .i.e1 l:tft-+l! is a crush­
ing blow.
b33) 25 �g2 .i.xg4! (25...hxg4 26 l:th1) 26
ltJxg4 hxg4 is dangerous according to Martin
and winning for Black according to Fritz, who
tends to be quite trustworthy in such positions.
Baklan examines the possibility of White trying
26 f4 here and gives 26...exf4 27 l:th1 (maybe
27 1i'd3) 1i'g3+ 28 �fl .i.d4 29 i&.ct l:tg7! 30 24 l:lc3
�e1 'iVg2 31 l:tft 1i'h3 32 1i'd3 l:lg3 33 'iVd4 White's play in this game was too ambitious.
'irxfl+ 34 'itd2 l:tg2 35 ljjcd 1 :l.c8! and Black Another game Sosonko-Hellers, Wijk aan Zee
WillS. 1986 also reached this position and there White
So, certainly food for thought. I didn't want preferred the more prudent 24 lLJh3 'ifh4 25
to recommend this as the main line as experi­ lLldf2 lfjf6 26 1i'dt .i.d7 27 aS l:taffi and after
ence is so limited but keep an eye out for future 28 iet the players agreed to a draw. Black
developments (or maybe create them yourself). obviously has good play for the pawn. A well­
19 fxg4 hxg4 20 hxg4 lbe8! timed ...£3 is the way to continue the attack.

33
Pla y the King 's Indian

Perhaps even immediately but it's difficult to tt:lf3 0-0 6 .te2 e5 7 0-0 tt:lc6 8 d5 tt:le7 9
work out the consequences. It is worth pointing /1},e 1 tt:Jd7 1 0 tt:ld3 f5 1 1 .td2 tt:lf6 1 2 f3 f4
out what could have happened if White had 1 3 g4
played 28 ..ib4. Black just replies 28...l:lh7 as 29
..ixd6? runs into 29.....ixf2+- 30 .!Dx£2 1i'h2+! 31
�ft �1+ 32 iDxh1 l:lxh1+ 33 �£2 .!Dxe4+ 34
�t1l:lxd1 35l:lxd1 iDxd6 with a decisive ad­
vantage for Black.
24- .. .td7 25 .tb5?
This move is asking for trouble. White is
planning to invade on c8 but it is too slow. The
bishop was needed for defence, especially
against the advance ...f4-t1 which is now going
to hurt in many variations.
25 ....txb5 26 axb5 Wh4 27 tt:lh3
27l:lc8 �2+ 28 �ft t1! is winning.
27 ...tt:lf6 28 lieS+?
Perhaps 28 gS was a better try. A possible We have seen in the preceding games how
variation is 28...lDg4 29 l:lc8t-l:lxc8 30 'ifxc8 all Black's kingside play revolved around play­
iDfB! and ...t1 is again on the cards. ing the move ...g4. If he didn't play this move
28 ...:Xe8 29 -.xeS+ �g7 30 tt:Jdf2 .txf2+! there was no attack. So White's logic is easy to
31 tt:lxf2 f3! understand. All he needs to do is stop Black
from playing g4 and let the queenside attack
decide the game in his favour. However, whilst
it is true that the move g4 slows down Black's
attack one can't advance a pawn two squares in
front of one's own king without creating weak­
nesses and new targets. The most important
factors in the position for Black now are the h­
file, the possibility to sacrifice on g4 and the
weakness of White's h4-square.
1) The h-file: Black will start by playing gS
and ...hS when White will have to play h3. Black
is now in a position to open the h-file, by play­
ing ... hxg4, at a moment of his choosing. It is
usually disastrous for White to touch any of his
The threats include 32...'ifg3 and 32...lDxe4 pawns there.
33lDxe4 £2+-. White is lost. 2) Sacrifices on g4: Black should always keep
32 •f5 tt:lxe4! 33 -.xf7+ maximum pressure on the g4-square. Sacrifices
33 'ifxe4 loses to 33...'iVg3!. here take two forms. In the first he just keeps
33 ...�xf7 34 tt:lxe4 -.xg4 35 l:l.e7+ tLl.e7 36 taking on g4 and ends up with two pawns and a
tt:lg5+ �g6 37 tt:lxf3 tt:lxd5 38 l:l.xb7 e4 39 strong attack for a piece. In the second scenario
l:l.bB �g7 40 l:l.b7+ � 0-1 Black sacrifices on g4 and when White plays
fxg4 he plays an annoying ...f4-t1.
Game 9 3) The h4-square: the h4-square is a serious
weakness in the white camp especially with the
Fioramonti-Cvitan
white king on g2 and a black knight on g6 (as is
Geneva 1995 almost always the case). White has to keep this
square covered with his bishop but again, at a
1 d4 tt:lf6 2 e4 g6 3 tt:le3 .tg7 4 e4 d6 5 moment of his choosing, Black may be able to

34
 Th e Classical Varia tio n : 9 ttJe 1 tiJd 7 with o u t 1 0 i..e 3

force White to give up his bishop for a knight. rook to the h-file as quickly as possible is an
The sequence ...l2Jh4+-, �xh4, ...gxh4 is not interesting alternative) 16 �e 1 l:thB 17 'it>g2
always in Black's favour as the h-file becomes lL!g6 (another idea is l:th6 and 'iVhB) 1B c5?! (1B
blocked but there are occasions when it is �f2 was better) 1B ...hxg4 19 hxg4 l2Jh5! 20
strong. For example, Black has a bishop on dB l:th 1 (20 gxh5 llxh5! is simply too dangerous
that is ready to take over the g1 -a7 diagonal for White) 20 l:th1 lL!g3! 21 �xg3 fxg3 and the
once White's bishop (usually on f2) has disap­ weak dark squares around the white king give
peared. Black good compensation for the pawn that is
You would be right to conclude from the about to drop off.
above points that White has to be careful. The
problem for him in this variation is that many
of his pieces are tied down. He usually needs a
rook on the h-file to prevent a Black invasion
there, a bishop on f2 to cover h4 and a bishop
on e2 and a queen on d 1 to defend g4. He must
take great care in how he conducts his queen­
side attack as opening the queenside prema­
turely may allow Black to lure some of these
pieces away from their defensive duty.
1 3 . . . g5
There are a couple of alternatives of dubious
value that deserve a quick look.
a) 13 ... fxg3 doesn't look so bad and the posi­
tion after 14 hxg3 is probably just slightly better 1 4...h5 1 5 h3 liJg6 1 6 �f2
for White but in practice Black has scored ter­ In the main game White, rather greedily,
ribly here - just 25% according to my database. went after the black a-pawn. He never even
A typical continuation is 14 ... c6 (14.)iJh5 1 5 managed to get in the important advance c4-c5
�g2) 1 5 �e3 h6 1 6 �g2 g5 1 7 liJ £2 cxd5 1 B which most White players would play either
cxd5liJg6 1 9 llh1 l:t f7 20 a4 with good play for now or on the next move. After 16 c5 l:tf7 we
White, Lutz-Fedorowicz, Porz 19BB. It's very have a couple of interesting examples:
difficult for Black to do anything.
b) 13 ...h5 14 g5! liJh7 1 5 h4 lL!xg5 (this is
forced as otherwise the kingside has been to­
tally sealed up and Black will die a slow death
on the queenside) 1 6 hxg5 lL!fS (taking the
knight leads to perpetual check; 1 6 ...liJxd5 is
less good as although Black gets more pawns
the white pieces become more active) 17 l:t£2
..xgS+ 1B l:tg2 liJg3 19 ltJ£2 'iff6 20 �fl and
White's position is to be preferred but Black is
not without his chances. This is the sort of line
which might not be quite sound at grandmaster
level but worth a punt at club level.
14 �e1
Although this move is not essential just yet, a) 1 7 b4 �f8 1B �g2 �e7 19 �£2 ..f8! (see
White can't get by without covering the h4- main game for explanation) 20 lL!b5 �dB 21 a4
square so he usually plays �e1 straight away. (21 cxd6 cxd6 22 lL!xa7 is no good as after
Instead, Gelfand-Kasparov, Reggio Emilia 1991 22...l:txa7 23 �xa7 b6 the bishop is trapped and
went 14 b4 h5 15 h3 'Ot>f7!? (in the main game White is going to die on the dark squares; in
Black plays differently but this plan of getting a general White is reluctant to exchange on d6 as

35
Pla y th e King 's Indian

it might let Black activate his dark-squared to sacrifice a piece on g4 an d regain it with
bishop) 2t....i.d7 22lbc:3 .i.e7 23tt::lb5 .i.d8 24 ...£3+. Additional points to .....fB are that it
lbc3 a6 (Black decides to play for a win and is brings the queen nearer to the h-file (...�6)
immediately rewarded by an error from White) and it also controls the c5-square. That may not
25 b5? axb5 26 llbt (after 26 axb5 llxal 27 be so important here but it was useful in the
1Vxal the white queen has been dragged away Gyimesi-Cvitan game given in the notes above.
from the defence of £3 so Black can play
27...hxg4 28 hxg4tt::lxg4! 29 fxg4 £3+) 26...bxa4
27 llxb7 a3! 28 c6 a2!? 29 lla7 (after 29tt::lxa2
hxg4 30 hxg4 Black has a choice of tactical
continuations such as 30....i.xg4 31 fxg4 llxa2
or 30...II xa2 31 cxd7 tt::lxe4! when 32 fxe4 fails
to 32...11xe2! 33 1Vxe2? £3+) 29...tt::lh4+l 30 �gl
llxa7 31 .i.xa7 hxg4 32 hxg4 (after 32 cxd7
both 32...gxh3 and 32...g3!? ensure Black of a
strong attack but he could have avoided these
complications by exchanging earlier on g4)
32....i.xg4! 33 fxg4 £3 34 .l:lx£3 (34 .i.x£3tt::lxe4Q
34..tt::lxf3+ 35 .i.x£3 tt::lxe4? (Black takes the
.

wrong pawn. 35...tt::lxd5! would have won eas­

ily) 36 .i.xe4 lift+ 37 ..xft ..xft+ 38 �xft 20 ll)b5 �dB 21 lbxa7?
at..+ 39 �g2 ..xc3 (39.....xa7 comes to the A mistake. 21 c5 is better with an unclear
same; Black's combination has resulted in an game after 21...a6. See the 16th move notes for
endgame with an overwhelming material advan­ examples of similar positions.
tage - queen and pawn against two pieces - but 21 . . .hxg4 22 hxg4 �xg4! 23 a&
unfortunately for him his king and bishop are 23 fxg4 tt::lxe4 is similar but the text at least
eternal prisoners to the pawn chain and it is obtains the c6-square for the knight.
impossible to make any progress) 40 .i.e3 1Vc2+ 23 ... bxa6 24 fxg4 ll)xa4 25 ll)c&
41tt::l£2 1Ve2 42 .i.a7 �fB 43 �g3 ..a6 44 .i.e3 The point of the combination is that the
1Ve2 1/z-1/z, Gyimesi-Cvitan, Chiasso 1994. white bishop must stay defending h4, e.g. 25
b) If White is worried about Black ever re­ .i.gt loses to 25...£3+ 26 .i.xf3tt::lh4+.
capturing with the bishop on d6 then he may 25 ...ll)xf2 26 ll)xf2
exchange straight away. Bareev-Belotti, Aosta White quietly returns the piece. 26 .l:lx£2 e4
1989 continued 17 cxd6 cxd6 18 a4 .i.fB 19 may be even worse
�g2 .i.e7 20 llct 1Vf8! (Black adopts the Cvi­ 26 . . .f3+ 27 �g1 fxe2 28 Wxe2 lBf4 29
tan set-up) 21 aS .i.d8 22 b4 .i.d7 23 .i.£2 We4 1i'h&
tt::lh4+l? 24 .i.xh4 gxh4 25 tt::l£2 ..e8 26 .i.d3
llg7 27 �h2 and now Black should have played
27...a6 and followed up with ...b6 in order to try
and take control of the g1-a7 diagonal. Bareev
considers the position to be better for Black.
1 6 ....1:1f7 1 7 a4 �fB 1 8 �g2
White needs to be in a position to meet
...llh7 with llhl.
18 �e7 19 a5 1i'f8!
.•.

I like these multi-purpose moves. The two

most important points are that it frees d8 for
the black bishop and creates x-ray pressure
down the f-file. For example, if White now
carelessly moved his <:JUeen Black would be able

36
 The Classical Varia tio n : 9 lt:'le 1 li:J d 7 with o u t 1 0 � e 3

It i s surprising that White i s able to put up

any resistance at all but Black needs his queen­ Game 10
side pieces to play their part in the attack as
Pinter-Nunn
well.
30 lla3! l:[h7 31 .:e1 'irh2+ 32 �1 �f6 33 Thessaloniki Ofympiad 1988
.:ee3 .:t8 34 .:g3 �h8 35 llaf3 .:ht7 36 b3
�g7 37 lL'lb4 1 d4 lL'lf6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5
Black can improve his position no further. �e2 0-0 6 lL'lf3 e5 7 0-0 lLlc6 8 d5 lLle7 9
When this occurs it's time to look for the lLle1 lLld7 1 0 f3 f5 1 1 g4
knockout blow.
37 ... lLlg6! 38 l:[h3?!

So another system where White tries to

block up the kingside. I sometimes wonder why
It looks tempting to trap the queen but these b'llYS are playing the Classical. If you don't
Black obviously had something in mind when like your king coming under fire then play the
he allowed this. 38 lL'lbd3 would, perhaps, offer Fianchetto Variation. This particular version of
more resistance although Black should win in White playing g2-g4 was very popular in the
the end. Starting with 38...lt:'lh4 looks like a 1980's but White eventually concluded that
good idea. there was too much tension in the position for
38...1i'xh3+! 39 .:Xh3 g2-g4 to be successful. They then started to
39 lt:'lxh3 l:lxt1+ 40 �e2 (40 �g2 lt:'lh4+ wait for Black to play ...f4 before playing g4 and
wins) 40...l:lxh3 is hopeless for White (I know that is what we have just seen.
he can take the knight but it's still hopeless after 1 1 ...�h8
41 'ii'xg6 l:lh6!). This flexible move was chiefly responsible
39....:xt2+ 40 �e1 ?! for the demise (partial) of 11 g4. The main idea
40 'iti>gt is a slightly better chance. The sim­ is to improve the position of the knight on e7
plest for Black is to play an ending after by playing ...lt:'lg8. This introduces the possibil­
40...lt:'lh4 41 lLld3 (41 l:lxh4 gxh4 offers no ity of playing .....th6 or ...lLlgf6. The knight on
hope) 41...l:lg2+ 42 'ii'xg2 lt:'lxg2 43 �xg2 which d7 may be able to occupy the other good
he should win comfortably enough after the square on cS. Before 1t ...'it>h8 became fashion­
reply 43...e4. able Black would often play 11...lLlf6 but he
40 .:f1 + 41 �d2 l:[8f2+ 42 �c3 .:e1 !
... struggled for equality.
A wonderful move. Black finally breaks the Before moving on, note that 11...f4? would
blockade and the King's Indian bishop has the be a serious mistake as after 12 h4! Black won't
final say. be able to create any play at all on the kingside.
43 .:e3 If he ever plays ... gS then White replies hS and
Or 43 'ii'xel e4 mate. if he plays ...hS White replies gS. The rule in this
43....:Xe3+ 44 'irxe3 e4+ 0-1 line is for Black to only consider playing f5-f4

37
Pla y the King 's Indian

when White has a bishop on e3. 1 2 . . .85

1 2 ltjg2 Black secures cS for his knight, a particularly
This is one of thtee moves that White usu­ good square now that White can no longer play
ally chooses from (others also occur from rime lbd3. With f6 earmarked for the other knight
to rime). The idea behind 12 lDg2 is to play h4 Black hopes to create strong pressure against
so that if Black tries to exchange his 'bad' e4.
bishop by playing ...lbg8 and ... ..i.h6 he just gets 1 3 h4 lt::lc5 1 4 �e3 lt::lg8
kicked back by gS. Once White has played h4
then Black will no longer be able to contem­
plate playing ... f5-f4 as then there will be no
hope of active play on the kingside. As we shall
see Black has a better plan whereby he creates
pressure against the white centre. The alterna­
tives:
a) 12 ..i.e3 lbg8 1 3 'iVd2 f4 (I won a couple
of games with 1 3 ... a6 but ... f4 is probably more
accurate) 14 J.t2 hS! 1 5 h3 (1 5 h4? gS!) 1 5 ... l:lt7
16 �g2 ..i.f6 17 lbd3 l:lh7 1 8 llh1 ..i.h4! with a
good game for Black, Jacimovic-Vukic, Kastel
Stari 1988. Black's position is better than in
Game 3 as he has managed to avoid playing
...gS. This means it is easy for him to exchange 1 5 l:.b1
or activate his dark-squared bishop and also he As 1 5 gS f4 1 6 J.t2 h6 leads to the opening
will be able to use the gS-square for his pieces. of the kingside in Black's favour White rums
for example a queen on gS creates a lot of his attention to the queenside. His long-term
additional pressure against g4. plan will be to try and expel the knight from cS
b) 12 lbd3 lbg8 13 'Oth1 aS (13 ... a6 is a more with a3 and b4. This position has been reached
flexible alternative) 14 llg1 l:lt7 1 5 gS?! (a on a number of occasions and White nearly
strange choice as this pawn gets cut off and always played a different move. Here are a cou­
becomes weak; 1 5 ..i.d2 was better when White ple of examples:
could justify his previous two moves by the fact a) 1 5 l:lct 'il'e7 (Black could also play
that 1 S.....i.h6 is now met by 1 6 gS) 1 5... f4 1 6 1 5 ... J.d7 so as to meet 1 6 a3 with 1 6 ...a4) 16 a3
..i.ft ..i.f8 17 b3 ..i.e7 18 tt'lbS b6 19 ..i.h3 tt'lcS fxg4 1 7 fxg4 :Xft+ 1 8 �xft lbf6 1 9 ..i.f3 ..i.d7
20 ..i.xc8 lbxd3 21 'il'xd3 'il'xc8 22 'iffl h6! 23 20 b4 axb4 21 axb4 tt'la4 with a comfortable
gxh6 gS with good play for Black, Vaganian­ game for Black, Markowski-Dolmatov, Po­
Uhlmann, Niksic 1978. lanica Zdroj 1993.
b) 1 5 'iVd2 b6 (Black must have spent some
time considering 1 5 ...fxg4 16 fxg4 lDf6 as 1 7
..i.xcS is the only way for White to avoid losing
a pawn) 16 exf5 gxf5 17 gS f4 1 8 J.xcS bxcS 19
lbe4 lDe7 20 ..i.d3 lbf5 was the dubious plan
selected by White in Chemuschevich-Rowson,
Bratislava 1993. White has the e4-square and a
stack of weaknesses in his camp which the
black knight on f5 is ideally placed to exploit.
1 5 ... �d7
Black is in no rush to play 1 5...lbf6 because
of the variation 16 exf5 gxf5 17 gS! lbhS 18 f4
when his bishop on c8 is no longer an active
piece.

38
 Th e Classical Varia tio n : 9 li:Je 1 li:J d 7 with o u t 1 0 fl.. e 3

1 6 b 3 b6 1 7 a3?! a4! 1 8 b4 lbb3 some chances to hold this inferior position.

White's inaccurate 17th move has enabled Now he loses by force.
Black to take over the initiative on the queen­
side. White's next move is played to prevent
Black &om playing ...li:Jd4.
1 9 lbb5 ltJf6 20 exf5 gxf5 21 ltJc3
White probably didn't enjoy retreating after
having just played li:JbS but he had to stop
Black &om playing ...e4 followed by ... f4. His
intention when he took on f5 must have been
21 gS but he now saw that 21...f4! 22 gxf6 fl..xf6
23 fl..£2 1:tg8 is a powerful piece sacrifice that is
likely to end with his king getting checkmated.
21 ...e4! 22 g5 ltJh5 23 fxe4
Black is attacking the knight on c3 but White
responds by attacking the black knight on hS.
23 . . .f4! ? 24 il..d 2 ltJxd2 25 'iVxd2 'ife8 26 40 J:lxf1+ 41 ltJxf1 J:lf4 42 'iVe1 �d4+ 43
. . .

�f3 ltJg3 27 J:lfe1 �e5 �g2 'ife5 44 lbg3 �g4!

Black could win the queen with 44 ..fl..f2 but
.

he doesn't want to allow White to get a rook

and a piece for it.
45 b5 �f2 46 'iVxf2 �h3+! 0-1
White resigned because of 47 �gt 'ii'at+l.

Summary
t) Black's main attacking idea is to play ... g5-
g4. He usually prepares this advance slowly,
with the aid of manoeuvres such as ...1:tf7-g7,
but once he feels things getting out of hand on
the queenside then he should not delay and just
play ...g5-g4. It can be used as a kind of panic
button, although hopefully we won't find our­
Black has sacrificed a pawn for total dark selves in that situation.
square control. His blockade of eS reduces 2) Black's main defensive task is to prevent
White to passivity. This ...e4, fxe4, ...f4 motif is White infiltrating on c7, at least until his own
an important theme in the King's Indian and attack is strong enough for this to be a mere
other examples can also be found in the book. irritation. He usually does this with the help of a
28 lbe2 ltJxe4!? 29 �xe4 f3 30 lbef4 fxg2 knight on e8, a rook on f7 and sometimes a
31 ltJxg2 'iVh5 bishop on d8.
In chess in order to keep the advantage one 3) In Games 9 and tO White tries to block
may have to transform it into another sort of up the kingside by playing g2-g4. This certainly
advantage. That is what has happened here. stops Black from breaking through with the
Now it's Black's two powerful bishops on an traditional ...g5-g4 but Black still has some dan­
open board that give him his compensation for gerous attacking ideas based on sacrifices on g4
the pawn. The chance to attack the white king and an invasion on the h-file.
is the bonus that makes his position preferable. 4) The King's Indian bishop can do worse
32 'ifd3 �g4 33 J:le3 'iVf7 34 'ifd2 'iVg7 35 than take up residence on d8 in these lines.
l:.d3 J:lf7 36 J:le1 J:laf8 37 ltJe3 J:lf4 38 ltJg2 From there it performs important defensive
J:4t7 39 ttJe3 �h5 40 :n ? duties as well as just being one small step away
White should have played 40 li:Jg2 again with from the diagonal of its dreams (a7 -g t ).

39
 CHAPTER THREE I
The Classical Variation :
The Bayonet Attack 9 b4

1 d4 lLlf6 2 c4 g6 3 tLlc3 .i.g7 4 e4 d6 5 b4. The old main lines, 9 lt:k1 and 9 ll'ld2, just
lLlf3 0-0 6 .i.e2 e5 7 0-0 tLlc6 8 d5 tLle7 9 disappeared. The results were overwhelmingly
b4 in White's favour. Even Gary Kasparov got his
Over the last eight or nine years the Bayonet fin!,>crs burned and the Bayonet Attack was the
Attack has been one of White's main success main reason for his recent loss of confidence in
stories in the King's Indian. Prior to this 9 b4 the King's Indian (he didn't play it once in his
was generally thought to be a poor relation to world title match with Kramnik, who happens
the main alternatives in this position, 9 lt:k1 to be one of the main advocates of 9 b4). Many
and 9 ll'ld2. It was rarely seen in top class chess. other King's Indian specialists, such as myself,
Strange, you may think, as 9 b4 is the obvious settled for giving up 7.. .li:1c6. A whole new
way for White to force through c4-c5 as quickly system with 7...lLla6 was born (see Chapter 5)
as possible. In this position the pawn strucrure almost solely due to White's successes in the
dictates that White will attack on the queenside. Bayonet Attack.
The reason why the move was not trusted was However, the tide finally appears to be turn­
that it allowed Black to play the active 9 ... lLlhS, ing. The main problem for Black was not the
whilst after moves like 9 lLle1 or 9 ll'ld2 the strength of the Bayonet Attack, but the fact that
black knight, in order to get out of the way of he was playing too ambitiously. The King's
the f-pawn, would have to retreat. Attitudes I ndian attracts players who are looking for a
began to change after White discovered the sharp struggle. Instead of trying to prove equal­
move 10 l:.e1 (in reply to 9 ... lLlh5). The simple ity they were trying to destroy the Bayonet At­
idea is to meet to ...lLlf4 with 1 1 i.ft . The tack. They didn't realise that the best way to
bishop on f1 is very weU place defensively and destroy the Bayonet Attack is to prove e<:juality!.
it turns out, that despite its active appearance, Once equality has been established White play­
that the knight on f4 is quite poorly placed. ers are bound to turn their attention elsewhere.
There is nothing for it to attack and it can even And that is exactly what has happened over the
get in the way of Black's kingside play. For last two or three years. We now see a lot less of
example, the traditional pawn storm with ... f5- the Bayonet Attack and a lot more of the other
f4 is not possible with the knight on f4 and variations in the King's Indian. We are also
Black wiU also have to be constantly on the seeing less of 7...lLla6 as players such as myself
lookout for White playing i.xf4 at a favourable are returning to the hean and soul of the King's
moment. It took quite a while, and cost an Indian, 7...lLlc6.
awful lot of points, before Black players appre­ 9 tLlh5
.. .

ciated this. Meanwhile, everyone was playing 9 Black has one other imponant move, 9 ... a5,

40
 Th e Classical Varia tio n : Th e Ba y o n e t A t tack 9 b 4

where there have also been interesting devel­ ambitious try) 13 ..td3 fxe4 14 lbdxe4 lLlxe4 15
opments so if the lines I am recommending get ..txe4 ..txct 16 l:lxc1 lLlfS 17 'ird2 lbd4 18
into trouble (chess theory is not static), then I lbe2 lLlxe2t 19 llxc2 ..tf5 20 f3 'irf6 21 l:.c4
suggest you tum your attention to 9 ... a5 'lz-1/z Gelfand-Shirov, Wijk aan Zee 1998.
Now the material is split up as follows: b) 1 1 c5 and now:
1 0 lle1 - Games 1 1 - 1 4. b1) 1 1 ...lLlf6 12 ..tg5 lLlxe4 (12 ... h6 was
10 g3 - Games 15 and 1 6. played by Topalov) 13 lLlxe4 fxe4 14 lbd2 h6
10 c5 (and others) - Game 17. 1 5 ..txe7 1i'xe7 16 lLlxe4 ..tf5 was about level
in Zontakh-Relange, Serbia 1 998.
Game l ! b2) 1 1 ...fxe4 12 lLlxe4 lLlf4 1 3 ..txf4 llxf4 14
lLlfd2 dxc5 1 5 ..tc4 lLlxd5 16 lbb3 c6 17
Malakhatko-Jenni
lbbxc5 �h8, Kramnik-Gelfand, Novgorod
Istanbul Ofympiad 2000 1 996. White has compensation for the pawn
but no more.
1 d4 lLlf6 2 c4 g6 3 lLlcl ..ig7 4 e4 d6 5 1 1 ...lLlf6!
lLlfl 0-0 6 ..ie2 e5 7 0-0 lLlc6 8 d5 lLle7 9 White was threatening 12 ..txh5 so the
b4 lLlh5 1 0 lle1 f5 knight has to move and this is the right S<:JUare.
There is no need for Black to put his knight There is a battle raging for control of the centre
offside on f4. Instead, he looks at the move and the knight clearly exens most influence on
lLlh5 in the same way he would look at lLle8 or the centre from f6.
lbd7 - as a means of clearing the way for The variation 1 t ...lLlf4 12 ..txf4 cxf4 13 l:.ct
Black's traditional King's Indian move f5. I t is has been tested extensively in practice and ap­
still better to play 9 ...lLlh5 (as opposed to pears to be slightly in White's favour.
9 ...lbe8, for example) as this induced White 1 2 ..ifl
into wasting a tempo with lie 1 . White defends his centre with pieces and
1 1 lLlg5 envisages the bishop on f3 becoming active
when the long diagonal opens. He can also
defend his centre with a pawn move. 12 f3 is
the subject of Games 3 and 4.
1 2 ...c6
The tension in the centre increases dramati­
cally after this move. Playing ...c7-c6 also gains
space for Black on the <:jueenside.
1 3 .tel

This is the critical test and the move that,

originally, put many Black players off playing
10 ... f5. The knight is heading for e6. When it
arrives there Black will have to give up his im­
portant light-s<:juared bishop for it. In return he
will have every chance to win the rather ran­
dom white pawn that will appear on e6. l ..ess
critical alternarives arc:
a) 1 1 lbd2 lLlf6 12 c5 ..th6 (this leads to Note how White doesn't rush with lbg5-e6.
simplification while 12...�h8 13 f3 f4 is a more He's waiting for Black to spend a tempo and

41
Pla y t h e King 's Indian

weaken his kingside with ...h7-h6 before carry­ he is offering his b-pawn in the process, this is a
ing out his plan. 1 3 .i.e3 is Krarnnik's move. It rather dubious idea. The e6, d7 duo may look
is an attempt to lure Black into playing 1 3... f4 visually impressive but in reality the bishop is
after which White would calmly rerum home totally out of play on d7. This idea would only
with 1 4 .i.ct . The move ... f5-f4 takes all the be dangerous if White could transfer his other
pressure off the white centre and White is quite bishop to the a3-f8 diagonal but this is very
happy to spend two moves to achieve this. difficult to achieve. In fact the white pieces
Practice has demonstrated that Black should sruck in the black camp resemble prisoners of
now instigate a forced series of exchanges to war more than anything else. It is so easy for
clarify the position. Black to step around them. They are hardly
Alternatives to 13 ..te3 are considered in the even an inconvenience. All Black has to do is
next game. keep his knight on e7 to maintain the blockade.
1 3 cxd5 1 4 cxd5 h6 1 5 lbe6 ..i..xe6 1 6
•.. In fact this line has been at the forefront of
dxe6 fxe4 1 7 lL!xe4 lL!xe4 1 8 ..i..xe4 d 5 1 9 King's Indian theory and White usually chooses
..i..c2 20 'ilfg4, a far more logical move which defends
19 ..tcS dxe4 20 'ilfxdB l:.fxdB 21 .i.xe7 leads the e6 pawn and attacks g6. Initially Black was
to the same ending we see in the 1 3th move playing 20 ...l:.f6?! but his pieces don't co­
notes in Game 1 2 Qine 'b'), the only difference ordinate so well after that move. The strongest
being that the white pawn is on b4 instead of move is 20...e4! in order to take control of the
bS. dark squares (pushing the other central pawn,
1 9 b6
.•. 20... d4, would be a positional disaster as it con­
cedes control of the central light squares). Black
is perfectly happy to exchange dark-squared
bishops as White is then left with his passive
bishop. Several games have continued 21 l:.adl
..c7 22 ..tb3 l:.fS!.

Black must stop White from playing .i.cS.

The position has settled down and is very
difficult to assess. White has two active bishops,
a safer king and a passed pawn on e6, which
although likely to die, may seriously annoy
Black in the process. In Black's favour is his This is an excellent move which defends the
strong centre and the fact that he is likely to pawn on dS, prepares to double rooks on the f­
emerge a pawn ahead. He must tread very care­ file and cuts communications between the
fully, though, as it will be easy to have a tactical queen and the pawn on e6. What more could
accident as White's rooks and bishops pile pres­ you ask? Practice has shown that White can
sure on the centre. only hope for equality in this position. A couple
20 ..i..a4 ?! of examples:
White comes up with the plan of parking his a) 23 l:.d2 1i'c3 (Black should avoid 23 ... .i.c3
bishop on d7 in order to ensure the survival of 24 l:.ct but Ponomariov has played 23 ... l:.af8)
his e6 pawn. Even if we leave aside the fact that 24 ..dt l:.dB 25 bS ..cB 26 .i.d4 ..txd4 27

42
 Th e Cla ssical Varia tio n : Th e Ba y o n e t A t ta c k 9 b 4

l:r.xd4 'i6'c5 (note how Black can't play �e2 0-0 6 ll:\f3 e5 7 0-0 ll:\c6 8 d5 lL\e7 9
27 ... 'i6'xe6 because of 28 l:r.dxe4!; in this line b4 ll:\h5 1 0 l:te1 f5 1 1 ll:\g5 ll:\f6 1 2 �f3 c6
Black should always make sure, before captur­ 1 3 �b2
ing on e6, that White doesn't have such a tactic) This hasn't been played too many rimes but
28 l:r.e2 l:r.dfB 29 a4 l:r.eS 30 h3 1/z-1/z Xu Jun-Ye may become more common after this game.
Jiangchuan, Shanghai 2001 . We shall see. White's other tries:
b) 23 ..td4 .ixd4 24 l:r.xd4 'i6'c3 (24 ...'i6'e5!?) a) 13 'i6'b3 enjoyed a brief spell of popularity
25 'i6'd1 l:r.afB 26 l:r.e2 'i6'c7 27 g3 'i6'c6 28 'i6'c2 last year but Black seems to be able to deal with
'lz- 'lz Iskusnyh-Motylev, Russian Cup 1 999. As it relatively easily. He should avoid 1 3 ... cxd5 14
in line 'a' Black cannot take on e6. I f he had exdS! and whilst 1 3...'it>h8 is quite playable the
wanted to play on then 28...l:r.c8 suggests itself. best line seems to be 1 3 ... h6 1 4 tLle6 .ixe6 1 5
20 . . .'iFd6 21 �d7 'i1Vxb4 22 l:tb1 'i1Vh4! dxe6 'i6'c8 and now:
Black plans a kingside attack.
23 f3 l:tf5!

a1) The position after 16 cS fxe4 1 7 cxd6

exf3 1 8 dxe7 l:r.e8 1 9 ..tb2 l:r.xe7 20 :XeS 1ic7
The black queen has a perfect square on f6 21 l:r.ee 1 was reached three rimes in 2003.
but if he is going to triple on the f-file he needs White has managed to keep his powerful pawn
one rook in front of the queen. The rook on f5 on e6 but at the cost of a seriously compro­
also does a fine job protecting Black's e-pawn. mised kingside. I doubt we will be seeing much
24 �f2 'i1Vf6 25 �g3 h5 26 h3 l:tf8 27 �h1 more of this &om White. One example is
'iFg5 28 �h2 e4! 29 fxe4 l:tf2 30 l:tg1 dxe4 Pelletier-Inarkiev, Istanbul 2003 which contin­
With White tied down to defending g2 and a ued 21 ...fxg2 22 h3 l:r.fB 23 l:r.e3 tLlhS 24 l:r.ae1
ridiculous bishop on d7 the passed e-pawn is 1i'f4 25 ttJd1 .ixb2 26 1i'xb2 'i6'f6 with advan­
going to decide the game in Black's favour. tage to Black.
31 l:tb5 l:t8f5 32 'iFa4 e3 33 h4 'i1Vf6 34 a2) 1 6 l:r.d1 :dB (and not 1 6...1i'xe6? 1 7
'iFe4 e2 35 l:tb3 'iFd4 36 l:te3 'i1Vxe4 37 :Xd6Q and now:
l:txe4 �c3 38 �d6 �8 39 �h2 l:td5 0-1 a21) 1 7 cS is similar to the above except
White lost on time but his position is hope­ even worse: 1 7 ... fxe4 1 8 cxd6 exf3 1 9 dxe7
less. For example, 40 ..tg3 e1'i6' 41 l:r.gxe1 ..txe1 l:r.xd1+ 20 'i6'xd 1 1i'xe6 21 'ii'dS+- Wh7 22 gxf3
42 l:r.xe1 :Xa2 43 l:r.fl+ tiJfS 44 ..tc6 l:r.d4. (22 1i'xa8 1i'g4-+) 22 ...1i'g8 23 1i'c7 1i'c4 24
..td2 l:r.eB and Black has a decisive advantage,
Game 12 Kallio-Kotronias, Batumi 2002.
a22) In Bacrot-Radjabov, Bled (Olympiad)
Shirov-Radjabov
2002 the young French star preferred 1 7 bS. I
Unares 2004 must say that I have some difficulty in accept­
ing that White has enough play for the pawn
1 d4 ll:\f6 2 c4 g6 3 ll:\c3 �g7 4 e4 d6 5 after 1 7...1i'xe6 but I suppose Bacrot may think

43
Pia y t h e King 's Indian

differendy. At any rate he quickly reached a �e8 2 8 e 7 1/2-11> Fressinet-Hebden, Lausanne

promising position after 18 i.a3 �h8 1 9 bxc6 2001. This looks riskier for Black than 'b1' but
ll'lxc6 20 exf5 gxf5 21 i.d5! so I suggest that Hebden assures me that Black has no problems
Black play 1 9...bxc6. at all and that this is an even easier route to
N.B. The very latest word: 17 b5 1i'xe6 1 8 equality than 21...l:.e8.
bxc6 ll'lxc6 19 exf5 1i'xf5 20 1i'xb7 ll'ld4 21 It was with regard to lines such as this that I
i.dS+ �h7 22 i.e3 l:.ab8 23 "W'xa7 l:.b2 24 talked about Black been too ambitious. Many
�h1 ll'lg4 25 ll'le4 l:.fB 26 h3 ll'lxe3 27 fxe3 players avoided this line just because White
ll'lt3 28 ll'lf6+ l:.xf6 29 i.xf3 e4 30 l:.d5 1i'e6 31 could force such an ending. I don't see the
i.g4 1i'e8 32 h4 l:.ff2 33 i.h3 1i'd8 34 h5 l:.xg2 problem. If it's a strong opponent then a draw
35 hxg6+ �h8 36 1i'a5 l:.h2+ 37 �g1 1i'e8 38 with Black is fine and if they are weak just out­
i.f5 l:.bg2+ 39 �ft i.xa1 40 "W'a7 i.g7 41 play them in the endgame.
l:.xd6 l:.xa2 0-1 Bacrot-Radjabov, HOE Wch 1 3 ...h6
KO, Tripoli (rapid) 2004. Keep an eye out for It is probably best not to give White the op­
future developments. tion of 13...cxd5 14 exd5!? here.
b) 1 3 b5 cxd5 14 cxd5 (1 4 exd5 is no longer 1 4 �6 -'.xe6 1 5 dxe6 fxe4 1 6 tt'lxe4
dangerous) 14...h6 15 ll'le6 i.xe6 16 dxe6 fxe4
(16...l:.c8!? is a more ambitious move which has
been played by Ponomariov) 17 ll'lxe4 ll'lxe4 18
i.xe4 d5 19 i.a3 (this is why White played b5)
19...dxe4 20 1i'xd8 l:.fxd8 (don't take with the
other rook as it will cost you an exchange) 21
i.xe7 and now:

16 i.xe4?! is not very convincing. Avrukh

now suggests 16...1i'b6!? with the trap 17
1i'xd6? 1i'xt2+! 1 8 �f2 ll'lxe4+ in mind but
16...d5 is the most logical reaction. Bareev­
Radjabov, Enghien-les-Bains 2003 continued
17 i.d3!? e4 18 i.ft 'iVb6 19 l:.b1 ll'lh5 and
now with 20 "W'd2 failing to 20...l:.xf2! 21 1i'xf2
b1) 2t...l:.e8 22 i.c5 l:.xe6 23 l:.xe4 a6! 24 i.d4 and 20 l:.e2 running into 20...ll'lf4, Bareev
b6 l:.c8 25 i.e3 i.f6 26 l:.d1 l:.d8 27 l:.b1 l:.c6 felt compelled to play the positionally repulsive
28 g4 h5 29 �g2 �f7 30 �f3 �e6 = Babula­ 20 c5. After 20..."W'c7 Black has a clear advan­
Degraeve, Istanbul Olympiad. 2000 ('/2-1!>, 37). tage although he eventually lost the game.
There have been quite a few other games which If White insists on playing this way then he
have been drawn in similar fashion. Some should at least try 17 cxd5 cxd5 18 i.c2 when
White players claim an edge in this ending be­ 18...e4 is the best move.
cause the pawn on e5 restricts the black bishop. 1 6 ...tt'lxe4 1 7 llxe4
This is true but the black rooks are active and This is played with the intention of sacrific­
his king closer to the centre. If there is an edge ing the exchange. Instead, 17 i.xe4 d5 18 cxd5
it is infinitesimal. cxd5 19 i.f3 (19 i.c2 'iVb6 takes advantage of
b2) 2t ...l:.d5!? 22 l:.ad1 l:.xb5 23 i.d8 l:.c8 the fact that the bishop is on b2 rather than e3)
24 l:.d6 l:.c6 25 l:.xc6 bxc6 26 h4 �fB 27 l:.ct 19 ...1i'd6 20 1i'e2 e4 21 i.xg7 �xg7 22 i.g4

44
 The Classical Varia tio n : Th e Ba y o n e t A t ta c k 9 b 4

l:lf4 23 g3 l:lf6 24 l:lad1 (threatening to take on 1 7th move alternative cenainly looks safer.
e4) 24...'itb6 leads to a roughly balanced game.
White has retained his passed pawn on e6 but
his bishop is reduced to a passive defensive role
and Black has a strong pawn centre.
1 7 ...d5
On the available evidence, 1 7...li::lf5 looks
less risky. Dautov-Kindermann, Nussloch 1996
continued 18 bS (White must soften up the
black queenside otherwise he might lose his e6
pawn for nothing) 1 8...l:lc8 1 9 l:le2 lieS 20
bxc6 bxc6 21 cS (the only real alternative avail­
able to White was to try 'ird1 -a4 on one of the
last few moves) 21 ...d5 22 �xeS �xeS 23 l:lxeS
'irf6 24 l:le1 '1>-'h. The game seems about equal
but of course Shirov may have a thing or two to 23 ... l:tf6! 24 l:te1 'ifxb4 25 a3
say in the future about this line. Black's fine defensive idea was to meet 25
1 8 cxd5 cxd5 1 9 l:txe5! hS with 25 ...:Xf3! 26 gx£3 l:tg8 with good play.
White must sacrifice an exchange to avoid a So Shirov drives back the queen to prevent this
worse position, but obviously he has planned line. There is a lot going on just beneath the
this all along. surface in top level chess.
1 9 ....be5 20 .be5 1Wb6 21 ..ib2! 25 ...'ifd6! 26 h5! l:taf8 27 'ife4 lL!c6
The first new move of the game, although The d-pawn is more imponant than the g­
Shirov and his trainer Lanka had already looked pawn. The black king was hoping to take refuge
at this position back in 1 997. In previous games behind the white pawn on g6 but Shirov has a
White had played 21 'ird2 but after 2 t ...'irxe6 fine sacrifice to prise him into the open.
(Radjabov had had this position before as well 28 hxg6+ �g7 29 ..ic1 ! 'ife71
but played the less accurate 21 ...�h7) 22 l:le1,
Black can equalise by returning the exchange,
e.g. 22...l:lxf3! 23 gxf3 li::l f5 24 �g3 'ir£7 25 l:lct
l:te8 26 l:lc7 l:le7 27 l:txe7 lt::lxe7 28 'irxh6
'iVx£3 29 �eS 'irdt+ '1>-'12 Kallai-Barbero, Bern
1 997.
21 ...�h7!?
Radjabov is not tempted by either of the
pawns that are on offer. Both 2t ...'iVxb4 22
l:tb1 ! and 2t...'iVxe6 22 'iVd4 <j;£7 23 'iVg7+ �e8
24 'irxh6 are better for White.
22 'ife2
The best square for the queen. White threat­
ens 'iVeS and avoids the type of exchange sacri­
fice we saw in the previous note. Black now, It's best to allow the sacrifice. With the e­
sensibly, closes the long diagonal. pawn blockaded Black is now threatening to
22 ...d4 23 h4! take on g6.
White switches his attack to the light 30 ..ixh6+! �xh6 31 'ifh4+ �xg6 32 ..ixc61
squares. The idea is obviously to weaken the Now the white rook gets to join in the fun.
black king position with h4-h5. 32 ... bxc6 33 lte5! 'ifxe6?
To be honest, even if Black seems just about According to Shirov this is the only real mis­
OK, this is not the son of position I would like take of the game. Instead 33 ...:Xe6! 34 'irhS+
to play. Those bishops frighten me and the 'l;g7 35. l:tgS+ 'irxgS 37 'irxgS+ l:tg6 38 'iVxd4

45
Pla y th e King 's Indian

l:.t7! 39 g3 l:.fg7! leads to a position where ble-edged llk6 that White is more or less
White can make no progress. committed to playing and that it weakens the
All I can say is that if Black is ei.Jual in this dark S(jUares around the white king.
line it is a lucky e�.juality. 1 2 ...llJh5!?
34 llxe6 llxe6 35 11'g4+! �7 36 1fxd4

How can we explain such a move? Black

Now White is winning as Black cannot set plays 9.)t�h5, 1 1 ...lDf6 and now 1 2 ... lDhS
up a fortress as in the previous note. The re­ again. I must admit that I didn't take this line
maining moves were: very seriously at first but it has grown on me.
36 . . .a6 37 g4 llgB 38 fl llf6 39 �2 :ea The first point is that Black only played lDf6
40 1fc4+ �g7 41 1fxa6 llefB 42 1fd3 c5 last move because he was forced to move the
43 a4 liaS 44 1fc3 �g6 45 1fxc5 llfa6 46 knight (White was threatening .i.xhS). With the
�g3 llxa4 47 1l'd6+ �7 48 g5 ll8a6 49 hS-s�.juare available again it is cenainly an op­
1fd7+ �g6 50 f4 lla1 51 1fd3+ �g7 52 tion to return there. From hS the knight has
"1l'd4+ �gB 53 �g4 ll 1 a2 54 1fd8+ �g7 55 many interesting options, as you will discover
11'c7+ �gB 56 f5 lla7 57 1fd8+ �g7 58 from the games and notes that follow. The
f6+ �h7 59 1fd3+ �hB eo �5 :as 61 reason that Black can get away with moving his
'lfhl+ �g8 62 �g6 ll2a7 63 1fe6+ �8 64 knight so many rimes is that White's last moves
1i'd6+ �gB 65 1fd5+ �hB 66 11'h1 + 1 -0 haven't really improved his position b4, l:.e1,
-

lDgS and f3 are just a collection of uncoordi­

Game 13 nated moves.
Before moving on a little word about the
Zhu Chen-Y e Jiangchuan
history of 12 £3. White players gave this varia­
Three Arrows Cup, Ji Nan 2003 tion up a few years ago because of the line
1 2...c6 1 3 .i.e3 .i.h6 as they couldn't find a
1 llJf3 llJf6 2 c4 g6 3 llJc3 �g7 4 e4 d6 5 satisfactory way to deal with the threat of .. .f4.
d4 0-0 6 �e2 e5 7 0-0 llJc6 8 d5 llJe7 9 b4 Due to 1 2 .i.f3's lack of success analysts re­
llJh5 10 lle1 f5 1 1 llJg5 llJf6 1 2 f3 turned to this position and discovered the
In the previous games White defended his threatened ... f5-f4 is not so serious after all. The
centre by 1 2 .i.£3 but this ultimately led to its position after 14 h4 cxdS 1 5 cxdS f4 1 6 .i.f2
destruction. Although White's active pieces .i.xgS 1 7 hxgS lDhS 1 8 .:.ct may actually be in
gave him some compensation it was insufficient White's favour (the knight on e7 is badly
to claim any advantage. This rime White prefers placed). It was after a couple of defeats here
to defend his centre more solidly with a pawn. that Black went back to the drawing board and
There are, however, cenain drawbacks to 1 2 £3. discovered 1 2 ... lDhS. Modem theory is in a
The principal ones are that this solid pawn permanent state of flux. Moves that get a '?'
move doesn't combine very well with the dou- one day are rewarded with an '!' the next.

46
 The Classical Varia tio n : The Ba y o n e t A t ta c k 9 b 4

In the next game we shall take a look at an 1i'xe2 l:.xc3! 0-1 Quinn-Shirov, European
alternative for Black, 12...�h8. Team (Leon) 2001.
1 3 c5 b) 13 �h1 f4! (this would have been a bit
White presses on with his queenside play slow after c5 but �h1 puts the black position
and frees c4 for the bishop. There are a couple under less pressure) 14 lLle6 .i.xe6 15 dxe6 .if6
of alternatives worth looking at: 16 c5 .i.h4 (White's king can easily get into
a) If White considers ...lLlh5-f4 to be a threat trouble here as ...lLlg3+ check is in the air; that is
then he can prevent it with 13 g3, although his why White decided to sacrifice material rather
results with this move have not been remark­ than play a move such as 17 l:.£1) 17 .i.c4 dxc5
able. The best reply is 13....i.f6!. Black wants to 18 lLld5 (18 bxc5 11i'd4 is good for Black) 18...c6
force lLlg5-e6 as it is difficult to live with this 19 bxc5 cxd5 20 exd5 'it'a5! 21 .i.b2 1i'xc5 22
permanently in the air, and this is a much better .i.b3 .i.xe1 23 1i'xe1 l:.fS! 24 l:.ct lLlg3+ 25
way than 13...h6 as the g7-square is now avail­ 11i'xg3 (25 hxg3 l:.h5 mate) 25...'it'xcl+ 26 .i.xcl
able for the knight. 14 lLle6 (White may try and fxg3 and Black's huge material advantage easily
improve on this in the future; 14 �g2 f4 15 defeated the passed pawns in Sargissian-Baklan,
iLlh3 doesn't inspire so he may have to try 14 European Ch., Ohrid 2001.
exfS) 14....i.xe6 15 dxe6 f4. This keeps the 1 3 . . .ll:lf4 1 4 ..tc4
centre closed and seriously reduces the threat Or:
from the white bishops. We now have: a) 14 'it'b3!? fxe4 15 fxe4 �h8 1 6 .ixf4 exf4
17 l:.ad1 iLlc6!? 18 iLlf7+ l:.xf7 19 dxc6 l:.tB was
rather unclear in Pelletier-Smirin, Biel 2002;
b) 14 .ixf4 exf4 15 l:.ct .i.f6 16 iLJe6 .ixe6
17 dxe6. If on move 11 Black had played
1t ...iLlf4 instead of 1t...iLlf6 then a similar posi­
tion would have been reached. The only differ­
ence is that here White has two extra moves, £3
and c5, and they don't improve his position. ln
fact, they harm it enough to change the assess­
ment from better for White to fine for Black. A
possible continuation: 17...dxc5 18 bxc5 .id4t
19 �h1 .i.e3 with an unclear game.
1 4 �h8 1 5 l:tb1 !?
. ..

a1) 16 �g2 c6 17 'ifb3? fxg3 18 hxg3 lLJxg3!

19 �xg3 .ih4+! 20 �g2 .i.xe1 21 .i.g5 .i.xc3
22 11i'xc3 lLld5! 0-1 Berkvens-T.Paehtz, Bun­
desliga 2002. Short and sweet.
a2) 16 g4 lLlg7 17 'W'b3? (now when a black
knight lands on d4 it is going to be with tempo:
better is 17 c5 lLlc6 18 cxd6 cxd6 19 .ic4 �h8
when Black will follow up with ...lLJd4 and
...N (g)xe6; White has some compensation for
the pawn but not quite enough as his king posi­
tion is compromised) 17...lLlc6 18 c5 1i'h8 19
.ib2 lLld4 20 'W'd5 lLlgxe6 21 cxd6 c6! (now
White doesn't even get control over d5) 22
11i'c4 11i'xd6 (Black is a pawn up with a good This new move was introduced in Bareev­
position but White accelerates the end by open­ Radjabov, Wijk aan Zee 2003. When asked in
ing the c-flle and running into a little tactic) 23 the press conference after the game if 15 l:.b1
b5 l:.ac8 24 bxc6 l:.xc6 25 1i'd3 lLlxe2+- 26 was the result of home preparation Bareev gave

47
Pla y th e King 's Indian

the enigmatic answer, 'Sometimes you find an Black may try 1 5....!Dexd5 16 lDxdS 'flxgS but
idea at the board, sometimes you find it at this is less good because of 17 .i.xf4 exf4 18
home'. Judging by his performance in Bareev­ tt:lxc7. By exchanging pawns first on e4 Ye has
Baklan, HOE World Championship (Moscow) ruled this variation out as he can now meet 1 8
2001 (1 5 g3 h6 1/2-1/2) I think you could safely .i.xf4 exf4 1 9 ltlxc7 with 19... £3!.
bet the house on him having found this one at 18 ...exf4 19 cxd6 �g4
home. Black invites an endgame but 19 ... .i.h3 was
Instead, 1 S lbe6 .i.xe6 1 6 dxe6 fxe4 1 7 fxe4 also possible.
lbc:6! was reached a couple of times by the 20 •d5 ••d5 21 �xd5 cxd6 22 .bb7
young Ukrainian talent Efimenko. l:lab8 23 �d5 �d4+ 24 �1 �d7
a) In Tukmakov-Efimenko, Lausanne 2001
he was just a pawn up for nothing after 18 .i.e3
lfri4 19 l:tct dxcS 20 bxcS tt:lfxe6.
b) In Berkvens-Efimenko, Hengelo 2001 he
emerged the exchange ahead after 1 8 cxd6 cxd6
19 tt:lb5 'fle7 20 ...xd6 ...gS 21 ...d2 a6 22 lbc:7
lfri4 23 �h1 l:tac8 24 e7 'flxe7 2S lfr!S tt:lxd5
26 .i.xdS lbc:2. The reason this position is good
for Black is that his knights are so active.
1 5 ...fxe4!?
This exchange looks more to the point than
the mysterious 1 5 ... a6 that was played in
Bareev-Radjabov. I suspect the main motiva­
tion behind 1 5 ... a6 was too simply play a move
that Bareev wouldn't have considered too Black has a slight initiative in the endgame
deeply in his preparation. For example, the although a draw is by far the most likely result.
most obvious-looking line for Black, 1 5 ... h6 1 6 The remaining moves were:
lbe6 .i.xe6 1 7 dxe6 fxe4 1 8 fxe4 ltlc6, would 25 :d1 �b5+ 26 �e1 �e5 27 �b2 f3 28
surely have been poured over in minute detail �xe5+ dxe5 29 �2 fxg2+ 30 �xg2 :t4
by Bareev and &om a practical point of view 31 h3 :bf8 32 :d2 g5 33 :b3 g4 34 hxg4
must, therefore be avoided at all costs. :xg4+ 35 :g3 :gf4 36 :g5 �d7 37 :xe5
Still, after 1 5 ... a6 1 6 �h1 Radjabov could :g4+ 38 �h2 :h4+ 39 �g3 :g4+ 40 �h2
find nothing better than 16 ... h6 17 lbe6 tt:lxe6!? :h4+ 41 �g1 �h3 42 :g5 h6 43 :t2 :xf2
(in similar positions Black has usually taken 44 :g8+ �h7 45 �xf2 :Xe4 46 �xe4+
with the bishop in order to preserve his active �xg8 Yz - Yz
knight on f4, e.g. 1 7....i.xe6 1 8 dxe6 fxe4 1 9
fxe4 lbc:6) 18 dxe6 lbc: 6 (the pawn on e6 has Game 14
considerable nuisance value but if White loses
Ponomariov-Radjabov
the initiative Black will easily win it) 1 9 bS (19
exf5! is best according to Bareev) 19...lfri4 Wijk aan Zee 2003
(after 1 9...axb5 20 tt:lxbS Black can no longer
play 20...lfrl4 because of 21 tt:lxd4 exd4 22 1 lZ:lf3 lZ:lf6 2 c4 g6 3 lZ:lc3 �g7 4 e4 d6 5
exf5) 20 bxa6 bxa6 21 .i.a3 l:te8 22 .i.dS! with a d4 0-0 6 �e2 e5 7 0-0 lZ:lc6 8 d5 lZ:le7 9 b4
difficult game for Black, 1 -0 in 32. lZ:lh5 1 0 :e1 f5 1 1 lLlg5 lZ:lf6 1 2 f3 �h8!?
1 6 fxe4 Just a few days after his defeat against
16 tt:lcxe4 h6! just wins for Black while 16 Bareev, Radjabov suddenly finds the FIDE
tt:lgxe4 tt:lf5 is not an inspiring way for White to World Champion taking aim at him in the same
play. variation. This is also a variation that Ponom­
1 6 ....!Dexd5! 1 7 lZ:lxd5 -.xg5 1 8 lZ:lxf4 ariov knows well but &om the Black side! So
It's interesting that Bareev originally said that what is Radjabov to do? He may well have

48
 The Classical Varia tio n : Th e Ba y o n e t A t ta c k 9 b 4

unearthed some improvements on his game fxe4 lt:lc6 1 7 lt:ld5 lt:lg8!

with Bareev but he made the prudent choice of Black obviously can't exchange on dS. This
avoiding 12...lbhS altogether. The reputation of retreat is an interesting extra possibility offered
t2...'�h8 suffered after Krarnnik inflicted a by his 1 2th move. From g8 the knight controls
heavy defeat on lvanchuk, but it seems to be the e7-square and now Black will want to play
back on the map now. ...lbc6-d4 and then caprure the pawn on e6 or
drive away the white knight with ...c7-c6
1 8 .i.d3 lDd4 1 9 'IFg4 g5!
And not 19 ...c6 20 'iVxg6! cxdS 21 exdS
when the weak pawn on e6 has suddenly rurned
into a monster. White has more than enough
play for the piece here.
20 1i'h3
20 bS must be worth considering to at least
ensure that the b-file is opened after Black kicks
the knight with c6 while Ftacnik suggests that
20 h4!? is an improvement
20 ...c6 21 lt:le3 'iff6 22 lL\g4
I t seems to me that within a couple of
moves White finds himself a pawn down for
1 3 l:b1 !? not much. Even parking his knight on a better
Could this be taking Bareev-inspired square looks insufficient though, e.g. 22 lbfS
prophylaxis a step too far? lLlxe6 (22...'iVxe6 23 ..ixgS) 23 l:fl lLlf4 24
a) The above mentioned Krarnnik-lvanchuk, 'iVg3 lbe7 is good for Black.
Monaco 2000 had gone 1 3 ..ie3 lbes 1 4 l:ct 22 . . .'1Fe7
c6?! 1 S cS! with a good game for White. But No need to allow 22...'ifxe6 23 ..ixgS.
Black can do better than that. For example, in 23 .i.e3 lt:lxe6 24 g3 '1Fd7 25 .l:f1 d5!?
Pogorelov-Nataf, Reykjavik 2004 he met 1 3
..ie3 with 1 3.....ih6!? not fearing 14 lbf7+ .:Xf7
1S ..ixh6 as after he blocks it up with 1S ... f4 he
is unlikely to miss his bishop too much. In the
game Black soon obtained good attacking
chances after the sequence 1 6 cS lbeg8 17 ..igS
h6 1 8 ..ih4 l:g7 1 9 ..i£2 gS 20 g4 fxg3 21 hxg3
lLlhS 22 �h2 g4 23 l:h1 'iVgS and 0-1 , S t .
And please note the very latest example
which followed the above game until move 1 S:
16 l:ct lbeg8 17 ..igS l:g7 1 8 g4 hS 19 h3 l:h7
20 �g2 ..id7 21 l:ht l:c8 22 'iVg1 l:a8 23 a4
1i'f8 24 aS lbh6 2S ..ih4 gS 26 ..i£2 a6 27 �ft
'lz-1/z, Bacrot-Radjabov, FIDE Wch KO, Trip­
oli 2004. Not the sort of move one normally recom­
b) Another possibility for White is 13 cS mends when the opponent has the bishop pair
when Black does best to play 13 ... h6 14 lbe6 but of course it all depends on the specifics .
..ixe6 1 S dxe6 dS! t 6 exdS lLlfxdS 1 7 lLlxdS 2S...lbd4 is the obvious and quieter alternative.
'i'xdS 1 8'iVxdS lLlxdS 19 ..ic4 lbxb4 20 l:b1 26 .l:f5?!
lbc6 21 .idS when White had enough play to Radjabov examines numerous other moves
hold the balance in Rechlis-Avrukh, Israeli in detail in his !tiformator notes. They are all
Team Ch. 2003. probably better than the text but none of them
1 3...h6 14 lt:le6 .lhe6 1 5 dxe6 fxe4 1 6 good enough for equality.

49
Pla y th e King 's Indian

26...dxc4?! often has to play awkward moves to collect the

The immediate 26...ltJd4 was stronger, one pawn. In recent times the main champion of 1 0
of the main points being that 27 �xd4 can be g3 has been the Dutch No. 1 Lock V an Wely.
met by 27 ... dxe4! Even when the rest of the world switched to 1 0
27 .bc4 llJd4 28 llxf8 llxf8 29 'ifh5 'ifd6 l:te1 he remained faithful to his pet variation.
30 �g2 b5! 31 �f7 lbc2 32 �c5 'ifd2+ 33 But something has happened recently. Suddenly
lbf2 Van Wely is playing a different system against
Or 33 �g1 c!De1 34 lt:lf2 lt:lf6 35 'W'g6 lt:\g4. the King's Indian in every game. 1 0 g3 is no­
33 ...lbf6 34 'ifg6 lbg4 where to be seen. As far as I can tell it is the
Black has whipped up a vicious attack out of two main games that I present here that have
nowhere. been the cause of Van Wely's disillusionment.
35 �g1 lbce3! 36 �xe3 lbxe3 37 h4 It seems logical to use the lines that forced Van
37 1i'h5 g4. Wely to give up 1 0 g3 as the basis of our reper­
37 ...'ife2 38 'ifh5 g4 0-1 toire.
White can't stop ...'W'£3 and ...'W'g2 mate. An 1 0 ...f5
effonless-looking victory although I'm sure it In an ideal world Black would like to play h6
wasn't quite so easy in practice. before f5 to prevent the white knight sortie into
e6. The problem is that after 1 o...h6 1 1 lt:ld2!
Game 15 White is threatening to take on hS and Black
has to retreat his knight before he can push his
Van Wely-Degraeve
f-pawn (he can throw in 1 1 ...�h3 1 2 lle1 first
Momiariz 2000 but it doesn't help).
1 1 lbg5 lbf6 1 2 f3
1 d4 lbf6 2 c4 g6 3 lbc3 �g7 4 e4 d6 5 White must support the pawn on e4. Black
lbf3 0-0 6 �e2 e5 7 0-0 lbc6 8 d5 lbe7 9 has now tried many moves, but as I have al­
b4 lbh5 1 0 g3 ready said, I am opting for the one that Van
Wely has found the most unpleasant.
1 2 ...f4

This is the old main line of the Bayonet At­

tack. White prevents lt:lf4 at the cost of weak­
ening his king position. As usual he plans to The direct approach. Black hopes to create
meet ...£5 with lt:lgS followed by lDc6. Again threats against the white king. Until recently
Black will have to take this knight and White Van Wely thought that 13 �g2 was the best
will recapture with dxe6. White appreciates that move and that is the subject of the next game.
in the long run the pawn on e6 is doomed First we look at the line he tried in his last cou­
(though there are tricks that Black has to avoid) ple of games.
but he will hope to develop compensation, 1 3 b5
both on the light squares and because Black The idea of this move is to prevent Hlack

50
 Th e Cla ssical Varia tio n : The Ba y o n e t A t ta c k 9 b 4

&om playing . . .c6. Although the knight o n g5 is So we have come to the end of a long forced
heading for e6 White prefers to wait for Black sequence set in train by Black's 13th move. This
to waste a tempo on h6 before completing his line was previously used as a means of agreeing
manoeuvre. Apart &om 13 �g2 (next game) a quick draw. Quite a few games have finished
and 1 3 b5 White has also tried 1 3 c5 (13 gxf4 is 19 lDxa8 ...xg3+ 20 �h1 �3+ 21 �g1 ...g3+
too risky as it exposes the white king). John 1/2- '12. This is a ridiculous line for White to play.
Nunn now likes the move 13 ... a5 (although he Black has at least a draw and maybe more. Van
himself had a very unclear game against Curt Wely obviously has something else in mind.
Hansen which went 13 ... dxc5 14 .i.c4!?) which 1 9 1lf2
attempts to show that the advance c5 was pre­ This avoids the perpetual check as
mature. There have been very few games with 19 .....xg3+ 20 llg2 just costs Black an ex­
this move. One of them was extremely enter­ change. Previously it was thought that Black
taining. Larsen-Dittmann, Reykjavik 1 957 con­ had to play 1 9 ...l:lac8 and earlier in the year Van
tinued 14 ltJb5 (White would like to play 14 a3 Wely had won a game against Golubev in pre­
but this is impossible whilst 1 4 bxa5 is met by cisely this line. After 20 :h2 ...xg3+ 21 llg2
14 ... dxc5 and 14 cxd6 by 14 ......xd6) 14 ... axb4 �3 (21 ...�4 is unclear according to old the­
15 "il'b3 h6 16 ltJe6 (or 16 cxd6 cxd6 17 ltJe6 ory but I'm sure Van Wely has something pre­
.i.xe6 18 dxe6 fxg3 19 hxg3 d5! with good play pared here) 22 ...xd6 :n 23 c5 White was on
for Black) 16 ltJe6 .i.xe6 17 dxe6 fxg3 18 hxg3 top. The hench Grandmaster Degraeve would
dxc5 19 .i.e3 b6 (White has given up two have noted this game during his preparation
pawns to maintain the jewel in his crown) 20 and started searching for an improvement. His
l:lad1 "il'b8 21 f4 c6 22 fxe5 (unfortunately for next move would have been a nasty surprise for
White 22 liJd6 loses to 22 ...l:la3!) 22 ... cxb5 23 Van Wely.
.i.f4 (23 exf6 ..xg3+) 23 ... ltJh5 24 .i.xh5 gxh5. 1 9 ...ltlxe4!
Could this be a unique pawn structure in the Now the theoreticians have to rewrite the
history of chess? The most relevant factor, theory books.
though, is Black's extra piece. The remaining 20 1lh2
moves were: 25 l:ld7 l:la3 26 ..d1 ...e8 27 ..d6 White avoids the critical line 20 fxe4 l:lx£2
..g6 28 ...xe7 l:lxg3+ 29 �h2 llg2+ 30 �h1 21 �x£2 :£St when Black has a dangerous
l:lg4 31 �h2 l:lg2+ 32 �h1 ..xe4 33 l:lg1 l:lf2+ attack for the sacrificed piece. Van Wely must
0-1 . have analysed this after the game, and judging
1 3 ...fxg3 1 4 hxg3 h6! 1 5 lLle 6 .be6 1 6 by the fact that he is no longer playing this line,
dxe6 "ifc8 1 7 tlJd5 not liked what he saw. Here are a couple of
Otherwise White loses the pawn on e6 for possible continuations:
nothing. a) 22 .i.f.3 � 23 �e3 ...xg3 24 ltJe6 :f6
1 7 ..."ifxe6 1 8 ltlxc7 "iFh3! 25 ltJxg7 �xg7 26 �1 g5 27 ...g2 ..e1+ 28
�d3 lbg6;
b) 22 �e3 (22 �e1 leads to the same posi­
tion) 22......xg3+ 23 �d2 :£2 24 "il'b3 ..g2 25
...e3 (25 ..d3 ltJ£5 26 ex£5 e4 is also strong)
25 ... h5!. This last move both threatens .i.h6 and
sets Black's most dangerous passed pawn in
motion.
In both cases the black attack looks well
worth the investment.
20..."ifd7
Of course not 20......xg3+ 21 :g2 when
Black loses his knight.
21 ltlxa8 ltlxg3!
The knight in the comer is going nowhere

51
Pla y th e King 's Indian

and can be collected later. First Black takes a

crucial kingside pawn. The white king is now
guaranteed a rough ride.
22 .bh6 .bh6 23 lhh6 �g7 24 l:l.h2 ltJef5
25 �2 :Xa8

1 3 ...c6
This is a desirable move as after the se­
quence lbe6, ... ..i.xe6, dxe6 White can no longer
play lbd5 (unless he sacrifices a piece as in the
game). It also means that if White plays b5
Black finally finds the rime to take the Black can seal up the centre with ... c5. True,
knight. Despite having just one pawn for the this concedes the d5-square after the subse­
exchange Black has the better game. There are quent exchanges on e6 but this is a price worth
excellent squares for his knights and his king is paying for blocking the position (see the notes
much safer. to move 14). Black has a large number of alter­
26 �d3 11'c7 27 �xf5 lLlxf5 28 1l'd3 l:l.c8 natives. I'd just like to take a quick look at two.
29 f4?! 11'c5+ 30 �3? a) 1 3 ... lLlh5 is interesting but I don't want to
This loses quickly. 30 �g2 exf4 would have recommend it as after 14 g4! ..i.f6 1 5 lbe6
lost more slowly. ..i.xe6 1 6 dxe6 lLlg7 1 7 c5 White has an im­
30... 84+! 31 11'xe4 11'a3+ 32 �g2 11'b2+ 33 proved version of Game 1 2. The only differ­
�3 11'c3+ 34 �g2 11'b2+ 35 �3 11'xh2 ence is that White has played �g2 instead of
Taking the other rook was even more con­ l:te1 but this means that his king is much better
vincing, e.g. 35 ...1i'xa1 36 1i'xb7+ �f6 37 1i'xc8 placed. The position is still not too bad for
'fi'fl+ 38 l:t£2 (after 38 �e4 dS+ Black also wins Black but I don't want to give a line against 10
the queen) 38...1i'h3+ 39 �e2 lbd4+ 40 �d2 g3, a supposedly less dangerous move than 1 0
'ifxc8. l:te 1, where Black ends up with a worse version
36 11'xb7+ �h6 37 11'xc8 1l'g3+ 38 �e2 of what he played against 10 l:te 1 .
11'e3+ 39 �d1 11'd4+ 40 �c2 11'xa1 0-1 b) 1 3 ...�h8!? (the black king is better off
here than on g8; the only question is whether
Game /6 he should spend a tempo on this move right
now) 14 c5 (other moves are less danger­
Van Wely-Fedorov
ous) 14 ... h6 1 5 cxd6 (1 5 lbe6 ..i.xe6 1 6 dxe6 d5
European Team Ch., Batumi 1999 1 7 exd5 lbexd5 1 8 lLlxd5 l2Jxd5 1 9 ..i.c4 c6 is
fine for Black; this is an important line to bear
1 d4 lLlf6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5 in mind as the ...d5 break is often a possibility
lLlf3 0-0 6 �e2 e5 7 0-0 lLlc6 8 d5 ltJe7 9 after White has played c5) 1 5 ...1i'xd6 (now
b4 lbh5 1 0 g3 f5 1 1 lLlg5 lLlf6 1 2 f3 f4 1 3 1 5 ... cxd6 1 6 lbe6 ..i.xe6 1 7 dxe6 d5 1 8 exd5
�g2 lLlexd5 1 9 lLlxd5 l2Jxd5 20 ..i.c4 is good for
White takes a time out to improve his king White as Black can't suppon his knight with c6)
position. The invasion on h3 that we saw in the 1 6 lLlb5 'ifb6 1 7 a4 lLlfxd5! (Black needs a
previous game is no longer possible. square for his queen as White was threatening

52
 The Classical Varia tio n : Th e B a y o n e t A t tack 9 b 4

aS) 1 8 exdS hxgS 1 9 aS 'W'f6 20 li:Jxc7 .l:b8 21 from similar positions but normally White gets
g4! Now in Van Wely-Nunn, Wijk aan Zee more direct play than he does here.
1992 Black got into trouble after 21.....td7 and 1 8 . . .cxd5 1 9 cxd5 g5
the line was more or less discarded. Nunn
points out, though, that both 21 ...1i'd6 and
21 ...e4 give Black an equal game
1 4 1Fb3
Alternatively:
a) 14 bS cS! 1 5 1i'd3 (1 5 li:Je6 ..ixe6 1 6 dxe6
li:Jes 17 li:Jds &7 18 li:Jxc7 'ifxc7 is good for
Black) 1 S...li:Je8 (threatening tilidS) 1 6 li:Je6
..txe6 1 7 dxe6 'W'c8 1 8 li:JdS 'ifxe6. White
doesn't have much play for the pawn as with
the position blocked his bishops are ineffective.
It will still be very difficult for Black to win as
his extra pawn is the backward one on d6.
b) 14 cS is very aggressive and probably best
met by 14 ...h6 1 5 ..ic4!? (1 5 li:Je6 ..txe6 1 6 dxe6 It is very hard to believe that White has
dS!) l S ...hxgS 16 cxd6 'it>h7 17 dxe7 li'xe7. I f enough for the piece. I'm sure Van Wely
White now plays 1 8 d6 he is likely to end up doesn't believe it either as it was after this game
losing the pawn (Black can start with 1 8 ... 1i'd7) that he gave up 1 3 �g2. Still, as the game
whilst 1 8 g4 is met by 18...1i'xb4 and other shows it is not easy for Black to make progress.
moves, such as 1 8 bS, by 1 8 ...g4 with good 20 .td2 .l:l.f8 21 g4 h5 22 h3 lLlg6 23 .l:l.dc1
attacking chances for Black. 1Fb8 24 .te1 hxg4 25 hxg4 lLle8 26 a4 .tf6
27 a5 .td8 28 •a4 lLif6 29 .l:l.a3 �g7 30
.tf2 .l:l.h8 31 .l:l.ac3 .te7 32 .l:l.c7 .l:l.h6 33
•a1 �8 34 .l:l.h1 lL!xg4!? 35 fxg4 f3+ 36
.txf3 lLif4+ 37 �g1 .l:l.xh1 +?!
Black got fed up with all the quiet manoeu­
vring and sacrificed a piece for an attack. Fe­
dorov should have contented himself with per­
petual check: 37 ... li:Je2t! 3S �g2 (38 ..txe2
.l:xht+) 3S ...li:Jf4t 39 �gl li:Je2+-. After his
choice he was lucky to escape with a draw.
The remaining moves were: 3S ..txh 1 'W'h3
39 1i'd1 .l:h8 40 .l:xe7+ �g6 41 ..if3 1i'h2+- 42
�fl .l:c8 43 .l:£7 'W'h3+ 44 �gl .l:c3 45 .l:xf4
gxf4 46 ..ig2 'W'hs 47 .tel .l:c7 48 .in a6 49
14. . . h6 bS axbS 50 li'bt b4 51 1i'xb4 .l:h7 52 ..ig2 'W'fB
\Vhen Black feels he is ready to face li:Je6 he 53 'ifa3 'ifbs 54 �fl 'irc7 55 'ifc3 'ifdS 56 ..tt2
plays this move; otherwise he just keeps on �gs 57 ..ts .l:h3 ss �g2 'ifhs 59 �gt 'ifds
improving his position. 60 �fl 'W'eS 61 1i'b3 'ireS 62 �g2 .l:hS 63
1 5 lL!e6 .b:e6 1 6 dxe6 -.c8 1 7 .l:l.d1 .l:l.d8 li'bt 'W'c7 64 'W'el 1i'c2 65 e7 �f6 66 e8'W'
Avoiding the trap 17 ...'W'xe6? 1 8 .l:xd6! :XeS 67 1i'h1 �g7 6S 1i'h4 .l:h8 69 1i'e7+ �g8
'ifxd6?? 1 9 cS+ which costs Black his queen. 70 1i'd8+ �g7 7 1 1i'b6 'ifct 72 ..igl 'ifd2+- 73
1 8 lLid5!? 1i't2 'W'xaS 74 1i'c2 �g6 75 1i'd1 bS 76 ..tt2 b4
White was faced with the choice of letting 11 gs 'iFa2 7s ..tg4 b3 79 ..tfS+ �g7 so 'W'n
his e6 pawn go for insufficient compensation or .l:hS Sl li'bs .l:h2+ 82 �xh2 'W'xf2t- 83 �hl
sacrificing a piece to ensure that it survives for 'ifel+ S4 �g2 1i'd2+- SS �fl 'W'dl+ 86 �t2
the rest of the game. This sacrifice is known 1i'd2+- 87 �fl '12-'12

53
Pla y th e King 's Indian

more comfortable for the white bishop) 1 1 ...f5

Game 1 7 (if Black does not feel comfortable with the
complications that follow he can play
Kamsky-Kasparov
t t ...lbxe2+ before playing ... f5) 1 2 ..t£3 gS 1 3
New York (rapid) 1994 exf5 ll'lxf5 1 4 g3 ll'lh3+ (Black can also sacrifice
a piece with 14 ...lbd4 but the latest evidence
1 d4 lt!f6 2 c4 g6 3 lt!c3 �g7 4 e4 d6 5 suggests it is unsound) 1 S 'it>g2 1i'd7! 1 6 ..te4
lt!f3 0-0 6 �e2 e5 7 0-0 lL!c6 8 d5 lL!e7 9 (16 �xh3? lbe3+! wins for Black) 1 6...g4 1 7
b4 lt!h5 1 0 c5 ll'lb3 1i'e7 1 8 1i'd3 h S with a double-edged
So far we have seen lines where White takes position. The game Mannion-Smirin, Las Vegas
the threat of ...lbf4 seriously. 10 l:.e1 enabled 1 997 should give you some idea of how Black
the bishop to drop back to f1 while 10 g3 pre­ can develop his game on the kingside: 19 cS
vented ll'lf4 altogether. In this game we shall ll'lf4+ 20 gxf4 exf4 21 f3 ..i.eS 22 l:.a2 1i'g7 23
look at lines where White ignores the threat of lbdt g3 24 l:.g1 gxh2+ 2S �xh2 ll'lg3 26 l:.gg2
...ll'lf4. 10 cS, the subject of the main game is by 1i'gs 27 ll'lf2 ¢>h8 28lbd4 'Wh4+ 29 'it>gt l:.f7
far the most common of these lines while be­ 30 lbe6 ..i.xe6 31 dxe6 l:.g7 32 ..tb2 l:.ag8 33
low we have a quick look at rwo of White's ..i.xeS dxeS 34 lbg4 'Wh1+ 3S 'it>f2 hxg4 36 fxg4
rarer alternatives: .:.Xg4 37 ..i.£3 e4 38 1i'd4+ l:.4g7 0-1 .
a) 10 'irb3 ll'lf4 1 1 ..i.xf4 exf4 1 2 l:.adl h6 1 0 lt!f4
...

(1 2.....i.g4 is also worthy of consideration) 13 cS 10 ... h6 and 1 o ... f5 are played from time to
gS 14 eS g4 (14 ...dxeS 1 S d6 is dangerous for time but there is no reason to avoid the main
Black) t S exd6 cxd6 16 lbd4 ll'lf5! (Nunn's line.
suggestion - we now follow his analysis) 1 7 1 1 �xf4
ll'lxf5 ..i.xf5 1 8 ll'lbS ..i.eS 1 9 ll'lxd6 (or 1 9 cxd6 1 1 ..tc4 is the only serious alternative. Byk­
£3) 19 ... £3 and Black has a good game, e.g. hovsky-Avrukh, Beersheba 1 996 now contin­
ued 1 1 ...¢>h8 (Black is planning f5 so he needs
to get his king off the sensitive diagonal;
t t .....tg4 1 2 h3 ..i.hS 1 3 l:.et �h8! leads to the
same position) 1 2 l:.et ..tg4 1 3 h3 ..i.hS! 1 4 ..tfl
(White defends h3 in order to threaten g3 or
g4) 14 ... f5 1 S ..i.xf4 (neither lS exf5 e4 nor lS
g3 fxe4 16 ll'lxe4lbxh3+ are playable) 1 S ... exf4
1 6 l:.ct aS 17 a3 axb4 18 axb4 ..i.x£3 19 gx£3 (19
1i'xf3 l:.a3 20 1i'd3 £3! is good for Black)
1 9 ... fxe4 20 l:.xe4 ll'lf5 with advantage to Black.
1 1 exf4
...

at) 20 gx£3 'Wh4 21 h3 gx£3 (2t .....i.h2+ 22

�xh2 'ifxh3+ 23 �gl g3 is a draw) 22 1i'xf3
..txh3 with good play for Black.
a2) 20 ..tx£3 ..i.xd6 21 cxd6 gx£3 22 1i'xf3
..tg6 and the three pawns are not as good as the
piece.
a3) 20 ..i.d3 'Wh4 21 h3 fxg2 22 ..i.xf5
gxf11i'+ 23 �xfl with advantage to Black
b) 10lbd2 ll'lf4 1 1 a4 (on the immediate 1 1
..t£3 Black achieves good play with 1 t ...lbd3 1 2
..ta3 aS; 1 1 a4 is designed to make the square a3

54
 Th e Classical Varia tio n : Th e Ba y o n e t A t tack 9 b 4

1 2 :c1 order to get a rook to the 7th rank.

White can also defend the knight with his 14 ...�g4 1 5 :c7 axb4 1 6 1Fd2
queen. 1 6 'iib3 can also be met by 1 6....i.xf3 with
a) 1 2 'iib3 can be met by 1 2 ... h6 as 13 l:tad1 similar play.
would then transpose to line 'a' to White's l Oth 1 6 ...�xf3! 1 7 �xf3 �e5
move, but 1 2... .i.g4 is also a good move. Black
wishes to take the knight on f3 so that he can
have total control over e5. Shneider-Gufeld,
Helsinki 1 992 continued 1 3 l:tad 1 .i.x£3 1 4
.i.x£3 g5 (this i s a key move i n a number of
variations; Black intends lbg6-e5 when his
knight would be extremely powerful) 1 5 .i.h5
lLlg6 16 .i.xg6 hxg6 17 lbbS l:te8 18 l:tfe 1 a6 1 9
lba3 g4 20 lbc4 .i.e5 with advantage to Black.
b) 12 1i'd2 .i.g4 13 l:tacl .i.x£3 14 .i.x£3 g5
15 .i.g4 lbg6 1 6 l:tfd1 1i'e7 1 7 cxd6 cxd6 1 8
l:te1 a6 1 9 a4 l:lae8 20 � fl lbe5 2 1 .i. £5 f3 22
g3 g4 and White was in trouble in Larsen­
Giigoric, Lugano 1 970.
1 2 ...a5!? 1 8 :xb7? !
More common is 1 2...h6 foUowed by ...g5 According to m y database this position has
and ... lbg6, but this is a good moment to chal­ been reached twice and both times White has
lenge White's supremacy on the queenside. The played this inferior move. Inferior because it
main points in 1 2 ... a5's favour are that it devel­ leads, without a fight, to a prospectless position
ops the rook on a8 and that Garry Kasparov where White might be able to hang on for a
plays it! draw or he might not. 1 8 l:tfcl is better after
which all White's pieces, bar one, are actively
placed. The problem is this one, the light­
squared bishop, has very little indeed to look
forward to. It is completely smothered by the
central pawns that are fixed on light squares. In
fact White would be better off if he simply
removed these pawns from the board. Black
also has a poorly placed piece, the knight on e7,
but this at least has the potential to become a
wonderful piece (see the main game if you
don't believe me). After 1 8 l:tfcl the plan of
...g5 and ...lbg6 is rather slow so Black should
initiate complications with the aim of exchang­
ing his passive knight for White's active one.
1 3 cxd6 Here are a few plausible variations: 1 8.. .lla5!,
A more solid continuation is 1 3 a3. Tukma­ and now:
kov-Smirin, Elenite 1993 then continued a) 1 9 :Xb7 'it'a8 20 1i'xb4 (20 l:tcc7 :Xa2 is
1 3... axb4 14 axb4 £5!? (14 ....i.g4 and 1 4... h6 also good for Black while 20 :Xe7 l:txb5 is also
come into consideration) 1 5 l:te1 .i.xc3 1 6 rather unpleasant for White) 20 ...1i'xb7 21
:Xc3 fxe4 1 7 lbg5 f3 1 8 gx£3 ex£3 1 9 .i.x£3 1i'xa5 l:ta8 22 1i'b4 (22 l:tc7 'it'b8! is worse)
lb£5 20 lbe6 .i.xe6 21 dxe6 1i'f6 22 1i'd2 lbe7 22 .. .llxa2 with an edge for Black.
23 cxd6 cxd6 24 l:td3 dS 1/z-1/z. b) 1 9 'ifxb4 l:txa2 20 l:txb7 l:tb2 21 'ifa3
1 3 ...cxd6 1 4 lt'lb5 1i'a8! 22 :Xe7 (22 'ifxa8 :Xa8 23 lLlxd6 looks
White is willing to give up his pawn on b4 in like a nice combination but 23 .. .llc2! is the

55
Pla y th e King 's Indian

refutation) 22...llxb5 and Black is better be­ Summary

cause his bishop has a juicy target on f2 whilst 1) The Bayonet Attack, or more specifically
W'hite's has nothing to attack. the 10 l:te1 variation, has become one of the
1 8 ...'ii'a5 1 9 ltXI4 'ii'xa2 20 'ii'xa2 llxa2 21 most important lines in the King's Indian.
llxb4 llfaB Games 1 1 -1 4 are, therefore, of particular im­
In Hemdl-Kindermann, Austrian League portance.
1 996 Black immediately played 21 .....i.xd4 and 2) The manoeuvre lbg5-e6 occurs very of­
after 21 .....i.xd4 22 l:txd4 g5 23 l:tb4 (23 ..i.h5 is ten. This leads to an unbalanced game with
better) � 24 ..ig4 lbe5 25 h3 l:tfa8 he also positions that are hard to assess. W'hite will
ended up triumphing on the dark squares. The obtain the bishop pair and a passed pawn on
remaining moves were: 26 l:tb6 l:t8a6 27 l:txa6 e6, which though doomed in the long run, has
l:txa6 28 l:tbt l:ta4 29 ..i.f5 h5 30 f3 cj;g7 31 considerable nuisance value. Black will aim to
l:tcl �f6 32 l:tc3 lbc4 33 ..i.d7 l:ta1+ 34 �f2 safely round up the pawn and then consolidate
l:ta2+ 35 �g1 lbe3 36 g4 h4 37 l:tc7 l:tg2+ 38 his position. Often W'hite will play lbg5 and
�h1 l:tg3 39 ..i.e8 l:txh3+ 40 �g1 l::tx£3 41 then the knight will then remain suspended
l:txf7+ cj;e5 42 l:te7+ �d4 43 e5 l:tg3+ 44 cj;h2 there for a few moves. This is because W'hite is
f3 0-1 waiting for Black to waste a tempo and weaken
22 �g4 his kingside with ... h6 before completing the
After 22 lbc6 lbxc6 23 dxc6 l:tc2 W'hite will journey to e6. In the end Black usually forces
lose his passed pawn but can at least activate his the issue as it is more unpleasant to play with
bishop by 24 ..i.d 1 l:txc6 25 ..i.b3. Kinder­ the threat of lbe6 hanging over him than it is
mann's 21.....i.xd4 would have avoided this for the move to actually happen. The old prov­
possibility. erb 'the threat is stronger than the execution' is
22 ...�xd4 23 llxd4 g5 very relevant here.
Black now threatens to achieve a strategically 3) Black must take great care when taking
won game by the manoeuvre ...lbg6-e5. We the pawn on e6 that W'hite hasn't prepared a
have already seen an example of this in the tactical refutation. For example, within the
Kindermann game above. W'hite could have notes to Game 1 1 we see ...1i'xe6 running into a
played 24 ..i.h5 to prevent ...lbg6 but then sacrifice on e4 while in Game 12 there is an
Black would have calmly improved his position example of 1i'xe6 running into l:txd6. Black
with .. .l:tb8 and ... �g7-f6. Kamsky decides to must take extra care when his king is on g8 and
allow the knight into e5 and instead pins his W'hite has a queen or bishop on b3.
hopes on breaking up the Black pawn structure. 4) When W'hite plays c5 after lbg5-e6 has
With best play the position may be a draw but it happened then the first thing Black should
is very unpleasant to defend. examine is whether he can play the move d5 or
24 h4 gxh4 25 �h2 ti)g6 26 �h3 llb2 27 not. If W'hite plays c5 with the knight still on g5
�f5 lDe5 28 �xh4 h6 29 �h3 �g7 30 then Black may be able to play h6, ..i.xe6 and
lldd1 llaa2 31 f3 ti)g6+ 32 �g4 �6 33 d5.
llb1 h5+! 5) Game 17 features a different sort of game
Suddenly the white king finds itself in a mat­ where White meets lLlf4 by playing ..i.xf4 and
ing net. then clears the long diagonal to limit the influ­
34 �xh5 liaS! 35 �g4 35 ...llhB 36 g3 llh2 ence of Black's dark-squared bishop. A key
W'hite stopped the mate but at the cost of manoeuvre for Black in this line is ... ..i.g4xf3 in
his bishop. order to gain complete control over e5. His
37 llh1 ll8xh3 38 llxh2 llxh2 39 gxf4 bishop may settle here but it is an ideal outpost
llg2+ 40 �h3 lLlxf4+ 41 �h4 �e5 42 llb7 for his knight. The usual route &om e7 is via g6
�d4 43 llxf7 �e3 0-1 after Black has played g5.

56
 CHAPTER FOUR I
The Classical Variation :
9th Move Alternatives

1 d4 lLlf6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5 sttcngth to the advance b4-bS-b6 which White

lLlf3 0-0 6 �e2 e5 7 0-0 lLlc6 8 d5 lLle7 likes to carry out. In fact after 9 �2 very few
In this chapter we shall examine the alterna­ Black players arc willing to get involved in a
tives to 9 �e1 and 9 b4. That means 9 �2 direct race. Against 9 � 1, for example, Black
and 9 ..igS. The first two games deal with 9 just gets on with game not minding too much
�2 and the third game concentrates on 9 ..igS that White's attack is a bit further along the line
with some of the rare 9th move options men­ than his. The point is that his prize at the end is
tioned in the notes. much greater - the white king - and is worth
suffering a bit for. But in this 9 �2 line, if
9 lbd2 Black decides to get involved in such a race he
can easily end up with his quecnside decimated
whilst his own attack is still just a glint in its
mother's eye.
To show you what I mean, take a look at Be­
liavsky-Solak, Europe Ch., Saint Vincent 2000.
After 9 �2 play continued 9...�7 10 b4 f5
1 1 f3 (White can also play 1 1 cS!? at once)
1 1 ...�f6 (now 1 1 ...f4 would certainly be met by
12 cS) 1 2 cS f4 1 3 �4 gS 14 a4 (a sttong case
can also be made for 14 ..ia3) 14 ...�h>6 1 S ..ia3
llf7 1 6 bS �8?! 1 7 aS ..iffi 1 8 �a4 hS 19 b6
..id7 20 bxc7 'ifxc7 21 a6 bxa6 22 c6 ..ic8 23
�ab2 llg7 24 �3 �h8 2S �f2 �f7 26 h3.
Beliavsky's last four moves really rubbed the
With 9 �2 White is taking prophylactic ac­ salt in Black's wounds. Instead of going for the
tion against Black's kingside attack. He will now kill on the queensidc he adopts the same defen­
be able to defend his centte, if need be, by play­ sive set-up we saw in Chapter 2, with the dif­
ing £3. But the move also has an aggressive side ference that instead of White just having attack­
and the knight on d2 is looking forward to ing chances on the quecnsidc Black has already
playing an important role in the quccnsidc at­ been wiped out in that part of the board. The
tack. Once White has played cS it will be able to remaining moves were: 26 ...�h6 27 llb 1 'it'd8
occupy the fine square on c4. From there it 28 ..ib4 �f6 29 llb3 aS 30 ..ixaS 1i'e8 31 c7
attacks the crucial d6-square and will give added 'ifg6 32 l:tb8 g4 33 hxg4 hxg4 34 llxa8 ..id7 3S

57
Pla y th e King 's Indian

c8'ii' i.xc8 36 llxc8 g3 37 lLlh3 lLlfl 38 i.b4 tiJf3 0-0 6 ..ie2 e 5 7 0-0 tlJc6 8 d 5 ttJe7 9
lLlgS 39 i.xd6 1i'hs 40 i-xfB lLlxh3+ 41 gxh3 tiJd2 a5 1 0 a3 ttJd7
1i'xh3 42 i.xg7+ �xg7 43 l:r.c7+ 'it>g8 44 J:r.f2 In the early 1 990's, before the Bayonet At­
gxf2+ 45 �xf2 1i'h4+ 46 <t>g2 1 -0. tack took over, this line was the main battle­
Okay, Black could have defended better but ground in the King's Indian. Black retreats his
I'm sure you get the picture. knight in order to play ... fS, and chooses the d7-
So that means that if we are not going to square (as opposed to e8) in order to hold up
challenge White to a race we have to take some cS for as long as possible.
action to slow down his attack on the queen­ The main alternative is 10 ...i.d7 which is
side. An added benefit of this strategy is that if not considered here but there is coverage in
Black can prevent White from playing cS the SO KID.
knight on d2 can end up looking quite stupid. 1 1 :b1 f5 1 2 b4 �h8
The most radical way to prevent cS is for
Black to play it himself. However after 9...c5
White has new targets on the queenside and by
playing J:r.b 1 and b4 he is able to develop the
initiative. The main line after 9...c5 runs 10 l:r.b1
(1 0 dxc6 bxc6 1 1 b4 is also interesting)
10 ...lLJe8 1 1 b4 b6 1 2 bxcS bxcS 1 3 lLlb3 f5 1 4
i.gS! with a n edge for White, e.g. 1 4...h 6 1 5
i.xe7 1i'xe7 1 6 lLla5.
I suggest that Black plays 9 ...a5!. This will be
no great surprise to those of you already famil­
iar with this position as it is the main line. The
point is that even after White plays a3 he is still
in no position to play b4 because of the pin on
the a-file. Therefore he needs another move A semi-waiting move that has displaced all
(usually l:r.b1) in order to be able to play b4. other moves. How did Black arrive at such a
Playing ...aS, therefore, saves Black a tempo. decision? Well, the two most obvious moves,
Even this might not be sufficient for Black to 1 2 ... f4 and 1 2...lLlf6 both have their drawbacks.
get involved in a straight race, but as we shall 1 2 ... f4 allows White to play 13 i.g4 while Black
see he no longer needs to. In some of the mate­ is also reluctant to move his knight from d7 just
rial below he actually challenges White for su­ yet as it is holding-up White's c4-c5. Therefore,
premacy on the queenside. This usually means by a process of elimination Black arrives at
playing ...c6 and Black is aided by the fact that 12 ...�h8 (he could also flick in 12 ...axb4 13
after a well-timed ...axb4 he can take control of ax b4 and then make his decision but delaying
the a-file. Having said this, Black's primary this exchange cuts down slightly on White's
objectives are still on the kingside and one of options). Of course this is not just a waiting
the reasons for playing ...c6 is to slow down, or move. Black now has the g8-square available
make less effective, the white attack on the for manoeuvres such as ... lt:)g8-f6 and the king
queenside. After 9... a5 10 a3 I suggest that is safer in the comer, especially if ... c7-c6 is
Black plays 10...lLld7 in order to prepare ...fS played.
and this is covered in games 1 8 and 1 9 below. 1 3 1i'c2
In fact another reason behind Black's previ­
Game 18 ous move was too pass the buck to White. He,
too, has a difficult decision to make. The obvi­
lputian-Dolmatov
ous way to suppon cS is with 1 3 lLlb3, but the
Rostov 1993 problem with this is that the knight will be mis­
placed once cS has been played. After cS White
1 d4 tiJf6 2 c4 g6 3 tiJc3 ..ig7 4 e4 d6 5 wants to play lLlc4. White normally chooses

58
 The Classical Varia tio n : 9 th M o v e A ltern a tives

between the text and 13 f3, which is the subject will assume that Black exchanged now. Any­
of the next game. way, this is probably the right moment to ex­
1 3 . . .itJf6 change to avoid the possibility of being sur­
1 3.)iJg8 was popular for a while but the po­ prised by White taking on aS.
sition after 14 f3 (14 exf5 gxf5 1 5 f4 is also 1 5 axb4 c6! ?
interesting) ltJg£"6 1 5 ..td3! has proved ctuite
difficult for Black. The only way for Black to
continue waiting is with 1 3...b6, but as I am
going to recommend a line which involves
playing ... c6 this is not possible for us.

With the centre still tense Black is not yet

committed to an ali-in assault on the white king.
The text fights for some space on the quecnside
and prevents White from playing cS. True, after
an evenrual exchange on dS Black will have
14 fl performed a task that White normally under­
This is the standard reaction to ...lbf6. If takes - opening the c-file - but he will have also
White doesn't play this move then Black may avoided the most dangerous lines where White
be able to ectualise by exchanging on e4. For keeps his pawn on cS supported by a knight on
example, 14 ..tb2 fxe4 (14 ... c6!?) 1 5lbdxe4 (1 5 c4. Then White will have the additional possi­
lbcxe4 axb4 16 axb4 c6) 1 5...lbf5 1 6 lLlxf6 bilities of c5-c6, b5-b6 or simply to exchange
lbd4 17 'ii'd2 'ii'x f6 should be fme for Black. on d6 at a moment of his choosing. Let's have a
Last year I had to face the unusual 14 l:tdt in look at a few examples where Black left the
Arlandi-Gallagher, Mitropa Cup 2003. Unfor­ queenside to fend for itself.
runately my form in this tournament was abys­ a) 1 5...f4 1 6 cS gS 1 7lbb5 (1 7lbc4 is more
mal (I can't think of worse, that's for sure) and accurate when Black should play 1 7...g4) and
I didn't react well. 1 4 l:tdt is obviously directed now:
against Black's ... c7-c6 but after a long (and I at) 1 7 ...lbe8? 1 8lbc4lbg8 19 ..tb2 lLlgf6 20
mean long) think I went ahead and played l:tal l:tb8 2t lba7 ..td7 22 c6 bxc6 23 lLla5! and
14 ... axb4 1 5 axb4 c6 only to end up with the White's attack is already decisive.
slightly worse position after 16 cS fxe4 1 7 a2) Instead of passively defending on the
lbdxe4 lLlxe4 1 8 lbxe4 � (5 19 �d3. I soon CJUeenside Black should have countered with
blundered and lost. 1 7...dxc5 1 8 bxcS c6!. Ftacnik-Nunn, Gronin­
Looking at the position in the cold light of gen 1988 concluded 1 9 d6 cxbS 20 dxe7 'ii'xe7
day it is possible to exchange at once on e4 (as I 21 l:xbS g4 22 �b2 gxf3 23 ..txf3 lbg4 24
suggested against 14 ..tb2) or, perhaps, more to ..txg4 ..txg4 25lbf3 l:tfc8 1/2-'12.
the point, start to advance on the kingside. b) 1 5...g5!? enables Black to combine both
14 ...g5 looks like a good move. pawn and piece play on the kingside and was
14 ...axb4 the choice of the then World Champion in
In some of the games below the exchange Vaganian-Kasparov, Manila Olympiad 1992.
on b4 occurred later but for simplicity's sake I Play continued 16 cS lbg6 1 7 lbc4 (17 exf5

59
Pla y th e King 's Indian

lbf4 18 �c4 dxcS 1 9 bxcS g4 is good value for prepares to attack on the kingside. This line is a
a pawn) 17 ...lbf4 18 cxd6 cxd6 19 �e3 g4 20 speciality of Dolmatov and he has been playing
�b6 'fke7 21 ex£5 lbxe2+ 22 lbxe2 gx£3 23 it successfutly for almost a decade now.
.:X£3 lbxdS with a very complex position. The
game eventually ended in a draw.
These lines look like decent alternatives to
1 5...c6.
1 6 �h1
One of the other points of 15 ...c6 is that
...'fkb6+ becomes a possibility. It is a sensible
precaution for White to remove his king from
the exposed diagonal. Let's have a look at the
alternatives:
a) 16 lLib3 fxe4 (I once played 1 6 ... lbh5
which is also interesting) 17 fxe4 cxdS 1 8 cxdS
'Wb6+ 19 �h1 �d7! (after 19 ...'ifxb4 20 �e3
the queen is in rrouble) 20 lbaS?! (20 'iVd3 is
better as by overprotecting the fl -square White 1 7 dxc6
would have avoided the fate that befeU him in In a more recent Dolmatov game White
the game) 20...lbg4! 21 h3 .:Xfl+ 22 �xfl l:t£8 tried 1 7 'fkb3, but eventually ended up getting
23 �gS? (and here White should have tried 23 crushed on the kingside. Here are the moves:
'ife2) 23...lbf2+ 24 �h2 lbxh3! and Black wins 17 ... cxd5 1 8 cxdS gS 1 9 lbc4 hS 20 �d2 g4 21
the queen after 25 gxh3 l:t£2+ and the king after l:ta1 l:tb8 22 lbbS lbe8 23 l:tfcl lDg6 24 �fl
25 �xe7 'fkgl+ 26 �g3 �h6! 27 �x£8 (there is �f6 25 lba7 �d7 26 bS �h4 27 �e1 �xe1 28
no defence) 27 ...'ife3+ 28 �h2 �f4+ with mate l:txe1 lbh4 29 fxg4 hxg4 30 lbd2 l:tf7 31 l:te2 £3
next move. In Lputian-Piket, Wijk aan Zee 32 l:te3 fxg2+ 33 �xg2 lbg7 34 �fl 'figS 35
2000, White played 25 �bS but his position 'fkd3 l:tf2 36 �e2 'ifhs 37 lLifl l:tb£8 38 l:te1
was hopeless after 25 ...lbxg5 26 �xd7 l:tf4 lLI£3 0-1 Ulibin-Dolmatov, Calcutta 1 999.
b) 16 dxc6 lbxc6 (this is the right way to re­ 1 7 ...lbxc6
capture; 1 6...bxc6 1 7 bS! is better for White) 1 7 Dolmatov has always played this but the al­
lbb5 lbh5!? (the solid 1 7... fxe4 1 8 lbxe4 lbxe4 ternative capture is also worthy of considera­
1 9 fxe4 l:txfl+ 20 �xfl lbd4 21 lbxd4 exd4 tion. In fact after 1 7 ... bxc6 1 8 bS gS we have
soon led to a draw in M.Gurevich-Ye Jiang­ transposed into some analysis by the French
chuan, Comtois 1 999) 1 8 l:td 1 1Vh4! 1 9 lbfl (19 grandmaster and King's Indian expert Igor
lbxd6? loses to 1 9...lbd4 20 'fkd3 lbf4) 1 9...fxe4 Nataf (he gives the move order 1 6...g5!? 17
20 'ifxe4 lbf4! 21 �xf4 (after 21 lbxd6 �£5! 22 dxc6 bxc6 1 8 bS f4). He fancies Black's chances
lbx£5 gx£5 23 'ife3 l:ta2 Black has dangerous in this position. A sample variation: 19 bxc6 (1 9
threats, e.g. 24 l:td2 �h6! 25 l:txa2 lbh3+ 26 b6 l:tb8) lbxc6 20 lbbS?! g4 21 'fkd3 g3! 22
gxh3 l:tgS+- 27 lbg3 �xe3 28 �xe3 f4 29 �f2 'ifxd6 'fke8 23 lbc7 'fkhs 24 h3 �xh3! and
fxg3 30 hxg3 l:txg3+ 31 �xg3 'ifxg3+ 32 �fl Black wins.
lbd4 and the powerful combination of queen 1 8 lZ'lb5
and knight leave White fighting for a draw) This both covers the d4-square, thereby
21 ...exf4 22 lbxd6? (this loses; 22 g3 is a better stopping Black from playing ...lbd4, and targets
chance) 22 ... �£5! 23 lbxf5 gx£5 24 'ifc2 (White the weak pawn on d6.
can't keep d4 under control as 24 'fkdS is met 1 8 ...'ife7
by 24...l:tad8) 24... �d4+ 25 l:xd4 (25 �h1 'iff2) Dolmatov calmly completes his develop­
25...lbxd4 26 'fkb2 'iff6 with a winning position ment, not rushing prematurely into a kingside
for Black, Nemet-Gallagher, Zurich 1 994. attack. Before attacking he must ensure that his
1 6 ...f4 weakness on d6 has sufficient protection. Later,
Back to basics. Black closes the centre and when his attack is further along the road he

60
 The Classical Varia tio n : 9 th Mo ve A lterna tives

may be able to jettison this pawn but for the 30 �e1 �e7!
moment it's best to consolidate. White has managed to defend along the sec­
1 9 l:td 1 �e6 20 lLif1 ond rank, so with no immediate breakthrough
White assigns his knight to defensive duty. Black just improves his position a bit more.
An alternative was 20 lLlb3 when Ruban­ There are two ideas behind ... ..te7. The first is
Dolmatov, Novosibirsk 1993 continued that the possibility ..td8-b6 is now in the air and
20...l:.fd8 (20...lLlxb4 21 'ifd2) 21 'ifd2 i.£8 and the second is simply to clear the way for the
instead of 22 lLla3 Ruban claims that White queen's rook to reach the kingside.
could have obtained the better game by 22 lLJaS 31 l:tb2 �g7 32 1i'g2 l:tcB
gS 23 lLlxc6 bxc6 24 lt:k3. This is debatable Not only is White suffering on the kingside
and, as Nunn points out, Black can even play but his queenside pawns are also weak. He
24...d5. would have more chances of organising a de­
20 ...l:tfd8 21 �d2 g5 22 �e1 �fB! fence if he could push his b-pawn back one
square.
33 �e2 lL\dB! 34 1i'f1 lLif7 35 �c3? �dB!
White's last move was a mistake as he can
no longer prevent the King's Indian bishop
getting to the lethal g1 -a7 diagonal.
36 l:tbd2 �b6+ 37 �h1 �xg4! 38 lLixd6
The simple point behind the sacrifice was
that 38 fxg4 loses to 38...lLlxe4.
38 ...l:txd6! 39 l:txd6 l:thB 0-1
White resigned as the only way to prevent
mate, 40 'it'g2, loses the queen to 40 ... i.h3.

Game 19
Ljubojevic-Kasparov
A very logical move. Black needs to defend
d6 but it is best to do so with his passive bishop Linares 1993
rather than his active queen. Swapping the posi­
tion of the bishop and queen will improve the 1 d4 lLif6 2 c4 g6 3 lLic3 �g7 4 e4 d6 5
fortunes of both. �e2 0-0 6 lLif3 e5 7 0-0 l2:\c6 8 d5 lL\e7 9
The immediate 22 ...g4 would have been a lL\d2 a5 1 0 l:tb1 lL\d7 1 1 a3 f5 1 2 b4 �hB
mistake on account of 23 ..th4! 1 3 f3
23 �f2 1i'g7 24 g4
A panicky reaction which doesn't stop
Black's attack. 24 l:.d2 is better according to
Dolmatov.
24. . . h5 25 h3 1i'h7
25...'it'h6! is more accurate as 26 lt:k7 loses
to 26 ... hxg4. White would have to play 26 �g2
when after 26...l:.d7 Black has an improved
version of the game.
26 �g1 l:td7
Black's long-term plan is to mate White
down the h-file. With this pawn structure the
sacrifice of a piece on g4, to break up the king's
defences, becomes almost inevitable at some
point. A more forthright move than the 1 3 1i'c2 of
27 lLih2 hxg4 28 hxg4 1i'h3 29 �f1 1i'h6 the previous game. What are the advantages of

61
Pia y t h e King 's Indian

1 3 £3 over 13 'ifc2? Well, firstly, it's not clear occasions when the queen on d 1 would have
where the white queen belongs yet. In some attacked the pawn on d6.
lines it may even be better off on its original
square so it is quite logical to first play a move,
£3, that White can rarely do without. Secondly,
by solidly defending the e4 pawn, White is tak­
ing prophylactic action against ... it)d7-f6. The
point is that White normally has to meet this
move with £3, thereby giving Black time to play
the lines with ... c6 that we saw in the previous
game, whereas now, with £3 already played,
White can meet lbf6 with c5. There are also
certain drawbacks to playing £3 so soon. Firstly,
it weakens the dark squares on the kingside
before it is essential to do so, and secondly, it
allows Black to play f4 and immediately start his
kingside attack (White no longer has .i.e2-g4). 1 4 ltJb3?!
The main way for Black to try and exploit the You have already learnt from the 1 3th move
dark squares is the manoeuvre ... lbg8-f6-h5-f4. notes to the previous game that this is a rather
In fact, Black has played 13 ...lbg8 in countless feeble move. Let's take a look at some of the
games but it was eventually discovered that alternatives.
after 14 'ifc2 lbgf6 1 5 i.d3! the attack to the f­ a) 1 4 lbb5 is often suggested but rarely
pawn means that Black has nothing better than played. The idea is to prepare c4-c5 by making
to play 1 5 .. .f4. The manoeuvre lbg8-f6 now the sacrifice too dangerous to accept. Black has
turns out to be useless as the f4-square is no the choice berween getting on with his kingside
longer available. In fact, the black knights are play or taking further precautions on the
just stepping on each other's toes and the queenside.
knight on f6 would actually be better off back
on e7. The point is that after Black plays f4 and
gS it would then have a very nice square on g6.
Therefore, I am suggesting that Black meets 1 3
£3 by immediately blocking the centre.
1 3 ...f4
There is one potential drawback to Black
declaring his hand (he is going to pawn storm)
on the kingside so soon. That is that, with the
centre blocked, White may be able to sacrifice a
pawn to accelerate his own queenside attack.
However, as we shall see in the concrete varia­
tions below, Black appears to be able to deal
with this.
And what about the plan with ... c6 that a 1) 1 4 ... b6 1 5 cS!? bxcS 16 bxcS lbxc5 17 a4
Black adopted in the previous game? Well, that occurred in Vladimirov-Temirbaev, Alma-Ata
is not so effective here as White has saved an 1 989. Black now played 17 ...g5?! and eventually
important tempo with his queen. In fact, not lost (only this move is given in NCO). In his
only has White saved the tempo he spent on notes Vladimirov says (according to Nunn) that
'ifc2 but, against the ...c6 idea, the queen is after 1 7 ... c6 1 8 dxc6 lbxc6 1 9 lbc4 it)d4 20
probably even better placed on d 1 . Remember lbcxd6 White would only be marginally better.
the lines where White had to worry about It is my experience that when players annotat­
...it)d4 attacking the queen? There were also ing one of their nice games give a variation in

62
 The Classical Varia tio n : 9 th Mo ve A lterna tives

the notes which is only marginally better, it is won't be able to resist for ever) 22...lLlg6 23
not better at all! Where is White's advantage lLlcb5 lLlh7 24 l:lfc t 1i'h4 25 h3 .ixh3! 26 gxh3
here? I don't see it. Can he even maintain the lbg5 27 .ift lLlx£3+ 28 Wh1 �5 (the f and g­
balance after 20...ltha4!? I'm not so sure. For pawns are enormous) 29 lLlxd6 £3 30 lLl£5 :X£5
example, after 21 .ia3 .ia6 22 1i'a4? loses to 31 ex£5 lLlxh3 32 .ixh3 1i'xh3+ 33 Wg1 f2+ 34
22...lik2t 23 Wh1 �3+!!. 'ii'x£2 gxf2+ 35 �xf2 'W'x£5+ and Black soon
a2) 14...g5 15 c5 axb4 16 axb4 lLlf6 17 1i'c2 won, Harestad-Rasmussen, Copenhagen 1996.
lLle8 (the usual response when White is threat­ A very typical King's Indian attack.
ening to invade on c7) 18 lLlc4 h5 19 .ib2 lLlg6 1 4. . .axb4 1 5 axb4 g5
20 l:la1 llxa1 21 llxa1 .id7 22 lLla5 'ti'bs 23
�h1 g4 with chances for both sides, Lukacs­
Xie Jun, Budapest 1992.
b) In the late 1980's and early 1990's there
were quite a few games that went 14 c5 axb4 15
axb4 dxc5 16 bxc5 lLlxc5 17 lLlc4. The open
lines on the queenside gave White good com­
pensation for the pawn. Black then made the
important discovery that he should refrain from
the exchange on b4. For example, after
14...dxc5 15 bxc5 lLlxc5 Black can meet 16 lLlc4
with 16... a4 when it is not easy for White to
generate threats. Black even has outposts of his
own on b3 and d4. In Hernandez-Frolov, Ha­
vana 1991 White preferred 16 a4 but Black had 1 6 �d2?
a little trick prepared against this: 16...lLle6! with White opts for the rather leisurely plan of
the point that 17 dxe6 1i'd4+ 18 Wh1 1i'xc3 is invading down the a-file. It is too slow. In a
good for him and that if the knight is not taken previous game, Dreev-Shirov, USSR 1988,
it will hop into d4. After 16...lik6 the game White was also in trouble after 16 c5 lLlf6 17
continued 17 .ia3 lLld4 18 .ic4 g5 19 lLle2 c6!? .id2 lbg6 18 c6 b6 19 .ie1 l:lg8 20 lLld2 h5 21
20 d6 lLlg6 with good play for Black. There is lLlb5 g4. It is clear that after one inaccurate
no need to worry about White grabbing an move (14 lLlb3?!) that the situation is already
exchange with 21 d7 as in rerum Black will becoming critical for White on the kingside.
have two pawns and a grip on the dark squares. Kasparov believes the only chance is to try and
c) 14 lLla4 is a rather crude way to supporr block it up with 16 g4.
the advance c5 and Dr. Nunn believes that this 1 6 ...tbg6 1 7 :S1 :Xa1 1 8 'ifxa1 ltlf& 1 9
decentralising move is best met by 14...axb4 15 'ifa7
axb4 c6. Now it's too late to block the kingside as af­
d) 14 1i'c2 also prepares the advance c5. ter 19 g4 fxg3 20 hxg3 lLlh5 21 Wg2 lLlgf4+1 22
Black can try and hold it up with 14...b6 but he gxf4 gxf4 Black has a crushing attack.
can also get on with his kingside attack, e.g. 1 9 g4 20 fxg4
.••

14... g5 15 c5 axb4 16 axb4 lLlf6 (it's too dan­ White can't allow the pawn to advance to g3.
gerous to take the pawn with the queen on c2) 1 9 ...ltlxg4 21 h3
17 lLlc4 h5 (17...lLlg6!?) 18 .ia3 (18 lLlb5 lik8 The pawn on h3 will rum out to be a fatal
transposes to 'a2' while 18 b5 is another try) weakness. White could have tried 21 .ixg4
18 ... g4 19 cxd6 cxd6 20 b5 l:lxa3! (rather than .ixg4 though after 22 1i'xb7 lLlh4 Black's at­
defending passively with 20...lik8 Black gives tack is decisive according to Kasparov.
up the exchange to kill off the white attack - 21 . . .ltlh6
now it's his rum) 21 lLlxa3 g3 22 b6 (White is The main reason for retreating here, as op­
reluctant to play h3 as he knows that Black will posed to f6, is to stay out of the queen's way.
be able to sacrifice a piece there one day but he 22 �e1 l:lgS 23 ltld2 �f&!

63
Pla y t h e King 's Indian

Clearing the g-file and preparing to play A fter 10 .U.ct f5 ( 1 0...c 5 1 1 dxc6 bxc6 1 2 b4
J..h4. Once the dark-squared bishops are ex­ lt:lc7 1 3 b5 d5 also offers good chances of
changed it will become easy for Black on the equality) we have:
kingside.
24 �h1
The last chance to put up any resistance was
offered by 24 it:l£3. Black should still play
24...J..h4!
24 ....i.h4 25 lBf3 .i.xe1 26 lBxe1
After 26 llxe1 it:lh4 27 l:tg1 it:lx£3 28 ..i.x£3
1i'h4 the game will also be decided by Black
sacrificing his bishop on h3.
26 ...lBh4 27 :t2 11fg5 28 lBf3 lBxf3 29
.i.xf3 .i.xh3 0-1
Ljubo, as he's commonly known, had seen
enough. A possible fmish: 30 1i'xb7 .i.xg2+ 31
.U.xg2 (31 J..xg2 f3) 3t...'iVh4+ 32 �g1 'ifet+ 33
�h2 .U.xg2+ 34 .i.xg2 1i'g3+ 35 �h1 f3 36 1i'c8+ a1) 1 1 lt:lg5 h6 1 2 lt:le6 .i.xe6 1 3 dxe6 'ifc8
lt:lg8 and Black wins. The expression 'taking 14 c5 (14 'iVb3 c6 1 5 f4 exf4 1 6 J..x f4 g5 17
candy &om a baby' seems quite appropriate in .i.g3 f4 1 8 .i.£2 1i'xe6 1 9 1i'xb7 is unclear)
relation to this game. 1 4...'ifxe6 1 5 cxd6 cxd6 1 6 it:lb5 1i'd7 17 J..b4
lt:lc6 1 8 .i.xd6 with an edge for White, Geller­
Game 20 Minic, Skopje 1 968. Ths is what most theoreti­
cal sources give but as Nunn points out,
Lukacs-Rajlich
16 ...1i'd7 is a strange move. Why doesn't Black
Budapest 2001 grab another pawn? After 16 ... fxe4 1 7 J..b4
.U.c8! it is not immediately obvious how \XIhite
1 d4 lBf6 2 c4 g6 3 lBc3 .i.g7 4 e4 d6 5 is going to win back his pawns.
lBf3 0-0 6 .i.e2 e5 7 0-0 lBc6 8 d5 !De7 9 a2) 1 1 exfS gxfS 12 lt:lg5 h6 13 lt:le6 .i.xe6
.i.g5 14 dxe6 'ifc8 1 5 1i'b3 c6 1 6 .i.h5 1i'xe6 1 7
Ths is \XIhite's 4th most common move af­ 'ifxb7 it:lf6 1 8 .i.e2 .U.£b8 1 9 1i'a6 llxb2 i s quite
ter 9 iDe 1 , 9 b4 and 9 lt:ld2. As we shall see it is good for Black, Taimanov-Fischer, Vancouver
not without danger for Black. There are a few (m/1 ) 1 97 1 .
others which occur from time to time. We shall a3 ) 1 1 1i'b3 b6 (Black prevents c5) 1 2 exfS
just take a quick look at them. gxfS 13 lt:lg5 h6 (13...it:lf6 14 f4 h6 15 fxe5
a) It has been quite a while since 9 .i.d2 was dxe5 16 c5 it:lfxd5 17 iDxd5 it:lxd5 18 cxb6
in fashion. 9...lt:lh5 is considered to be the main axb6 1 9 .U.c6 �h8 is the famous 3rd game of
line but as there are quite a few tricky sidelines the Taimanov-Hscher match; Taimanov
there for Black to avoid we are going to con­ claimed that 20 'iVh3 would have been very
centrate on 9 ... lt:le8. Ths line gained fame in good for \XIhite but subsequent analysis shows
the 1971 Candidates quarter-final match be­ that Black can hold the balance with 20....U.f6!)
tween Taimanov and Fischer which the Ameri­ 14 lt:le6 .i.xe6 1 5 dxe6 'ifc8 1 6 lt:ld5 1i'xe6 1 7
can won 6-0. Apart from the main line 10 .U.cl , lt:lxe7+ 1i'xe7 1 8 c 5+- �h8 1 9 cxd6, Taimanov­
\XIhite can play 1 0 iDe 1 f5 1 1 lt:ld3 it:lf6 trans­ Ma.Tseitlin, USSR 1 973. In the game Black
posing to Chapter 2 or 1 0 b4 f5 1 1 1i'b3 it:lf6 chose 19 ... it:lxd6 and after 20 .U.c6 .U.ad8 21
1 2 exfS gxfS 1 3 c5 �h8 14 cxd6 cxd6 1 5 .U.acl l:tfcl l:td7 22 .i.b4 White had pressure for the
J..d7 16 a4, Korchnoi-Geller, Moscow 1 97 1 , pawn. Perhaps it was better for Black to play
and now instead o f 1 6...lt:lg6 1 7 .i.b5! which 1 9 ... 1i'xd6 20 l:tfd1 c5. The bishop pair still
was slightly better for White, 16 ... a6 would have gives White compensation but Black has a solid
kept the bishop out and maintained the balance. position.

64
 The Classical Varia tio n : 9 th Mo ve A lterna tives

b) 9 �h1 is a strange idea of Larsen that was

also played a lot by the late Tony Miles. It is a
son of waiting move which frees the g1-square
for use by White's pieces. The variations
9...lbd7 10 g4! �h8 1 1 llg1 and 9...ltlh5 1 0
ltlg1 ltlf4 1 1 �£3 f5 1 2 g3 illustrate a couple of
White's ideas. After 9...lbe8 it is not so easy for
White to justify his king move. Sulava has tried
10 llg1 (Black replies 10 ... £5 of course) while
Miles preferred 10 lbe1 f5 1 1 ex£5 (1 1 £3 f4
would be like Chapter 2 except that White has
wasted a move on �h1) 1 1 ...lbxf5 (1 1 ...gxf5 1 2
f4 ltlg6 i s also playable) 1 2 ltl £3 ltlf6 13 �d3
lbd4 14 ltlxd4 exd4 1 5 lbe4 lbxe4 1 6 �xe4
'ii'h4 17 1i'c2 �£5 1 8 �x£5 l:tx£5 and Black's 9 . . . �h5
active pieces compensated for his slightly infe­ White's basic idea when playing 9 �g5 is to
rior pawn structure, Miles-Tirard, Cappelle Ia deploy both his queen's bishop and king's
Grande 2000. knight actively. Thus, after 9...lbd7 or 9...lbe8
c) 9 a4 shocked the chess world when it was White plays 10 lbd2 f6 (or 10... £5 1 1 £3 f4 1 2
introduced in the game Korchnoi-Kasparov, �h4) 1 1 �e3 f5 1 2 £3 with the ideal set up for
Barcelona 1989. The move appears to be anti­ a queenside attack. His bishop is active on the
positional as after 9...a5 it will be more difficult g1 -�7 diagonal and his knight on d2 is ready to
for White to open the queenside. 9 ... a5 was come to c4 after White has played c5. There are
indeed Kasparov's choice and after 10 lbe1 two main ways for Black to disrupt White's
lbd7 1 1 �e3 f5 1 2 £3 lbc5 (Black can also plan. The first is to play 9 ...ltlh5 as now 10
transpose to Game 6 by 1 2... f4 1 3 �£2 g5 but lbd2 is met by 10...ltlf4. That is the subject of
Kasparov wants to take advantage of White's our main game. The other is the most princi­
strange move order) 1 3 lbd3 b6 14 b4 lbxd3 1 5 pled reaction, 9...h6. White now has to give up
1i'xd3 axb4 16 ltlb5 �h8! (in order to improve his bishop (the bishop is not well place on h4
the position of the knight on e7) 17 1i'b3 ltlg8 and 10 �e3 is well met by 1 0...lbg4) but in
1 8 1i'xb4 Kasparov played the move 1 8 ...ltlf6 return he will force through c5 quicker than
and eventually won but afterwards he said that usual. After 10 �xf6 �xf6 1 1 b4 Black can
18 ... fxe4! 19 fxe4 llxf1+ 20 l:txfl �h6! 21 �£2 play for ... £5 at once or he can delay this until he
ltlf6, with advantage to Black, was more pre­ has his pieces on their optimum squares.
cise. a) 1 1...�g7 12 c5 f5 13 lbd2 fxe4 (Black is
Later White tried to rehabilitate this line by too far behind to play for an attack) 14 cxd6 (14
replacing 1 1 �e3 with 1 1 lbd3. After 1 1 ...£5 1 2 lbc4 i.£5 and 14 lbdxe4 ltl£5 15 �g4 h5 16
£3 (12 ex£5 gx£5 1 3 l:ta3, Larsen-Gabriel, Bad i.h3 a6 are thought to be OK for Black)
Homburg 1998, should be met by 13 ...f4 ac­ 14 ... cxd6 1 5 lbcxe4 ltl£5 1 6 ltlc4 (the white
cording to GM Daniel King) 1 2...�h8!? (It's not knights are ideally placed) 1 6 ... b5 17 ltla5 1i'h4
so bad for Black to block the position with 1 8 1i'd3 lbd4 1 9 lbc6 �£5 20 £3 with an edge
12 ... f4 here as White's bishop is less actively for White, Baikov-Odeev, USSR 199 1 .
placed; in that case White would try to open the b) 1 1 ...�g7 looks slow but Black wants to
queenside with b3, �a3 and b4 while Black play improve his minor pieces with ...ltlg8 and
would get on with the kingsidc attack) 1 3 �e3 ...�e7 (or maybe ... �g5). The knight on g8 will
b6 14 b4 ltlg8!? (14... axb4 1 5 ltlxb4 lbc5 is then be well placed to jump into f6 after Black
supposed to be about equal) 1 5 bxa5 l:txa5 16 has played f5 and the bishop on e7 will force
ltlb4 �b7 17 ltlb5 �h6 with a satisfactory White to come to an early decision on how he
position for Black, Bergsson-Steffansson, Reyk­ is going to conduct his queenside attack. 1 2 c5
javik 1990. ltlg8 and now:

65
Pla y t h e King 's Indian

for equality) 12 l:te1 f5 13 lLlh4 lLlf6 14 exf5 g5

15 lLlg6 lLlxg6 16 fxg6 �f5 17 �f3 'ife8! fol­
lowed by 1i'xg6 with a level game.
1 0 .�f4 1 1 �d3 �xe2+
..

11...ltJxd3 12 'ifxd3 f5 13 f3 f4 14 �h4 is

quite promising for White.
1 2 'ifxe2 h6
This is what nearly everyone plays, but per­
haps there is no need to chase the bishop away.
According to established theory 12... f5 should
be met by 13 f4 (after 13 exf5 both 13...gxf5
and 13...�xf5 lead to a roughly equal struggle)
13...h6 14 �xe7 (after 14 �h4 exf4 White wiU
have to take on e7 anyway) 14...'ifxe7 15 l:tae1
b1) 13 a4 �e7 14 c6 (14 cxd6 is met by c6!? (15 .. . fxe4 16 lLlxe4 exf4 17 lLlxf4 �f5 18
14...�xd6!) 14...bxc6 (better than 14...b6 as g4! may lead to a slightly better endgame for
now Black gets more room for his pieces) 15 White) 16 exf5 �xf5 17 fxe5 dxe5 18 ltJe4
dxc6 f5 16 �c4 (on 16 lbd2 Black should play cxd5 19 cxd5 l:lad8 20 d6! 'ifd7 21 lbdf2,
16...�g5) 16...lbf6 17 �d5 (on 17 lbd2 Black Cvetkovic-Simeonidis, Ano Liosia 1998, with
can contemplate 17...d5!?) 17...f4 (17...lLlxd5 18 an edge for White. But 19...l:tad8 is a clear mis­
lLlxd5 fxe4 19 lbd2 �f5 20 'ife2 is supposed to take which allows White to hold onto his d­
be very good for White, but I'm not so sure; pawn. After 19...1i'd7! Black has no problems.
e.g. 20...�e6 21 'ifxe4 �g5 looks like an inter­ 1 3 .i.d2
esting battle between knights and bishops while 13 �e3 has been a speciality of Pelletier. He
on 22 l:tfd1 Black can take the two knights with just wants to have a traditional queenside attack
a roughly level game) 18 lbd2 g5 (the pawn versus kingside attack race with his bishop well
sacrifice 18...f3!? is an option) 19 'ifb3, Vlas­ placed on the g1-a7 diagonal. However, I be­
kov-Umanskaya, St Petersburg 1994, and now lieve the game Pelletier-Fedorov, European
Black should play 19...'ife8 20 b5 'ifg6 in order Team Ch., Plovdiv 2003 has dented his confi-
to prevent White from exchanging bishops with dence in this variation. After 13... £5 14 f3 Fe-
21 �e6. This now runs into 21...lLlxe4! dorov produced the novelty 14...g5! This is
b2) There are various other alternatives more flexible than 14...f4 against which Pelletier
thought to be less critical than 13 a4, e.g. 13 had already won several games. Play continued
lbd2 �e7 14 c6 bxc6 15 dxc6 �g5 =; 13 Act 15 c5lLlg6 16 l:tfct l:t£7 17 l:tc2 (17 exf5 �xf5
�e7 14 'iib3 lLlf6 (14... £5 looks like the obvi­ is level according to Pelletier) 17 ...lLlf4 (this is
ous move) 15 cxd6 cxd6 (15...�xd6) 16 lbd2 the main reason that Fedorov held back with
lLle8 17 lLlc4 �g5 18 l:tc2 f5 with equal f5-f4; he can now exchange of the knight that
chances, Ftacnik-Kolev, Budapest 1993; 13 often just gets in the way of the other black
1i'c2!? �e7 14 lLlb5 �g4 15 lLlfd4 �xe2 16 pieces) 18 1i'd2 lLlxd3 19 1i'xd3 g4! 20 fxg4
lLlxe2 a6 17 lLla3 a5 18 b5! dxc5 19 lLlc4 �d6 (otherwise Black takes on f3, and perhaps also
20 b6 cxb6 21 l:tab1 gave White good compen­ on e4, opening up even more lines on the king­
sation for the pawns in Pelletier-Polzin, Bun­ side) 20...f4 21 �f2 �xg4. Black has already
desliga 2000. Black should be able to do better achieved a great deal on the kingside - he has a
than this. 14... £5 looks critical. positional advantage to go with his attacking
10 lDe1 chances - before the white queenside attack has
As previously stated 10 lbd2 lLlf4 is quite even got out of the starting blocks. He went on
good for Black but in the early days of this to win the game.
system White used to play 10 g3. This is now 1 3 . . .f5
considered harmless because of 10...h6 11 �d2 I'm giving this as the main line even though
�h3 (the immediate 11...£5 is also good enough it looks untrustworthy. That is because we are

66
 The Classical Varia tio n : 9 th Mo ve A lterna tives

going to re-write existing theory. Unfonunately, �b5) appears very dangerous for Black the
though, not in Black's favour. The alternatives: main line ran 20...1i'd8 21 11'£3 �h3 22 1i't7+
a) 1 3...g5 prevents White from playing f4. �h8 23 hxg3 �xft 24 �e3! �e5 (24...�xe3??
Now 14 h4? g4 1 5 f4 gx£3 16 1i'xf3 f5 was 25 ...g7 mate) 25 �f6 �xf6 26 1i'xf6+ �g8 27
good for Black in OU-Shirov, Tilburg 1 992, so 'ife6+ <lr'fB 28 'ifxh6+ �e8 when both 29 e5 and
White usually plays 14 g4 to prevent Black from 29 1i'h5+ lead to a draw but no more.
playing fS. After 14...�g6 1 5 £3 �f4 1 6 �f4 But then GM Nataf suggested a very inter­
exf4 17 tt'xl1 c6 1 8 �c3 White probably has a esting improvement for White: 2 1 �b5!? �h3
very faint edge as with such a blocked pawn 22 �d4! g2+ 23 1i'xg2! �xg2+ 24 �xg2. White
structure it is better to have a knight than a has just two pieces for the <:JUeen but with
bishop. moves such as �6, �f6+ and �c3 about to
b) 1 3 ... c6!? and now: swamp Black he has a massive attack. Nataf
b1) 1 4 f4 exf4 (14 ... f5 is all right for Black suggested that Black might be able to try
according to Nunn) 1 5 �xf4 cxd5 1 6 cxd5 f5 is 21 ...�xb2 but this appears to lose to 22 11'£3!
unclear because 17 e5 can be met by 17 ... g5 1 8 (see Line 2 in next note).
exd6 �g6 1 9 �e3 lieS. Why, then, you may ask have I awarded 20
b2) 14 llact f5 1 5 f3 and now Ftacnik gives �b5 an '!' if 20 lift is so strong. The reason is
1 5...g5 1 6 �t2 �g6 1 7 g3 as unclear. that in a game between Prakash and Konguvel
1 4 f4 exf4 (Na!,rpur 1999) Black met 20 lift with the
14 ... fxe4 1 5 �e4 �fS 1 6 fxe5 (16 �c3!?) amazing novelty 20...�h3!. White must take the
16 ... dxe5 1 7 �c3 but Nunn suggests that Black queen but after 21 llx£8+ l:r.xfB it turns out that
try 1 5...b5. he cannot hold on to his own <:JUeen. In the
1 5 ll:lxf4 g5 1 6 ll:lh5 game Prakash played 22 �xg3 and after
16 �6?! �xe6 17 dxe6 f4 18 g3 �g6 1 9 22...llt2 23 'ifxt2 �xt2 24 tt'x11 �d4 25 �e3
gxf4 �xc3! 20 bxc3 gxf4 21 'it>h1 ...f6 22 llg1 �xe3 26 �xe3 <lr't7 27 'it>g1 �g6 he had to
<lr'h7 was good for Black in Blees-Klarenbeek, fight hard to avoid defeat. 22 hxg3 may have
Heraklio 1 993. been better, though here too 22...llt2 regains
1 6 . . .�d4+ 1 7 �h1 f4 1 8 g3 the <:JUeen. So that's why White was in need of
18 ...�h3 19 gxf4! is a powerful exchange an improvement and why 20 �b5 is deserving
sacrifice. of an '!'
1 8 ...fxg3 1 9 l:l.xf8+ "ifxf8 20 ..."iff2!
The bishop on d4 is attacked but it's too
dangerous for Black to move it:
a) 20 ... �b6 and now 21 lift �h3 22 llx£8+
llxfB 23 �xg3 llt2 24 1i'xt2 �xt2 25 �xc7 is
good for White (though it is still complicated
after 25 ...�g6) but the simple 21 hxg3 looks
very strong. The black kingside is in trouble
with the bishop off the long diagonal.
b) 20...�xb2 21 lift 'ifd8 22 ...£3 �h3 23
1i't7+ �hB 24 llf6 wins for White. This was not
possible in the similar variation in the previous
notes (20...1i'd8) as with the bishop on d4 Black
could play 24...g2 mate here.
21 "ifxf2
20 ll:lb5! And not 21 �xd4?? g2 mate!
A very interesting moment. Previously 21 ...�xf2 22 ll:lxc7!
White used to play 20 lift and, <:Juite under­ The right pawn to take. In Prakash-Kong­
standably it was thought that Black had to uvel, Indian Ch. 2000 (yes, them two again)
move his <:JUeen. As 20......e8 21 llf6 (or 21 White played 22 hxg3? and after 22...�b6 23

67
Pla y th e King 's Indian

i.c3 a6 24 �4 �f7 25 �g2 �g6 26 l:lht i.d7 25 �xh3 .J:.xc4 26 .tel

Black was better and eventually won. White retains some advantage as his pieces
22 ... ..ih3! 23 hxg3 are more active.
Of course not 23 tt:'lxa8?? g2 mate. 26 .....td4 21 :n ?!
23 ...:c8 Was it really necessary for White to allow his
queenside pawns to be weakened 27 i.xd4
llxd4 and now either 28 llct or 28 lift look
more natural.
27 .....txc3 28 bxc3 g4+ 29 �g2 l:txc3 30
�f6+ �h8 31 �xg4 l:tc2+ 32 �3 �g7 33
�f6 :xa2 34 �8+ �g6 35 �xd6 .J:.e2 36
.l:.f4 b6 37 �f7 �g7 38 �8 �c8 39 �6+
�g6 40 .J:.f8 �6 41 �f4+ �g7 42 .l:.d8
�f7 43 �5+ �h7 44 l:td7 �g8 45 :xa7
�5+ 46 �g4 �e4 47 �3 �3 48 �f6+
�8 49 l:tc7 l:te1 50 �h7+ �g8 51 d6
.l:.f1 + 52 �g2 l:td1 53 d7 b5 54 �f6+ �8
55 .J:.xc3 �e7 56 l:tc8 .J:.d2+ 57 �h3 1 -0
So although the main line of 9...tt'lh5 now
24 �h2?! seems quite good for White there are, between
This seems good enough for an edge but the moves 9 and 1 4, plenty of alternatives in the
critical line must be 24 tt:'lb5 l:lxc4 25 �h2! notes that Black can choose from. Take your
Now Black would love to keep his powerful pick.
bishop but its obvious retreat squares are
mined. Both 25...i.g4 and 25... i.d7 lose to 26 Summary:
tt:'lf6+. That leaves: 1) Not so long ago 9 tt:'ld2 became the main
a) 25 ... i.c8. Now White has many tempting line of the whole King's Indian with many of the
continuations, e.g. 26 tt:'lxd6 llc2 27 tt:'lxc8 world's top players taking part in a theoretical
tt:'lxc8 28 llct ! llxd2 29 :XeS+ �f7 30 �h3 battle that raged here during the late 1 980's and
and the central passed pawns should prove early 1 990's. Those days have gone and now it is
decisive; or 26 i.c3 i.c5 27 b3 llxc3 28 tt:'lxc3 lying relatively dormant but who knows what the
i.d4 29 l:lct i.g4 30 tt:'lb5 i.e5 31 tt:'lxd6! future holds for this rich variation.
i.xh5 32 tt:'lxb7 and White should win. 2) Black should not challenge White to the
b) 25 ...g4 locks the bishop on h3 out of the usual direct kingside attack v queenside attack
game. White should not play 26 tt:'lxd6? on race as White is clear favourite in this line. The
account of 26.. .llc2 but both 26 llct and 26 reason for this is that the knight on c4 is in­
i.c3 are good for him. Note that Black can't credibly powerful. Instead Black should take
play 26.. .llxe4 because this too loses to 27 precautions on the queenside (9 ... a5, maybe
tt:'lf6+. What an annoying creature that knight ... c7-c6 at some point) whilst still, of course,
on h5 is. pursuing his ultimate aim of delivering check­
c) 25 .. .llc2 26 �xh3 llxd2 27 tt:'lxd6 is also mate to the white king.
very good for White. 3) 9 i.g5 is in my opinion an underrated line
If these variations prove correct then the where Black has to be quite careful. It is very
whole line seems unsound for Black. He should easy to slip into a bad position if you are unfa­
choose one of the alternatives suggested earlier. miliar with its nuances as many Black players
24...:Xc7 are. The existing theory is also quite poor and
The only way to keep the bishop was 24...g4 some of it has been re-written in Game 20. It's
(don't forget his retreats are mined) but 25 worth taking a close look at this game, not only
tt:'lxb5 llxc4 transposes to line 'b' above, which for its theoretical value but because it also in­
was good for White. cludes some fascinating variations.

68
 CHAPTER FIVE I
The Classical Variation :
7 ltJa6
. . .

1 d4 tt:lf6 2 c4 g6 3 tt:lc3 �g7 4 e4 d6 5 play: ...c7-c6 followed by ...lbc7 and ...lbe6. The
tt:lf3 0-0 6 �e2 e5 7 0-0 lba6 move 7 ...lba6 is also more flexible than the old
If the main lines with 7 ... 0JC6 arc too sharp 7 ...lbbd7 line as it doesn't interfere with the
or too theoretical for you then this chapter is development of the rest of Black's queenside.'
your refuge. 7 . ..lt�a6 has completely displaced White's most popular response to 7 ...lba6 is
the old traditional lines such as 7 ... lbbd7 and 8 ..ie3 and that is tackled in Games 21 -25. This
7 ... exd4 as Black's main alternative to 7 ... lbc6. position is important as whilst you may prefer
The line simply didn't exist until the early the variation 7 0-0 lbc6, I am also recommend­
1 990's but its popularity grew as many players ing that Black meets the Gligoric System, 7
became weary of suffering in the Bayonet At­ ..ie3, with 7 ... lba6 which then transposes here
tack. It was also part of a general revolution in after 8 0-0. Nowadays the vast majority of
King's Indian thinking as Black started to play Games reach the positions examined in Games
lba6 in many lines where he had not done so 23-25 where I examine both the old (if we can
previously (Four Pawns Attack for example). call anything old here) 1 t ... h6 and the new
In SOKID I asked the question, 'Isn't the 1 1 ...f6. Both seem viable.
knight badly placed on the edge?' and I an­ In Games 26 and 27 we look at the main al­
swered myself as follows: 'On a superficial ternative 8 l:r.et (others are briefly mentioned in
reading of the position the knight is indeed the Game 21 notes). Black's main choice here is
badly placed on the edge of the board. The whether to exchange on d4 or encourage White
move would appear to go against the basic to block the centre with d5. The former (C �arne
chess principles of controlling the centre. How­ 27) often leads to a drawish position while the
ever, once we look a little more closely we can latter (Game 26) can lead to a dour struggle.
see that this is not the case. White will not be
able to hold the tension in the centre for ever
Game 21
and at some point he is either going to take on
Rogers-Gallagher
e5 or play d5. In both cases the knight on a6 is
poised to jump into c5, one of its best squares
Bundes/iga 1997
in the King's Indian. On other occasions Black,
himself, may relieve the central tension by play­ 1 d4 tt:lf6 2 c4 g6 3 tt:lc3 �g7 4 e4 d6 5
ing ... e5xd4 and here, too, the square c5 be­ tt:lf3 0-0 6 �e2 e5 7 0-0 lba6 8 �e3
comes available to the knight. I f for some rca­ Otherwise:
son Black is unable to play ... lbc5 then there is a) 8 dxe5 dxe5 9 'ifxd8 llxd8, as usual, of­
an alternative method of bringing it back into fers White nothing.

69
Pla y th e King 's Indian

b) 8 d5 ttk5 is considered in Game 29. White has a choice of evils and settles for
c) 8 J..g5 has earned a few supporters. That defending the e-pawn in this manner. The
is because the variation 8 ... h6 9 ..i..h 4 'iVe8 1(J problem with 1 3 f3 is that having already played
i.xf6 i.xf6 1 1 c5! is quite unpleasant for Black. h2-h3 this would leave the kingside full of
It is much simpler to play 9 ...g5 10 ..i..g3 (1 0 holes.
dxe5 lt::Jh 5 will transpose) 10 ...lt::Jh 5 when 1 1 d5 1 3 ...�h7!
transposes into a harmless variation of the Pet­ It is because of this fme move that White
rosian and 1 1 dxe5 lbxg3 1 2 hxg3 dxe5 is level. players nearly always play 10 dxe5 these days.
d) 8 l:te 1 , the main alternative, is covered in That in itself can be considered a small feather
C:rames 26 and 27. in Black's cap as White would prefer to main­
8 ...llJg4 tain the status quo in the centre.
Disruption is the name of the game. White 1 4 1le1
should not be allowed to develop his forces White players are reluctant to play the ugly
harmoniously without a snuggle. 1 4 h4 but it seems as good as anything else. The
9 ..ig5 Wee position after 1 4 ...'ire7 1 5 g3 c6 is probably
In the early days of this variation Black often about equal.
played 9 ...f6 but this was discarded because of 1 4...�g5
the continuation 10 J..c t c6 1 1 h3 lt::!h 6 12 c5!. A strong case could also be made for start­
10 h3 ing with 1 4...'ire5!?, e.g. 1 5 ..i..e3 lbg5 1 6 'ird2
10 dxe5 is the subject of Games 22-25. c5! (but not 1 6 ... J..xh3 1 7 i.e2!) looks quite
1 o . . h6! 1 1 ..ic1
. good while 1 5 lt::!b3 lbg5 1 6 ..i..g4 ..i..xg4 1 7
Also: hxg4 'ire6 1 8 'iVe2 l:tae8 was a t least equal for
a) 1 1 i.h4 exd4 t 2 lt::Jxd4 lt::J f6 and now: Black in Lassila-Yrolja, Finnish Team Ch. 2002
at) 13 ..i.. f3 lt::!h7 1 4 l:te1 lt::!g5 1 5 ..i..xg5 hxg5 1 5 ..ig4 ..ixg4 1 6 hxg4 �5 1 7 ..ixg5!?
1 6 'ird2 c6 1 7 l:tad1 'iVe7 1 8 ..i..g4 ..i..xg4 1 9 After the game the Australian grandmaster
hxg4 i.e5 was about level in Nakamura­ Ian Rogers made an interesting comment. He
Perelshteyn, Bermuda 2003. This is similar to said he couldn't decide whether to play 1 7 f3 or
the main game. 17 ..i..xg5 but settled on the text as King's In­
a2) 13 l:te 1 g5 14 ..i..g3 lt::!xe4 1 5 ..i..d3 lt::!xc3 dian players tend to overestimate the impor­
16 bxc3 'iVd8 and it's doubtful that White has tance of the dark-squared bishop (well, we cer­
enough play for the pawn, Groszpeter-Galla­ tainly respect it). In other words he was trying
gher, San Bernardino 1 990 (I won quite easily). to provoke me into playing recklessly for a win.
b) 1 1 hxg4 hxg5 (1 1 ...exd4!?) 12 dxe5 dxe5 Well, he succeeded but to be honest I consid­
13 lt::Jxg5 'ire7 14 lt::Jh3 c6 gives Black more ered the position to be about level around here.
than enough compensation for the pawn. 1 7 . . . hxg5 1 8 Wfd2 c6 1 9 1lad1 lld8 20 �f3
1 1 ...exd4! 1 2 �xd4 �f6 1 3 ..if3 Wfe6!

70
 The Classical Varia tio n : 7 . . .&iJ a 6

Searching for counterplay on the queenside

rather than defending the position passively Game 22
21 'ifxg5 'ifxc4 22 'ifh4 tlJe6 23 .l:le3 .l:lfe8!
Jackelen-Gallagher
It's crucial to give the king a square on £8.
24 g5?! d5! Bundesliga 2002
Black must break out in the centre before
White plays l2Jh2-g4. 1 d4 lLlf6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5
25 ltJd2? lLlf3 0-0 6 �e2 e5 7 0-0 lLla6 8 �e3 lLlg4 9
It's only after this move that White is in real �g5 'ife8 1 0 dxe5 dxe5 1 1 ltJd2
trouble. 25 l2Jh2? d4 26 e5 dxc3! 27 'it'xc4 1 1 h3 is the subject of the next three games.
l:lxdl+ 28 llJfl cxb2 is not good either but 25 1 1 .. .f6!?
exd5! 'it'xh4 26 l2Jxh4 l2Jxg5 27 l:lxe8t l:lxe8 28
dxc6 bxc6 29 l2Jf3 should lead to a slightly
worse but tenable endgame for White.
25 ...'ifb4 26 .l:lh3 'ifxb2!
It's just a check on h7, nothing to worry
about.
27 exd5 cxd5 28 lLldb1 'ifb6 29 .l:lhd3 lLlc7

The latest fashion and it is now beginning to

look like Black's most reliable move. In fact it
looks like the majority of White players are just
giving up 1 1 l2Jd2 because of it. The older line
is 1 1 ...h6 1 2 ..ih4 l2Jf6 13 l2Jd5 with Black now
having the choice between 1 3...g5 and the
queen sacrifice 1 3 ... 'tlfd8 1 4 f4 l2Jxd5.
Black has won a pawn and now he has to 1 2 �xg4
defend for a few moves I don't like this anti-positional move. White
30 lLld2 Wc6! needs his bishop to cover the sensitive dark
Preventing l2Je4. squares. However, 1 2 i.h4 doesn't appear too
31 .l:lc1 dangerous for Black either, e.g.
31 l2Jde4 fails to 31 ...dxe4 32 l:lxd8 ..ixc3. a) 1 2...l2Jh6 1 3 a3 c6 (1 3 ...l2Jc5 14 b4 l2Jc6 1 5
31 . . .d4 32 lLle2 'ifa6! 33 lLlf4 ltJd5 34 .l:lf3 l2Jb3 c6 i s similar) 1 4 b4 tDc 7 1 5 l2Jb3 lD fl 1 6
lLlc3! 35 .l:lh3 'tlfc2 (16 l:la2!?) 1 6. . .l2Je6 1 7 l:lad1 l2Jd4! 1 8
Or 35 g3 l2Je2+ 36 l2Jxc2 'tlfxe2 37 'tlff4 l:le7 'tlfb 2 (18 l2Jxd4 exd4 1 9 l:lxd4 g5 20 ..ig3 f5
and wins and Black threatens both .....ixd4 and ... f4)
35 ...tiJe2+! 36 lLlxe2 'ifxe2 37 .l:lc7 1 8...l2Jxb3 1 9 'it'xb3 ..ie6 with a level game,
This loses but so docs everything else. Huzman-Gallagher, Crete (rapid) 2002.
37 . . .'ifxd2! 38 'ifh7+ lii>f8 39 .l:lxf7+ b) 1 2 ... h5 1 3 a3 ..ie6 (probably better than
39 l:lf3 'it'et+40 �h2 'tileS+ 1 3. .. c6) 1 4 b4 l:ld8 and Black can follow up
39 ...�xf7 40 .l:lf3 + �e6 41 'ifxg6 + �d7 with l2Ja6-b8-c6 and hopefully into d4.
0-1 1 2 ...fxg5
White resigned as the checks will run out in At first glance the pawn structure may look
a few moves. favourable for White but the reality is different.

71
Pla y t h e King 's Indian

Not only is there the usual hole on d4 for easy plan

White to worry about (remember Black can 1 8 1We2 'ilfe6 1 9 lL\a4 ..if8 20 c5
play c6 to cover d5) but the f-fi.le has also been White is dreaming of playing lLla5-c4 but
opened for Black's rooks. In addition, having this wiU remain just a dream as the c5 pawn is
doubled g-pawns means that Black will be able in need of protection.
to use the front one aggressively without weak­ 20 . . ...ie7 21 fll?
ening his own king position. Black's pawn A slightly panicky reaction due to dissatisfac­
structure is very dynamic and it might only tion with his game. He didn't want to sit there
begin to tum in White's favour in a static end­ doing nothing while Black built up his position.
game position. 21 . . .gxf3 22 llxfl llxfl 23 1Wxf3 1Wc4!
1 3 lL\b3 This is the problem with f3: the queen was
1 3 a3 tlX5 1 4 b4 .ixg4 1 5 1i'xg4 lL!e6 1 6 c5 needed on e2 to prevent this invasion. I sup­
h5 1 7 1i'e2 lL!d4 1 8 1i'c4+ �h7 1 9 a4 g4 20 pose White could have played 23 gxf3 but that
lL!b3 c6 with good play for Black, Sasikiran­ is just awful.
Thipsay, Nagpur 2002. StiU there was no need 24 1ld7
for 21 lLle2?? 1i't7! after which Sasikiran re­
signed.
1 3 ...c6 1 4 ..ixc8 'ifxc8 1 5 'ilfd6
1 5 1i'd2 1i'e6 1 6 1i'xg5 1i'xc4 is supposed to
be about level but I would prefer to play 1 5 ...g4
with similar play to the main game.
1 5 ...1lf7 1 6 llad1 'ife8!
Black's last two moves show that the white
queen is redundant on d6 and there is nothing
better for her than to beat a hasty retreat
1 7 'ilfd2
1 7 l:ld3?? .i£8 and 1 7 tlX5?? .i£8 are not
recommended.
1 7 ...g4!
White was relying on this move whilst I ac­
tually thought that the best chance was 24
1i'g4!. After 24 ... .ixc5+- 25 lL!a.xc5 lL!xc5 26
lLixc5 1i'xc5+- 27 �h1 1i'c4 28 b3 1i't7 29 h3
l:te8 Black has an extra pawn but wiU have to
overcome considerable technical difficulties.
24. . ...ixc5 + ?
Based on an oversight otherwise I would have
played 24.. .lle8! as 25 1i'g4 1i'xa4 26 1i'e6+ �£8
27 1i'xe5 and now 27....ixc5+ is an ending with
an extra pawn but 27 ...1i'xb3!! is a killer as 28
axb3 .ixc5+ regains the queen with an extra
piece, while after both 28 1i'h8+ and 28 1i'f4+
Black can interpose his queen to block the check.
The white knights are passive and it is much 25 lL\axc5?
easier for Black to improve his position than The obvious knight, but the wrong one! I
White. For example, he can move his queen to had seen the variation 25 lL!bxcS! lL!xcS 26 b3
e7 or e6, double rooks on the f-file, advance his 1i'ct+ 27 l:ldt 1i'c2 but stopped my analysis
h-pawn, activate his bishop on g7 and bring his here, not noticing that 28 lift! leads to a
knight on a6 to the tasty outposts in the centre roughly level game.
and on the kingside. White, meanwhile, has no 25 ... lL\xc5 26 lL\xc5 llf8!

72
 The Classical Varia tio n : 7 . . .li:J a 6

The point. will discover that 1 4 lbd5 is not terribly worry­

27 1i'd1 ing for Black whilst in this game we shall look
The only move. 27 "iVd3 "ifxcS+ 28 �h1 at the other main line where White advances his
"ifct+ 29 "ifd1 l:.ft+, 27 1ib3 l:tft mate and 27 <:JUeenside pawns.
lbd3 l:tx£3 28 gx£3 "ifxa2 all win for Black. 1 4 a3 c6 1 5 b4 lt:lh5
27 ...1i'xc5 + 28 �h1 1i'c2! Black's previous two moves have solidified
A neat way to force the exchange of <:JUeens. and co-ordinated his position. Now he is ready
Black would be unable to win with <:JUeens on for the standard ...lLlh5-f4 manoeuvre. Black's
as his king is too exposed. move order is <:juite imponam. The text is pref­
29 h3 1i'xd 1 + 30 lbd1 llf7 31 lld8+ erable, for example, to 1 5...l:.d8 16 "iVc2 lLlhS
Otherwise ...�f8-e7. when 17 l:tfd 1 gives White chances of an edge.
31 ...�g7 32 �g1 � 33 lla8 a6 34 �2 1 6 lle1
Falling for my trap but 34 l:td8 is met by The other move White plays here is 1 6 cS
34...l:td7 and other moves by 34... ..t>e6. and now Black should avoid both 16 ... lLlc7,
34 ...�e7+ 35 �e3 llf8 36 lla7 llb8! 0-1 which allows 17 1i'd6, and 1 6 ...ll:Jf4, which is
A rather sad end to the white rook. better for White after 1 7 .i.xa6, in favour of
16 ... l:td8! 17 "iVct ..t>h7 with about e<:jual
Game 23 chances. Black is no longer worried about
.i.xa6 because of the passive white queen on c1
Van Wely-Gallagher
(it normally goes to a4 after �xa6 to attack the
Bie/2000 weak pawns).
1 6 ...lt:lf4 1 7 �xf4
1 d4 lt:lf6 2 c4 g6 3 ltlc3 �g7 4 e4 d6 5 A double-edged move. The solid 17 �ft is
ltlf3 0-0 6 �e2 e5 7 0-0 ltla6 8 �e3 lt:lg4 9 played more often and now:
�g5 WeB 1 0 dxe5 dxe5 1 1 h3 h6 a) Many Black players play 17 ..."iVf6 and then
1 1 ... f6!? (Game 25) is currently fashionable. 1 8 l:ta2! l:td8 1 9 l:td2 may indeed be slightly
1 2 �d2 better for White.
The position after 12 hxg4 hxgS has already b) But 1 7 ...l:.d8! is a much better and after
been discussed in the 1Oth move notes to 1 8 "ifc2 it is probably best to play 18 ...lLle6
Game 21 (line 2). (1 8 ..."iVf6 is also interesting) and then into d4.
1 2 . . .ltlf6 1 3 �e3 We7! For example, 19 l:.ad 1 lbd4 20 �xd4 exd4 21
lLla4 cS! looks like an edge for Black.
1 7 . . .exf4 18 e5 �f5
Black must stop lLle4. Of course 1 8... -i.xeS
runs into a pin after 1 9 .i.d3.
1 9 c5 lt:lc7!

Black has tried many moves in this position

(and so have I) but over the last few years the
text has finnly established itself as the most
respectable. That is because it makes it difficult
for White to play c4-c5. In the next game you

73
Pla y th e King 's Indian

Better than the 1 9 ...l:lad8?! that was played Certainly not the horrible 14...tt::\xd5 1 5 cxd5
in Gallagher-Nwm, Bundesliga 2003 (yes, I was after which the pawn structure is very much in
also most surprised to find myself on the other White's favour. There is no need to take the
side of this variation). The reason Nunn re­ knight. it is much better to drive it away with
frained from 19 ...lbc7 was 20 1i'd6 but after ... c7-c6. Meanwhile, Black is threatening to take
20...J:.fe8 he failed to appreciate that 21 l:lad1 is the pawn on e4.
just met by 21...�xe5 while 21 �c4 l:lad8 22 1 5 lDxf6+
1i'xe7 l:lxe7 23 ii:k4 �xe4 24 l:lxe4 g5 is com­ Alternatively:
fortable for Black. a) 1 5 1i'ct �h7 (1 5 ...tt::\xe4 1 6 �xh6 is quite
20 .i.c4 11ad8 21 'ifc1 �6 risky for Black) 1 6 l:ld 1 tt::lxe4! 17 tt::lb6 axb6 1 8
Black would like to exchange off the knight l:lxd8 l:lxd8 gives full compensation for the
on £3 (by playing either tt::lg5 or tt:Jd4) which is queen. Therefore, in Wells-Gallagher, Bundes­
doing a good job protecting the white centre liga 1 999 White tried 17 c5 and after 1 7...�d7
and kingside. 1 8 c6 bxc6 19 �xa6 cxd5 20 l:lxd5 c6 21 l:ld1
22 lDe4 1i'c7 22 'iVc4! f5 23 l:lacl he probably had just
22 �xe6 fxe6 23 1i'xf4 �xh3 is nothing to about enough compensation for the pawn ('/2-
worry about. '12 in 49 after many adventures).
22 .....be4 23 :Xe4 lljg5! 24 lDxg5 'ifxg5 b) 1 5 1i'c2 c6 1 6 tt::lx f6+ 1i'xf6 1 7 l:lfd1 1i'e7
25 'ife1 :tea 26 e6 �h7! is very comfortable for Black. I annotated my
Now the draw is inevitable. nice game with Tukmakov in SOKID so I'U
27 11d1 :Xd1 28 'ifxd 1 fxe6 29 :Xe6 :Xe6 just give the moves here. Play continued (after
30 .i.xe6 'ife5 31 .i.cB b6 32 cxb6 axb6 33 1 7...1i'e7) 1 8 c5 (the rather pathetic 1 8 1i'd2
'ifc1 1/:z-'/z �h7 1 9 1i'd6 occurred in Gofshtein-Tkachiev,
Paris (rapid) 2001 and White managed to hold
Game 24 the ending with some difficulty) 1 8 ... tt::lc7 19 b4
'ith7 20 a4 f5 21 b5 fxe4! tt:Jd5 23 tt::lxe4 �f5
Knott-Gallagher
24 l:la3 l:lad8 25 £3 tt::lxe3 26 l:lxe3 l:lxd 1+ 27
British Ch., Torquf!} 2002 �xd 1 l:ld8 28 bxc6 bxc6 29 �e2 h5! 30 �d3
�h6 31 l:le1 l:ld5 32 �h1 'iVh4 33 l:le2 'iVd8!
1 lDf3 d6 2 d4 lDf6 3 c4 g6 4 lDc3 .i.g7 5 34 �c4 l:ld1+ 35 �h2 1i'd4! 36 l:l£2 l:lct ! 37
e4 0-0 6 .i.e2 e5 7 0-0 lDa6 8 .i.e3 lDg4 9 1i'b3 �xe4 38 �g8+ �h8 0-1 Tukmakov­
.i.g5 'ife8 1 0 dxe5 dxe5 1 1 h3 h6 1 2 .i.d2 GaUagher, Baste 1 999.
lDf6 13 .i.e3 'ife7 1 4 lDd5 1 5 ...'ifxf6 1 6 c5 lDb8
The most obvious move which some theo­ The knight is repositioning itself on c6 and it
retical sources claim gives White an edge must take this route as 1 6 ... tt::lb4? 1 7 1i'd2 costs
14 .....'ifd8! a pawn.
1 7 b4
1 7 1i'b3 tt::\c6 is about level so the text is de­
signed to prevent the knight &om settling on
the c6-square
1 7 ...86!?
A novelty I found over the board. At first I
wanted to play 17 ...tt:Jc6 1 8 b5 l:ld8?! (the im­
mediate 18 ... tt:Jd4 is better) 1 9 1i'ct tt:Jd4 but I
wasn't too happy with the position after 20
tt::lxd4 exd4 21 �xh6 d3 22 �xg7 �xg7 23
�£3. I realised, though, that the possibility to
open the a-file would make all the difference to
this variation so that is how I found 1 7 ...a6.
18 a4

74
 The Classical Varia tio n : 7 . . J[J a 6

1 8 1i'a4 tDc6 1 9 b5 tL\d4 20 .i.xd4 exd4 21 24...'ii'xe5 because o f 25 .i.xd3! l:.xd3 26 l:.aet!,
bxa6 1i'f4! is fine for Black. but instead 24...'ii'e7! 25 .i.£3 axb5 26 axb5
1 8 ...ttJc6 l:.xa 1 27 l:.xa1 'ii'xe5 wirh a clear advantage.
21 ...exd4 22 e5 'ife7 23 bxa6
White tries to avoid opening rhe a-file (23
'ii'c2 axb5 24 axb5 l:.xal 25 l:.xal .i.f5 is excel­
lent) but runs into a powerful pawn sacrifice.
23 . . .d3!

1 9 b5
A year later in AI.Spielmann-Gallagher,
French League 2003 I reached rhe same posi­
tion again. Play continued 1 9 l:.bt l&?h7!?
(1 9 ... l:.d8 20 1i'ct tL\d4 21 l2Jxd4 exd4 22 .i.xh6
is good for White) 20 1i'ct 1i'e7 (now Black is 24 axb7 .bb7 25 �d1 d2!?
ready to play ...t7-f5 against quiet moves; I Maybe it would have been simpler to just
didn't want to play ...tDc6-ttJd4 until White play 25....i.xe5 as 26 l:.et is impossible on ac­
pushed me and 20....i.e6? 21 l2Jg5+! hxg5 22 count of 26....i.xf3 followed by 27...d2.
.i.xg5 is a typical trap for rhis variation rhat 26 'ifa3
Black has to avoid) 21 b5 axb5 22 axb5 tL\d4 23 My idea was to meet 26 l2Jxd2 wirh
.i.xd4 exd4 24 .i.d3 and alrhough rhe position 26....i.xg2 (and not 26....i.xe5 27 l:.el .i.h2+ 28
isn't quite as tremendous for Black as I first �ft .i.xg2+ 20 l&?xg2 'ifxel 30 lt:J£3! when
rhought (open a-file, c3-square, bishop pair and White escapes into an ending) 27 �2 'ifg5+
weakness on c5) he is at least not worse. I even­ and 28 .. .llxd2 which is very good for Black.
rually won in 54 moves. After rhe text it is time to cash in and take rhe
1 9 ...l:l.d8 exchange rhat is on offer. Black has a decisive
It was also possible to exchange on b5 and advantage. The remaining moves were:
at before playing ...tDc6-d4. 26...�a6 27 'ife3 .bf1 28 �xf1 .l:l.a5 29
20 'ifc1 ? .l:l.a2 'ifxc5 30 .l:l.xd2 .l:l.xd2 31 'ifxd2 �xe5
Knott was going for rhe variation I gave in 32 'ifxh6 'ifc4+ 33 �e2 'ifxa4 34 g3 'ife4
the 17rh move notes (which he told me after­ 35 'ifd2 .l:l.a3 36 'ifd8+ �g7 37 We7 .l:l.a1 +
wards had been recommended somewhere as 38 �g2 'ifxe2 39 'ifxe5+ Wxe5 40 �xe5
good for White) but doesn't appreciate how rhe �6 0-1
position of rhe a-pawns has changed every­
rhing. He should have played 20 1i'c2 wirh an Game 25
unclear/equal game after 20 ... tt:Jd4 21 .i.xd4
Grooten-Motylev
exd4 22 bxa6 bxa6 23 l:.ad 1 .
20. . ..!tJcl4 2 1 �xd4 Hoogovens 2003
Now he realised rhat his intended 21 l2Jxd4
exd4 22 .i.xh6 is just good for Black after 1 d4 �f6 2 c4 g6 3 �c3 �g7 4 e4 d6 5
22 ... .i.xh6 23 'ifxh6 d3 because of rhe pin on �f3 0-0 6 �e2 e5 7 �e3 �6 8 0-0 �4 9
rhe a-file. Note rhat on 24 e5 Black doesn't play �g5 'ireS 1 0 dxe5 dxe5 1 1 h3 f6!?

75
Pla y th e King 's Indian

a) It was assumed that White had some ad­

vantage after 16 lt:lxc7 lt:lxc7 1 7 ...xc7 until
Black discovered 17 ... ltld8!. This excellent
move allows the black pieces to find new
squares that suit the altered structure. The
knight is going to c6, the rook to f7 and the
bishop to e6 and everything will co-ordinate. A
couple of the latest examples:
a 1) 1 8 �c4+ �e6 1 9 l:lac1 l:lfl 20 �xe6
...xe6 21 ...c4 ...xc4 22 l:lxc4 lt:le6 23 l:lfc1
lld7 24 �fl �fl 25 �e2 �£8 26 a4 b5 27 axb5
l:lb8 28 g3 :Xb5 29 �c3 l:ldb7 30 l:la 1 '12- 'lz,
Eljanov-Efimenko, Ukrainian Team Ch. 2004.
a2) 1 8 l:lfc1 �e6 1 9 �e3 l:lfl 20 1i'c3 lt:lc6
This move has become more and more 21 b4 l:ld7 22 b5 ltld8 23 a4 lt:lfl 24 ltld2 f5 25
popular over the last couple of years and it is f3 ltld6 with an active game for Black ('12-'/z in
cenainly more fashionable than 1 1 ...h6 at the 64 moves), Volkov-Nakamura, HOE World
moment. Somewhat surprising, as White has Ch., Tripoli 2004.
never proved anything against 1 1 ...h6. Probably b) The other critical line is 1 6 �b4 after
this line just suits the style of King's Indian which 1 6...c6 1 7 �x£8 (17 lt:le7+ �h8 1 8 ltlxc8
players better. c5 is a key line when 19 �xeS :Xc8 20 �x£8
1 2 �d2 llJh6 1 3 �e3 l:lxc1 21 �xg7+ �7 22 l:laxc1 ltld6 is at least
White's main idea in this line is to play c4-c5 equal for Black) 1 7......xfB 1 8 lt:le3 ltld6 1 9 1i'c2
which he very often follows-up with �xa6 to leads to an unbalanced position with roughly
cripple Black's queenside pawn structure. In the equal prospects. Of course I would prefer to
early days of the variation they played the text play Black but fully accept that this is because
in order to prepare c4-c5 until some bright of my King's Indian bias. In fact, I don't even
spark realised that 1 3 c5 could be played imme­ consider Black to be material down here as I
diately as a pawn sacrifice. For a while Black work to the rule of thumb that dark-squared
players just took White at their word and de­ bishop + pawn = at least rook in a typical
clined the offer. But this wasn't altogether satis­ King's Indian middlegame. Let's take a look at a
factory as the tempo saved on 1 3 �e3 was couple of examples:
enough to make life unpleasant. But then some b1) 1 9 ... f5 has occurred in a couple of Sha­
even brighter spark worked out that Black balov games. In the first, Shabalov-Perelshteyn,
could take on c5 after all. Let's take a look: Vermont 2003 he played 20 lt:lc4 but after
13 c5 lt:lxc5! 14 1i'c1 ltlf7 1 5 ltld5 lt:le6 and: 20... fxe4 21 lt:lfxe5 ltld4 22 1i'd2 c5 23 f4 �e6
24 lt:le3 l:ld8 25 ...e1 ltl6f5 Black had the better
game. So in Shabalov-Movsesian, Bermuda
2004 he was ready with an improvement, the
interesting 20 exf5 gxf5 21 �c4!?. Normally
one worries when the opponent has the bishop
pair but here Shabalov is trying to destroy the
knight pair. His idea worked well after 2 1 ...�h8
22 ..txe6! �xe6 23 l:lad1 as Black's knight on
d6 is unstable. Better, as A vrukh has pointed
out, would have been the continuation
21...ltlxc4 22 ...xc4 (22 lt:lxc4 e4) 22 ... 'iti>h8 23
l:lad1 e4 24 lt:le1 ltlf4 with a good game.
b2) In Mikhalevski-Oratovsky, Israeli Team
Ch. 2004 Black chose 1 9 ...lt:lf4. He also wanted

76
 The Cla ssical Varia tion : 7 . . .Ci:J a 6

to take the white bishop but without opening l:td1 .:n.

the position. After 20 l:tfd 1 lt:'!xe2+ 21 'iVxe2 b) 17 lt:'!d2?! f5! 18 exf5 gxf5 19 l:tfe1 �e6
.i.e6 22 ...d3 lt:'!f7 23 ltk4 'iVb4!? 24 b3 .i.f8 25 20 f4 l:txb2 21 lt:'!c4 .i.xc4! 22 'iVxc4 exf4 23
l:tact aS! 26 'iVc3! 'ifbs 27 a4 'ifb4! 28 ...xb4 .i.xf4 'iVd8 24 .i.d6 l:tc2! 25 l:te3 f4! 26 l:t£3
.i.xb4 Mikhalevski considers the position to be 'iVgS 27 g4 hS! 28 l:te1 (28 .i.xfB hxg4 29 h4
about level. That seems like a reasonable as­ 'iVxh4 30 'iVxf4 .i.d4+! 31 'iVxd4 � 32 �ft
sessment. �1+ 33 'iVg1 ...x£3+ 34 �e1 ...xc3t 35 �ft
1 3 ... c6 14 c5 'iVxa1 mate) 28 ... hxg4 29 �e7 gx£3+ 30 �xgS
14 a3 lt:'!f7 15 b4 ...e7 is very comfortable f2+ 31 �ft fxe1'iV+ 32 �xe1 l:txc3 33 ...xf4
for Black. In Barkhagen-Gallagher, Stockholm lt:'!es 34 ..,e4 li:'!£3+ 35 �d1 lt:'!xgS 36 ..,xc6
2003 I actually went on to lose this position but l:td8+ 37 �e2 li:'!£3 38 'iVe6+ �h8 39 �£2 l:td2+
only because I was lulled into a false sense of 0-1 Volkov-Volokitin, Halkidiki 2002. An ex­
security. After 16 'iVb3 f5 17 l:tad1 f4 1 8 .i.ct cellent attacking game from another fine young
lt:'!c7 19 cS lt:'!e6 20 lt:'!b1 ! (White continues to talent.
go backwards - but correctly so) 20...b6 1 5 ...f5 1 6 f3 1i'e7
(20...lt:'!d4 21 li:'!xd4 exd4 is fine for Black) 21 16 ... f4 1 7 .i.£2 gS had been played before af­
cxb6 axb6 22 .i.b2 I now started my downhill ter which \Vhite can play lt:'!c4-d6. The text is
trip with 22...�h8 when 22 ... c5! would have much better, targeting the c-pawn and making
been about equal. it more difficult for \Vhite to carry out this plan,
1 4 ...�f7 at least if he is intent on refraining &om �xa6.
Black can just play 1 4... ltk7 here if he is 1 7 llc1 f4 1 8 �f2 lidS 1 9 1i'e1 �e6
concerned about 1 5 .i.xa6.
1 5 �d2!?
\Vhite usually plays 15 .i.xa6 bxa6 16 ...a4
but recently he has not fared well after
16...l:tb8!.

20 �xa6
A rather damning indictment on \Vhite's
play. He did everything to avoid this move and
ends up playing it anyway.
20 . . .bxa6 21 1i'e2 h5 22 llfd 1 g5 23 �1
lbis little rook move is a significant im­ A clear sign that things are not going accord­
provement on the previously played 16 ... f5 as ing to plan. The open centre is no sanctuary for
now the rook will be able to swing to the king­ the white king, so presumably he was just look­
side or centre via the b7-square. A couple of ing to free g1 for the bishop.
examples: 23 ...g4 24 hxg4 hxg4 25 1i'xa6
a) 17 b3 f5 1 8 exf5 gxf5 19 l:tfe1 l:tb7! 20 25 fxg4 lt:'!h6 won't be fun for \Vhite.
l:tad1 l:te7 21 �4 l:te6 22 lL'lgS li:'!xg5 23 �xgS 25 ...1i'f6!?
h6 24 .i.ct �h7 25 l:td6 Y:>-Yz Savchenko­ Black transfers his queen to the h-file whilst
Damljanovic, Halkidiki 2002. I f anything Black indirectly defending the pawn on c6.
has an edge after 25 ...l:txd6 26 cxd6 ...g6 27 26 �e2?!

77
Pla y th e King 's Indian

He should have at least gone straight to el. 35...'ifg2+ 36 ..i.t2 c!L!e6 37 l:tc4 (or 37 'ifg1
26 ...g3 27 �g1 "ih16 28 �81 1i'h1 29 'iff1 lL!d4+ 38 'it>e1 gxf2+ 39 'ifxt2 'itb t+ 40 c!L!fl
ttJg5 30 b3 lld3! ! ..i.f6) 37 . .lL!d4+ 38 l:txd4 exd4 39 �d3 gxt2 40
.

1i'h1 'it'g3 Black should still be winning.

35 ....:Xd2+! 36 'ifxd2 'ifxf3+ 37 �81 1i'h1 +
38 �82 'iff3+ 39 �81 'ifg2 40 'ifd8+ �h7
41 �d2 gxf2 42 .:n .!bf3+ 43 �e2 .!bd4+
0-1
As 44 �d2 'ifg3 and 44 �d3 'ifxfl+ 45
:Xfl ..i.xft+ 46 �d2 c!LJ£3+ 47 �cl ..i.h3 are
both totally hopeless.
A tremendous game from one of Russia's
numerous rising stars.

Game 26
Soffer-Mittelman
Wonderful! Black plans to increase the pres­ Israeli Team Ch. 2003
sure by doubling on the d-file taking advantage
of the fact that the knight on d2 can't move 1 c4 .!bf6 2 .!bc3 g6 3 e4 d6 4 d4 �g7 5
because it must prevent a crushing sacrifice on �82 0-0 6 .!bf3 85 7 0-0 lDa6 8 ll81
£3. But isn't the rook simply en prise? A quiet, prophylactic continuation. White
31 'ifxd3 decides to reinforce his e-pawn and await de­
31 tbdb1 c!L!x£3+ 32 gx£3 l:tx£3 is the crush­ velopments.
ing sacrificed I mentioned. 8 . . .c6
31 ...'ifxg2!! This move has several points. It covers the
Another great move. The g2 pawn is more d5-square and opens a path to the queensidc
important than the bishop. for the black queen. It also allows Black to meet
32 'iff1 1i'h1 d4-d5 with ... c6-c5 without having to worry
Threatening ..ih3. about White capturing en passant, as well as
33 �82 �h3 34 'ii'8 1 lld8! giving him the chance to exchange on d5 as he
Making sure that the white king cannot does in the main game.
evacuate the danger zone via d3. 8 ...•e8 is a major alternative. After 9 ..i.fl
35 �f2 ..i.g4 1 0 d5:

The last chance was 35 lL!cb 1 in order to a) 10 ...c!L!b4 has occurred countless times.
shore up the defences of £3. However, after Black now threatens 1 1 .....i.xf3 as 12 'ifx£3 will

78
 The Classical Varia tio n : 7 . . . !D a 6

cost White an exchange and 12 gx£3 wreck his after 1 2 l:tb1 lLlcs 1 3 b4 lL!a4 1 4 lLlxa4 .txa4. I
kingside pawn structure. White usually defends am a little concerned, however, about the un­
against this with 1 1 .te2 and Black smugly tested 1 2 a3 as now 1 2...lLlc5 1 3 b4 lL!a4 can be
plays 1 1 ... a5, his little trick having gained the met by 1 4 lL!bS!. This was not possible with the
time to improve his fortunes on the queenside. rook on b 1 because of ... .txbS followed by
It seems to me, however, that after 1 1 a3! .tx£3 ...ltJc3.
1 2 gx£3 Black has paid to high a price for the So Black probably should meet 12 a3 with
damage he has inflicted on White's kingside. 12 ... £5 when White can play 13 b4 locking the
Can he really get away with giving White the knight on a6 out of the game. Maybe the Black
moves a3 and b4 for free, not to mention part­ position isn't so bad but its definitely first blood
ing with the precious light-squared bishop. to White.
P.Nielsen-De Ia Riva, Bled Olympiad 2002 9 .i.f1
suggests not. Play continued 12.)Da6 1 3 b4 On 9 l:tb1 Black should probably exchange
lLld7 (to hold up c4-c5) 1 4 .te3 f5 1 5 l:tct l:tf7 in the centre. After 9... exd4 1 0 liJxd4 l:teS 1 1
(It is surprisingly difficult for Black to create .tn (1 1 .t£3 can be met 1 t ...lLlg4 and 1 1 £3 by
threats on the kingside. If, for example, he plays 1 t ...lLih5) 1 t ...'W'b6 12 lLla4 (12 h3?! ltJcs is
... f5-f4 then White just drops his bishop back to good for Black) 1 2 ... 1i'a5! the game is level
d2 and there is no obvious way for Black to according to Dautov who backs up his claim
continue. Note that the light-squared bishop with the following analysis: 13 £3 (13 b4? lL!xb4
will be able to deploy itself very actively on h3. 1 4 .td2 cS 1 5 lL!bS lL!xe4 16 a3 .tg4!) 13...d5!
Paradoxically, the white king seems safer on an 14 b4 (14 cxdS lL!xdS!=; 14 .td2 'iVdS 15 cxdS
open g-file than in the more traditional lines.) lL!xdS!=) 1 4 ...1i'dS! 1 5 cxdS lL!xdS 16 .txa6
16 lha4 lL!b6 1 7 lLib2 lLld7 1 S cS dxcS 1 9 bxcS lLlc3! 1 7 lL!xc3 .txd4+ 1 S .te3 .txc3=.
f4 20 c6 fxe3 21 cxb7! ex£2+ 22 �x£2 l:tbS 23 9 ....i.g4
.txa6 l:tf6?! 24 d6? (Nielsen's powerful play has 9 ... exd4 is Game 27.
netted him a clear advantage but now he 1 0 d5
couldn't resist a further combination that Black's last move was designed to force this
turned out to be a false trail. I f he had seen one advance and White doesn't really have a choice.
move further he would have settled for 24 'W'a4 Exchanging on eS would be too feeble and the
and would most likely have won the game.) move he would like to play, 10 .te3 runs into
24...l:txd6 25 1i'xd6 cxd6 26 l:tcS lLlcS! 27 tO... ..tx£3 1 1 1i'x£3 lLig4! 12 1i'xg4 exd4 13 .tgS
:XeS+ :Xes 2S .tc4+ �£8 29 .tds lLlxb7! f6 1 4 .td2 f5! 1 5 'iVh3 dxc3 1 6 .txc3 .txc3 17
(White loses the pride and joy of his position. 1i'xc3 fxe4 1 S l:txe4 lLlcs 19 l:te2 1i'f6 with a
Perhaps Nielsen had only considered 29...l:tbS level game, Portisch-Cramling, Marbella 1 999.
when 30 l:tb1 ! saves the pawn.) 30 lLld3 (30 1 0. . .cxd5!?
.txb7 l:tbS 31 lLlc4 l:txb7 32 lL!xd6 is equal) It's not so common for Black to exchange
30 ... lLlc5 31 lL!xcS dxcS 32 l:tb1 (White has on dS, but as it was the choice of Shirov it is
enough activity to compensate for the missing certainly worth taking seriously. Let's take a
pawn) 32 ... .tf6 33 �e2 �g7 34 a4 c4 35 .txc4 look at some of the alternatives.
l:1e7 36 aS l:tc7 37 'ifi>d3 .te7 3S .tdS 'lz-1/z a) 1 0 ...lL!b4 is the move Black has played the
b) I have played 10...ltJh5 myself a couple of most. We have already seen this idea in the
times but I'm not totally convinced by it. After analysis of S ...'iVeS (see above). This time 1 1 a3
1 1 h3 there is: .tx£3 12 gx£3 is not good for White because he
b1) 1 t ....tx£3 1 2 'W'x£3 f5 1 3 ex£5! gxf5 14 doesn't have the same queenside initiative as in
.te2 liJf6 1 5 .td3! (better than 1 5 1i'xf5 lLixd5) the previous example. Instead he usually plays
1 5 ... f4 (15 ... lLlc5 1 6 .tx£5 e4 17 1i'g3; 1 5...e4 16 1 1 .te2 and after 1 1 ...a5 he has played a large
'iVxf5) 1 6 .t£5 with some advantage to White, number of moves but 12 .tgS h6 1 3 .te3
I.Sokolov-Gdasi, Antwerp 1 995. seems to give him the best chance of an edge.
b2) 1 1 ....td7 is, perhaps, the more natural b) Blocking the centre with tO... cS is Black's
continuation when Black has been doing fine second most popular choice. After 1 1 h3 .td7

79
Pla y th e King 's Indian

White, again, often plays 12 �g5 h6 13 �e3. didn't want to have to give it up for the knight,
Black would much prefer not to have played e.g. 12... f5 1 3 h3 fxe4 14 hxg4 ex£3 1 5 gx£3
the move ... h7-h6 as he may well feel the weak­ seemed in White's favour to me. White now
ness of g6 after he has played ...f7-f5. produced the interesting 1 3 g3 f5 1 4 �h3 and I
c) Therefore, recently quite a few Black agreed to an exchange of light-squared bishops
players have staned playing 10 ... lLle8. with 1 4.. .f4 as I thought I would have good
attacking chances against his king. His monarch
did eventually perish after a tough and interest­
ing manoeuvring game: 1 5 �xd7 11Vxd7 1 6 �g2
h6 17 h3 g5 1 8 lLlh2 lLlf6 1 9 g4 l:.f7 20 �d2
i.ffi 21 'iVa4 'ifc8 22 £3 �e7 23 l:.h1 i.d8 24
�e1 �g7 25 i.t2 �b6 26 l:.hct lLlb8 27 'ifd1
lLlbd7 28 lLla4 'iVc7 29 11Vg1 l:.c8 30 iLJfl?
(White needed his rooks to stay in touch with
the kingside) 30...h5 31 lLld2 hxg4 32 hxg4 l:.h8
33 'ifd1 lLlffi! 34 b4 'iVd7 35 bxc5 lLlxg4! 36
fxg4 f3+ 37 lLlx£3 11Vxg4+ 38 �g3 l:.h3 0-1
1 1 cxd5 �8 12 .b2
This is directed against 1 2 ... f5 which is now
unplayable on account of 1 3 lLlg5. A couple of
The idea is still to defend on the queenside other examples:
with ... c6-c5 (this lLla6 and pawn on c5 combi­ a) 1 2 h3 �d7 1 3 a4 h6 14 lLld2 f5 1 5 lLlc4
nation is one of the toughest set-ups for White lLlc5 1 6 exfS gxfS 17 �e3 b6 1 8 b4 lLlb7 1 9
to breakdown) and attack on the kingside with l:.ct f4 20 �d2 lLlf6 2 1 �e2 � fS 22 � £3 �h8
... f7-f5, but with this move order White no with excellent chances on the kingside for
longer has the annoying �g5. Here are a couple Black, Haritakis-Khamatgaleev, Ikaros 2002.
of examples from my own practice: b) 1 2 �xa6 bxa6 13 11Vd3 �c8 14 lLld2 f5 1 5
cl) Gelfand-Gallagher, Biel 2000 went 1 1 h3 lLlc4 f4 1 6 a4 'iVh4 17 �d2 g5 1 8 £3 g4 1 9 fxg4
�d7 1 2 dxc6!? (an attempt to punish Black for �xg4 with good play for Black, Candela­
his move order) 12 ...bxc6!? (12 ... �xc6 13 l:.b1 Comas Fabrego, Burgos 2003.
lLlac7 14 b4 b6 is a more solid continuation) 1 3 White's play can no doubt be improved on
�g5 f6 14 �e3 c5! (preventing White from in these examples but they do serve to demon­
playing c4-c5 and although Black has ceded strate typical plans for Black.
control of the d5-square he will be hope to 1 2 ..id7 1 3 a3 !i'Jac7 14 'iib3 b5
. . .

manoeuvre a knight of his own into d4) 1 5 Black defends actively on the queenside.
lLld5 lLlac7 1 6 b4 (16 lLlxc7 11Vxc7 1 7 11Vd5+ 1 5 ..ig5 f6 1 6 ..id2 f5
�h8 18 11Vxa8 �c6) 1 6 ... lLle6 17 l:.b1 �h8 1 8
lLld2 (1 8 11Vct !?) 18... f5 1 9 exfS gxfS 20 f4 cxb4
21 l:.xb4 �c6 22 lLl£3 11Va5 23 l:.b1 o!LJ&7! and
now best play is 24 �d2 11Vc5+ 25 �e3 11Va5
with a draw, but Gelfand preferred to gamble
with 24 fxe5 dxe5 25 lLle7 even though he is
close to lost after 25 ... �e4! From a sporting
point of view his decision was entirely vindi­
cated as he won due to my rime trouble errors.
c2) In Akesson-Gallagher, Gausdal 2001
White chose not to attack the bishop and
started queenside operations at once. After 1 1
l:.b1 c5 1 2 a3 I still felt the need to retreat my
bishop with 1 2 ... i.d7 before playing ... f7-f5 as I

80
 Th e Classical Varia tio n : 7 . .liJ a 6
.

1 7 l%ac1 e4 0-0 6 .i.e2 lt:)a6 7 0-0 e 5 8 lle1 c6 9

A new move. In Fracnik-Shirov, Bundesliga .i.f1 exd4
2001 , White played the unconvincing 1 7 lUgs The reason that 8 l:.e 1 is less popular than 8
'it'e7 1 8 lt:le6 and after 1 8... lt:lxe6 1 9 dxe6 .i.xe6 .i.e3 is because of this move. The most that
20 ..xbS lt:lf6 Black's superiority in the centre White can hope for is a faint edge in a rather
gave him an edge. dry position and it is unlikely that he can even
1 7 . . .lt:)f6 1 8 lt:)g5 �h8 1 9 .i.d3 .i.h6 20 achieve that.
'iVb4 The reason that 9 ... exd4 is not even more
20 lt:\f7+ l:.xf7 21 .i.xh6 fails to 21...lt:\g4 popular amongst Black players is that it is very
while 20 lt:le6 lt:lxe6 21 i.xh6 lt:lcS 22 'it'd1 l:.f7 hard for Black to play for a win in this line.
is rather unclear. 1 0 lt:)xd4 lt:)g4 1 1 h3
20 . . .lt:)g4 21 f3 lt:)e6! ? White's choice is limited because of Black's
kingside threats. He has tried 1 1 lt:\£3 and 1 1
l:.e2 bur both of these are stiU weU met by
1 1 ...1i'b6.
1 1 . . .'ifb6
Only Mark Hebden has punted 1 1 ...lt:lxf2 1 2
�xf2 1i'b6 but that was 1 S years ago and I
haven't seen him do it again.
1 2 hxg4 'ifxd4
1 2 ... .i.xd4 is considered inaccurate as after
1 3 .i.e3 White gets to exchange bishops. StiU,
it's not that bad for Black.

My first reaction was that this must be a

misprint, but no it really is a playable move. It's
aU based on ...'it'd8-b6+- and smothered mate
ideas. Another interesting possibility is 21 ...aS
22 1i'xd6 l:.f6 23 fxg4 l:.xd6 24 lt:\f7+ �g7 2S
lt:lxd8 .i.xd2 26 lt:lb7 .i.xc l ! (26...l:.b6 27 lt:lcs
.i.c8 28 exf5±) 27 lt:lxd6 .i.xb2 28 ltJcxbS
lt:lxbS 29 i.xbS .i.xa3 with equality.
22 fxg4
Or 22 dxe6 'iVb6+- 23 .i.e3! lt:lxe3 24 �h 1
with unclear play.
22 ... lt:)xg5 23 .i.xg5 11'b6+! 24 �h1 .i.xg5 1 3 ..f3
25 llc2 a5 26 'ifb3 fxg4 27 .i.xb5 % - % White has tried quite a few moves here. Let's
To be honest, I'm not sure what Black have a look at some of them:
should play on move 10 (or move 9, see next a) 1 3 gS is played quite often when Black can
game) as I keep changing myself. You'U have to choose between 13 ...lt:\cS and the endgame
make your own mind up. after 1 3 .....xd 1 1 4 l:.xd 1 .i.eS where I see no
advantage for White. For example,
Game 27 at) l S .i.e3 lt:lc5 1 6 f3 i.e6 17 l:.acl aS 1 8
l:.c2 a4 1 9 'itt2 f6 20 gxf6 l:.xf6 21 i.d4 gS
Gelfand-Markowski
(Hoelzi-Naumann, Austrian Team Ch. 2003)
Po/anica Zdroj 1998 and if anyone is better here it's Black
a2) 1 S lt:la4 (White is simply getting out of
1 d4 d6 2 lt:)f3 lt:)f6 3 c4 g6 4 lt:)c3 .i.g7 5 the way of a possible capture on c3 by Black)

81
Pla y t h e King 's Indian

15 ...lLlb4!? 16 a3 (16 f4 �g4) 16 ...lLlc2 17 .:.a2 army. Black has a potentially nice position but
lLld4 1 8 b4 b6 1 9 �e3 '1•-'1> Huzman-Comas he is lagging in development and not ready for
Fabrego, Istanbul 2003. an open centre. That is why it was so costly to
b) 1 3 1i'e2 1i'e5 1 4 �e3 1i'e7 (or 1 4... lLlc5 1 5 release the blockade of eS, even if it was only
£3 aS) 1 5 .:.ad1 .:.e8 1 6 £3 lLlc5 1 7 1i'd2 �eS 1 8 supposed to be momentarily.
.i.£2 lLle6 1 9 g3 gS with a level game, Osten­
stad-Trygstad, Fredrikstad 2003.
c) 1 3 .i.e3 gives Black the choice between
t3 ...1i'e5 and 1 3 ...1i'xd1 14 .:.axd1 �xg4 1 5 £3
.i.e6 1 6 .:.Xd6 :res with a roughly level end­
game.
1 3 ...-.es
The queen doesn't belong in the middle of
the board so she starts to make her way back to
her ideal home on e7.
1 4 gS
14 �f4 1i'e7 1 5 1i'g3 �eS 16 gS f6 is similar
and perhaps slightly easier for Black.
14 ...-.e7 1 S •g3 ll:lcs 1 6 �f4 �es 1 7
llad1 f6! 22 . . .dxeS
Gelfand had aimed for this position as White Or 22...lbxe5 23 lLle4 .:.m 24 lLlxd6 with a
had already won it several times after poor play clear advantage to White.
from Black. The text seems to be the best. 23 ll:le4 llf7 24 cS �g7 2S �c4 lle7 26 a4
lieS?
26 ... b6 offered more resistance.
27 llf1 lle7 28 lld2 bS 29 axbS cxbS 30
�dS 1 -0

Summary:
1) For those of you who prefer a strategical
battle to a theoretical battle then 7 ... lba6 is an
excellent alternative to the main lines. I know as
I have played it a lot myself in the last few
years. It was virtually unknown a decade or so
ago but has now developed into one of the
major systems of the King's Indian.
2) After 8 �e3 Black, as long as he played
1 8 gxf6 :Xf6 1 9 �xeS -.xeS 20 b4 -.xg3? carefully, was doing fine in the older lines with
A serious error although it is not easy to ap­ ...h7-h6 (1 3 ...1i'e7Q but that wasn't enough for
preciate that without seeing Gelfand's idea. some players and new lines with ... f7-f6 have
Instead, 20...lbd7! maintains the balance, e.g. been developed. Black has also started out fine
a) 21 1i'xe5 lLlxe5 22 cS dxcS 23 bxcS bS. there and whilst I feel pretty confident in the
b) 21 1i'e3 can be met by 2t ...a5 22 a3 (22 bS 1 t ...f6 of Game 22 I would say that the jury is
lLlcS) 22...g5 with no problems for Black. still out on the 1 1 .. f6 of Game 25.
.

c) 21 cS 1i'xg3 22 fxg3 lbe5! 23 cxd6 �g4 24 3) 8 .:.e1 is less popular but slightly more
.i.e2 .i.xe2 25 .:.xe2 .:.ds and, as Gelfand says, unpleasant for the typical King's Indian player
White may have an extra pawn but the advan­ to face. That is because he may be reluctant to
tage is with Black as he controls the centre. enter into the solid equalising lines of Game 27.
21 fxg3 ll:ld7 22 eS!! I am not saying that Black is worse in Game 26,
An excellent move that liberates the white just that it feels like we are playing Black.

82
 CHAPTER SIX I
The Classical Variation
White's 7th Move Alternatives

In this chapter we examine \Vhite's alternative from moving his knight on f6 and playing ... fS.
tries to the extremely natural 7 0-0. Black can stiU arrange ... fS but only by playing
some awkward-looking moves or making posi­
Classical 7 d5 tional concessions.
Black's two main replies to 7 dS are 7 ... a5
1 d4 lL!f6 2 c4 g6 3 lL!c3 �g7 4 e4 d6 5 and 7...l2Jbd7 and then after 8 .i.gS we have
lL!f3 0-0 6 �e2 e5 7 d5 entered the main lines of the Petrosian System.
I am, however, suggesting that we take two of
Petrosian's main characteristics - his wiliness
and his prophylactic nature - and use them
against his own system. By playing 7...lL!a6
Black renders the Petrosian System hannless as
he has a relatively simple way to equalise against
8 .i.gS. I am amazed at how many \Vhite play­
ers are totaUy ignorant of this line which we
cover in Game 28. So if \Vhite is going to re­
frain from 8 ..tgS what does he play instead?
WeU, he has several options and these are cov­
ered in Games 29 and 30. StiU, they are not the
most dangerous lines in the King's Indian. This
is because with his 7th move White made a
The Petrosian System is the name given to slight concession by blocking the centre with­
the variation where \Vhite blocks the centre and out getting anything in return (such as a tempo,
foUows' up with 8 .i.gS. It is named after one of for example).
the aU time greats, the 1 963-69 World Cham­
pion Tigran Petrosian. Vladimir Kramnik also Game 28
played this variation extensively in his formative
Golod-Gallagher
years. The wily Armenian, as Petrosian was
often called, was famous for his deep, strategic,
Zurich 2003
and often negative play. His philosophy was to
stop the opponent's threats before they even 1 d4 lLif6 2 c4 g6 3 lL!c3 �g7 4 e4 d6 5
thought of them. The reason he played ..tgS lLif3 0-0 6 �e2 e5 7 d5 lL!a6!?
against the King's Indian was to prevent Black I shaU not b e covering the main line 7...a 5 i n

83
Pla y t h e King 's Indian

this book but for those of you who prefer that b) 13 ltJd2! is an improvement for White as
it is examined in SOKID. As I mentioned his pieces co-ordinate better with knights on c3
above the text is a relatively unusual move that and e4 than on d2 and e4. Now I would dearly
seems to neutralise the Petrosian System. love to play 1 3 ... e4 but on close inspection it
8 .i.g5 doesn't seem adequate. Therefore, 13 ...ltJcS 14
There are three main alternatives. 8 0-0 and ltJde4 and now I think Black should play
8 lbd2 are covered in Game 29 whilst 8 ..i.e3 1 4... ..i.f5 1 5 �d3 fxg3 (it is also possible to
transposes to the Gligoric System (1 ..i.e3) and delay this capture, e.g. 1 5...ltJxd3 16 'ii'xd3 a6)
that is covered in Game 30. 1 6 hxg3 ltJxd3+ 17 'ii'xd3 and now I like 1 7...a6
8 . . h6 9 .i.h4
. best. For the time being White has a relatively
Obviously it doesn't make sense for the secure blockade of e4 but his position is some­
bishop to go back to e3 now as White can't what unwieldy as if the blockade is ever broken
meet 9...lbg4 with 10 ..i.gS. he will be in trouble. 17...a6 is a semi-waiting
9 . . g5 1 0 .i.g3 �xe4!
. move which prepares the possible advance
... b7-b5. What is Black waiting for? Well, he's
waiting to see where the white king is going. If
White plays 1 8 0-0-0 then 1 8...b5 looks good
whilst if White plays 18 0-0 then 1 8...1i'e8, in­
tending ...'ifg6 looks like a good idea. And if
White plays 18 f3, intending to follow up with
g3-g4, then Black should play 18 ...g4! himself.
The one thing he should never do is exchange
off his light-squared bishop for one of the
knights. I consider the position to be dynami­
cally balanced.
1 2...fxe4 1 3 �e4 .i.f5 14 .i.d3
Or 14 ltJc3 ltJcS 1 5 0-0 aS (I spent a long
tome thinking about 1 5...e4 and it is certainly
Tactical ideas based on ...ltJxe4 followed by playable as well) 16 f3 e4 17 fxe4 ..i.xe4 1 8 ..i.£2
... £5 and ... f4 are quite well known in the King's 1i'e7 19 .i.d4 '1>-'12 Rahman-Gallagher, British
Indian but they are usually not so good as Ch. (Edinburgh) 2003. Another easy game with
White often ends up with a powerful blockade Black against a grandmaster.
of the e4-square. This time it is different as 1 4. . . g4!
although White may still be able to blockade e4
it is of a more &agile nature.
1 1 �xe4 f5 1 2 �fd2
The first decision White has to make is
whether to part with a knight or the bishop on
g3. Although the text is the most common I
believe it is slightly more accurate to give up the
bishop on g3 which can easily become a prob­
lem piece for White. Therefore 1 2 ltJc3 f4 and
now:
a) 13 ltJe4 ..i.£5 14 ltJfd2 ltJcS! (White is now
forced to weaken his position in order to main­
tain the blockade of e4) 1 5 f3 fxg3 16 hxg3 g4
17 'ii'c2 c6! 1 8 fxg4 ..i.g6 1 9 ..i.d3 cxdS! 20
ltJxcS e4 21 ..i.xe4 dxe4 22 ltJe6 'iVb6 23 ltJxf8 This is the key move as White is going to
.l:.x£8 with a clear advantage for Black, Olsen­ have to weaken his kingside to save his bishop
Kindermann, Reykjavik 1998. on g3.

84
 The Classical Varia tio n : White 's 7 th M o v e A lterna tives

15 0-0 h5 16 f3 ..ba4! 1 7 ..ba4 lL!c5 1 8 suggested way for Black to play is still untested.
'li'e2
White would prefer to keep his bishop but Game 29
this means lifting the blockade of e4. After 1 8
Legky-Gallagher
�c2 e4! 1 9 ..txe4 (19 fxg4 .:.Xfl+ 20 1i'xfl
1i'g5!? looks better for Black) 1 9...lbxe4 20 fxe4 French League 2001
�xb2 21 l:b1 l:xfl+ 22 1i'xfl ..td4+ 23 �h1
1i'g5 Black had some advantage in Jacimovic­ 1 d4 lL!f6 2 c4 g6 3 lLlc3 �g7 4 a4 d6 5
Maki, Pula 1 997. lL!f3 0-0 6 -'.e2 e5 7 d5 lLla6 8 0-0
1 8 ....lL!xa4 8 lbd2 has some supporters. I imagine they
1 8 ... h4 1 9 �e1 .!Dxe4 20 fxe4? .:.Xfl+ 21 justify this ugly move along the foUowing lines:
�fl 1i'f6+ 22 �g1 l:£B eventually led to a OK, I don't like 8 ..tgS and 8 ..te3 and I'm not
Black victory in Wildig-Hebden, Nuneaton that keen on 8 0-0 as this just transposes into an
1 999 but 20 1i'xe4 just transposes to the notes old and relatively harmless line after 8...ltx5 9
below. 1i'c2 aS (see main game). So why not overpro·
1 9 'li'xa4 teet the e-pawn before Black attacks it. Now
1 9 fxe4 1i'g5 is good for Black. lLla6-c5 is out of the question because of the
1 9 ....gxf3 Y.z - Y.z reply b2-b4. That means it won't be so easy for
Black to bring his knight on a6 into the game.
And if Black foUows his most obvious plan of
playing for ... f7-f5 with 8...lL\e8 then we can
exploit the fact that we haven't casded yet by
playing 9 h4!.
And what should Black think on seeing such
a move: What? lbd2. Surely that can't be any
good with his bishop still on ct. OK, I see what
he wants but we don't have to fall in with his
plans. In fact I have a few interesting ideas of
my own which may highlight the drawbacks of
8 lbd2.
Firsdy 8.....th6 is quite possible and a lot de­
pends on the assessment of the pawn sacrifice 9
Accompanied by a draw offer (early morn­ h4 lLlc5 1 0 hS ..txd2+- 1 1 1i'xd2 lLlfxe4 12 lLlxe4
ing game) which my opponent accepted. lLlxe4 and now either 1 3 1i'h6 or 1 3 1i'e3.
I expected the game to continue along the The move I like the most, however, is 8...h5.
foUowing lines: 20 ..th4! (20 .:.X.£3 .:.X.£3 21
'ifx£3 1i'g5 is pleasant for Black) 20...1i'e8 21
l:x£3 l:x£3 22 gx£3 (22 'ifx£3 e4 23 1i'e2 1i'e5
24 l:e 1 l:e8 25 b3=) 22 ...1i'f7 23 �f2 l:£B 24
l:g1 1i'f5 25 1i'xf5 l:xf5 26 �e2 �f7 with
equality.
1 9 ... h4 may look like a tempting alternative
but just weakens Black's position. After 20 .tel
l:f4 21 1i'g6!? White is certainly not worse.
To sum up, I like this way of playing against
the Petrosian System. White players, even
grandmasters, seem to be quite unfamiliar with
it and it is very easy for White to end up in a
poor position. The only critical variation seems
to be line 'b' in the 1 2th move notes as my Now there is no need to worry about White

85
Pla y th e King 's Indian

advancing on the kingside (we got there first) king.

and we can play ... lLlf6-h7 followed by ... f7-f5. b) 1 1 ...tLlhS!? 1 2 g3 .i.h3 13 l:lfe1 b6
And take a look at the white position, What (1 3 ... £5!? 14 lLlh4 leads to sharp play) 14 lLld2
move do you feel like playing now? lLld2-f3, tLlf6 (the solid choice, 14 ... £5 is still sharp) 1 5 f3
perhaps. tLlh7 16 lLld 1 .i.d7 1 7 lL\£2 f5 1 8 exf5 gxf5 19
Here are a few examples of how play can f4 1i'f6 with a comfortable game for Black,
develop after 8 ... h5!?: Piket-Hemandez, FIDE Wch KO, New Delhi
a) 9 a3 cS (it's best to stop b2-b4) 10 tLlf3 2000
tLlh7 1 1 h4 f5 (Black can delay this in favour of c) 1 1 ... b6 (in many lines it's useful to have
1 1 ...lLlc7 has Hebden has done in the past) 1 2 the knight on cS properly defended) 1 2 lLld2
ex£5 .i.x£5 1 3 .i.e3 e4 14 lbgs lLlxgS 1/2-1/2 tLlh7!? (12 ...lL\g4 and 1 2 ....i.g4 are more com­
Komarov-Sakaev, St Petersburg 1 997. mon) and now:
b) 9 lLlb3 cS 10 .i.gS 1i'e8 1 1 lLld2 lLlh7 1 2 c1) 1 3 a3?! f5 14 f3 a4! 1 5 .i.xcS bxcS 16
.i.e3 h4 1 3 tLl f3 1i'e7 1 4 g3 h3 with a good lLlxa4 hS! (a typical King's Indian sacrifice:
King's Indian for Black, Lawton-Hebden, Black has excellent attacking chances on the
Swansea 1995. weakened dark squares around the white king)
c) 9 h3 lLlh7 10 g4 hxg4 1 1 hxg4 .i.f6! 1 2 1 7 lLlc3 .i.h6 1 8 l:lae 1 h4 1 9 lLld 1 lLlf6 20 l:l£2
lLl fl .i.gS 13 .i.e3 1i'f6 1 4 a 3 c S 1 5 dxc6 bxc6 .i.d7 21 tLlb 1 lLlhS 22 tLlbc3 .i.f4 23 .i.d3 .i.g3!
1 6 1i'd2 lLlcS 17 .l:ld 1 .l:lds 1 8 lbg3 .i.f4 with an (Nunn relates how the teenage Anand took all
edge for Black, Stojanovic-Arsovic, Belgrade of five seconds to choose this course of action)
2002 24 hxg3 hxg3 25 l:ld2 1i'h4 26 lLle2, Palatnik­
8 lDc5 9 .i.g5!?
... Anand, New Delhi 1 986. Black has a winning
Quite a controversial move as White is will­ attack and the best way to continue was
ing to give up his important bishop in return 26 .. .'�g7 in order to play .. lth8 and bring the
for a few tempi on the queenside. rook into the attack. The moves ...1i'h1+ and
The old, old main line, rarely seen today, ...lLlf4 are likely to be played at some point as
runs 9 1i'c2 aS 1 0 .i.gS h6 1 1 .i.e3 when Black well.
has a wide choice of playable moves. It's not c2) 1 3 b3 is a better move. White spends a
easy to suggest one (even for myself) so I'll take tempo to avoid the positionally undesirable
a look at a few of the most interesting lines: continuation of the previous note. The game
Naumkin-Gieizerov, USSR 1 986 now contin­
ued 1 3 ... £5 14 f3 f4 1 5 .i.£2 gS 1 6 a3 lLla6 17
1i'b1 hS 1 8 b4 g4 1 9 �h1 g3 20 hxg3 fxg3 21
.i.xg3 h4 with good compensation for Black.
9 h6 1 0 .i.xf& •xf&
...

I prefer recapturing with the queen as after

1 O... .i.xf6 Black will have little choice but to
return the bishop to g7, whilst after the text the
queen has the option of dropping back to e7.
Which square is better for the queen is unclear,
at least to me, as can be seen &om the fact that
I once played ...1i'd8 and the other time ... 1i'e7.
1 1 b4 tDd7 1 2 tDd2
I had never faced this line before and then I
a) 1 1...lL\g4 1 2 .i.xcS dxcS 1 3 h3 lLlf6 1 4 had to face it rwice in a relatively short period
lLlxeS lLlxdS 1 5 cxdS .i.xeS 1 6 f4 .i.d4+ was of time. The plan I used successfully against
played by Fischer and is obviously a very critical Legky I had already found over the board in
variation. However, it is rarely seen as most Kraai-C:rallagher, Bundesliga 1 999. In that game
Black players deem it to be unnecessarily risky. White played 12 l:lc1 and play continued
It is already unclear where White shbuld put his 1 2 ...1i'd8 1 3 lLld2 hS!? 1 4 lLlb3 .i.h6 1 5 l:lc2

86
 The Classical Varia tio n : White 's 7 th M o v e A lterna tives

ltJf6 1 6 c5 ltJg4! 17 h3 ltJf6 18 ltJa5 lDe8 19 That was a nice surprise. White is trying to
�b5?! ltJg7 20 ltJc4 a6 21 .i.a4 h4 22 ttJe2 ltJh5 clear the third rank so that his rook defends the
23 l:.c3 b5! 24 cxb6 cxb6 25 .i.c6 l:.a7 26 a4 f5 vulnerable h3-square but this was not a good
27 exf5 gxf5 28 'it>h1 l:.g7 with a large advan­ way to go about it.
tage for Black which I eventually converted. 22 ...�xh3!
1 2 ...-.e7 1 3 lt:lb3 h5!
Of course Black can also play ... f7-f5 but the
text just felt right to me. With his opposite
number departed the dark-squared bishop
needs to be activated. Control of squares in the
white camp such as c1 and d2 may also prove
annoying for White.
14 a4 �h6 1 5 a5 lt:lf6 1 6 c5 lbg4!
The idea is to encourage White to create a
weakness by playing h2-h3. That is well worth a
couple of tempi.
1 7 h3
Could Black have ignored the knight? Well,
certainly not if he wants to win the game. For
example, 1 7 c6 bxc6 1 8 dxc6 ltJxh2 19 Wxh2 23 lt:lcb1
�f4+ 20 Wg1 'Wh4 21 g3 �xg3 22 fxg3 'ii'xg3+ Legky preferred to take his chances in the
23 'it>h 1 and Black has a draw or may even play middlegarne rather than the bad ending after 23
on with 23...'ii'xc3!? gxh3 'ii'g5+ 24 Wh1 1i'xd2 25 1!6'xd2 �xd2.
1 7 ...lt:lf6 23 . . .�c8 24 -.a4
White is hoping to cause trouble with 1i'a7.
24...�f4 25 l:ld1
25 'ii'a7 loses to 25....i.g4 26 t3 'Wh4!.
25 ...�g7
I feel sure there was a nice point to this
move - I just can't remember it. Maybe I was
just waiting for White's next move. It certainly
didn't help his cause.
26 �f1 ?
He had to try 26 liJfl though Black should
still be winning.
26 ...�g4! 27 lt:lf3
27 t3 loses to 27 ... 'Wh4 and after 27 l:.e1
Black can, at the very least, play 27 ... h4. My
Now I wanted to play ltJh7-g5 and sacrifice only worry in this position was that the rook
something on h3. could get trapped on b8 due to the lack of
18 c6 l:lb81 1 9 b5 b6 20 axb6 axb6 21 l:la3 S<.Juares for the bishop on the c8-h3 diagonal.
White has just one target to aim at on the Once the move ... h5-h4 has been played then
queenside - c7. I had spent ages trying to work the bishop will just be able to drop back to h5 if
out whether he could cause any trouble with attacked.
l:.a7, followed by ltJc3-a2-b4-a6 but concluded 27 . . . lt:lg5 28 �e2 l:laS! 29 "ibaS l:lxa8 30
that my kingside attack was too quick for this l:lxa8 lt:lxe4
lengthy manoeuvre. Legky must have reached a 30...ttJxf3+- 31 .i.xt3 .i.xt3 32 gxt3 'ii'gS+ 33
similar conclusion and opted for a more defen­ Wfl h4 is also good but the text is crushing.
sive rook move. 31 �1 h4 32 l:la3 h3 33 g3?! �xg3 34
21 ...lt:lh7 22 lt:ld2? fxg3 lt:lxg3+ 35 �2 lt:lxe2 0-1

87
Pia y th e King 's Indian

Classical 7 �e3 (Giigoric Systeml Bayonet Attack chapter) 1 1 ...'iVe8! (1 1 ...�f6 12

�xf6 lbxf6 1 3 ex£5 �x£5 1 4 d5 lbe7 1 5 lbg5 is
1 d4 �f6 2 c4 g6 3 �c3 �g7 4 e4 d6 5 a lirtle better for White) 1 2 dxe5 (12 d5 can be
�f3 0-0 6 �e2 e5 7 �e3 met by 1 2...lbe7 or 1 2... lbd8, intending lbd8-
f7) 12 ...dxe5 13 h3 (13 lbd5 'iff?=) lbf6 1 4
�d3 i.e6 1 5 l:le1 1i'f7 1 6 c5? (1 6 �xf6 �xf6
1 7 lbd5=) 1 6 ...lbd7 17 �b5 lbd4! 18 �xd7
�xd7 19 i.h4 i.c6 20 lbg5 1i'f6! (Black threat­
ens to win a piece with ... h7-h6 and there is no
good discovered attack with the knight) 21 ex£5
gx£5 22 lbe2 l:lad8 23 lbxd4 l:lxd4 24 "ii'h5? h6
0-1
b) 7...h6!? was a favourite of John Nunn
who raised its profile by using it against top
class opposition. The idea is to play ...lbf6-g4
but without allowing �g5. The main line runs 8
0-0 lbg4 9 i.cl lbc6 10 d5 lbe7 1 1 lbe1 f5 1 2
�xg4 fxg4 with a most unusual kingside pawn
The coverage of the Gligoric System wiU be structure. Black will now play ...g6-g5 and
light in this book as I am recommending that ... lbg6-f4 while White will seek his forrune on
Black play 7 ...lba6. White now usually plays 8 the queenside. Most theoretical sources agree
0-0 which takes us into Chapter 5. The only real that White has slightly the better chances.
alternative is 8 d5 and that is covered in Game c) 7 ... exd4 is not such a bad move as Black is
30 below. I will just take a quick look at the often able to create counterplay by a quick ...c7-
main 7th move alternatives for Black. c6 and ...d6-d5. for example, 8 lbxd4 l:le8 9 f3
a) 7 ...lbg4 is probably the most important. It c6 10 1i'd2 (10 i.£2!?) 10 ...d5 1 1 exd5 cxd5 1 2
leads to very complex strategical play with posi­ 0-0 lbc6 1 3 c 5 l:lxe3!? is the famous exchange
tions that are rather difficult to handle for both sacrifice that first occurred in Karpov­
sides. After 8 �g5 f6 there is: Kasparov, New York 1 990 (1 1 th game of the
a1) 9 �h4 lbc6 (9 ...g5 10 �g3 lbh6 is the world championship match).
other main line) 10 d5 lbe7 1 1 lbd2 lbh6 1 2 f3
g5 (12...c5 has been played a lot but White Came JO
seems to have an edge here after 1 3 dxc6 bxc6
Berkes-Jobava
14 b4) 1 3 �£2 f5 1 4 c5 and now Black usually
chooses berween 14 ...lbg6 and 14 ... g4. The play European Team Ch., P/ovdiv 2003
is sharp but the practical results seem to be
slightly in White's favour. 1 d4 �f6 2 c4 g6 3 �c3 �g7 4 e4 d6 5
a2) 9 �cl lbc6 10 0-0 (10 d5!?) £5. Black has �f3 0-0 6 �e2 e5 7�e3 �6!? 8 d5
played many moves here but the text seems to Most White players accept the invitation to
offer the best chance of a level game. I shall transpose to Chapter 5 by playing 8 0-0.
now take a quick look at I.Sokolov-Shirov, 8 llJg4
...

FIDE Wch KO, Las Vegas 1 999. It should be 8...lbc5?! 9 lbd2 would allow White his ideal
of interest mainly to those of you who do not set-up but 8...lbh5 looks like a plausible alterna­
wish to meet 7 0-0 lbc6 8 �e3 with the accu­ tive. At any rate Black must play actively.
rate but boring 8 .. .:e8! (see Chapter 1), prefer­ 9 �g5 f6 1 0 �h4 h5!?
ring instead to mix it up with 8...lbg4 9 �g5 f6 I myself have favoured 10...lbh6 in this posi­
10 �cl f5 and a direct transposition here. The tion, winning one very nice game against Smirin
game continued 1 1 �g5 (1 1 d5 lbe7 1 2 lbg5 and losing a rather less nice one against
lbf6 is thought to be okay for Black: it is similar Korchnoi. However, most other grandmasters
to material we have already looked at in the seem to prefer the text and as they always seem

88
 The Classical Varia tio n : Wh ite 's 7 th Mo ve A lterna tives

to win with Black in this variation I am begin­ White players although it seems to me that the
ning to see their point of view. opening of the h-file actually favours Black.
1 1 �2 c5 1 4. . .hxg4 1 5 hxg4 lZ'!f7 1 6 1Wb3 1We7

This is a standard defensive reaction in such Now Black is ready to play the three key
positions. It will now be much harder for White moves, .....tg7-h6, ...�g8-g7 and ...l:lf8-h8 that
to make progress on the queenside and there ensure he won't have to face any serious threats
wiU be occasions when Black can take over the on the kingside.
initiative there as weU. For example, if White We already saw this same manoeuvre above
castles queenside then Black may be able to in Volkov-Nakamura.
organise the advance ...b7-b5 to open lines 1 7 lZ'ld1 �h6 1 8 1Wh3 �g7 1 9 ll'le3 llh8
against his king. In the most recent game played With his grip on the kingside dark squares
in this line, however, Black delayed the advance Black has at least equality in this position.
... c7-c5 for some time. Volkov-Nakamura, 20 1Wg2 �d7 21 0-0-0 lZ'lc7 22 lldg1
FIDE World Ch., Tripoli, 2004 went 1 t .....td7 Perhaps White should sacrifice a pawn with
12 h3 tt:lh6 13 g4 hxg4 14 hxg4 'W'e7 15 tt:lft 22 g5!? He doesn't get anything concrete in
ltlfl 16 ltle3 ..th6 17 ..td3 �g7 1 8 'ii'e2 l:lh8 return, just some open lines. A possible con­
19 0-0-0 c5 20 a3 tt:lc7 21 �b1 a6 with good tinuation could be 22 g5 ..txg5 23 l:ldg1 l:lag8
...

play for Black. He went on to win in 65 moves. 24 ltl£3 ..txe3+ 25 fxe3 �£8 and White has
1 2 a3 some compensation.
White, rather half-heartedly I feel, begins to 22.....i.f4 23 �1 ll'lg5 24 ..i.xg5 fxg5 25
prepare b2-b4. llxh8 :Xh8 26 llh1 llxh1 + 27 1Wxh1 �xe3
In Razmyslov-Matamoros, Coria del Rio 28 fxe3 ll'le8 29 lZ'lf1 lZ'lf6 30 ll'lh2 1Wd8
2004 he played an over-ambitious pawn sacri­
fice: 12 h3 tt:lh6 1 3 g4 hxg4 1 4 hxg4 tt:l£7 1 5
g5?! tt:lxg5 16 ..txg5 fxg5 17 ..tg4 tt:lb4 1 8 ..te2
(a slightly embarrassing retreat) 1 8.. .l:tf4 19 a3
tt:la6 20 'W'b3 ltlc7 21 llh2 l:lh4 22 tt:l£3 l:lxh2
23 ltlxh2 1i'f6 with a clear plus for Black.
1 2 ...-.eS
Black prefers to wait for h3 before retreating
his knight as 12 ...tt:lh6 can be met with 13 £3.
13 h3
On 13 tt:lb5 Black just plays 13...1i'e7 and
foUows-up with ...tt:la6-c7.
1 3 ...lZ'lh6 1 4 g4
This kingside advance is a favourite amongst

89
Pla y th e King 's Indian

Black has a strategically won game but one 9 ltJxe5? i s just a mistake. After 9 ...ltJxe4! 1 0
could have expected some tough resistance ltJxe4 .i.xe5 Black has the better game after
&om White. Instead he committed suicide and both 1 1 .i.g5 l:ld4! and 1 1 0-0 ltJc6. His forces
lost in a few moves. are more effectively placed and the d4-square
31 .ttl 'ih18 32 'ifg1 .ta4 JJ b4 b5 34 makes a nice home for any number of Black
�b2 bxc4 35 �c3 cxb4+ 36 axb4 a5 37 pieces.
bxa5 'ifb8 0-1 9 . Jie8
.

9...c6 is a more modem line which can lead

Classical 7 dxe5 (Exchange Variation) to initially sharp play but can also fizzle out
easily. 9 .. .l:tf8 is also a good move but I prefer
1 d4 ltlf6 2 c4 g6 3 ltlcJ .i.g7 4 e4 d6 5 to stick with the old reliable variation that was
ltlfJ 0-0 6 .i.e2 e5 7 dxe5 dxe5 8 'ifxd8 played by Bobby Fischer and that I have been
:Xd8 recommending for years.
1 0 ltld5
10 0-0-0 is the subject of the following
game.
1 0 ... ltlxd5 1 1 cxd5 c6!

I have written many scathing words on this

variation in the past. I won't repeat them all
here but just say that most of the players who
play this variation are playing 'scared chess' and
are just looking to draw the game. Make them An essential move as Black doesn't want to
work for this draw. be left with a weak pawn on an open file.
After the almost universal choice 9 .i.g5 I 1 2 .i.c4
am still recommending that Black play the old 12 d6?! would just leave the pawn weak and
main line 9 . l:le8. In Game 31 we take a look at
. . isolated deep in the enemy camp. It would be
10 lbd5 (I have very little to add to what I most likely to drop off.
wrote in SOKID here) and Game 32 fearures 1 2 . . .cxd5 1 3 .txd5 ltld7!
10 0-0-0. 13 ...ltJc6 and 13 ...ltJa6 have also been played
but against these moves White does have the
Game 3 1 chance of a nagging edge. Black now threatens
to gain the bishop pair by ... lbd7-f6.
Salgado-Gallagher
1 4 ltld2!
L'Hospita/et 1992 The only way to maintain equality. By de­
fending his e-pawn White renders the ...itJf6
1 d4 ltlf6 2 c4 g6 3 ltlcJ .i.g7 4 e4 d6 5 idea harmless. A mistake that White plays quite
ltlfJ 0-0 6 .i.e2 e5 7 dxe5 dxe5 8 'ifxd8 often is 14 l:lcl?!. What could be more narural
:Xd8 9 .i.g5 than to put the rook on an open file? The prob­
White now threatens to win material with 1 0 lem is that after 14 ... h6 1 5 .i.e3 (15 .i.h4 g5 1 6
lbd5. ..tg3 itJf6 i s also good for Black) 1 5 ...itJf6

90
 The Classical Varia tio n : White 's 7 th Mo ve A lterna tives

White would be in trouble if he allowed Alternatively, the position after 16 0-0 i.e6
... lbxd5. That means he has to go in for 1 6 1 7 .i.xe6 :Xe6! may appear dead drawn at first
i.b3 lbxe4 1 7 l:.c7 .i.e6 1 8 .i.xe6 l:.xe6 1 9 sight but a closer inspection will reveal a size­
l:.xb7 l:.a6 20 a3 lbd6! when Black has an edge able initiative for Black here. His rooks are
because he has a mobile central pawn majority more active, White's bishop is offside on g5
and White has still not completed his develop­ and the knight on c4 will soon be hit by ... b5.
ment. This position was reached in Teschner­ Acebal-Gallagher, Candas 1992 continued 1 8 f3
Escher, Stockholm 1962. The young Bobby b5 19 lbe3 h6 20 .i.h4 lbd3 21 lbds l:.c8 22 b3
failed to win but other players, including myself, l:.c2 23 l:.fd 1 lbb4! 24 lbxb4 .i.xb4 25 .i.£2 a6!
have since registered the full point in this end- 26 a3 .i.d2! (White is totally paralysed and can
ing. only watch while Black calmly improves his
14 ...lDc5 position by bringing the king to the centre and
playing ... f5) 27 .i.cS aS 28 �ft :C6 29 b4 a4
30 l:.ab1 <t;g7 31 l:.at f5 32 �g1 �f6 33 l:.ft
'iti>e6 34 l:.£2? (this loses material but passive
defence would also have lost) 34...l:.6xc5! 35
exf5+ gxf5 36 l:.xd2 l:.xd2 37 bxc5 l:.c2 0-1 .
1 6 . . ..i.e6 1 7 'iPb1 :ace
1 7 ...lbxe4 1 8 .i.xe4 .i.xc4 19 .i.xb7 l:.ab8 is
fine for Black but I wanted more. I saw a
sneaky way to improve this line.
1 8 .i.e3?
Thank you very much. White falls for the
trap. He should have played 1 8 l:.he 1 when
1 8 ... .i.xd5 1 9 exd5 should be slightly better for
Black as White's d-pawn is more likely to tum
1 5 lbc4?! out weak than strong.
A slight mistake as it allows Black some
tricks based on ... lbxe4, while the knight may
also get booted by ... b7-b5 at some point.
White is also not out of the woods after 1 5 0-0
i.e6 16 i.xe6 :Xe6 (16 ...lbxe6 1 7 .i.e3 lbd4 1 8
lbb3! should be a draw). Mark Hebden has won
this position a couple of times against grand­
masters.
The best move is 1 5 0-0-0! and after
1 5 ...lbe6 (not 1 5 ...lbd3+ 1 6 �b1 lDxf2 1 7 l:.dft !
and White wins but 1 5....i.e6 is again a possibil­
ity) 16 .i.e3 lbf4 1 7 .i.xf4 exf4 both 1 8 �bl
.i.e6! 19 ..i.xe6 l:.xe6 20 f3 f5! and 1 8 f3 .i.e6 1 9
lbb3 .i.xd5 20 l:.xd5 f5 ! lead to equality. Note
how in both cases Black plays ... f7-f5 to activate 1 8 ...lbxe4! 1 9 .i.xe4 .l:.xc4 20 .i.xb7 .l:.b8
his rooks. 21 .i.d5 .i.f5+ 22 �a1 .l:.c2
1 5 ....i.f8 Now White reached for his bishop to bring
This keeps the knight out of d6 and also it to b3 when he suddenly spotted my trick: 23
transfers the bishop to a more active post. .i.b3 :Xb3! 24 axb3 l:.c6! and there is no way
There is not much life on the long diagonal. to stop ... l:.a6 mate. The only chance to resist
1 6 0-0-0 was 23 l:.bt (although Black wins after 23...i.c5
On the last move castling queenside was 24 i.xc5 llxc5 25 llbd 1 llc2) but White just
best. Now it is rather risky. picked up his other bishop and played...

91
Pla y th e King 's Indian

23 ..ba7? :bxb2 24 .i.e3 .i.b4 25 g4 .i.c3! f3 lLJe6 14 ..te3 ..i f8 1 5 a3 b6 1 6 .:td2 ..ta6 17
0-1 <J;b1 1/2-1/2 Akobian-Yermolinsky, Agoura Hills
Now that's what I like to do to people who 2004.
exchange queens in the King's Indian. 1 1 .i.e3
White can also play 1 1 ..th4 but it doesn't
Game 32 really change the character of the game.
1 1 ......c6
0 . Moor-Ekstroem
There is no need for Black to complicate life
Ziirich 2002 with 1 t ...lbg4 12 lbds lLla6 13 c5. The text
emphasises that Black has control over his d5-
1 d4 iDf6 2 c4 g6 3 lbc3 .i.g7 4 e4 d6 5 square whilst there is a big hole on White's d4-
.i.e2 0-0 6 iDf3 e5 7 dxe5 dxe5 8 'frxd8 square. This can turn out to be an imponant
:Xd8 9 .i.g5 :es 1 o 0-0-0 positional advantage.
1 2 h3
I don't believe that Black should play ...lLlf6-
g4 even if given the chance. For example, the
position after 1 2 lLle 1 lbg4 1 3 ..ixg4 ..txg4 can,
no doubt, be assessed as level but Black has
given White the chance to rid himself of what
usually turns out to be the worse minor piece.
After 12 lLle1 the game Ludden-Nijboer,
Amsterdam 1 997 went instead 12 ... ..te6 1 3 lLlc2
..tfB 1 4 f3 a6!? 1 5 .:td2 b5 16 b3 lLlbd7 17
.:thd 1 .:teeS with a slighdy better position for
Black. White has no entry squares for his rooks
on the d-file and Black eventually took over the
initiative on the queenside. Normally in such
This is played considerably less often than positions Black plays with ... a7-aS in order to
the 10 lbdS of the previous game although it is secure the c5-square but Nijboer's plan looks
not inferior. It is just equally hannless. It usually more dynamic.
leads to a lengthy manoeuvring game. 1 2 .. .. ...i.f8
10 .. .. . h6 I shall repeat my comment &om the previ­
Black can also play 10. ..lt�a6 here when: ous game - there is not much life on the long
a) 1 1 lLlxe5 has usually been met by 1 t ...lLlc5 diagonal.
with a roughly level game after 12 ..txf6! ..txf6 1 3 iDd2 .i.e& 14 g4 iDbd7 1 5 f3 a6!?
13 f4 c6 14 b4 ..txe5 1 5 fxe5 lLJJ7 1 6 e6 .:txe6
17 ..ig4 .:te7 1 8 b5 <J;g7, Haik-Spassky, French
Ch. 199 1 . However, Uhlmann, a lifelong devo­
tee of this line recendy played 1 t ....:txe5!? 1 2
.:td8+ .:te8 ( 1 2...lLJe 8 13 f4 .:te6 14 ..tg4 i s sup­
posed to be good for White although I
wouldn't bet the mortgage on this verdict in
view of 14 ....:tb8Q 13 ..txf6 .:txd8 14 ..txd8
..ixc3 1 5 bxc3 ..th3 16 .:td 1 ..txg2 and he went
on to win this rather unbalanced endgame in
Wronn-Uhlmann, Dresden 2002. Food for
thought.
b) If White doesn't take on e5 then we
should see similar play to the main game. A
recent example is 1 1 lLld2 c6 12 lLlb3 lLlc7 13 The player conducting the black pieces in

92
 The Classical Varia tio n : White 's 7 th Mo ve A lterna tives

this game is an experienced international master 32...� 33 �e1 1

but a newcomer to our favourite opening. His Covering the b4-square so that White can
King's Indian career took off after he borrowed meet 33 ...We8 with 34 l:lc2.
a copy of SOKID from me during the last 33 ...lL!cS 34 tl:lxc5
Olympiad and then didn't want to give it back White can also play 34 l:ld8t i.xd8 35 lt:'lxc5
at the end (this may even have been his debut �e8 36 o!t:'lxb7 l:lxb7 37 l:ld1 although it is not a
game). Note how he sticks religiously to the cast iron draw after 37 ...l:ld7 38 l:lxc4 l:ldt+ 39
Nijboer plan that was mentioned in that book. l:lct l:lxct+ 40 �xcl f6!.
1 6 lbb3 .:ecs 1 7 h4 b5 1 S �b1 h5 1 9 g5 34 ... �xc5 35 .l:[dS+?!
lL!eS 35 i.g3! looks better when 35...l:le7
I suppose it is easy to say with hindsight but (35 ...i.d4 36 l:lxd4) 36 l:lct looks close to
White, by advancing his kingside pawns for no ettuality.
particular reason, has just laid the seeds of his 35 ...�e7 36 �c3 .l:[d7! 37 .l:[1 xd7+ .:Xd7
own defeat. The weakness of the pawn on h4 3S .:cs
will cost him the game. The point is of course that after 38 l:lxd7+
20 �f1 lbd6 21 lbas .:abS 22 �a7 .:as 23 �xd7 39 i.xe5 i.f2 the pawn on h4 drops.
�g1 .:abS 24 �a7 .:as 25 �g1 lbxc4! Now the endgame is very difficult for White.
Ekstrom had obviously also taken on board The remaining moves were:
the following advice given in SOKID, 'On no 3S ...�d6 39 �e1 �d4 40 �c2 .l:[b7 41 b3
account should the wimps who play this varia­ aS 42 .l:[dS+ �e7 43 .:cs c5 44 .:as cxb3+
tion be given a draw until every last possibility 45 axb3 c4 46 bxc4 .l:[b2+ 47 �d1 .:b3 48
has been exhausted. Perhaps they will achieve �xa5 .l:[xf3 49 .i.b4+ �d7 50 c5 .l:[b3 51
their objective, if they play extremely well, but .:S7+ �c6 52 .i.e1 .l:[f3 53 �e2 .l:[f4 54
they should at least be made to suffer.' �g3 .:Xe4+ 55 �3 .l:[g4 56 .:e7 �xeS 0-1
26 lbxc4 bxc4 27 lba4 .:SbS 2S �h3 �xh3
29 .:xh3 .:c7 30 .l:[h2 �e71 Summary:
1) 7 ...o!t:'la6 is an excellent move against
someone hoping to play the Petrosian System.
In my experience very few White players know
anything about this slightly unusual move and
even if they do it still looks like a comfortable
game for Black.
2) The Gligoric System (7 i.e3) can be one
of the toughest lines for Black to face. I suggest
that we sidestep this line by playing 7...o!t:'la6
(that move again) after which White has noth­
ing better than 8 0-0 and a transposition into
Chapter 5.
3) I used to face the Exchange Variation all
the time but ever since I staned to slag it off in
Better than 30...llb4 31 l:lhd2!. Black is now public people have stopped playing it against
ready to walk his king back to the centre and me. The endgame is, of course, just equal but
the £8-square becomes available for the knight. make sure you play it on till the very end as the
31 .l:[hd2 .l:[bb7 32 �f2 son of person who plays this variation is liable
Now if White tries 32 l:lc2 l:lb4 33 l:lcd2 to crack at some point. They are, with some
Black can play on with 33.)t:'lf8!. exceptions of course, psychologically weak.

93
 CHAPTER SEVEN I
The Samisch

1 d4 itJf6 2 c4 g6 3 itJc3 ..ig7 4 e4 d6 5 f3 recommend for Black, 6 ... c5, White will often
0-0 look to break through in the centre.
Our coverage of the Samisch, a variation 3) The move f3 secures the square e3 for
named after the German grandmaster Fritz White's bishop. It can now take up residence
Samisch, starts after the moves 1 d4 tt:'!f6 2 c4 there without having to worry about ...tt:'!g4.
g6 3 tt:'!c3 i.g7 4 e4 d6 5 f3 0-0 although it is The move f3 also has certain drawbacks, the
the move 5 f3 which characterises the variation. main one being that the price for this strong
What are the ideas behind this non-developing centre is clumsy kingside development. There is
move? also some potential for trouble on the dark
squares that were weakened by f3.
How docs Black combat this popular varia­
tion? Well there are three main methods.
1) He can strike in the centre in Benoni fash­
ion with c5
2) He can strike in the centre in traditional
King's Indian fashion with 6...e5
3) He can delay his central strike in favour of
queensidc play. The most common way to do
this is the Panno variation with 6...tt:'!c6.
I covered all these methods in SOKI D but
in this book I have decided to base the reper­
toire around Plan 1 , striking in the centre with
c5. This also happens to be Black's most popu­
1) It defends e4 so creates a secure and sta­ lar choice as it gives him the best chance of
ble centre. exploiting the weakened dark squares in the
2) With his centre so stable White can start white camp.
an attack on the kin!,>Side. i.e3, 'ii'd2, i.h6 as
well as g2-g4 and h2-h4 are typical moves when Recent Developments in the Simisch
White wishes to carry out this plan. On other In the diagram position White usually chooses
occasions White may attack on the queenside. between the moves 6 i.e3, 6 i.g5 and 6 tt:'!�-,rc2.
The Samisch is a flexible variation which does The fact that he even thinks about his 6th move
not commit White to any particular course of is a relatively new phenomenon. In the past he
action. In fact, in the line that I am going to always used to play 6 i.e3 because this was

94
 The Samisch

thought to prevent Black from playing 6...c5. you're opening repertoire involves immediate
True, Black can still play the move but it loses a transposition into an endgame a pawn down.
pawn and an impressive-looking white per­ Otherwise things could possibly get rather em­
formance from Karpov in the mid 1970's had barrassing.
convinced everyone that the sacrifice was un­
sound. The general feeling was that cS needed 6 .i.e3
further preparation and that Black had to pre­
pare the advance with either 6...b6 or 6 ...lDbd7. Gambit Accepted
But both of these moves have their drawbacks 1 d4 ttlf6 2 c4 g6 3 &3 .i.g7 4 e4 d6 5 f3
(b6 wastes a tempo and d7 is not always the 0-0 6 .i.e3 c5!
best square for the knight in these lines and it
also allows White to play lDh3-f2) and they
never really caught on. So until the early 1 990's
Black generally met 6 ..i.e3 with either 6 ... e5 or
6... lDc6. But there then occurred a revolution
that changed the whole landscape of the
Siimisch. Certain Black players started playing
6...c5 again and it became clear that the lines
which were supposed to cause serious prob­
lems for Black were actually quite harmless. A
theoretical debate on the merits of the pawn
sacrifice raged for a number of years but Black
eventually emerged triumphant (i.e. he proved
equality). White players simply stopped accept­
ing the pawn sacrifice. Most of them gave up I have already explained much about this
the Siimisch and played another variation but move above but I shall just stress that this is the
others decided to try their luck in the Benoni most positionally desirable way for Black to
positions that arise when White meets ...cS with combat the Siimisch and it was only because it
dS. Results here were also mixed and then lost a pawn that it took so long to become
White players began to realise that e3 might not popular. Once it became clear that the end­
be the ideal square for their bishop in these games a pawn down were fine for Black King's
variations. And why were they still playing 6 Indian players flocked to this variation in their
..i.e3 anyway? The main reason for the move, to droves after years of suffering in the other
stop cS, was clearly redundant. So many White variations of the Sarnisch.
players switched to 6 ..i.gS as they considered 7 dxc5 dxc5 8 "iFxdB
the position after 6 ...c5 7 dS to be a slight im­ Others:
provement on 6 ..i.e3 cS 7 dS, whilst others a) For a while White tried to make the line 8
preferred 6 lDge2 keeping all the options open eS lDfd7 9 f4 work but this has virtually disap­
for the bishop. Nowadays the three moves (6 peared from practice now. It is essential that
..i.e3, 6 ..i.gS and 6 lDge2) are regularly seen in Black immediately challenges White's grip in
international chess. I am suggesting that Black the centre. This is done by 9 .. .f6 and after 10
meets them all with 6 ... c5 and the Benoni posi­ exf6 he should offer the c-pawn again with
tions that arise from each are of course quite 10 ... lDxf6! and after 1 1 1i'xd8 l:xd8 1 2 ..i.xcS
similar (and there are even quite a few transpo­ play 1 2.....i.f5!. The most important factor in the
sitions from one line to another). I shall cover position is that White, having played c4 and f4,
each variation in tum although the lion's body has no pawn control over the imponant central
of theory is still concentrated on 6 ..i.e3. The squares e4 and d4. This was borne out quite
players who accept the pawn sacrifice may be splendidly by the continuation of Campos Mo­
few and far between today but it is still essential reno-Mortensen, European Club Champion­
to have a thorough understanding of why ship 1 991: 1 3 lD£3 lDe4! (the point of ... ..i.£5 is

95
Pla y th e King 's Indian

revealed; the knights are removed &om the lbeS looks at least equal for Black.
long diagonal to increase the scope of Black's 8 . . ..1:1.xd8 9 �xc5 tOe&
bishop on g7) 14 lbxe4 .ixe4 1 5 .ia3 lbc6 1 6
.ie2 lbd4 1 7 lbxd4 .ixd4 when Black's bish­
ops simply radiate power. He should have won
this ending but White managed to escape with a
draw.
b) A few players have taken the c-pawn
without exchanging queens. But this is rather
risky, for example 8 .ixc5 lbc6 and now the
following are possible:

Don't worry if you are rather confused by

the fact that Black has lost a pawn and allowed
the queens to be exchanged. It took the cream
of world chess about 50 years to understand
this position. If you are unfamiliar with this
position feel free to be a little sceptical. I sus­
pect that the more you look at the variation the
more it will grow on you.
Anyway, here is an attempt to define exactly
b 1) 9 .ie3 lbd7! 10 l:.ct (Diugy was later what compensation Black has for the pawn?
successful with 10 lbge2 lbdeS 1 1 lbf4 lbaS!? Well, firstly, the long diagonal has been opened
12 'ifa4 but after 12 ....id7 13 'ifb4 l:.c8 Black and the King's Indian bishop has a very bright
has good compensation) 1 1 lbh3 l:.d8 (the future. Secondly, Black has a decent lead in
drawback of retaining the queens can be seen development. He has four pieces in play to two
clearly here; whilst the black queen is aggres­ of White's. Furthermore, the presence of a
sively posted on aS, White's is struggling to pawn on f3 seriously hinders the development
escape the black rook's line of fire) 12 lbf2 of White's kingside which will, at best, be labo­
(best, according to Burgess, is 1 2 lbf4 with an rious. Thirdly, White has a big hole on d4 as
unclear game) 12 ...lbc5 13 .id2 .ixc3! 14 bxc3 well as several other squares on the queenside
.ie6 1 5 'li'c2 lbe5 1 6 .if4 lbxc4 17 .ie2 g5! 1 8 which Black's knights will be eager to occupy.
.ixgS lbd3+ 1 9 .ixd3 'li'xg5 and Black has a A fourth factor is that the white king can easily
winning position, Dlugy-Gelfand, Baleares find itself exposed to enemy fire. All these
1989. points put together would be worth far more
b2) 9 .!Llge2 'ifas t o .ie3 l:.d8 1 1 'ifct lDhS than a pawn if it wasn't for the fact that White
12 �f2 f5 1 3 exf5 lbb4 14 lbg3, Graf­ has one big trump of his own - the possibility
Zulfugarli, Dubai 2003, and now in this posi­ to play lbds.
tion I like the straightforward continuation We shall now examine the main lines
14 ...lbxg3 1 5 hxg3 .ixf5 best with good play through a couple of recent games (recent by the
for the pawn. standards of this variation anyway). Game 33
b3) The inventive grandmaster Jacob Murey features 10 .ia3 as well as a few other 10th
has tried to solve the problem of Black's active move possibilities whilst in Game 34 we shall
queen on aS by getting their first with 9 'ifa4!?. see White playing the immediate 10 lbds, the
However, the position after 9...lbd7 10 .if2 move which was once thought to be the refuta­
.ixc3+!? 1 1 bxc3 'ifaS 12 'li'xa5 lbxa5 13 0-0-0 tion of 6 ... c5.

96
 Th e Siimisch

...lLlb4; this idea is going to become familiar to

Game 33 you) 1 3 lLle3 (White did manage to survive a
more recent game after 1 3 lLle2 lLlb4 14 .ixb4
Fritz 6-Har Zvi
axb4 1 S lLJct but it was hardly a good adver­
Israeli Team Ch. 2000 tisement for the first player) 1 3 ...lLlb4 1 4 lLlh3
lLlcS (I don't see how White can hold his
My apologies for presenting the game of a queenside together) 1 S lLJ£2 e6 1 6 ..i.e2 b6 1 7
computer but with so few humans willing to lLlfd 1 lLlxa2 1 8 lLlc2 ..i.a6 19 ..i.xcS bxcS 20
accept the pawn sacrifice these days this was lLla3 lLlc1 ! 2 1 lLlbS llb8 22 ..i.ft a4 23 �d2
the most thematic recent game I could find. lLlb3+- 24 �c2 lld8 2S lLlbc3 lld2+ 26 �b 1
What is incredible is that Israeli league teams lLlaS 27 g3 a3 28 bxa3 ..i.xc4 29 f4 ..i.b3 0-1
are allowed to have a computer in their side. I b) t o lLlge2 tDd7 and now White has a
hope that they are soon banned. choice of bishop moves:
1 d4 �f6 2 c4 g6 3 �cl �g7 4 e4 d6 5 fl b1) 1 1 ..i.a3 tDdeS 1 2 lLlf4 e6 1 3 1ld1 llxd1+
0-0 6 �el c5 7 dxc5 dxc5 8 Wxd8 .:Xd8 9 1 4 lLlxd1 aS!, intending ...lLJb4, gave Black good
�xc5 �c6 1 0 �al play in Murey-Degraeve, Cappelle Ia Grande
1 993.
b2) 1 1 ..i.e3 tDdeS 1 2 lLlf4 lLlb4 13 �£2 (1 3
lld1 fails to the nice variation 1 3 ...lLlxf3+! 1 4
gx£3 ..i.xc3+- 1 5 bxc3 lLJc2+ 1 6 'Ote2 llxd 1 1 7
�xd1 lLlxe3+- with a clear advantage for Black)
1 3. ....i.c6! 1 4 lLJcdS (1 4 lLlxe6 fxe6 is good for
Black) 1 4.....i.xdS 1 S lLlxdS (1 S cxdS lldc8)
1 S...lLlc2 1 6 l:tct lLlxe3 1 7 �xe3 ..i.h6+ 1 8 f4 e6
1 9 lLlc3 gS! 20 g3 lLlg6!, Gunawan-Gelfand,
Minsk 1 986 and Black wins the pawn on f4 as
21 f5 g4+ picks up the rook on c1 and 21 lLle2
allows 21 ...gxf4 22 gxf4 eS.
1 0 ...a5!
Black has a much quieter system where he
What is the point of this bishop move? Well, plays moves such as 1 O... b6 and .....i.b7 and
one of Black's main ideas is to play ... tt:xJ7. The relies on long-term pressure. The text takes the
knight is not doing very much on f6 so it often fight to White at once by introducing the possi­
retreats to d7 in search of a new home (cS or eS bility of ...lLlb4.
for example). In doing so it also gets out of the 1 1 l:ld1
way of the bishop on g7. ..i.a3, therefore can be White almost always plays this move. 1 1
considered a prophylactic retreat so that the tLkJS lLJxdS 1 2 cxdS lLlb4 1 3 0-0-0 e6! was fine
bishop on cS will no longer be attacked if Black for Black in the game Beliavsky-Nunn, Am­
plays ... tt:xJ7. White chose the a3-square in or­ sterdam 1 990.
der to protect the pawn on b2. He is still plan­ 1 1 ...�e6
ning to play tt:xJs. This is better than exchanging on d 1 as that
Let's have a quick l(XJk at some of the lesser would only help the white king to a better
played alternatives. home on the queenside, e.g. 1 t ...llxdt+ 1 2
a) to lld1 was quite popular at one point but �d 1 lLlb4 1 3 lLlge2 b 6 14 lLlct e 6 1 S lLla4
after the hammering it took in Razuvaev­ tDd7 1 6 b3 ..i.a6? (after 1 6.....i.b7 the situation
Shirov, Bundesliga 1 992 it more or less disap­ would have been far from clear as 1 7 .ib2
peared. Play continued to .. .:Xd 1+ 1 1 lLlxd 1 (1 1 could then have been met by 1 7 ... lLle5! 1 8
�xd1 might be an improvement though Black lLlxb6 lld8+ 1 9 �e2 ..i.h6!) 1 7 ..i.b2! with a clear
is at least equal after 1 t ...tt:xJ7 1 2 ..i.a3 ..i.xc3 1 3 advantage for White, Graf-Nunn, Manila 1992.
bxc3 tDdeS) 1 t ...tDd7! 1 2 .ia3 aS! (preparing 1 2 �5 �b4!

97
Pla y th e King 's Indian

a) 17 ltJh3 (17 cxb5 lLlxd5 1 8 exd5 ..i.xd5

must be good for Black) 1 7...l:lac8 1 8 ..i.e2 bxc4
19 lLlhf4 lLlxd5 20 lL!xd5 ..i.xd5 21 llxd5 lLd5
22 exd5 ..i.d4t- 23 �g3 c3 with a clear advan­
tage to Black, Bigler-Har Zvi, Bicl 1993.
b) 17 a4!? is an interesting novelty from
Rowson or, more likely, from his computer
programme. I almost played this move myself a
few years ago after Fritz kept telling me it was
good for White but in the end I didn't really
believe it promised anything. At any rate it is
certainly an improvement on the above varia­
tion. The point of 17 a4 is to try and force
Black to capture on c4 before he would like to.
This excellent discovery of Shirov was basi­ Is White better after 1 7 ... bxc4 1 8 ..i.xc4 l:lac8
cally the death knell of White accepting the 1 9 ..i.b3 ..i.xd5 20 exd5 ..i.£B? I'm not sure he is
gambit in top class chess. That doesn't mean but in the game Rowson-Kotronias, Hastings
that Black is better after ... lL!b4, just that White 2003/04 Black preferred 1 7 ... bxa4 which does
has no hope of obtaining an advantage. When at least keep open the option of winning the
that happens the theoreticians move away en game. A fter the further moves 1 8 lL!e2 l:lab8 19
masse in search of fresh pastures. ltJc1 l:lb2+ 20 ..i.e2 l:lc8 21 ltJe7 l:le8 22 ltJd5
In earlier games Black had played 1 2 ... ..i.xd5 l:lc8 23 ltJe7 Rowson suddenly decided to take
13 cxd5 lL!b4 but 14 ..i.b5! lLlc2+ 1 5 �f2 lLlxa3 a draw by repetition. He could have played on
16 bxa3 e6 1 7 d6 e5 1 8 lLle2 ..i.£B 1 9 d7 ..i.xa3 with something like 23 �e3 but it wouldn't be
20 g4! was shown to be good for White in the risk free.
game Kramnik-Nunn, Manila Olympiad. 1 992. 1 4. . ..:Xd8 1 5 lt:ld5
The logic behind 12 ...lL!b4 is not difficult to
understand. Black has a massive lead in devel­
opment so he seeks to complicate the game
before White can consolidate. There is also the
little matter of ... ltJc2+ on the cards, which will
deprive White of his dark-squared bishop and
ruin his pawn structure.
1 3 lt:lxe7+
White accepts the challenge. 1 3 ..i.d3 is also
possible when 13 .....i.xd5 14 cxd5 lLlxd3+ 1 5
lLd3 e6 leaves Black with full compensation
for the pawn. If White plays d6 (with or with­
out .i.e?) Black will round up the pawn while
White is trying to get his kingside out.
lines that White should definitely avoid are White is two pawns up but miles behind in
13 lLlxb4 l:lxd 1+ 14 �xd 1 axb4 1 5 ..i.xb4 l:lxa2 development. He will now try and complete his
and 13 ..i.xb4?! axb4 14 lLlxb4 lLld7! 1 5 l:ld2 development under the cover of his one good
lLlc5 as in both cases his position is on the piece, the knight on d5. I f Black takes the
verge of collapsing. knight then White will recapture cxd5 and ob­
1 3 ...�h8 1 4 l:lxd8+ tain a strong passed pawn. The centre would
White can also keep rooks on the board and still be blocked so White should have enough
play 14 ltJd5. After 14 ...ltJc2+ 1 5 �£2 lLlxa3 16 time to get his pieces out before anything seri­
bxa3 b5! (the standard idea to break White's ous can happen to him. Black has another plan.
grip on d5) there is: He is going to undermine the defenders of this

98
 Th e Siimis c h

knight so that it won't be able to maintain its take the pawn and so must retreat his bishop to
dominant position in the centre. a passive SCJUare. White could have played 25
By the way, the position after 1 5 .i.xb4 axb4 l:.cl but then his bishop would be pinned and
16 tDd5 l:.aB 1 7 lbxb4 tDd7! is promising for after an exchange of rooks (all White can play
Black despite his three-pawn deficit! for now) the game would be completely drawn.
1 5 . . .llJc2+! The computer program playing White is not
Black starts with the obvious move. Once he interested in such a variation. Despite the ad­
takes on a3 White will only be able to claim, at vances made in computer play they still tend to
most, a one and a half pawn advantage. be very materialistic. It was too much for Fritz
1 6 �2 to resist the chance to go three pawns up in an
16 'itd2 was the choice in the prototype ending.
game, Kramnik-Shirov, Bundesliga 1 992. After 1 9 llJxf6? ..txf6 20 cxb5 :c2+! 21 �e3
16 ... tt'lxa3 17 bxa3 b5! (the star undermining l:.xa2
move) 1 8 tt'lh3 a draw was agreed in view of White's problem is that once a3 drops
the variation 1 8 ... bxc4 19 ..ixc4 tt'lxd5 20 exd5 Black's a-pawn will be tremendously strong as it
..ixd5 21 .i.xd5 l:.xdS+. I would be tempted to receives great support from the bishops.
continue a little with Black as his strong bishop White's extra pawn on the kingside is virtually
looks more important than the useless extra a­ irrelevant in such a position.
pawn. 22 f4 ..td8!
1 6 ...llJxa3 1 7 bxa3 b5!
Again this undermining move. White's con­
trol of d5 is soon to be history. Kramnik has
also suggested the more ambitious 17 ... tDd7 if
Black is playing for a win . The idea is to take
control of the dark SCJUares.
1 8 llJh3 :c8

Excellent chess. From b6 the bishop will

firmly blockade White's passed pawn and also
menace the white king.
23 ..td3 ..tb6+ 24 �3 :Xa3 25 :d1 a4!
As the saying goes, passed pawns have to be
pushed.
26 liJf2?1 :b3! 27 :b1 a3 28 :xb3 ..txb3
1 8 ... bxc4 19 .i.xc4 l:.cB 20 ..ib3 a4! is also 29 ..tb1a2 30 ..txa2 ..txa2
good enough for e<juality. The rest of the game is a simple technical
The move played by Har Zvi, 1 8...llc8, was exercise for the grandmaster playing Black. A
analysed by Shirov in the notes to his game knight and two pawns are rarely a match for
with Kramnik. He now gave the variation 1 9 two bishops and certainly not here.
tt'lhf4 bxc4 20 tt'lxe6 fxe6 21 tt'lb6 tt'lxe4+l 22 31 lbg4 �g7 32 llJ&3 �6 33 g4 �e6 34
�e3 l:.c6 23 �xe4 l:.xb6 24 .i.xc4 as level. At g5 ..txe3 35 �xe3 ..tc4 36 b6 ..ta6 37 �d4
first I was not so sure as White's pieces l(x>k ..tb7 38 �e3 �d6 39 �d4 �c6 40 h3
more active. However Black has a strong con­ �xb6 41 h4 �c7 42 �e5 �d7 43 �d4
tinuation 24.. .llc6 25 .i.b3 a4! when White can't �d6 44 �d3 �c5 0-1

99
Pla y th e King 's Indian

Game 34
Graf-Guseinov
Dubai 2003

1 d4 lLlf6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5 f3

0-0 6 �e3 c5 7 dxc5 dxc5 8 •xd8 llxd8 9
�xc5 lt'lc6 1 0 lLld5 lt'ld7!

a) 14 �d1 l:le8 1 S �c2 �g7 1 6 �d6 �S!

17 g3? �d7! 18 llxb7 �c6 19 l:lb4 aS! (by of­
fering the b-pawn Black gains rime to start a
dangerous artack: note how he steered clear of
19 ...ti)xf3? 20 ti)xf) �xe4+ 21 �d3 �x£3
which wins his two pawns back but also solves
the problem of White's undeveloped kingside
for him) 20 l:lb6 l:led8 21 cS (after 21 �xeS
�a4+! 22 'ifi>b2 �xeS+ 23 �a3 �c2 the white
It was the discovery of this possibility that king is in a mating net) 21 ...�a4+ 22 'iti>b1 ti)d7!
alerted Black to the fact that 6...cS might be a 23 l:lbS and now instead of 23 ...�cS?! (which
good way to meet the Samisch. After the game also won) Black should have just played
Karpov-Barle, Ljubljana-Portoroz 1 97S, which 23 ... �xbS 24 �xbS ti)xcS! 2S �xeS l:lac8
went 10 ... ti)xdS 1 1 cxdS �xb2 1 2 l:lb1 �c3+ when White can't move his bishop from cS on
13 �f2 with advantage to White, the general pain of checkmate.
consensus was that 6...cS was a dubious pawn b) I t is not an improvement for the king to
sacrifice. During the next ten years only a few head to the kingside as after 1 4 �f2 �d4+ 1 S
die-hards tried it but after the discovery of �g3 (1 S �e1 l:le8) 1 S. ..l:te8 1 6 �gS ti)f6! it
10...ti)d7! just about everyone jumped on comes under fierce artack. Gil-Howell, Gausdal
board. 1 986 didn't last long: 17 ti)h3 ti)hS+ 1 8 �h4
1 1 �xa7 'iti>g7! 1 9 g4 h6 20 �xh6t 'iti>xh6 21 gxhS fS! 22
White gives up his important bishop, in the �g3 fxe4 23 �g2 gxhS 24 f4 l:lg8+ 2S ti)gs h4+
hope that his knight will prove to be dominant 0-1 as it is mate next move.
on dS. 1 1 ...lLlxa7 1 2 ll'lxa7+ �8 1 3 lt'ld5
After 1 1 ti)xe7+?! it is doubtful whether 1 3 ti)xc8? �xb2 1 4 l:lb 1 �c3+ 1 S 'ifi>f2 �d4+
White can even maintain the balance, e.g. 1 6 'iti>e1 l:taxc8 and Black has an extremely dan­
1 1 ...�e7 1 2 �xe7 �xb2! 1 3 l:lb1 (the position gerous initiative. 1 7 l:lxb7 fails to 1 7 ... �b6.
after 13 �xd8 �xa1 is very promising for 1 3 . . .�xb2 1 4 llb1
Black as he is able to co-ordinate his forces 1 4 l:ld1 has been played a few times. After
quicker than White. For example, 1 4 ti)h3 1 4...ti)cS (14...ti)b6!?) 1 S �2 both Timman­
�c3+ 1 S �d1 �eS 1 6 ti)f2 ti)b6 1 7 ti)d3 �d4 J.Polgar, Paris (rapid) 1 992 and Brenninkmeier­
1 8 �1 ti)a4 1 9 �aS �e6 20 ti)b3 �eS 21 Troyke, Groningen 1 992 continued 1 S ...�d7
ti)d2 b6 22 �b4 aS 23 �a3 �3+ 24 �c2 1 6 �c3 l:tac8 1 7 �e2 �e6 1 8 'it>f2 and now
ti)xa2 and Black was well on top in James­ Polgar played 1 8...l:tc6 and Troyke 1 8 ... l:ld6 but
Wells, Irish Open 1 993) 1 3...�c3+ and now against either move White could have obtained
possible are: a good game with 1 9 l:lb 1 ! Black should proba-

1 00
 The Samisch

bly have retreated his bishop before it got cut side doesn't solve White's problems. The game
off on b2, e.g. 15 .....tg7 16 tt:kc3 ..td7 17 ..te2 Levitt-Watson, London 1 990 continued 1 5 ...b6
l:tac8 1 8 �f2 ..te6 1 9 <lr'e3 l:td6 and White 1 6 g5 ..tb7 1 7 h4 l:tac8 1 8 f4 (White looks as if
hasn't really gained by having his rook on dt as he possesses one of the loosest positions of all
opposed to b 1 . time) 1 8....!Dc5 1 9 e5 ..txd5 20 cxd5 l:txd5 with
14 .i.g7
• . . advantage to Black.
c) 1 5 h4 has the intention of bringing the
rook on h 1 into play as quickly as possible. The
most logical reaction is 1 5 ... .!bc5, meeting 1 6 h5
with 1 6 ...g5 and 1 6 lbh3 with 1 6 ... ..te6. In
Ehlvest-Gelfand, Polanica Zdroj 1 997 Black
preferred 1 5 ... lbb6 and after 16 h5 ..te6 1 7
lbh3 l:tac8 t 8 lbg5 ..txd5 1 9 cxd5 l:tc2 20 ..td3
..tc3t 21 �ft l:td2 22 hxg6 hxg6 23 ..te2 1:txa2
an extremely unclear position had been reached
which Black evenrually won in 63 moves.
1 5 ...lLlc5 1 6 lLlf2 .i.e& 1 7 .i.e2
17 lbd3 l:tac8 18 ..te2 (after 18 lDxcS l:txcS
1 9 l:txb7 l:taS Black will win the pawn on a2
and gain a powerful passed a-pawn; with 19
For his pawn Black has an extremely strong l:tb3! however White should be able to achieve
dark-squared bishop and a couple of weak a draw) 1 8...b6 (18...lba4 is another possibility)
white pawns on the queenside to take aim at. 1 9 0-0 lbb7!? (19 ...lbxd3 20 ..txd3 l:tcS is an-
He also has a fine ourpost on c5 for a knight. other idea) 20 l:tfct lbaS 21 l:tb4 l:tc6 22 l:tc2?
True, White also has a strong knight on d5 but (a tactical oversight; after 22 �ft l:tdc8 23
Black is often able to play around this piece. lb3f4 ..th6 24 g3 l:tcS Black has strong pressure
The knight on d5 will only be taken if Black has for the pawn) 22 ... ..txd5 23 exd5 l:txdS! 24 cxd5
a concrete follow up in mind (such as an inva­ l:txc2 25 <lr'fl 1;e7 and Black was better in
sion to the 7th rank. There has to be a good Georges-Gallagher, Ziirich 1 994 as his queen­
reason to allow White to straighten out his side pawns are more dangerous than the white
pawn strucrure with cxd5. I suspect the posi­ d-pawn.
tion is dynamically balanced but in practice 1 7 :aca 1 8 o-o b6
...

Black has done well. It is quite notable that Black decides to manoeuvre his knight to aS
even Karpov or Korchnoi have struggled to in order to exert maximum pressure on the
make anything out of the extra pawn. white c-pawn (as he did in Georges-Gallagher
1 5 lLlh3 above). The other main idea, which we have
The text gets the Karpovian seal of ap­ already seen in line 'a' in the 1 5th move notes,
proval, but there are a number of alternatives: is to quickly transfer the rook to the a-fLie. for
a) 1 5 tt:k2 .!Des 16 tt:kc3 (16 .!Dct ..te6 1 7 example, 1 8...lld6 1 9 l:tfd 1 l:ta6 20 l:td2 l:ta3
lbd3 transposes to the note to move 1 7 i n the was fine for Black in Korchnoi-J.Polgar, Ro­
main game) ..td7 (16 ... ..te6 1 7 ..te2 ..txc3t 1 8 quebrune 1 992.
lbxc3 lbd 3t is given as equal by Hazai) 1 7 ..te2 1 9 .l:l.fe1
l:tac8 1 8 0-0 l:tc6! (transferring the rook to the Graf may have been concerned about Black
a-file is a standard way for Black to increase his playing ... fS, which is another one of his typical
pressure in this line) 1 9 l:tfdt l:ta6 20 f4 ..txc3 ideas in this position. However, Black must
(Black rids himself of his bishop before e5 ren­ time this advance well and he is often better off
ders it redundant) 21 lbxc3 l:ta3 22 l:tbct l:tc8 holding it in reserve until he has tied White
23 e5 ..te6 and Black had full compensation for down on the queenside.
the pawn in Wells-Lamoureux, Oakham 1993. 1 9 . . .lLlb7 20 .l:l.ec1 .i.h6 21 .l:l.d1
b) 15 g4. 1bis immediate action on the king- Obviously not 21 l:tc2 ..txd5 22 exd5 l:txd5

101
Pla y th e King 's Indian

21 ...lba5 22 l:l.b4 l:l.c5 l:l.b2+ 37 �g3 lL!b6 38 l:l.c7 lL!xd5 39 l:l.xa7

Preparing to double on the c-ftle. Graf was lLle3 40 �e4 g5 41 h4 l:l.g2+ 42 �h3 l:l.xg4
obviously not feeling too comfortable as he 43 �c2 gxh4 44 �b3 l:l.g3+ 45 �h2 l:l.xf3
now elected to return the pawn without alleviat­ 46 l:l.xf7+ �h6 47 �h1 l:l.h3+ 48 �g1 l:l.g8+
ing any of the pressure. 49 �2 l:l.g2+ 0-1
23 f4?! �xd5 24 exd5 An impressive performance from Black. It is
24 l:.xdS might be better. not an easy task to despatch the German no. 1
24 . . .�xf4 25 lL!e4 l:l.c7 26 �2 in such fashion.
26 d6 just loses a pawn.
26 . . .lL!b7? 6 �e3 c5
A desirable move but tactically flawed. In­ Gambit Declined
stead 26... £5 27 g3 l:.e7! 28 gxf4 (28 tBc3)
28...l:.xe4 29 �f3 l:.de8 30 ..i.f1 tBb7 leaves Due to their lack of success in the Gambit Ac­
Black with a clear positional advantage which cepted more and more White players turned to
may weU prove sufficient to win the game. the Benoni positions that usually arise from the
27 l:l.a4?! Gambit Declined. I shall now look at the vari­
White misses his chance: 27 cS! bxcS 28 ous possibilities through these games. The first
ltJxcS tBxcS 29 l:.xf4 looks like a draw. two feature 7 tBge2, the most common move,
27 . . .lLld6 28 lL!f6 whilst in the third White just plays the immedi­
I wouldn't wish the long term suffering in­ ate 7 dS.
volved after 28 tBxd6 ..i.xd6 on anyone but it
may be a better chance. Game 35
28 ...�e51 29 lL!xh7+
Bogdanovski-Kempinski
29 tt)g4 ..i.g7 just leaves the knight horribly
placed.
Ha/kidiki 2002
29 ...�g7 30 lLlg5 �f6! 31 c5
I f the knight moved (or is captured after 3 1 1 d4 lL!f6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5 f3
h4) then 3 t ...tBe4+- foUowed b y 3 2. . .tBc3 picks 0-0 6 �e3 c5 7 lLlge2 "ifa5!?
up an exchange.
31 ... bxc5 32 lL!f3 c4

A very tricky move. The normal continua­

tion is 7 ...tBc6 8 dS tBeS 9 tBg3 and we have
Black has a decisive advantage as his pieces, entered a heavily analysed variation. The point
including the respective passed pawns, are far of the text is to wait for 8 'ii'd2 before playing
superior to their counterparts. He eventually 8 ... tBc6. This transposes into a similar variation
translated this superiority into a mating attack. which is normally reached via the move order
The remaining moves were: 7...tBc6 8 'ii'd2 'ii'aS. But most White players
33 l:l.c1 c3 34 �d3 lL!c8 35 l:l.c4 l:l.b7 36 g4 discarded this line, long ago, in favour of 8 dS.

1 02
 The Samisch

It seems that the inclusion of the queen moves and he has space problems.
is not unfavourable for Black. In Game 36 we B b5!
. . .

shall take a look at this line while here we shall

examine the 8th move alternatives for White.
8 d5
The only other possibility (apan from 8
'it'd2) is 8 ttJct and this is a good moment for
Black to exchange in the centre. After 8...cxd4:
a) 9 .i.xd4 (a strange choice) 9 ...itJc6 to lbb3
'it'd8 (I suppose Black could also have swung
his queen over to the kingside) 1 1 .i.e3 .i.e6 1 2
l:ct lbe5 1 3 lbd5 .i.xd5 1 4 cxd5 e6 (Black
must play actively as if White can complete his
development before anything nasty happens to
him his space advantage and bishop pair will
give him the edge) 1 5 dxe6 fxe6 1 6 .i.e2 (on 1 6
'it'd2 Black can play 1 6...d5) 1 6...lbxe4! ( I cer­ The correct response to White's premature
tainly approve of that one) 1 7 fxe4 1i'h4+ 1 8 closing of the centre. Black now gets a good
�d2 'it'xe4 1 9 l:c3 1i'xg2 (why is Black dis­ version of the Benko Gambit.
tracted by this pawn: after 1 9 ... d5! I consider 9 cxb5 a6 1 0 lbgJ
him to be clearly better) 20 �cl l:ac8 21 l:et to lbct , heading for b3, is the move that
l:xc3+ 22 bxc3 l:c8 23 .i.d4 'irxh2 24 �b 1 White would like to play but this has the draw­
'it'g2 25 lbd2 1i'd5 '1•-'l• Bischoff-Stellwagen, back of leaving the rook on at undefended. So
Pulvermuehle 2004. what, I hear you say. But it does matter. After
b) 9 lbb3 is more logical. Now I recommend to ...lbxe4! 1 1 fxe4 .i.xc3+ 12 bxc3?! 'irxc3+
the queen sacrifice 9 ... 1i'xc3+ to bxc3 dxe3. Black wins.
Only joking! I would love to but Black can't Perhaps the sad-looking 10 b6 is White's
have enough here after 1 1 1i'd3. But Dautov best move when Black should probably just
has pointed out that 9 ... 1i'h5 to lbxd4 itJc6 reply to...1i'xb6.
gives Black an easy Mar6czy type position. 1 0 . axb5 1 1 �xb5 �a6 1 2 �xa6 lLlxa6 1 3
..

There are no practical tests but a possible con­ 0-0 lLld7

tinuation could be something like 1 1 1i'd2 (1 1 Black has more pressure than usual in the
..te2 1i'h4+ is quite irritating for White after Benko, mainly because the knight on g3 is out
both 12 g3 1i'h3 and 12 .i.£2 1i'g5) t t ...lbxd4 of the game. Kempinski now quickly polished
12 ..txd4 .i.e6 and now: off some pretty inept play from White.
b 1) White could try and embarrass the black 1 4 .l:l.c1 lLle5 1 5 'ife2 c4!
queen with something like 13 h4 l:fc8 14 c5 (or
14 b4 ..txc4 1 5 g4 lbxg4 with plenty of play for
the piece) but then 1 4...:Xc5! 1 5 g4 (obviously
Black has more than enough play for the ex­
change if White takes on c5) lbxg4 1 6 fxg4
..txg4 should not be worse for Black and looks
like fun to play.
b2) 1 3 lbd5 is the solid choice when either
capture on d5 leads to a balanced game, e.g.
1 3 ... lbxd5 14 .i.xg7 �xg7 1 5 cxd5 .i.d7 or
13.....txd5 1 4 cxd5 lbd7 1 5 .i.xg7 'iPxg7 16
..tb5 lbf6 with equality. These Mar6czy type
positions can be unpleasant for Black but only
when there are more minor pieces on the board

1 03
Pla y th e King 's Indian

The d3-square is the advanced base camp point of view) so it is a shame to waste such
that will enable Black to destroy the white posi­ material.
tion. b) It is stiU very relevant as there have been
1 6 a3 tt:ld3 1 7 ltc2 lbac5 1 8 llJd 1 'ifb5 1 9 very few games in this line in the intervening
�d2 lba4 years.
1 9...lLlb3 foUowed by ...lLld4 also looks 9 d5
good. Obviously White is not forced into this ad­
20 �c3 ll:\xc3 21 ll:\xc3 'ifb3 22 llJd1 l:ab8 vance but other continuations promise nothing:
23 f4 a) 9 dxcS (this anti-positional caprure is usu­
White fmaUy plays an active move but it just aUy only good when it wins a pawn, and even
offers Black an additional target. then ...) 9 ... dxc5 10 lLldS (otherwise the white
23 ...e6! 24 dxe6?! queen is just exposed on the open file)
Understandable, as otherwise Black will just 10 ... 'irxd2+ (10 ...lLlxd5 1 1 'irxaS lLlxe3! is a
exchange on dS and pick up the weak white typical and strong queen sacrifice but 1 1 cxdS
pawn on that square, but after the text White lLlb4 1 2 lLlc3 is less clear) 1 1 ii.xd2 lLld7! fol­
loses his f-pawn. lowed by ...e6 with an easy game for Black.
24...fxe6 25 l:lf3 ll:\xf4! 0-1 Note that this is stronger than winning a pawn
White resigned as the f-pawn IS JUSt for with 1 t ...lLlxdS 12 exdS �xb2 1 3 l:tb1 lLlcs 14
starters, e.g. 26 'ird2 and 26 'ir£2 both lose to lLlc3! when White has the advantage.
26 ...lLlh3+! whilst 26 l:txb3 lLlxe2+ 27 lLlxe2 b) 9 0-0-0 is most simply met by 9 ... cxd4 10
cxb3 is obviously hopeless as weU. lLlxd4 lLlxd4 11 �xd4 �e6 1 2 �b 1 l:tfc8 13 b3
a6 and in view of his king position, it's probably
safest for White to liquidate into an equal end­
Game 36
ing with 1 4 lLld5 'irxd2 1 5 l:txd2 �xdS 16 cxdS
Lehtivaara-Gallagher
lLld7 1 7 �xg7 �xg7 1 8 �e2 aS.
Neuchate/ 2004 Another decent possibility for the more am­
bitious Black player is 9 ... a6 10 �b 1 e6! and
1 d4 ll:\t6 2 c4 g6 3 ll:\c3 �g7 4 e4 d6 5 f3 now:
0-0 6 �e3 c5 7 ll:\ge2 'ifa5 8 "ifd2 ll:\c6

b1) 1 1 dS exdS 12 lLlxdS lLlxdS 1 3 cxdS

This game may be hot off the presses but we 'irxd2 14 �xd2 lLleS with a pleasant game for
have now reached a position which I covered Black;
extensively in my 1 995 book on the Siim.isch. I b2) 1 1 �h6 bS 1 2 �xg7 �xg7 1 3 dxcS b4!?
shaU be drawing quite heavily on this material (there is nothing wrong with 1 3. ..dxc5 e.g. 14
for a couple of reasons: lLlc1 l:tds 15 'ire1 .:.xd1 1 6 'irxd1 lLld4) 1 4
a) Hardly anybody bought the book (proba­ lLld S exdS 1 5 cxdS dxcS!? 1 6 dxc6 �e6 with
bly my best one as weU, at least from a technical good attacking chances for Black, Gheorghiu-

1 04
 The Samisch

M.Piket, Lugano 1 989. 1 2 a4 was the choice in Kramnik-Gelfand,

b3) 1 1 dxcS dxcS 12 ttk1 l:ld8 1 3 1i'f2 llxd1 Linares 1 993. After 1 2 ... 1i'b4 it is dubious to
14 tDxd1 (Schneider-Hazai, Espoo 1 988) and play 1 3 lD1a2 on account of 1 3 ...lDxc4! 14 'iVd3
now Hazai gives 14 ...1i'd8 1 S ttk3 b6 as Black's when Black has the choice between an interest­
best. White would struggle to claim equality in ing queen sacrifice, 14 ... lDxe3!? 1 S lDxb4
this position. lDxg2+ 1 6 �£2 lDf4, or an ending with an ar­
9 . . .�e5 1 0 �c1 mada of passed pawns after 14...lDxb2 1 S 1i'c2
With the black queen on aS, the text makes (1S lDxb4 tDxd3+ 1 6 lDxd3 lDxe4!) 1 S...i.xa4
more sense than 1 0 tDg3. At any rate, its what (1S ...lDxdS!?) 1 6 lDxb4 bxc2 1 7 lDxc2 lDd7.
everybody plays. That's what I wrote 9 years Therefore Kramnik defended c4 with 13 b3
ago. We do now have an interesting example of and now:
10 tDg3. Dautov-Kempinski, Munich 1 997 a) 1 3 ...lbfg4 is the move crying out to be
continued 1 0... a6 1 1 f4?! (1 1 .i.e2 looks more played, but unfortunately it's unsound: 14
solid) 1 1 ...tDed7! (there is no need for Black to lD1a2! lDxe3 1 S lbxb4 lDxg2+ 1 6 �£2 cxb4 1 7
get involved in 1 1 ...tDeg4 1 2 .i.g1 as after the lDa2 lDh4 1 8 f4 f5 ! 1 9 �g3 gS! is given a s un­
text White has serious problems with his cen­ clear by Kramnik, and I would like to add the
tre) 12 h3 (Dautov points out the variation 1 2 following possible continuation: 20 fxgS (20
i.e2 hS! 1 3 0-0 b S 14 cxbS h 4 1 S tDh 1 axbS 1 6 fxeS .i.xeS+ 21 �£2 fxe4+) 20... fxe4 21 �4
i.xbS when the standard King's Indian trick (21 llaft tDeg6! 22 .i.g4 .i.eS+ 23 �h3 lDf3
16 ... tDxe4! leaves White in bad shape) and now wins for Black) 2t ...llf3! 22 h3 llafB 23 llag1
Black should have played 12 ... bS! 1 3 cxbS hS 1 4 ll8f4+ 24 llg4 (24 1i'xf4 � 24...llxg4+ 2S
bxa6 h4 1 S tDge2 .i.xa6 with excellent compen­ hxg4 lDg6t 26 �hS lDf4+ 27 �h4 lbg2+ 28
sation for the pawn. �hS .i.e8+ 29 g6 .i.xg6t 30 �gS
1 0 ...a6
Black prepares to play ...bS.
1 1 .b2
White has also played 1 1 a4, intending to
meet 1 1 ... .i.d7 with 1 2 lla3 when Black can no
longer play ... bS. Therefore it is best to counter
in the centre at once with 1 1 ... e6. This occurred
in Spassky-J.Polgar, Budapest (m/3) 1 993 and
Black achieved good play after 1 2 lla3 exdS 1 3
cxdS tDhS! 14 .i.e2 f5 1 S exf5 gxf5 1 6 .i.h6
'iVb4 17 .i.xg7 lDxg7 1 8 f4 ttk4 19 .i.xc4 1i'xc4
20 aS .i.d7 21 lD1e2 and now, according to
Nikitin (Spassky's second in the match), the
most accurate continuation was 2t...lDhS!, im­
mediately improving the position of the worse I shall claim the diagram posltlon as my
placed piece. After 22 0-0 lDf6 23 llf3 llae8 study. Black mates with 30....i.f6+l (the immedi­
Black would have a good game. ate 30... �f7 allows 31 llxh7!) 31 �h6 �f7!! 32
1 1 ...�d7 .i.xf3 .i.g7+ 33 �gS h6t 34 llxh6 .i.f6 mate.
1 t ...bS looks premature but 1 1 ...e6 is a pos­ Chess can be a beautiful game.
sible alternative. However, Kramnik provides the cold
1 2 f4?! shower: 1 7 �xg2! bxc3 1 8 1i'xc3! and whilst
Lehtivaara, who was unfamiliar with the po­ Black is not completely lost, he doesn't have
sition, became nervous about the potential anything like as much play as in some of the
tactics on the queenside and decided to change other typical Slimisch queen sacrifices.
the course of the game (I was playing very b) 1 3 ... lDxf3+ 14 i.xf3 lbxe4 1 S lDxe4
quickly but in reality I was struggling to re­ 1i'xd2+ 1 6 �xd2 .i.xa1 is interesting but proba­
member stuff during his thinking rime). bly in White's favour as Black will just create

1 05
Pia y th e King 's Indian

weaknesses if he advances his pawns. 1 5 . . . ..ih6! 1 6 ..ixg4

c) 13 ...e6 was played in the game. It contin­ Of course the point of ... i.h6 is to make the
ued 14 lb1a2 'iVaS 1 5 dxe6 (1 5 0-0 exdS 1 6 eS-s<:juare available for the black knight. White
cxdS b S was not to White's liking s o h e decides probably has nothing better than this capture,
to grab a pawn) 1 5 ... �xe6 1 6 1i'xd6 lbfd7 1 7 which wouldn't have helped much if I'd played
�£2!? (1 7 0-0 may appear more natural, but 14 ...i.h6 as then Black could recapture with the
sometimes it can be useful to have the bishop knight.
on e2 protected and, in the case of an ending 1 6 .....ixg4 1 7 ..ie3 ..ixe2
arising, f2 is somewhat nearer the centre than I was already running shon of time (it
g1) 17 ...lbc6 1 8 l:tacl lbd4 1 9 b4 1i'b6 doesn't take much at 36 moves in 1 .5 hours) so
(19...1i'd8!? could be an attempt to punish opted for what I though was going to be a
White for his 1 7th move) 20 1i'xb6 lbxb6 21 slightly better ending. My opponent surprised
bxcS lbxe2 22 lbxe2 and now Gelfand played me, however, with...
22 ... lbxc4?! and went on to lose but Kramnik 1 8 �xe2?!
has pointed out that 22 ... lbxa4! would have A brave man but he should have played 1 8
given Black a very comfonable game, e.g. 23 lbxe2.
lbf4 l:tfc8 24 lbds �£8! 25 lbab4 aS 26 lbd3 1 8 ...ll:lf6
lbxcS 27 lbxcS i.xcS 28 l:tb 1 ! is e<:jual. 1 8 ... £5 is good but I had another idea in
Before gerting back to the Lehtivaara­ mind. Of course the immediate threat is 1 9... b4.
Gallagher game it's worth pointing out that the 1 9 �3
attempt to trap the black knight in the centre No choice as 19 e5 lbg4 is good for Black.
with 12 h3? bS 1 3 b3? just loses material to 1 9 ...lLld7!
13 ...lbe8! or 1 3 ... lbh5! (I once played
1 3...lbfg4?? but still won).
1 2 ...lDeg4 1 3 ..ig1 b5 1 4 ..if3

The knight is heading for b6 and the c4-

S<:Juare. I am trying to force White to exchange
on bS and open lines rather than playing bxc4
Of course 14 h3 is met by 14 ... b4, but now myself, which may win a pawn but reduces
with e4 defended White is threatening to play Black's possibilities of active play.
h3. I looked at several interesting continuations 20 g4?!
but probably failed to find the most precise The gung-ho approach turns a difficult situa­
move order tion into a desperate one.
1 4...ll:lh5 20 ...ll:lb6 21 cxb5
Black should just play the immediate 21 lbbl lbxc4 is just a clear pawn.
14 ... i.h6! when it is difficult to suggest a good 21 . .lLlc4 22 'ife2 axb5 23 lLld1
.

move for White. After 23 a4 1i'b4! 24 axbS the <:Juiet 24...i.g7

1 5 ll:l1e2 is probably the most unpleasant for White.
15 h3 lbg3 was the idea. 23 . . ...ig7 24 h4 e6! 25 h5

1 06
 The Siimisch

25 dxe6 fxe6 is equally hopeless. hope for after 7 dS is transposition to another

25 ...exd5 26 exd5 l:aeS variation (there are several possibilities). White
26 ... l:.fe8 is also good but I wanted to vacate normally waits for ...llJc6 before playing this
a8 for the queen and had seen that there was a advance in order to play similar positions but
simple way to deal with White's attack. with a black knight already committed to the
27 ..h2 double-edged eS-square. Play can transpose but
Or 27 hx�:,>6 fxg6 28 'ifh2 h6. it is up to Black whether or not he wishes to
27...g5! 2S h6 .bb2! 29 l:b1 -.as! move his knight to eS.
You may also compare this position to the
one after 6 .i.gS cS 7 dS and you will notice that
with the bishop more actively deployed on gS
Black usually feels the necessity to weaken his
kingside with h6. The bishop then retreats to e3
and White can hope to gain a tempo, in com­
parison with this variation, as at some stage
Black is going to have to deal with the threat to
his h-pawn.
7 ... 86 s ..d2
l .ess common moves are:
a) 8 dxe6 .i.xe6 9 .i.d3 liJfd7! with a good
game for Black.
b) 8 llJge2 exdS 9 cxdS liJbd7 10 llJg3 is
30 l:xb2 ••d5+ 31 �g3 similar to Game 38 but with White already
Or 31 �£2 lLlxb2 32 llJxb2 l:.xe3 33 'it>xe3 having committed his bishop to e3.
l:.e8+- and Black mates. S . . .exd5 9 cxd5 a6
31 ...lbxe3 32 -.d2 •xh1 0-1

Game 37
Mihajlovic-Kovacevic
Yugoslav Ch. 1996

1 d4 ltJf6 2 c4 g6 3 ltJc3 ..ig7 4 e4 d6 5 f3

0-0 6 ..ie3 c5 7 d5

One of the standard ways for Black to seek

active play in these Benoni positions is to ad­
vance on the queenside with ... bS. White nearly
always prevents this by playing a4 but, on bal­
ance, it is in Black's favour to have the moves
... a6 and a4 included. There are a few positions
where liJbS could be annoying and Black still
has the possibility to play ... b7-b5 as a Benko­
style pawn sacrifice.
This is the least played of White's seventh 10 a4 .:as
move options as the most that he can really I\ good altemative is to ... hS transposing into

107
Pla y the King 's Indian

Game 42. His idea is to achieve the ideal set-up for his
A less good alternative is 1 0...iDbd7 as this knights by playing ltJec3 and ltJt2. Note how
allows White to develop his king's knight fa­ this has takes four moves (not counting the
vourably to h3. The knight will then settle on t2 original ltJc3) with the knights whilst if Black
from where it can keep an eye on a number of had played ...ltJbd7 before White played ltJge2
important squares, as well as not clogging up then by ltJh3-t2 White gets the same set-up
the kingside as it does on e2. having played only two knight moves. Now you
A good rule of thumb for Black in these can appreciate why Black shouldn't rush with
lines is don't play ...ltJbd7 until White has ...ltJbd7 (see above). The downside to 1 2 ttJd1
played ltJe2. is that it does nothing to speed up White's
1 1 llJge2 rather sluggish pace of development.
White should give up on the idea of liJh3, as Here is an example where things didn't go
he hasn't got a good alternative to the text, e.g. too well for Black: 12 ...ltJeS 1 3 lLlec3 1i'a5?! 1 4
a) 1 1 .i.e2 1i'c7 1 2 �d1 lLlbd7 1 3 lLlh3 ltJes .i.e2! (White i s right not t o fear . . .bS, and con­
and Black threatens ...ltJc4 centrates on development in order to be able to
b) 1 1 aS bS 12 axb6 1i'xb6 13 �d3 liJbd7 1 4 counter in the centre as quickly as possible)
lLlh3 ttJes 1 5 ltJ t2 1/2-'12 i n Schneider-Szvia, 1 4... b5 1 5 0-0 liJfd7 1 6 ltJt2 ltJc4 1 7 �xc4 bxc4
Budapest 1991, even though Black already has 1 8 f4 l:.b8 1 9 eS dxeS 20 lLlfe4 1i'b6 21 fS!,
the upper hand. Meulders-Douven, Holland 199 1 . White has
c) 1 1 �d3 lLlbd7 12 lLlh3 ttJes and ex­ obtained a powerful attacking position by em­
changes nearly always suit Black in the Benoni ploying a standard Benoni trick, eS, ...dxeS, fS,
structure. which blocks the long diagonal and thereby
1 1 ...llJbd7 restricts the black bishop, while vacating the e4-
1 1 ...1i'a5 1 2 l:.a3 is not a favourable devel­ square for White's knights.
opment for Black. I think that Black's troubles in this game can
1 2 llJc1 be traced back to 1 3...1i'a5. I would prefer to
The knight has to move again to allow prepare for ... bS with 1 3 ... .i.d7 14 �e2 l:.b8
White to complete his development. The ques­ with what appears to be a reasonable game for
tion is where does it go? Here are a couple of Black. Alternatively, Black could also consider
alternatives: 13 ... ltJhS with the idea of striking in the centre
a) 12 lLlg3 hS 13 �e2 h4 14 lLlfl lLlh7 leads with ... fS.
to an inferior version, for White, of Game 39. See also Game 42 for an example of the
The inclusion of the moves 1i'd2 and l:.e8 same plan from White.
should be slightly in Black's favour. 1 2 ...llJe5 1 3 ..ie2
b) 1 2 ttJd1 is a typical move from White in
such positions.

1 3 ...llJh5!?
Black gets ready to play ... f7-f5. The alterna-

1 08
 The Siimisch

rive is to play for bS with 1 3...i..d 7 14 0-0 l:b8 6 .!Dge2

when White should play 1 5 aS. Now, in prac-
tice, Black has played both 1 S ... i..c 8 and Game 38
1S ... 'ifc8, neither of which I like, as well as the
Dreev-Bologan
pawn sacrifice 1 5 ... b5. I believe Black has rea-
sonable compensation after 1 6 axb6 l:xb6 1 7 Shanghai 2001
l:xa6 l:xa6 1 8 i..xa6 'ifaS 1 9 i..e2 l:b8 but I
would still prefer to play for ... fS than give up a 1 d4 .!Df6 2 c4 g6 3 .!Del �g7 4 e4 d6 S fl
pawn. Note that Black also has 1 3 ... h5!? fol­ 0-0 6 ltjge2
lowed by ... lLlh7 as another means of preparing
fS.
14 0-0 fS 1 S f4 .!Dd7!?
Interesting decision. Black allows his king­
side pawns to be shattered just as in Game 40.
15 ...lLlg4 1 6 i..xg4 fxg4 also looks quite play­
able and quite difficult to assess.
1 6 �xhS gxhS 1 7 eS
Even if he wanted to take on f5 White
couldn't because of the tactical response
17 ...l:xe3! 1 8 'ifxe3 i..d4 (a standard trick in
these positions).
1 7 ...dxeS 1 8 .!Dd3 b6
Black prefers to keep control of d4, e.g.
18 ... c4 1 9 lLlxeS lLlxeS 20 fxeS i..xeS 21 i..d4 This time White delays the development of
looks good for White. his queen's bishop. He will decide later where it
19 fxeS .!DxeS 20 .!DxeS �xeS 21 �gS 'ifd6 should go (if anywhere) and first concentrates
22 �f4 �xf4 23 llxf4 �d7 24 l:laf1 lieS on the king's knight. This is often a problem
piece for White in the Siim.isch as the move f3
has deprived it of its natural home. It takes
more than one move to develop the knight as it
can't stay on e2 where it clogs up the rest of the
white kingside. The knight is heading for g3 in
order to restrain (or is it to encourage) Black's
kingside play. As we shall see this knight is not
in for a quiet life. The main exponent of this
line is the Russian grandmaster Dreev who
features in our two games. It was, therefore,
most interesting to do battle with him this year
in Gibraltar in a variation which I have played
with both colours (see next game). Black has
the usual assortment of replies but I see no
It's one of those posltlons that looks as reason to differ from our chosen response to
though it should be go<xl for White but some­ White's other 6th moves.
how isn't (Black controls the centre and has a 6 ... cS 7 dS e6 8 .!Dgl
mobile queenside majority). In fact it may even 8 i..gS transposes to Game 4 1 .
be better for Black. At any rate he won pretty 8 ...exdS 9 cxdS hS!?
quickly. Quite often Black flicks in the moves 9 ...a6
2S ll4f3 bS 26 llgl+ �h8 27 axbS axbS 1 0 a4 before deciding on his plan of action.
28 .!De2 h4 29 lldl b4 30 .!Df4 �bS 31 llc1 However, in conjunction with a quick ... hS it is
.bdl 32 'ifxdl lle4 0-1 better to delay these moves. The reason will be

1 09
Pla y th e King 's Indian

seen later. The knight on g3 makes a tempting then White would be able to play a4-a5 after
target for the h-pawn. It has just spent two ltJh7 and when the black queen retreats to c7
moves getting there and it is soon to be on the then he could indeed play i.h6 in order to
move again. Of course there is some risk in­ exchange off the bishops. That is why Black is
volved in advancing the h-pawn but if Black delaying the move ... a6 in this variation.
wants to play ... f5 it is necessary to first kick the 1 1 ... "ifc71 1 2 .i.e2 a6
knight &om g3.
In Game 40 I examine an alternative for
Black, 9 . ..ltJh5!?

1 3 0-0
An interesting moment. Most players would
play 1 3 a4 without much thought but Dreev
10 .i.g5 decides to allow Black to play ... b5 as he will
For the time being White stops Black &om then hope to undermine the queenside with a
playing ...h4. The other move that White plays subsequent a4. Black probably does best to
with about equal frequency is 10 i.e2. See delay ... b5 for the time being and get on with
Dreev-Gallagher (Game 31). his kingside play, as Bologan did in the game. In
1 0 ...'iib6 fact Dreev had already reached this position
Black normally breaks such pins by playing more than once. The ftrst time, against Topa­
... h6 but with the pawn on h5 that is no longer lov, Elista Olympiad. 1 998 he did play 13 a4
possible. He must resort to other means to and wasn't very successful. Play continued
break the pin. The b6-square is not an ideal 1 3. ..ltJh7 14 i.e3 'ii'e7 1 5 0-0 ltJd7 (Black
home for the queen but developing it here avoids the immediate ... f5 as White has a little
gains Black a tempo by attacking the b2 pawn. trick that wins a pawn: 1 5 ... f5 16 i.xc5 dxc5?
1 1 'iib3 17 d6+ 'ii'f7 1 8 i.c4 costs Black his queen) 16
This is a specialiry of Dreev. The comments f4 i.d4 17 i.£2 h4 1 8 ltJh1 gS (this is a stan­
to ...'i'b6 are equally applicable to 'i'b3. This is dard idea in such positions; Black will now take
not a good square for the queen but White control of the eS-square which will make a ftne
gains time by offering an exchange. In this posi­ outpost for a minor piece and also ensure that
tion White has the sounder strucrure (pawn White won't be able to smash through in the
majoriry in the centre, weakness on d6 to aim centre) 1 9 'ii'd 1 'ii'f6 20 fxgS '/z-1/z.
at) so Black must seek his chances in dynamic 1 3 ...tt:lh7 1 4 .i.e3 h4
play. As the most dynamic piece is the queen The game Dreev-Tkachiev, Cap D'Agde
Black should keep them on the board. 2000 went 14 ... 'W'e7!? 1 5 f4 h4 16 ltJh1 bS 17
1 1 'ii'd2 is also possible when Black would ltJ£2 ltJd7 1 8 i.£3 gS 19 ltJe2 f5 with a very
just reply 1 t ...ltJh7. Note that 12 i.h6 (12 i.h4 promising position for Black. Why is Dreev
is best met by 1 2... ltJd7) is not possible because playing the same line again? Has he simply for­
after an exchange of bishops Black takes on b2. gotten this game? And why has Bologan not
If the pawns had already been on a6 and a4 played 14 ...'ii'e7? Well, Dreev, one of the

1 10
 The Siimisch

world's top players, has certainly not forgotten 0-0 6 lt:\ge2 c5 7 d5 e6 8 lt:\gl exd5 9 cxd5
this game. What has happened is that he has h5 1 0 ..lte2
worked on it at home and found an improve­
ment for White. Bologan avoids 14 ...1i'e7 as he
doesn't wish to become another victim of
Dreev's homework.
1 5 lt:\h1 f5 1 6 exf5
16 f4 ..td4!? may well be the subject of a fu­
ture game in this line.
1 6 ....bf5 1 7 ..ltf2?!
The start of an over-ambitious plan. It is
quite common for White to retteat his knight to
the comer but very rarely does he leave it on
this sad square for any length of time. 17 ltl£2
would have been more prudent.
1 7 ...g5 1 8 f4 gxf4 1 9 ..ltxh4 ..ltd4+ 20 lt:\f2
t:Dd7 21 �h1 lt:\e5 Until a few months ago I was responsible
for running a King's Indian website for Chess
Publishing. This involved a monthly update
with analysis of the most recent King's Indian
!,>ames. No doubt I was slightly biased in Black's
favour but my successor, the Israeli grandmas­
ter Mikhalevski (who plays the King's Indian
about as often as Karpov) has taken things too
far the other way. He called Black's last move,
for example, 'unfortunate' and describes 10
..te2 as 'killing'. A slight exaggeration, I feel (by
the way, his stuff on Chess Publishing is still
quite interesting). 10 ..te2 may or may not pose
more problems for Black than other moves but
it is certainly not killing.
Dreev was probably hoping to show that the 1 0 . . .lt:\h7
pawn on f4 is weak but he has allowed the Black can also play 10...h4 1 1 itlft ltlh7 but
black pieces ttemendous activity and anyway, I preferred to avoid this move order as I once
the pawn can be defended easily enough. obtained a good game with White by playing 12
22 lt:\te4 lt:\g6 23 ..ltt2 .l:.ae8 24 -�ttl lt:\e5 25 ltle3.
'it'd1 lt:\xfl 26 'it'xfl b5 27 l:tae1 ? 1 1 .\tel
A blunder in a difficult position. Dreev doesn't fall for 1 1 0-0?? ..td4+ 12
27 . .b4 28 ..ltxd4 cxd4 29 t:Dd1 ..ltxe4! 30
. �h 1 h4 when the customary h 1-sctuare is no
.l:l.xe4 lt:\g5 31 _.g4 .l:l.xe4 32 -.xg5+ _.g7 longer available for the knight But your oppo­
33 �5 dl 34 lt:\f2 .l:l.e5 35 'it'd1 .l:l.xd5 36 nents might do!
lt:lxdl _.g6 37 .l:l.fl .l:l.e8 38 hl .l:l.el 39 .l:l.xf4 1 1 . . . a6
.l:l.exdl 40 'ire2 .1:1.5d4 41 .l:l.f1 .l:l.h4 0-1 Dreev's adversaries have been ctuite keen on
playing 1 1 ...h4 12 itlft f5 but since they all lost
Game 39 (fkachiev, Radjabov and Obodchuk) I pre­
ferred another approach. In general I like to
Dreev-Gallagher
have more pieces developed when I play ... f5.
Gibraltar 2004 1 2 a4 h4
Black can also delay this move and play
1 d4 lt:\f6 2 c4 g6 3 lt:lc3 ..ltg7 4 e4 d6 5 fl 12 ...ltld7 when 13 0-0 h4 14 ltlh1 f5 leads to a

1 1 1
Pla y t h e King 's Indian

fairly typical sort of position. In fact this was The point of Dreev's novelty is to bring the
reached in Dreev's next game in this line which knight back into play via e3. But first he pre­
was, rather amazingly, against Anatoly Karpov. vents Black from playing ... f4.
The black side of the King's Indian is not a 1 6 . . .l:le8!?
place where one expects to fmd Karpov but he This leads to fascinating complications but
acquitted himself rather weU for his debut with just a few weeks after our game in Malakhatko­
the opening. Play continued 1 S 1i'd2 1i'f6 Damljanovic, Montenegro 2004 Black played
(Mikhalevski suggests 1 S...h3 1 6 g3 fxe4 1 7 fxe4 more solidly with 1 6...1i'f6. The idea is to meet
tt'lhf6 1 8 tt'lf2 ltkS, which does look fine for 1 7 tt'le3 with 'iVh6! So Malakhatko changed
Black, but Karpov prefers to force White to plan and played 1 7 lbd2 but after 1 7...11Vh6 18
deal with the threat of ... f5-f4) 1 6 f4 (16 exf5 is 0-0 1i'xf4 19 .i.xcS 'iVh6 20 �f2 tt'leS (20...h3!?)
the alternative when Black wiU recapture Black had good squares for his pieces. Mik­
16 ... gxf5 with a reasonable game) 16 ... fxe4! 1 7 halevski suggests 1 8 tt'lc4 as an improvement,
tt'l f2 1Ve7 1 8 tt'lcxe4 tt'ldf6 1 9 tt'lxf6+ 1i'xf6 with the point being that after 1 8...1i'xf4 White has
an unclear game ('/•-'/•, 70 after Karpov messed 19 .i.e3. He gives 1 8...l:lb8 1 9 aS bS 20 axb6
up a completely winning position) tt'lxb6 21 tt'laS 'ifxf4 22 tt'lc6 as best and claims
1 3 ttlf1 llJd7 an edge for White here. I'm not so sure but it is
Topalov played 1 3 ... f5 here against Dreev a definite improvement on 1 8 0-0.
and he lost too. My comment &om move 1 1 is 1 7 ttle3 ..td4
stiU applicable. The logical foUow-up.
1 4 ..tf2! 1 8 ttlxf5!?
When he played this almost instantly I
started to get worried but relaxed a little when
he took over an hour on his 20th move. I was
planning to meet 1 8 tt'lc4 with 1 8... .i.xf2+ 19
�xf2 tt'lb6 20 llk3 1i'f6 with a satisfactory
game. The queen may even jump into d4 in the
near future.
1 8 . . . ..txf2+ 1 9 �xf2 llf8!

A novelty, which in my opuuon White

needed quite badly. My preparation for this
game had homed in on this position, and with
this move order, as the best way to create prob­
lems for White in this line. This position had
been reached previously in the game Korchnoi­
Xie Jun, Roquebrune 1998 which continued 14
tt'ld2 f5 1 S exf5 gxf5 1 6 f4 tt'ldf6 1 7 h3 (White
wants to rule out ...tt'lg4 possibilities) 1 7...1Ve7 Better than 19 ...1i'f6 when 20 g4! hxg3t 21
18 �f2 l:le8 1 9 tt'lc4 l:lb8 20 aS bS 21 axb6 and hxg3 l:lxe2+! (2t ...'ifxf5 22 �d3) 22 1i'xe2
now Black should have just recaptured with 1i'xf5 23 1i'e6+l? (23 1i'e8+ lbd£8 24 l:lae1 is an
21 ...l:lxb6! as after 22 tt'lxb6 'ifxe3t 23 �f1 alternative) 23...1i'xe6 24 dxe6 tt'ldf6 2S l:lae1
1i'xf4+ 24 .i.f3 tt'le4 she would have had a very b6 26 l:lh4! is not a lot of fun for Black.
strong attack. 20 g4!
1 4...f5 1 5 exf5 gxf5 1 6 f4 After his marathon think Dreev opts for a

1 12
 Th e Siimisch

sacrificial attack. The main point behind Black's by now) and I had to play something. The right
play is that after the solid 20 �d3, defending move, 28 ...'ith2!, simply didn't occur to me.
the knight and threatening 1i'g4+, Black has the Now after 29 1i'e7+ 'it>g8 30 'itxf6 1i'h3+ 31
excellent reply 20.)tJeS! with a good game. I 'it>£2 1i'h2t White cannot escape the checks so
suppose it may be possible to grab the d-pawn his only winning try would be 29 1i'h 1 . Black
but Black has a strong initiative after 20 lLlxd6 shouldn't have too many problems in the end­
llxf4+. ing though, e.g. 29 ... 1i'xh1+ 30 ..z:txh1+ �g7 31
20 ... hxg3+ 21 hxg3 l:txf5 22 :Xh7! aS (31 lib 1 bS) 31 ...llc8 32 lib 1 lieS 33 .:.Xb7+
Stronger than 22 �d3 lLldf6! 'it>g8 34 llb6 lLlxdS 3S ltJxdS l:txdS 36 l:txa6
22 .. .<�'xh7 23 �d3 tiJf6?! c3=.
I thought this was forced as I have to pre­ 29 •e7+ l:tg7 30 l:th1 +! �g8 31 Wd8+ �7
vent "it'hS+ but I missed a fascinating possibility 32 Wc7+ �8 33 WeB+ 1 -0
here in 23...1i'gS!. It seems that White has to On account of 33...lLle8 34 llh&t- and
play 24 1i'f3 and after the game we thought that 33 ... �£7 34 1i'e6+.
both 24 ...1i'g6 2S g4 lLleS! 26 .i.x£5 ltJxg4+ 27
'i'xg4 .i.x£5 28 llh 1+ �g7 29 1i'h4 �£7 and Game 40
24...1i'g6 2S fxgS .:.X£3+ 26 �xf3 lLleS+ 27 �e3
Svetushkin-Romero Holmes
ltJxd3 28 �xd3 .i.fS+ 29 'iti>d2 'it>g6 led to
roughly level positions. Bled Ofympiad 2002
24 g4
White regains the rook. 1 d4 tiJf6 2 c4 g6 3 tlJc3 �g7 4 e4 d6 5 f3
24 . . . c4! 0-0 6 tlJge2 c5 7 d5 e6 8 tlJg3 exd5 9 cxd5
Avoiding 24 ... 1i'b6 2S gxf5! c4+ 26 'it>f3 cxd3 tlJh5!?
27 1i'h1+ 'it>g7 28 l:tg1+ �£7 29 1i'g2! when
White has a winning attack.
25 �xf5+ �xf5 26 gxf5 'ifb6+ 27 �3
••b2 28 •e1 !

This crazy-looking move is the latest fashion

but it does come with a health warning. Black
allows his kingside pawn structure to be shat­
tered so that he can play f5 and !,Tt:t active play
28 .. Jig8?? for his pieces on the kingside as soon as possi­
A sad conclusion to a wonderful game. I had ble. In the famous 1 972 work! championship
seen that Black is losing after 28...lle8 29 1i'h4+ match Fischer played ...lLlhS in a similar posi­
and was about to try 28 ...'it>h8 when I saw that tion and the chess world was shocked. In mod­
it too lost in brilliant fashion to 29 lla2 1i'b3 30 ern chess virtually anything goes but one should
llh2t lLlh7 31 llxh7+ 'it>xh7 32 'ii'e7+ 'it>h8 33 srill be careful. There are many more positions
'i'f6+ 'it>g8 34 1i'g6+ 'it>fB 3S f6 1i'xc3+ 36 �g4. where the rules can't be broken than there are
And then my flag was about to fall (I was down exceptions.
to 30 seconds a move for the rest of the game 10 tlJxh5 gxh5 1 1 �d3

1 13
Pla y th e King 's Indian

Everyone plays this except Chris Ward who l:[b 1 .i.d4 23 l:[b3! (the point behind the sacri­
likes 1 1 .i.f4. He has reached the position after fice; White's rook will be very active on the
1 1 ...£5 12 1i'd2 1i'f6 13 .i.g5 1i'g6 14 ..id3 a open 3rd rank) 23 ...1i'e7 24 lbg3 1/2-1!2.
couple of times. The unconvincing 14 ...lba6 I'm afraid that in this most unclear position
was played the first time but in the second the players chickened out and called it a draw.
game Black found a better plan: 14 ... fxe4 1 5 Most likely a great time scramble was looming.
..ixe4 (or 1 5 lbxe4 .i.f5) 1 5 ....i. f5 1 6 0-0 tbd7 GM Stohl analysed this position and considered
1 7 :ae1 tbe5 1 8 .i.f4 .i.xe4 19 :Xe4, Ward­ the chances to be about equal. Both sides have
Snape, 4NCL 2001 , and now Black should dangerous pawn majorities and although
expand on the queenside with 19 ... b5! as 20 Black's king is more exposed he still has the
lbxb5? fails to 20...tbx£3+! and 20 .i.xe5 .i.xe5 more active pieces.
21 lbxb5 :Xf3! 22 :Xf3 1i'xe4 is fine for Black. 1 2 f4
11 ... li)d717 In the two games I have where Black played
The latest twist. Black is getting clever with 1 t ...tbd7 White has replied both times with this
his move order. After the usual 1 1 ...£5 1 2 0-0 double-edged move. I doubt it is better than 1 2
0-0 when Black can play 1 2...tbe5 13 .i.c2 f5
transposing into line 'c' in the previous note.
1 2 ...f5

Black has tried:

a) 12 ... f4. This is what Black would really like
to play if he could maintain the blockade whilst
slowly transferring his pieces to the kingside. 1 3 1Wxh5
But he can't as White can play 1 3 tbe2 .i.e5 14 A brave move. Instead 1 3 0-0 ltlf6 and now:
g3!. It's still complicated but good for White. a) In Agrest-Kazhgaleyev, Nice 2000/01
b) 12 ...ltla6 when it seems that White does White played the passive 14 h3 and Black soon
best to play 1 3 a3, preparing 14 1i'c2. obtained an excellent game: 14 ... fxe4 1 5 tbxe4
c) 1 2 ...tbd7 and now most White players fa­ tbxe4 1 6 ..ixe4 1i'f6 17 �h1 .i.f5 1 8 .i.x£5
vour the prophylactic 1 3 .i.c2 even though 1i'xf5 1 9 .i.e3 :aes 20 .i.g1 :e4 21 l:[b1 h4 22
Black seems to be doing all right after 1 3...tbe5 .i.h2 l:[fe8. He has a positional bind and went
14 tbe2 lbg6!?. For example, Khenkin­ on to win the game.
Reinderman, European Ch., Ohrid 2001 con­ b) The critical line must be 14 e5 (14 exf5
tinued 1 5 exf5 .i.xf5 1 6 .i.xf5 :xf5 17 lbg3 lbg4 looks fine for Black) when 14 ...lbg4 looks
:n 1 s ltlxh5 .i.d4+ 1 9 �h 1 1i'h4 20 f4 :es. better than the immediate exchange on e5 (both
Quite a fitting position for this line One of the for concrete tactical reasons and the fact that
weak black pawns has dropped off but in rerum Black must be careful about opening lines for
he has completed his development most har­ the white pieces with his king so exposed).
moniously whilst White's queenside still lan­ Now 1 5 e6 would be a terrible blunder on ac­
guishes at home. White decides to rerum the count of 1 5 ... .i.d4+ 16 'iti>h1 1i'h4 17 h3 1i'g3!
pawn in order to rectify this. 21.i.d2 ..ixb2 22 when White gets mated and 1 5 h3 dxe5! 1 6

1 14
 The Samisch

hxg4 e4 17 gS exd3 appears quite satisfactory .i.a3

for Black. Perhaps White has to settle for some­ The wonderful point behind Black's last
thing like 1 5 'ife2 and be content with some move is that 23 gS loses to 23 ... .i.xf5!! when 24
compensation for the pawn Black will now take 'ifx£5 aUows 24......£3 mate and 24 .i.x£5 'ifdt+
on eS. 25 �g2 l:.e2+ 26 �h 1 ...dS+ is also the end.
1 3 ... b5!? 23 ...hxg4 24 llxg4 llh8
One would expect nothing less from the With some very powerful play Romero has
imaginative Spanish grandmaster playing Black built up a winning position and now, with time
in this game. For those of you who relish a no doubt extremely short he opted for a very
quieter life I don't see anything wrong with good ending (obviously 25 "i1Vxh8 ...£3+ 26 �e1
13. ..ltJf6, e.g. 1 4 ...£3 fxe4 1 S ltJxe4 .i.£5 16 0-0 ...xg4 is hopeless for White) instead of the even
c4 17 ltJxf6+ "i1Vxf6 1 8 .i.x£5 ...x£5 with full stronger 24 ... .i.b7!. This threatens l:.h8 with a
compensation for the pawn. decisive attack and if White plays 25 l:.d 1 then
14 exf5?! 25 ... 11Vf3+ wins material.
After this there is no question that Black has 25 "ifg2 "ifxg2+ 26 llxg2 .i.b7 27 lle2 llae8
an exceUent game. The critical response appears 28 llae1 llxe2 29 llxe2 .i.f3 30 llf2 �g4!
to be 1 4 .i.xbS. I'm not sure what Romero had 31 lld2
in mind apart from the general opening of lines 31 .l:.x£3 ltJxh2+ 32 �f2 ltJx£3 33 'it>x£3
in aU sectors of the board. Perhaps he was go- l:th3+ 34 'it>e4 l:.xc3 and Black wins.
ing to try and win material with 1 4... .i.xc3+ 1 5 31 ...�xh2+ 32 �2 d5
bxc3 'ifaS 1 6 l:.b1 ! l:.b8 (not 1 6...'ifxc3+ 1 7
.i.d2 'ifc2 1 8 l:.b3!) 1 7 a4 'ifxc3+ (I don't trust
17 ... a6 18 0-0! axbS 19 c4) 18 .i.d2 'ifc2 19
"iiVgS+ �h8 20 l:.ct "i1Vxe4+ and now 21 �fl
l:.g8 22 .i.c3+ ltJeS 23 .i.xeS+ dxeS 24 'iff6+
l:.g7 25 'if£8+ l:.g8 26 ...f6+ is a draw while I'm
not sure what is going on after 21 �d1 !?
14 ... c4 1 5 .i.c2
Maybe 1 5 .i.e2 is better.
1 5 ...lieS+ 1 6 �2 .i.xc3! 1 7 bxc3 'ifb6+ 1 8
� 1 �f6 1 9 Wl'h4 �7!

Black has a decisive advantage an went on to

win as foUows: 33 .i.cS a6 34 .i.d 1 .i.e4 35 .i.e2
l:.g8 36 .i.d4 .i.x£5 37 .i.hS+ �e6 38 .i.eS .i.e4
39 ie2 l:.g2+ 40 �e3 .i.d3 41 .i.b8 .i.xe2 42
l:.xe2 ltJg4+ 43 'ifi>£3+ l:.xe2 44 �xe2 'it>£5 45
�£3 d4 46 cxd4 ltJf6 47 a4 b4 48 dS ltJxdS 49
.i.eS b3 50 'it>e2 ltJxf4+ 0- 1

6 .i.g5

Game 4 1
Of course everyone knows that a knight is
Hauchard-Krakops
the best defender for the king. With his bishops
hemmed in by his own pawns White's only Cappel/e Ia Grande 1997
chance is to advance on the kingside.
20 g4 1fc5! 21 llg1 "ifxd5 22 'ih13 h5! 23 1 d4 �f6 2 c4 g6 3 �c3 .i.g7 4 e4 d6 5 f3

1 15
Pla y th e King 's Indian

0-0 6 �g5 Because the knight on f6 is pinned White

I have already discussed earlier in the chap­ has the option of recapturing with the knight.
ter why 6 .i.g5 became quite popular, but it has An important difference but, as we shall see,
to be said that its popularity has declined Black should not lose any sleep over it.
slightly over the last few years as more and The main line is 9 cxd5 and that is the sub­
more White players prefer 6 �ge2. Against 6 ject of the next game.
.i.g5 I am remaining faithful to the Benoni set­ 9 �e6 1 0 tbe2 tbc6!?
...

up and suggest that Black plays 6...c5. Inciden­

tally, if you decide to go your own way just
make sure you don't play 6...e5?? as it loses
immediately.
6 ... c5 7 d5 e6

Black usually plays 10 ....i.xd5 so that White

has to take back with a pawn. For example,
Yusupov-Gelfand, Moscow 1 992 continued 1 1
cxd5 �bd7 1 2 �3 a6 1 3 a4 1i'c7 1 4 .i.e2 c4
15 0-0 b5 with an unclear game. The text seem­
8 1i'd2 ingly allows White to maintain a knight on the
Occasionally White plays 8 �ge2 and now important d5-square but Black has a tactical
8 ... exd5 9 �d5 ..ie6 1 0 �c3 is a slightly im­ solution to the position.
proved version of the main game for White. 1 1 tbec3 �xd5 1 2 tbxd5 h6! 1 3 tbxf6+
Therefore, Black should play 8 ..h6
. and ex­ The first point is that after 1 3 .i.xh6? �xe4!
change on d5 only after the bishop has dropped 14 fxe4 1i'h4+ Black regains the piece with ad­
back to e3. Then play is likely to transpose vantage.
elsewhere in the chapter. 1 3 ...�xf6 1 4 �xh6 �xb2!
For example, I reached the position after
8... h6 9 .i.e3 exd5 10 cxd5 a6 1 1 a4 �bd7 twice
with Black against Kelecevic.
a) The first time went 1 2 1i'd2 �5 1 3 � 1
(13 ..ixh6 �xe4) 1 3. . .�h7! 1 4 .i.e2 f5 1 5 0-0 g5
1 6 exf5 ..ixf5 and Black was already better
because of the misplaced knight on c t .
b) The second time went 1 2 �g3 h 5 1 3 .i.e2
and now 1 3 ... h4 14 �ft �h7 is Dreev­
Gallagher and 13 ... �h7 14 0-0 h4 1 5 �h1 f5 is
Dreev-Karpov (this is what I played).
Incidentally, don't worry too much about all
these confusing move orders. I get pretty con­
fused myself! Just make sure you have a good
understanding of the underlying plans. 1 5 l:lb1
8 . . .exd5 9 tbxd5 Black's idea was, of course, to meet 1 5

1 16
 The Samisch

'ii'xb2 with 1 5 ...1i'h4+. White tries to avoid this

exchange but after Black's next move he has no
choice but to acquiesce.
1 5 ... ..tc3! 1 6 11Fxc3 11Fh4+ 1 7 g3 11Fxh6 1 8
1Vd2 11Fh 5 1 9 ..tg2 ilJd4 20 0-0 b6
From a strategic point of view the position is
very interesting. At first glance it looks as
though Black is doing really well because he has
a massive knight on d4 and White a passive
bishop on g2. However, that is not the whole
story. Black's pawn snucture is slightly inferior
and White may also be able to create play on
the queenside by pushing his a-pawn. Black will
also have to keep an eye out for eS as this
would open the position for the white bishop. There are a few players who don't consider
He can prevent this easily enough by playing f6. . .. bS to be a serious threat and just play 10
The game is about level. ltJgc2. The position after something like 10...b5
1 1 ltJg3 ltJbd7 1 2 .i.e2 c4 1 3 0-0 ltJcs is highly
complex with chances for both sides. I'll just
say that, playing Black, I always feel more com­
fortable with the move ...bS in than without it.
1 0 ... h6
It is desirable for Black to break the pin and
this is the best moment to attack the bishop.
1 1 .tel
So what happens if White takes the pawn.
After 1 1 .i.xh6 ltJxe4! 12 ltJxe4 1i'h4+ 1 3 g3
1i'xh6 1 4 'ifxh6 .i.xh6 1 5 ltJxd6 l:.dB 1 6 ltJxcB
:XeS Black has full compensation. This conclu­
sion was reached after a number of games in­
cluding one where I was White. Funnily
21 f4 l:tae8 22 l:tbe1 �g7 23 h3 f6 24 a4 enough, if the a-pawns were still at home the
l:te7 25 a5 bxa5!? 26 11Fxa5 ltlc2 27 g4 position would be more promising for White.
11Fh8 28 l:te2 ltld4 29 l:tef2 1IFh4 The point is that with the pawns on a2 and a7
I'm slightly puzzled as to why Black didn't White can castle queenside quite safely (first he
go here straight away. has to play f4 to block out the bishop on h6)
30 11Fa6 l:td7 31 e5 fxe5 32 fxe5 l:txf2 33 but now if he castles long Black will be able to
:Xf2 dxe5 34 11Fc8 11Fe7 35 ..td5 l:td8 36 start an attack with ... bS. That is the reason for
"iWa6 l:tf8 37 l:txf8 11Fxf8 38 11Fxa7+ �h6 39 not playing 9. .h6 as then 10 .i.xh6 is certainly
.

�g2 ltle2 40 h4 ltlf4+ 41 �g3 ltle2+ 42 an option for White.

�g2 ltlf4+ 43 �g3 ltle2+ % -% 1 1 ... h5
Of course Black is not forced to move the
Game 42 pawn yet but at some point in the near future
White is really going to be threatening to take it.
Rufener-Gallagher
Black will then have to choose between defend­
Bern 1997 ing it with ...'Ot>h7 (passive) or advancing it. I
think it's best to move it at once as the other
1 d4 ltlf6 2 c4 g6 3 ltlc3 ..tg7 4 e4 d6 5 f3 two natural moves, 1 t ...ltJbd7 and 1 1 ...l:.e8
0-0 6 ..tg5 c5 7 d5 e6 8 11Fd2 exd5 9 cxd5 both have slight drawbacks
a6 10 a4 a) 1 t ...ltJbd7 allows White to play 12 itJh3!

1 17
Pla y th e King 's Indian

and after 12 ...lL!e5 13 lLJ£2. As I have already You see what I mean about gerring con­
pointed out this is thought to be the ideal fused. I am not sure why I didn't play 14 ...lLJh7
square for the king's knight in this line. as I did in the game with Klauser (see below).
b) The drawbacks to 1 1 ...l:e8 are less obvi­ Probably I thought I was following it. The main
ous and this move is often played. However, reason for playing ...l:e8 in such positions is
with Black's main plan being to play for ... fS it that occasionally White can meet ... f7-f5 with
is not really clear yet if the rook is better on £B £3-f4, attacking the knight on eS, followed by
or e8. There are also occasions when Black may e4-e5 and after Black plays d6xe5 then ..ie3xc5,
retreat his knight from f6 to e8 (see Rogers­ which is obviously stronger if it hits a rook on
Gallagher in the notes below) so, all in all, £8. However, that doesn't seem much of a
1 1 ...h5 feels the most accurate. threat here but it is worth bearing in mind for
1 2 lBge2 lBbd7 1 3 ttxl1 similar positions. Anyway, K.lauser-Gallagher,
I refer you to the 1 2th move notes in Game Silvaplana 1 997 had gone 1 4 ... lL!h7 1 5 ..ie2 f5
37 for a discussion of the ideas behind this. 16 0-0 ..id7 17 lLJ£2 'i1Vh4 1 8 exfS gxfS 19 ..if4
An alternative for White is to play 1 3 lL!f4. l:ae8 20 l:ae1 'ilff6 21 �ht �h8 with an un­
This awkward-looking move is of a purely de­ clear position that I eventually won.
fensive narure. The idea is to prevent Black's 1 5 .i.e2 lBh7 1 6 lBf2 f5 1 7 exf5 gxf5
active plan of ...lL!h7 and ... fS. With the knight Although there are quite a few exceptions in
on f4 it will take an awful lot of arranging to general it is better to recaprure with the pawn in
play fS. In SOKID I annotated my nice win order to keep control of e4.
against Ian Rogers. Here, I'll just give the 1 8 0-0 -..,4 1 9 l:fe1?!
moves in brief: White got into trouble o n the e-file so i t is
1 3. .. lL!es 1 4 ..ie2 ..id7 1 5 0-0 lL!e8 (1 5...b5!?) easy, with the benefit of hindsight, to criticise
16 l:fb1 aS! (a most unusual way for Black to this move.
play in this line but it made sense to me to pre­ 1 9 ... .i.d7 20 .i.f4 l:e7! 21 lBh3?! :Sea
vent b4 and leave the rooks looking rather silly) Black is threatening ... lL!x£3+ so White de­
1 7 lL!bS ..ixbS 1 8 ..ixbS lL!c7 1 9 .in 'ii'd7 20 cides to defend his rook on el.
l:d1 �h7 21 h3 b6 22 b3?! fS! 23 exfS?! 'ilfxfS 22 �1 b5! 23 axb5 lBxf3!!
24 l:ac1?! l:ae8 25 ..ibS l:e7 26 lL!e6? lL!xe6 27
dxe6 'ilfxe6 28 ..igS lL!xf3+l 29 gx£3 ..id4+ 30
�h 1 (30 �g2 l:x£3! 31 'iii>x£3 'ilfxh3+ 32 'iii>f4
'i1Vg4 mate) 30 ...:X£3 31 .in 'ii'e4! 32 �h2 (32
..ig2 l:xh3 mate) 32 ... l:ef7! 33 'ii'g2 l:£2 34
'ith1 :Xg2 35 ..ixg2 l:£3! 0-1 Rogers-Gallagher,
Lugano 1 999.
1 3 ...ttJe5 1 4 �eel l:e8!?

I must have felt good playing that one.

24 .i.xf3
24 gx£3 'ilfxh3+
24 ... .i.xc3! 25 'ifxc3
25 l:xe7 ..ixbS+ 26 ..ie2 ..ixd2 27 l:xe8+
..ixe8 28 ..ixd2 is hopeless for White.
25 ....i.xb5+
The point of 21 ...b5 is revealed. The check is

1 18
 The Samisch

crushing as the white king cannot desen the llf1 �g4 53 h5 .i.c4 54 llh1 lleJ 55 .i.c5
rook on el. llhJ 56 llxhJ �xhJ Y.z - Y.z
26 .i.e2 :Xe2 27 :Xe2 :Xe2 28 'iFgJ+
'iFxgJ 29 .i.xgJ lbf6 30 �g1 lDe4 31 lDf4 Summary
llxb2 32 lbxh5 1) Black should meet the Siim.isch with
6 ... c5. It is playable against all three of White's
6th move options.
2) The endgame a pawn down after 6 .i.e3
cS 7 dxcS dxcS 8 1i'xd8 l:xd8 9 .i.xcS holds no
fears for Black and very few White players ac­
cept the pawn in master chess nowadays. Most
players are happier to play chess a pawn down
with a good position than a pawn up with a bad
position.
3) 6 .i.e3 cS 7 lbge2 'it'aS is a very interest­
ing, almost totally unknown move order, which
encourages White to transpose into a variation
that he discarded. I see no profitable way for
White to avoid this transposition.
Black has a completely won endgame and I 4) 6 lbge2 is probably the most difficult
only wish that I could now write that I won move order for Black to face but with precise
a&er a few more moves. But I can't as I some­ play he can achieve a good game. 9 ... lbh5
how contrived to draw! I must have just (Game 40) is the more controversial way to
switched off a&er the beautiful combination. I play and we will have to see if it stands the test
can only hope the editor cuts out the remaining of time. 9 ... h5, despite the comments of Mik­
moves but in case he doesn't here they are. halevski, is a much more respectable line.
32 ...a5 33 .i.f4 a4 34 lDgJ lDxgJ 35 .i.xgJ 5) In these Benoni positions Black would,
lld2 36 .i.xd6 :Xd5 37 .i.f4 c4 38 �2 �7 primarily, like to expand on the queenside. As
39 h4 �g6 40 gJ cJ 41 llc1 llc5 42 .i.d6 White usually prevents this his most common
llc4 43 �eJ c2 44 .i.f4 aJ 45 �d2 .i.a4 46 method of connterplay comes in the form of
lla1 .i.bJ 47 �c1 lla4 48 .i.d6 lld4 49 .. f5. Black cannot play this variation passively.
.

.baJ lldJ 50 �b2 llxgJ 51 .i.e7 �h5 52 If he does he will lose.

1 19
 CHAPTER EIGHT I
The Fianchetto Variation

1 d4 lL!f6 2 c4 g6 3 lL!f3 ..tg7 4 g3 0-0 5 ers want to control the game and they begin to
..tg2 d6 6 0-0 feel uncomfortable when they feel this control
The Fianchetto Variation is characterised by slipping away. Make them dance to your tune
White developing his king's bishop to g2. It is a and they won't like it - even if the position is
solid, restraining line where it is difficult for objectively in their favour. This system which I
Black to obtain active play by the standard helped to develop actually ended up bearing my
means we see elsewhere in this book. In fact name. I didn't invent it but did most of the
the normal plan of attacking on the kingside is early work on it. I remember playing a never­
simply no longer viable, at least in the early ending tournament in Baku in 1 988 where in­
stages of the game. Firstly, because the bishop stead of spending my time preparing for the
on g2 affords extra protection to the white king rather dangerous Soviet opposition I just im­
and secondly because, unlike in other lines, mersed myself in this fascinating variation. On
White has no intention of blocking the centre. my return home I started to play it in tourna­
Attacking on the wing is a risky business when ments (I didn't get the chance in Baku) with
the centre is fluid. White's plan in this variation amazing results. In those early, heady days of
is to stifle Black's active play and gradually use the variation I won game after game, and often
his space advantage. in crushing style, against rather confused oppo­
sition. Information traveUed much slower then
Gallagher Variation and it took sometime for the rest of the world
Although White's play is initially quiet, the ri­ to cotton on. But when they did the variation
anchetto is not a hannless system. It is, in fact, suddenly became the height of fashion! Every­
an extremely dangerous line for the typical one started playing it including some of the
King's Indian player. The positions can be very world's top players. This was of course very
difficult to handle. In my early King's Indian pleasing but it had an inevitable downside.
days I suffered horribly in this line. Thin!,>s got White players began to take the variation seri­
so bad that I just felt like resigning when I saw ously and, in the quiet of their studies, started
my opponent reaching for the g-pawn. These to work out the optimum set-up for their
problems continued for many years until I dis­ pieces. Suddenly the ... cS and ... bS attack was
covered a way to create chaos on the board. In no longer blowing White away and the variation
order to create this chaos Black has to take began to look rather dubious. The crowds de­
great positional risks but these risks seemed paned almost as suddenly as they had arrived
justifiable against the son of player who plays and even I became increasingly unfaithful to the
the Fianchetto Variation. Above all these play- variation. But there seems to have been a mini-

1 20
 The Fia n c h e t to Varia tion

revival recently with the new brand of players discussion of this move order, as weU as the
not so willing to give up their d6 pawn. Nowa­ independent lines it can lead to, foUows later
days I tend to reserve the variation in its purest (Games 51 -52).
form (i.e. ...cS and ... bS against whatever White 8 . .exd4
.

does) for rapid play tournaments or selected It is anti-positional for Black to concede the
opponents, and rely on a closely related but less centre in this fashion but he does so with a
committal system (8...a6) for more serious concrete method of developing counterplay in
tournaments. That second system will form the mind. The alternatives are:
backbone of the repertoire I am suggesting. a) 8 ... c6, the old main line which I am steer­
However, it would be impossible to play that ing weU clear of as that is where I used to suffer
line properly without a thorough understanding horribly in my youth.
of the ideas behind the Gallagher Variation b) 8 ... a6 wiU be dealt with in depth in the
(and, also, it does seem rather appropriate for second half of the chapter.
me to address this topic). Therefore, the first 9 tt'lxd4 :tea 1 o hJ
part of this chapter details the rise and fall fol­ In the past it was thought that tO l:r.e1 was
lowed by the semi-comeback of my variation inaccurate because of the reply 1 o...lt'lg4 but in
before moving on to the study of 8...a6 in Game 48 you can discover that this is not the
Games 47-52. case. After 1 0 l:r.e1 Black can just play 1 0 ... a6
Games 53 and 54 deal with early deviations with a likely transposition to the main lines after
from White. Let's take a look at some moves. Wll ite plays h3. A few White players have tried
to save a tempo by omitting h3 altogether but
1 d4 tt'lf6 2 c4 g6 3 tt'lfJ �g7 4 gJ 0-0 5 this means they have to constantly worry about
i.g2 d6 6 0-0 tt'lbd7 a Black piece coming to g4.
Black prepares to counter in the centre with 1 0 . .a6!?
.

...e7 -eS. lbis is known as the Classical Variation

while the other main line, 6. ..lbc6, is known as
the Panno Variation. In this book we are going
to concentrate on the Classical. Not because it
is a superior line but because I have played it all
my life and should be able to impart its secrets
better than a line I have hardly ever played.
7 tt'lcJ
7 'ifc2 is covered in Game 53.
7 ...e5 8 e4

The first sign that Black has something spe­

cial in mind. In the old days he would play
moves such as ... c6, ...lbcS and ... aS and try and
hold the balance, often unsuccessfuUy. The
problem was that it was very difficult to gener­
ate active play. 1 O. .a6 reveals Black's aggressive
.

intentions. I t is the first preparatory move in

the queenside pawn storm that Black is plan­
ning. His next move will be 1 1 .. .l:r.b8 and he
will then foUow up with ... c7-c5 to kick the
Certain players prefer 8 h3, although this of­ knight from d4 and then ... b7-b5. Sometimes
ten just transposes back into the main lines. A Black throws in lbeS before advancing his

121
Pia y th e King 's Indian

pawns, on other occasions he just leaves it on White on the long dark diagonal while the
d7. 1bis can be considered as the basic starting move f4 means that the bishop on e3 is no
position of the Gallagher Variation. White now longer protected.
has various ways to deploy his pieces and these White could also have retreated his knight to
are covered in Games 43-46. c2. 1bis is what happened in Miralles­
Gallagher, Bern 1 991 where Black achieved a
Game 43 very promising position in similar fashion to
the main game: 1 3 tbc2 ltJe5 1 4 b3 b5 1 5
Hohler-Gallagher
cxb5?! (there is no need for White to exchange
Bern 1994 of a c-pawn for an a-pawn - the immediate 1 5
f4 ltJed7 1 6 ...xd6 is the better option)
1 d4 lLif6 2 c4 g6 3 lLifl �g7 4 gl 0-0 5 1 5 ...axb5 1 6 f4 liJed7 1 7 1i'xd6 b4 1 8 e5 l:tb6!
�g2 d6 6 0-0 lLibd7 7 li:\cl e5 8 e4 exd4 9 19 1i'd1 (19 1i'd3 .ia6) 1 9 ... bxc3 20 exf6 .i.xf6.
lLixd4 .l:leB 10 hl a6 1 1 .l:le1 .l:lbB White is close to lost but I managed to spoil a
Black continues the preparation for his beautiful game in time trouble and Miralles
queenside offensive. even emerged with the full point.
1 2 �el c5 1 3 ...lL\e5
1 3 ... b5 appears less convincing for tactical
reasons. Here is a key variation. 1 4 'ifxd6 b4 1 5
ltJa4 ltJxe4 1 6 .i.xe4 l:txe4 1 7 ltJxc5 l:txe3 1 8
ltJxd7! l:txe2 1 9 l:txe2 l:tb7 (1 9....i.xd7 20 l:td1
l:tb7 21 l:te7 wins the bishop) 20 liJf6+! .ixf6
(or 20...1i'xf6 21 1i'xf6 .i.xf6 22 l:te8+-) 21 'ifxd8
.ixd8 22 l:te8+- and White wins material.
14 bl b5 1 5 f4 lL\ed7
Black must retreat here as 1 5 ...ltJc6 IS
strongly met by 16 e5!
1 6 "irxd6 b4

1bis is the move that shocks the positional

men. Not only does Black give himself a back­
ward d-pawn on an open file but he also cedes
control of d5. Modern chess players are not so
worried about such static details. They are more
interested in the dynamics of the position.
Black wants to grab the initiative and if that
means making sacrificing material or making
positional concessions - so be it.
1 3 lDde2
Black has an interesting decision to make
now. He can either play 1 3 ... b5 at once (yes, The reason that Black can get away with
that's right, ignoring the threat to the d-pawn) such opening play is that he is not really attack­
or he can throw in the moves 1 3 .. .lbe5 1 4 b3 ing on the queenside at all - White's central
b5 1 5 f4 liJed7 (1 5 ... ltJc6? 1 6 e5!). In both cases play would crush an attack there - but he is
Black has the same position but in the latter attacking in the centre himself. The queenside
case he has given White the moves f4 and b3 demonstration is just the means employed to
for free! The claim is that these moves actually destabilise White's e4-point. It can no longer be
harm the white position. The move b3 weakens held.

1 22
 The Fia n ch e t to Varia tion

1 7 e5!
White can't play as in the 1 3th move note as Game 44
after 17 li:)a4 li:)xe4 18 itxe4 l:xe4 both the
Drasko-Gallagher
rook on a 1 and the bishop on e3 are hanging.
The text makes better use of the moves, f4 and Aosta 1990
b3, that White has been given for free.
17 ...bxc3 1 8 exf6? 1 d4 lLlf6 2 c4 g6 3 lLlf3 �g7 4 g3 0-0 5
But this is a serious mistake after which �g2 d6 6 0-0 lLlbd7 7 lLlc3 e5 8 e4 exd4 9
White is in trouble. He should have played 1 8 lLlxd4 lleB 1 0 h3 a6 1 1 �e3 llbB 1 2 'ifc2
lbxc3! li:)hS 1 9 g4 � £8 20 1i'd2 li:)g7 when in At the time of this game the system was still
return for the piece White has a couple of in its infancy and I'm not even sure that Drasko
pawns and a strong grip in the centre. The posi­ understood what Black was planning? Perhaps
tion can be considered unclear. not, as now he doesn't get to win the pawn on
18 ...llxe3 1 9 fxg7 llb6! 20 'ifd1 llbe6 d6.
The fine rook manoeuvre has left White 1 2 ... c5 1 3 lLlde2 b5 1 4 llad1 'ife7 1 5
stuck in an awk-ward pin on the e-file. �f4?!
21 llc1 'iff6 22 'ifc2 'ife7 23 �2 lLlf6 24 White is in an ambitious mood. His plan is
llcd1 �b7 25 �d5 �xd5 to lure the black knight into eS so that he can
It was also possible to play 25 ...li:)xd5!? 26 attack it with a later f4. However, it is all going
cxdS l:6e4 as after 27 d6 Black has a stunning to explode in his face. A more solid continua­
combination: 27 ...l:£3+! 28 q;g2 (28 'iti>gt loses tion is 1 5 cxbS axb6 1 6 li:)f4 with the plan of
to 28...1i'd7 with the threat of 1i'xh3) occupying dS. The game is very unclear after
28...l:ee3!! 29 dxe7 (White is also getting mated 1 6...�b7!.
after 29 'ith2 .l:£2+ 30 'itgt l:g2+ 31 'itft 'itd7) 1 5 ...lLle5 1 6 cxb5 axb5 1 7 �g5 'iff8
29 ...l:xg3+ (don't forget it's double checkQ 30 It's best to get out of the pin. Now Black
�h2 l:g2+ 31 'ith 1 l:xh3 mate. doesn't have to worry about a knight landing
26 cxd5 ll6e4 27 lld3 c4! on dS.
1 8 f4 lLlc4 1 9 'ifc1 �a6!
I'd seen that the tactics were going to work
in my favour.
20 e5 dxe5 21 fxe5 lLlxe5! 22 �xf6 b4

Opening the a7-g1 diagonal so that the black

queen can get at the white king.
28 llxe3 llxe3 29 d6 'ifa7 30 'ifc1 ?
White had to try 30 bxc4 though after
30...�4+ 31 �g2 both 31 ...li:)xd6 and 31 ...1i'a8 Black regains the sacrificed piece as the
win comfonably for Black. knight on c3 can't move on account of �xe2.
30...lLle4+ 31 �g2 llxg3+! 0-1 23 �xe5 llxe5 24 �d5 llxd5! 25 Ld5
In view of 32 li:)xg3 1i'f2+ 33 'Oti>ht li:)xg3 bxc3 26 lLlxc3 �d4+ 27 llf2 �b7
mate. The black bishops are monstrously powerful

1 23
Pla y the King 's Indian

and the white king too exposed. bad but when grandmaster after grandmaster
28 l:l.d7 'ife8 29 l:tc7 'ife51 30 l:l.cxf7 'ifxg3+ abandoned their normal repertoire in order to
31 �1 .i.a6+ 32 �e1 l:l.e8+ 0-1 take me on in this line I felt it was time to lay it
to rest.
Game 45 1 2 l:l.b1
Yin Hao-Ye Jiangchuan
Shanghai 2000

1 d4 Q\f6 2 c4 g6 3 Q\t3 .i.g7 4 g3 0-0 5

.i.g2 d6 6 0-0 Qlbd7 7 Qlc3 e5 8 e4 exd4 9
Qlxd4 l:l.e8 1 0 h3 a6 1 1 l:te1
For a while it became fashionable for White
to withdraw his knight from d4 before it was
hit with cS. The moves 1 1 tLlb3 and 1 1 tLlde2
are usually described as prophylactic but I pre­
fer to call them cowardly. White is just trying to
avoid complications at all costs. The knight
was, after all, perfectly weU placed in the middle
of the board and could only be dislodged by This innocuous-looking move has caused
Black taking positional risks. Not surprisingly, Black quite a few problems. I think White
Black is able to find adequate counterplay by found this move by a process of elimination. It
'normal' means. doesn't do a lot but on the other hand it doesn't
a) 1 1 tLlb3 tLleS 1 2 cS (the only way to justify do any harm either. It doesn't interfere with the
his previous move as 12 1i'e2 is weU met by defence of the e-pawn (12 i..e3), it doesn't
12...i..e6) 12 ...dxc5 13 tLlxcS (13 f4 tLld3 14 eS expose White on the long diagonal (12 b3) and
tLlhS 15 l:tf3 c4 gets the White centre rolling at it doesn't deny White the opportunity to play
the cost of a pawn but the complications fa­ 1i'xd6 (12 1i'c2).
voured Black in Yrjola-Cramling, Nordic zonal Not only did White players try and make the
1992) 13. ..1i'e7 14 tLlb3 i..e6 1 5 1i'c2 l:tad8, ... cS and ...bS advance less appetising (12 l:tb1)
Spassov-Van Wely, Munich 1992. Black's active but some of them decided to prevent it alto­
pieces more than compensate for White's cen­ gether. The most obvious way to do this is 12
tral pawn majority. The availability of the c4- a4. At first White players were reluctant to play
square for his knight or bishop is an important such an anti-positional move. The point is that
factor. if Black meets 1 2 a4 with 1 2 ... a5 (12 ... tLlc5 and
b) After 1 1 tLlde2 Black has a choice be­ 12 ...tLle5 arc also possibilities with ... aS some-
tween a complex rniddlegarne with 1 1 ... l:tb8 or times corning later) he · obtains outposts for his
a tactical trick which wins a pawn but quickly pieces on cS and b4. Ho"'Z_ver, just as Black can
leads to a drawish endgame. Pr.Nikolic­ get away with his anti-positional advance ...cS,
Gelfand, Wijk aan Zee 1 992 continued (after 1 1 White, too, can make a case for playing a4. The
tLlde2) 1 1 ...tLlb6 1 2 b3 tLlxe4! 1 3 tLlxe4 :Xe4! justification lies in the fact that the time spent
(13...i.xa 1 14 i..gS) 14 i..xe4 i..xa 1 1 5 i..g2 on the moves 10 ... a6 and 1 1 ...l:tb8 has now
i.f6 16 tLlf4 l:tb8 1 7 i..e3 i..d 7 (Black obvi­ been lost so White obtains a sizeable lead m
ously felt that keeping the extra pawn by development
17 ...tLJJ7 was too risky) 1 8 i..xb6 cxb6 1 9 1 2 . . .ttJe5
1i'xd6 i..c6 20 1i'xd8 l:tfxd8 2 1 tLld S 1/z-1/z. Black has the usual choice between playing
1 1 ...l:l.b8 ... cS and ... bS at once or flicking in ...tLleS first. I
When I became disillusioned (see notes to don't really trust 1 2...c5 if only because of the
Black's 14th move) with the variation I started foUowing variation: 13 tLlf3!? (I also have my
to play 1 t ...tLlc5 here. My results were not too doubts after 1 3 tLlc2 bS 14 1i'xd6) 14 ... b4 1 5

1 24
 The Fia n c h e t to Varia tion

�5 lbxdS (1 5 ...lt:Jxe4 16 "W'f4 is awkward for This clamps down on the intended ...bS but
Black) 1 6 'ii'xdS .ib7 1 7 1i'd3 lt:Jf6 1 8 1i'xd8 at the same time creates holes on the white
.:tbxd8 1 9 e5 lt:Jd7 20 .igS l:lc8 21 b3 .ix£3 22 queenside. These are well covered at the mo­
.ix£3 lt:Jes 23 .idS with an unpleasant end­ ment but as the game goes on they may be­
game for Black. Of course this line is not come more relevant. 1 5 f4 is more aggressive
forced but it's not easy to find anything signifi­ and also prevents the immediate ...bS on ac­
cantly better for Black. count of the reply eS. Gavrikov-Volokitin, Bad
13 b3 c5 1 4 lbc2 Wiessee 2000 continued 1 5 ... .ie6 1 6 1i'd3 bS!
Or 14 lt:J£3 lbxf3t 1 5 'ii'x£3 .ie6 16 .if4. 17 .id2 (Black exploits the unprotected posi­
What Black would like to do now his pick up tion of the white queen to force through ... bS;
his knight on f6 and place it down again on c6. after 1 7 eS dxeS 1 8 1i'xd8 lt:Jxd8 19 fxeS lt:Jd7
That would give him maximum control in the Black has a very good position) 1 7 ...lt:Jd7 1 8
centre and chances to occupy the tasty outpost cxbS lt:Jd4! (more tactics from the young
on d4. White's last move is designed to prevent Ukrainian - you'll be hearing more of Volokitin
the manoeuvre ...lt:Jd7-e5-c6 as now 1 6 ...lt:Jd7 in the future) 1 9 lt:Je3 (1 9 bxa6 c4! 20 bxc4 lt:Jc5
would be met by 17 .ixd6. What Black must forces White to sacrifice his queen to avoid
do is play 16 ...lt:Jh5 and after 17 .id2 come losing the knight on c2) 1 9... a.xb5 20 1i'f1 b4 21
straight back with 17 ... lt:Jf6. He is again in a lt:Je2 f5 with a very active game for Black, al­
position to carry out his knight manoeuvre. though White eventually won.
White must allow the manoeuvre or agree to a 1 5 . . .�e6! 1 6 �b2 h5 1 7 Wd2
repetition of moves after 1 8 .if4 lt:Jh5. 1 7 lt:Jtl5 was an alternative.
14...lt:lc6! 1 7 ... lt:lh7! 1 8 lZ.ed1 .i.e5! 1 9 f4 .i.g7 20
�h1
20 1i'xd6? .id4+ cuts communications be­
tween the queen and rook. White is forced to
give up the exchange with 21 l:lxd4 but doesn't
get sufficient compensation.
20 ...Wa5 21 lt:ld5 Wxd2 22 :Xd2 .i.xd5 23
cxd5

Of course Black would like to play 1 4...b5

but it simply doesn't work in this position. A f­
ter 1 5 cxbS a.xbS 1 6 f4 lt:Jed7 17 1i'xd6 there is
no good way for Black to continue. I became
quite disillusioned with Black's chances against
12 .:tb1 as it takes the sting out of Black's
counter-attack. I had assumed that if Black
could not force through ...bS then he would live White should have played 23 exdS. Although
to regret having played ... cS. However, it, seems his centre may look impressive the pawn on e4
that even without the hand to hand fighting of is just a weakness which Black can attack.
the previous games Black is able to live with his 23 . . .lLla5 24 .i.xg7 �xg7 25 b4 cxb4 26
dubious pawn structure. His plan is still to play l:txb4 lt:lf6 27 lLle3 l:tbc8 28 �g1 l:tc3 29
bS but at a more convenient moment. �2 lt:lb3 30 lt:ld1 l:txg3! 31 l:tc2 h4 32 l:tc3
1 5 a4 lt:lc5 33 e5?

1 25
Pla y t h e King 's Indian

White blunders but even 33 l:txg3 hxg3+ 34 I have always been amazed by how many
�xg3 lDfxe4 is losing in the long run. White players unthinkingly exchange on b5 in
33 ...:Xc3! 0-1 similar positions when it is obviously in White's
White resigned because of 34 exf6+ �xf6 35 favour to keep the c4 and a6 pawns on the
lDxc3 tDd3+. board.
1 4. . .b4 1 5 lDB4 ltlxe4 1 6 .be4
Game 46
S. Williams-Gallagher
British Ch., Scarborough 2001

1 d4 ltlf6 2 c4 g6 3 ltlf3 �g7 4 g3 0-0 5

�g2 d6 6 0-0 ltlbd7 7 ltlc3 e5 8 e4 exd4 9
ltlxd4 :es 1 0 h3 a6 1 1 �e3 :be 1 2 b3

1 6 ...:Xe4
I spent ages wondering whether to take the
rook or not. It seems that White is better after
1 6 ... i..xa 1 1 7 :Xa 1 l:txe4 1 8 lDxc5 l:te8 (I ex­
amined both 18 ...l:tb6 and 1 8 ...:Xe3 but didn't
like either) 19 l:td1 'ii'e7 20 'ii'c7! lDxc5 21
1i'xb8 as Black has insufficient compensation
for the pawn.
It seems that this set-up also casts doubt on 1 7 :ad 1 •es 1 8 ltlxc5 :xe3 1 9 ltlxd7
the immediate ... c5 and ... b5 advance and a few :xe2
players have mentioned to me that this is the 1 9 ... i..xd7 20 fxe3 i..xh3 21 l:t£2 does not
refutation of the Gallagher variation. offer Black enough for the exchange
1 2 ...c5 20 ltlxb8 �xh3 21 ltld7?
So it is certainly worth considering This looks very strong but Black can main­
1 2 ...'ii'e7!?. The idea is still to play ...c5 and ... b5 tain the balance through a queen sacrifice. The
but without sacrificing the d-pawn. The game best continuation is 21 1i'b6! i..fB 22 l:td8 1i'e4
Ruck-Vajda, Hungarian Team Ch. 2003 contin­ 23 'ii'c6! 'ii'xc6 24 tDxc6 i..x fl 25 �xfl l:txa2
ued 13 l:te1 c5 14 tDde2 b5 1 5 cxb5 axb5 1 6 26 tDxb4 when the powerful white c-pawn will
lDf4 'ii'£B 1 7 a4 with unclear play although I ensure Black has a difficult rime in the ending.
would prefer 16 ...b4 on positional grounds. 21 . . .�xf1
Maybe White can improve on the above, but Not 21 ...1i'a8 22 1i'b8+ with advantage to
refutation? I don't think so. White.
1 3 ltlde2 b5?! 22 ltlf6+! �xf6 23 •xf6 :e1 !
It is better to defend the d-pawn with 23...l:te6 24 .:td8 l:txf6 25 :Xe8+ �g7 26
13. ..'ii'c7 and after 1 4 a4 (to stop 14 ... b5) b6! 1 5 �xfl is certainly not easier to hold.
.:ta2 i..b7 16 f3 l:tbd8 17 l:td2 tDe5 Black had a 24 :de �xc4+ 25 �h2 �d5?
perfectly playable game in Shengelia-Mamodev, After 25...i..b5! 26 l:txe8+ l:txe8 Black is in
Barumi 2003. There are weaknesses in both no danger as White will only be able to win the
camps. pawn on b4 at the cost of a2.
14 •xd6! 26 :xeS+ :xeS 27 •d6?!

1 26
 The Fia n c h e t t o Varia tion

27 'ifxa6 l:te 1 28 g4 should be winning for of a pin on the a6-f1 diagonal.

White. White is srill better after the text but Game 48: White plays 9 l:tet . Now 9...b5 is
Williams just kept playing instantly to try and not playable but by 9 ...exd4 10 ltJxd4 lbg4!
take advantage of my horrendous time trouble. Black obtains a level game.
27 ...�f3 28 'ii"xb4 l:te2 29 1Wc5?! h5 30 Games 49-50: White plays 9 'ifc2 (and oth­
g4?! hxg4 31 �g3 l:tx82 32 �4 l:te2 33 ers). This has become quite popular and Black
�g5 l:te6 34 �h6 g5+ 35 �xg5 �g7 % - % can revert to a sort of delayed Gallagher Varia­
A tough game but this second round en­ tion.
counter must have put me in good stead for the Games 5 1 -52: 8 h3 a6 (White delays c4).
rest of the tournament as I went on to become Quite often this will just transpose into other
British champion that year! lines but there are alternatives for both sides.
That concludes my direct coverage of the
Gallagher Variation. Game 47
Renet-Gallagher
8.. 86
.

French uague 2003

1 d4 l!Jf6 2 l!Jf3 g6 3 c4 �g7 4 g3 0-0 5
�g2 d6 6 l!Jc3 l!Jbd7 7 0-0 e5 8 e4 (8 h31 1 d4 l!Jf6 2 l!Jf3 g6 3 c4 �g7 4 g3 0-0 5
8 86
... �g2 d6 6 l!Jc3 l!Jbd7 7 0-0 e5 8 e4 a6 9
h3
White is very fond of the move h3 in these
Fianchetto systems and we shall sec in the next
game that allowing ...lbg4, and on other occa­
sions ... ..tg4, can be irritating for White. It
should be said that this position is often
reached via the alternative move order 8 h3 a6
9 e4.
9 ... b5

The move 8... a6 was played <:Juite a lot in the

1 970's and 80's by the East German grandmas­
ters Vogt and Knaak but then gradually disap­
peared from practice. It is now back in vogue.
As far as I can see White has considerable diffi­
culty in proving any advantage against the semi­
waiting move, 8...a6. The basic idea is the same
as in the Gallagher Variation - to aim for active
play on the <:JUeenside with ... b5. This time,
though, Black delays his decision on whether to This may look like a sacrifice but the pawn
exchange on d4 or not. It all depends on what can't be taken. After 10 cxb5 axb5 1 1 lbxb5?
White does next. The basic rule is that if Black ..ta6! Black wins material. For example, 12 a4
can play ... b5 without having to concede the ..txb5 or 1 2 'ife2 c6. After 9...b5 the central
centre (...exd4) then he will do so. The material situation has become very tense. \Xfhitc must
is split up as follows: now decide what pawn exchan1-,>es, if any, he is
Game 47: White plays 9 h3. This can be met going to make.
by 9...b5 as White cannot win a pawn because There was of course still the chance to

127
Pla y t h e King 's Indian

ttanspose into the G·allagher Variation with lDdS) 1 2 l:td1 'W'e7 1 3 lbe1!? (the knight wasn't
9...exd4. doing a great deal on f3 so it heads off towards
10 dxe5 the queenside) 1 3...ltlb6 14 cS!? (I must admit
Or: that I hadn't seen that one coming at all)
a) 10 cxbS can be considered a slight posi­ 14...ltlbd7 (14...'W'xcS?? 1 S .i.e3 wins a piece)
tional concession as White gives up a central 1 S b4 aS 16 .i.a3 (White would be able to claim
pawn for a wing one and at the same time an edge if he could suppon his pawn chain with
opens the a-file for the black rook. f'ilippov­ 1 6 a3 but this is of course impossible because
Lyrberg, Minsk 1 996 continued lO ...axbS 1 1 b4 of the pin on the a-file) 1 6 ...ltlb6! (back we go!
(An attempt to blockade the position on the the knight wishes to install itself on c4 and 17
queenside. It is perfectly feasible for Black to cxb6 axb4 is not good for White) 1 7 bxaS lbc4
just play 1 1 ...c6 and aim, in the long run, to 1 8 .i.b4 lLlxaS 19 lbc2 .i.e6 20 1i'e1 ltlc4 21
occupy the square on c4 with a knight. In the lbe3 lbxe3 22 1i'xe3 l:tfd8. The game is about
game, though, Black prefers to challenge White level and f'edorowicz-Gallagher, Chicago 1999
at once on the queenside.) 1 1 ...cS!? 12 bxcS was drawn in 48 moves.
(Safer was 12 dxeS dxeS 1 3 bxcS b4 14 lbds b) 1 1 1i'c2 c6 (preparing ... 'W'e7 again and
ltlxcS with an equal game.) 12 ... b4 1 3 lbe2 (13 better than the greedy 1 1 ... bxc4? which wrecks
ltla4 exposes the knight and is well met by Black's queenside for a pawn that he won't
13 ...1i'aS while 1 3 lbds ltlxdS 1 4 exdS e4 is even be able to hold on to in the long run) 12
good for Black.) 13 ... -i.b7! (The main point .i.e3 'ire7 and now:
behind Black's play as now there will be no b1) 1 3 cxbS axbS 14 l:tfd1 .i.b7 1 S b3 l:tfc8
tactical problems on the long diagonal like there 1 6 a4 ltlcs 1 7 aS .i.£8 1 8 lbe1 lbe6 1 9 lbd3
would have been after the inunediate lbd4 with a good game for Black, Vaulin­
13...ltlxe4.) 1 4 cxd6 ltlxe4 1 S dxeS lLlxeS 16 Smimov, Tomsk 1 998. The manoeuvres in this
lbxeS .i.xeS 17 .i.h6 l:te8 1 8 l:tb 1 (On 1 8 d7 game are very typical for this type of position.
Black just plays 1 8.. ..J:te7 and takes the pawn The hole on d4 makes a ftne home for a black
next move.) 1 8 ... 'W'xd6 1 9 'irb3 (Exchanging knight. Here the queen's knight made it via cS
queens was certainly an option though Black and e6 while in other games the king's knight
has a slight advantage in the ending as White's heads for d4 via e8-c7-e6.
pawn on a2 is weaker than Black's on b4.) b2) 13 a3 .i.b7 14 l:tfd 1 l:tfc8 1 S lbd2 trans­
19...lbd2 20 .i.xd2 .i.xg2 21 �xg2? (21 .i.xb4 poses into Renet-Cnillagher.
'W'c6 probably looked to risky for White as 1 1 ...c6
Black gets to keep the powerful light-squared 1 1 ...b4 12 lbds ltlxe4 1 3 lbd2 ltixd2 14
bishop, but the game just looks like a draw after 1i'xd2 l:tb8 1 S ltlxb4 is better for White.
22 l:tfcl 1i'b7 23 .i.c3 1i'xb3 24 axb3 i.xh3.) 1 2 Wc2 We7 1 3 lbd2 .i.b7
21 ...'W'xd2 22 lbc1 ? (White's previous error The bishop doesn't really belong here but
should have just cost him his a-pawn but he the idea, in conjunction with the next move, is
must have planned all along to defend it in this to be ready to meet an early b4 by White with
fashion.) 22 ... l:ta3! (Only now did White realise ...c6-cS or ...a6-aS.
that if he moves the queen 23 ...l:txg3+! leads to A reasonable alternative could be 1 3...lbe8,
a mating attack.) 23 lbd3 l:txb3 24 l:txb3 .i.c3 intending lbc7-e6. Black could, for example,
0-1 . meet 1 4 ltlb3 lbc7 1 S ltlaS with 1 S...ltlb8 as the
b) 10 'W'c2 c6 1 1 .J:td1 'W'e7 12 .i.e3 exd4 1 3 white knight is not particularly well placed on
ltlxd4 .i.b7 14 cxbS axbS 1 S a3 l:tfe8 was ftne aS. The move ... £7-£5 may even be on the cards
for Black in P.Nielsen-Hillarp Persson, Malmo one day but this desirable move is difficult for
2003. Black to arrange once he has played ... b7-bS as
1 0 ...dxe5 1 1 .i.e3 there may be trouble on the long diagonal after
Others: White plays ex£5. I think I will try that next
a) 1 1 'W'e2 c6 (this secures a home for the time.
queen on e7 without having to worry about 1 4 a3 l::tfc8 1 5 1lfd1

1 28
 Th e Fia n c h e t t o Varia tion

may be best not to touch the queenside at all,

waiting for White to play tDb3 before playing
... aS. How about 1 5 ... h5!?. That was the move
Renet was most concerned about and after­
wards he revealed that he was not too happy
with his position around here. Another idea is
1 5 ...1ff8 intending to follow up with ... i.h6. I
consider the chances to be about level although
the line is not as straightforward for Black as I
once thought it to be (and not only me - a lot
of annotators were awarding 9 h3 the '?!' sym­
bol until recently).
1 6 b4 lL\e6 1 7 lLib3 bxc4 1 8 lL\cS!

1 S ... lL\cS?
This just plays into White's hands and allows
him to set up his ideal position.
15 ... a5 is a better move which I had actually
played before. I am afraid that I was totally
shocked to see that game on the database af­
terwards. My memory isn't what it used to be!
Perhaps it was because it occurred via a differ­
ent move order and was actually agreed drawn
after 16 lDe2 (Cvitan-Gallagher, Pontresina
2000). I had assumed the critical test to be 1 6
cxbS cxbS 1 7 1i'b3 attacking the pawn o n bS.
The natural move is 1 7...b4 but this runs into
18 ll:hls. Now exchanging on dS gives White a For some reason I had thought he was going
strong passed pawn but because of his 1 7th to play 1 8 tDaS when 1 8...c5! is a good reply.
move Black is no longer in a position to play 1 8 ...as
18 ...1i'd6 as this just runs into 1 9 lbc:4. After 1 7 I had just assumed, without any calculation,
1i'b3 Black does best to sacrifice a pawn with that I would be able to play 1 8 ...ll:hl4 here but
1 7...i.c6 1 8 tDxbS (1 8 llhls 1i'd6) 1 8 ...lbc:s 1 9 on closer inspection 1 9 i.xd4 exd4 20 l:txd4
i.xcS 1i'xc5 as his bishop pair and open lines tDdS 21 exdS i.xd4 22 d6 1i'f6 23 tDxb7 i.xc3
on the queenside provide adequate compensa­ 24 l:td1 persuaded me against this line. Reluc­
tion. tantly, I decided to sit tight and try and soak up
Interestingly enough, since I wrote that there the opponent's pressure. The text at least gets
did occur a game with 15 ... a5 where the strong rid of one of my weak pawns.
grandmaster Predrag Nikolic ground out a win 1 9 lLI3a4 axb4 20 axb4 lL\xcS 21 lL\xcS
in typical fashion (he once did the same to me l:tdS 22 l:txd8+ 'ifxd8 23 :as 'ilea 24 'ifa2
in this line): 16 l:tacl i.a6 (16 ... lbc:S? 1 7 cxbS l:tb8 2S 'ifxc4 lL\d7
cxbS 18 llhls tDxdS 1 9 i.xcS) 1 7 i.ft 1i'e6 The move I had been relying on to ease the
(1 7 .. .b4 1 8 tDa4 cS 1 9 tDb3 i.b7 20 i.g2 looks situation.
a bit better for White but 1 7 ...1ff8 may be a 26 lL\d3 'ifd8 27 .i.f1 'ife7 28 l:ta7 'ifd8
better move) 1 8 cxbS cxbS 1 9 1i'b3! 1i'xb3 20 Both players were extremely short of time
li:)xb3 b4 21 axb4 i.xft 22 ..t>xft axb4 23 ll:hls here. We were down to 30 seconds per move
with an unpleasant ending for Black, until move 40. Don't ask me why we played so
Pr.Nikolic-Stevic, Istanbul 2003. slowly.
So perhaps life is not that easy after 15 ... a5 29 �h2 hS 30 h4 lLib6 31 'ifa2 lL\cS 32
after all. What should Black play then? Well, it :as lL\d6?!

1 29
Pla y th e King 's Indian

A bluff which worked but 32...1i'e7 was It was only now that Renet spotted the
more sensible. variation 49 �e3 .i.h6+ 50 �d3 .i.£5 mate!!
33 tllc 5?! This brought on a fit of panic and he retreated
I was planning to meet 33 ltlxe5 with 33...c5 his king. He should still have played this as 50
but 34 :XeS .i.xe4 (34.. 1i'e8 35 .i.f4 lla8 36
. f4 (instead of 50 �d3) 50...llc3+ (50...exf4+ 51
�3) 35 lDc6 .i.xc6 36 llxc6 ltl£5 (36.. .llxb4 gxf4 .i.xf4+ 52 �xf4 .l:.£2+- 53 �e5 :Xfl 54 b6
37 'iVd2) 37 .i.g5 is obviously good for White. llb1 looks like a draw as well) 51 .i.d3 e4 52
33 . .-*.cS 34 :as 'ifc7 35 l:l.a7 'ifd8 36 'ifa5
. lbe7+ �ffi 53 �xe4 .i.g7 54 b6 llb3 55 .l:.c7!
.i.g4 37 'ifxd8+ l:l.xd8 38 l:l.c7 l:l.c8 39 l:l.a7 should be a draw.
l:l.b8 40 tlla6 l:l.d8 41 �g2 49 ...e4!
The rest is a massacre.
50 fxe4 icl+ 51 �d1 l:l.d2+ 52 �c1 .i.b3
53 l:l.b8+ �g7 54 l:l.d8 l:l.c2+ 55 �b1 l:l.a2
0-1

Perhaps this is not the greatest advert for the

system with 8 ... a6 but note the suggested im­
provements in the opening for Black.

Game 48
Tregubov-Gallagher
French uague 2002

I was quite pleased with myself for wriggling 1 d4 tllf6 2 c4 g6 3 tllfl .i.g7 4 gl 0-0 5
out into this roughly level endgame and was .i.g2 d6 6 0-0 tllbd7 7 tllcl e5 8 e4 a6 9
looking forward to splitting the point when my l:l.e1
team captain suddenly informed me that I had This move prevents 9...b5 as after 10 cxb5
to try to win! a.xb5 1 1 ltlxb5 .i.a6 the knight on b5 is not
41 ...tllb5 42 l:l.b7 pinned to the rook on f1 and so can just go
42 .i.xb5 cxb5 43 lDc7 .i.e2 44 llb7 .i.ffi is back to c3. However, Black now has a simplify­
level. ing manoeuvre which equalises the game.
42. . .tlld4 43 ixd4 9 ...exd4 1 0 tllxd4 tllg4!
43 .i.c4 .i.f3+ 44 �h2 ltle6! was the main
idea as after 45 .i.xe6 lld1 White gets mated
while 45 llb8 llxb8 46 ltlxb8 .i.xe4 leaves him
fighting for a draw.
43 ...:Xd4!?
The winning attempt. I certainly hadn't
imagined it would go quite so smoothly. Instead
43 ... exd4 44 .i.d3 .i.c8 45 lla7 .i.xa6 46 llxa6
llb8 is a draw.
44 tllb8 .i.e6 45 tllxc6 l:l.xe4
White has a dangerous pawn but, as we shall
see, the white king is not immune from attack.
46 fl l:l.e1 47 �2?!
Renet's attempt to centralise his king ends in
disaster. He should have just pushed his b­ 1 1 hl
pawn. Let's have a look at the alternatives:
47 . . .l:l.c1 48 b5 l:l.c2+ 49 �e1 ? a) 1 1 ..xg4 .i.xd4 1 2 ..e2 .i.g7 leads to a

1 30
 Th e Fia n c h e t t o Varia tion

roughly level game. If possible Black will look

to continue his development with ...lL!e5 and
... i.e6. Note that 1 2 �h6? would be a mistake
on account of 12 ... lL!e5! when White doesn't
have a good square for his queen, e.g. 1 3 1i'f4
lbd3, 1 3 1i'd1 �x£2+! 1 4 'it>xf2 lL!g4t wins a
pawn and 13 1i'e2 lL!g4! is good for Black.
b) 1 1 £3 lLJge5 12 b3 c5!? 13 lbde2 b5 gives
Black good play. Baburin-Robovic, liechten­
stein 1 993 continued 14 cxb5 axb5 1 5 i.e3 b4
16 lL!a4 lL!c4! 1 7 bxc4 �xa 1 1 8 lL!xc5 J:txa2 1 9
lbd3 �g7 20 lL!xb4 and with just one pawn
White doesn't have enough compensation for
the exchange.
c) 1 1 J:tfl lL!ge5 1 2 b3 lL!c5 1 3 h3 b5 14 Much better than 13 ...1i'xc4 14 i.ft with
cxb5 axb5 1 5 f4 lL!ed3 1 6 �e3 b4 1 7 lL!d5 �b7 advantage to White. In fact the insignificant­
led to a promising position for Black in Chi­ looking text (13 ... c6, that is) was quite instru­
burdanidze-Su.Polgar, World Championship mental in Black taking up 8 ... a6 again. The
match 1 995. Not surprisingly White can't get point is that in the similar position that can arise
away with playing 9 J:te1 and then two moves after 8 ... exd4 (instead of 8 ... a6) 9 lLlxd4 J:te8 10
later going back to fl . J:te1 lLJg4?! 11 h3 1i'f6 1 2 hxg4 1i'xd4 1 3 1i'xd4
d) 1 1 f4?! c 6 1 2 h3 1i'b6 1 3 lL!ce2 lLigf6 1 4 .txd4 1 4 lL!d5 is very annoying for Black -
1i'c2 J:te8 1 5 �e3 c5! 1 6 lL!b3 lL!xe4!! 1 7 �xe4 14 ... c6 is not playable as 1 5 lL!c7 forks the black
lLif6 1 8 lL!c3 lL!xe4 1 9 lLixe4 i.f5 20 i.f2 1i'c6 rooks. With the rook on f8 White cannot play
21 lL!bd2 i.xh3 (Black swaps one pin for an­ lL!c7 as after J:tb8 the knight on c7 is stuck.
other) 22 J:te3 J:te6 23 J:tae1 J:tae8 and White is 1 4 liJe7+
caught in a terrible bind, Wojtkiewicz-Vogt, 14 lL!e3 1i'xd1 15 J:txd1 lL!e5 16 J:txd6 lL!xg4
Altensteig 1 995. 1/2-'12Grabarczyk-Wojtkiewicz, Polish Team
1 1 ...1i'f6! 1 2 hxg4 Ch. 1 997 is not very troubling for Black.
The alternatives: 14 ...�h8 1 5 1i'xd4 �xd4 1 6 �e3!
a) 12 lLI£3 lLJge5 13 lL!d5 1i'd8 14 lLlxe5 This was a novelty from Tregubov as he had
lL!xe5 1 5 1i'e2 J:te8 is about level. 1 5 ... .i.e6 is played all his moves up to here instantly. Per­
also a good move. Black's ...lL!g4 trick has en­ haps he believes White has an endgame edge. I
abled him to exchange off a minor piece. This don't think so but it is cenainly the best try.
means that his space disadvantage is hardly a 1 6...�xe3 1 7 l:l.xe3 lt:\e5 1 8 lt:\xc8 l:l.axc8
disadvantage at all as he has enough squares for 1 9 b3
all his pieces.
b) 12 lL!f5 is another important variation but
Black has the strong reply 12 ... lL!xf2! (12 ... gxf5
13 hxg4 is better for White) 1 3 'it>xf2 and now
not 1 3 ...gxf5 1 4 lL!d5 1i'd4t 1 5 1i'xd4 i.xd4t 1 6
�fl which is promising for White, but
13. ..lL!b6! to stop White from playing lL!d5. The
knight on f5 is pinned and can be collected
later. The game Obukhov-Pugachev, USSR
1990 continued 14 .tf4 gxf5 1 5 e5 1i'd8! 1 6
exd6 cxd6 1 7 .txd6 lL!xc4 1 8 i.x£8 'ii'x£8 with
excellent attacking prospects for Black in return
for the exchange.
1 2...1i'xd4 1 3 tt:ld5 c6

13 1
Pla y th e King 's Indian

He still hadn't paused for thought. :e7+ 40 �d2 :xb3 41 ..id5! h3 42 �c2
1 9 . . .c5 % -%
I decided the most important thing was to I could have continued a little with 42 ... l:1a3
secure d4 for the knight in order to try and get 43 �b2 l:lxa4 44 �b3 l:la1 45 l:lxh3 but I don't
a good knight via bad bishop position. even think Black is better here, and this time I
Both 1 9...o!t:lxg4 20 l:ld3 l:lcd8 21 l:lad 1 and just needed a draw to secure an important vic­
19 ... f6 20 l:ld1 li:Jfl are a little uncomfortable tory for my team.
for Black, though still tenable.
20 :d1 :cdB 21 g5 Game 49
21 f4 li:Jc6 22 eS li:Jd4 23 exd6 .:.Xd6 24 'it£2
Rogozenko-Gallagher
b6 is fine for Black.
21 . . .lL!c6 Bundesliga 1999
This allows White to activate his bishop. It
would have been safer to play 21 ...b6 before 1 d4 lL!f6 2 c4 g6 3 g3 ..ig7 4 ..ig2 0-0 5
manoeuvring the knight into d4. lL!c3 d6 6 lL!f3 lL!bd7 7 0-0 e5 8 e4 a6 9
22 e5! lLld4 23 ..bb7 dxe5 24 �g2?? 'irc217
An unbelievable error. The best way to deal
with the threatened ...lbe2+ was 24 �ft when I
was planning 24...l:lfe8 25 .i..xa6 e4 with some
compensation for the pawn. The main point is
that after 26 .i..b7 Black wins the exchange with
26 ... li:Jc2 although this time after 27 .:.Xd8
li:Jxe3t 28 �e2! l:lxd8 29 �xe3 there is no dan­
ger of White losing the game.
24......lL!c2!

This became the main line for a while as it

more or less prevents Black from playing 9... b5
(see next note for explanation). Black must,
therefore, put plan B into operation, i.e. ex­
change in the centre and aim for a quick ..cS
followed by ...bS. Let's take a quick look at
White's other 9th move possibilities which
don't merit a game of their own.
a) 9 dxeS dxeS 10 1i'c2. This is acrually
Black wins the exchange, as 25 l:led3, which played quite often but doesn't have much inde­
would have been the response if he had played pendent significance as play will be very similar
24 �ft , is no longer possible. It is still not that to Game 47. For example, to ..c6 1 1 l:ld1 1i'e7
.

easy for Black to win but I certainly missed a 12 h3 bS 1 3 .i..e3 .i..b7 1 4 a3 l:lfc8 1 S li:Jd5 aS
few chances. The remaining moves were: 1 6 lbe2 li:JcS was the acrual move order of the
25 :ee1 lL!xe1 + 26 :xe1 f6 27 gxf6 :xf6 Cvitan-Gallagher game considered in the notes
28 :xe5 :d2 29 ..if3 :t5 30 :eB+ �g7 31 to move 1 5.
g4 :t7 32 a4 :b2 33 :ca h5! 34 gxh5 b) 9 dS can be met by 9 ... b5 although the
gxh5 35 :xc5 h4 36 :g5+ �6 37 :h5 position after to cxbS axbS 1 1 b4 is less trust­
:g7+ (37...l:lxb3! 38 l:lxh4 l:lg7+ looks win­ worthy than the similar one we saw in the notes
ning) 38 �1 :b1 + (38...llxb3) 39 �e2 to Game 47. The difference is that Black can-

1 32
 The Fian c h e t t o Varia tion

not play ...c6. In fact after White has blocked In Stohl-Kindermann, Bundesliga 1999 he
the centre with dS there is no reason why Black played 12 .i.e3 and after 12... b5 13 cxbS axbS
shouldn't fall back on the trusty old King's 1 4 lbd4 lDeS 1 5 f4 lL!fg4 16 .i.c1 cS 17 lL!dxbS
Indian plan of playing for ... £5. lL!c6 1 8 h3 Black should have played 1 8 ...lbd4
Hiibner-Gallagher, Baden 1999 continued 1 9 lL!xd4 .i.xd4+ 20 �h 1 lLif6 with good com­
9...lbe8 10 .i.d2 cS (this makes it difficult for pensation for the pawn.
White to create active play on the queenside) 1 1 1 2. . .&51
a4 aS (sealing up the queenside is logical as
Black is inferior on this side of the board; now
the action will take place on the kingside) 12 h4
lbdf6 (it's too early to play 12 ... £5 as White has
the reply 13 o!Dgs, threatening iDe6, so Black
brings some more pieces to the kingside) 1 3
..e2 lLihS 1 4 iDe1 iDef6 1 5 lL!c2 tbg4 (15 ... £5
would have led to a more complicated game) 16
lDe3 .i.h6 1 7 lL!xg4 .i.xd2 1 8 ..xd2 .i.xg4 19
ltJbS 1i'e7 with an equal game.
c) 9 l:tb1 bS 10 cxbS axbS 1 1 b4 (1 1 lL!xbS
i.a6 12 a4 c6 1 3 lL!xd6 .i.xfl is unclear but
Black can also consider 1 1 ....i.b7) 1 1 ...c6 (cS is
not available to Black with the white rook on
b1 so she must adopt the more solid plan) 12 White's last move prevented Black from
'l'c2 (the boring 1 2 dxeS comes into considera­ playing ...bS so he had to find another idea. The
tion) 12 ... exd4 13 lL!xd4 lL!e5! 1 4 h3 (14 f4 lL!c4 first thing is to secure the cS-square for his
15 lL!xc6? fails to 1 5 ...1i'b6+) 14 ....i.d7 1 5 l:td1 knight.
'l'c8 16 �h2 hS with active play for Black, 1 3 b3 ll:X:5 1 4 .b3 b6 1 5 .l:lad1 h5!?
Pr.Nikolic-J.Polgar, Monte Carlo (rapid) 1 995. Advancing the h-pawn to soften up the
9.. exd4
. white king position is not an uncommon way
9 ... b5?! is not so promising as the position for Black to treat such positions. 1 S....i.b7 and
after 10 dxeS dxeS 1 1 l:td1 ! (better than 1 1 1 5...1i'e7 are more solid alternatives.
lL!xbS) is uncomfortable for Black. He would 1 6 tl:ld4 .i.d7 1 7 .l:[fe 1 h4!?
like to play 1 1 ...1i'e7 but this is strongly met by When playing this move I had obviously
12 lbd5. seen White's central breakthrough but hadn't
1 0 tl:lxd4 .:es 1 1 tl:lde2 appreciated the sort of gymnastics I would have
It's not completely clear if Black is threaten­ to perform to stay in the game.
ing to take on e4, e.g. 1 1 h3 lL!xe4 1 2 lbxe4 1 8 e5 tl:lg4!
i.xd4 1 3 .i.gS gives White play for the pawn My original idea had been to play 18 ...lL!h5
but this tactic is usually prevented by the natural but I couldn't find a satisfactory defence to the
move 1 1 l:tdt. Rogozenko prefers to beat a powerful move 1 9 .i.dS!. The threat is 20 1i'xg6
prophylactic retreat with his knight. He knows and 1 9.. �ffi 20 e6! didn't inspire confidence.
.

that Black wants to play ...cS and ...bS so he Still, I wasn't too unhappy with the text as
gets his knight out of the way of cS and plays a4 White can't play 1 9 exd6 on account of
on his next move to prevent ... bS. Tigran Petro­ 1 9...l:txe 1+ and 20... .i.xd4.
sian, World Champion 1963-69, would have 1 9 f4 hxg3 20 hxg3
been proud of him. His approach to chess was This game was played in the German Bun­
to stop the opponent's aggressive ideas before desliga and at this point I thought for so long
they even thought of them! that my captain started to give me a lot of
For 1 1 l:td 1 see Game 50. funny looks. I eventually moved when I had
1 1 ....:bs 1 2 a4 less than a quarter of an hour left to reach
White continues the prophylactic approach. move 40. I had assumed that I would be able to

1 33
Pla y th e King 's Indian

give up an exchange for an attack but now 24 llxe1 lLif2+ 25 Wxf2?

realised this wasn't possible. I was going to I think my opponent was in a state of shock
have to give up a lot more. by what had just happened on the chessboard
20 ...dxe5 21 lL!c6 exf4!! and he played this losing move almost instantly.
He probably just saw that 25 'ifi>h2 lbg4+ 26
'ifi>h 1 lbf2+ was a draw by perpetual check and
played the text by a process of elimination.
Remember I hardly had any time left so he
wanted to keep the game going at all costs. In
fact White can avoid the draw by 25 'iti>g1

Yes, a whole CJUeen! This move made it into

grandmaster John Emms' book on the most
amazing moves of all time. The alternatives are
hopeless.
22 lL!xd8
What now?
22 . . . id4+! 23 �h1 as with no black rook on the e-ftle 25 ...lbg4+
The beautiful point is that 23 l:.xd4? loses to 26 �ft lbh2+ is no longer mate. True, Black
23...l:.xe1+ 24 .ift l:.xfl+l 25 'ifi>xfl llk3+ 26 still has a vicious attack after 27 'ifi>e2 ..ig4+ 28
'it>g1 lbxc2 forking the rook on d4 and the <t>d2 l:.xd8 but it is hard to believe that it can
bishop on a3. After 27 .ixc5 lbxd4 28 .ixd4 succeed with only one piece for the CJUeen. This
l:txd8 Black wins the ending. The other nice variation was not what I had in mind when I
point is that the white king is forced into the sacrificed the ctueen. On 25 'iti>g1 my idea was
comer as 23 �fl lbh2! is checkmate. to play 25 ...llk4+. Now 26 'ifi>f1 lbxg3 is mate
and 26 <t>h 1 lbf2+ the same draw by perperual
check we have already seen. Therefore if White
wants to play for a win he must try 26 'it>h2
fxg3+ 27 �h1 lbf2+ 28 �g1 with the difference
being that after 28...llk4+ 29 �fl Black can no
longer play 29...lbg3 mate as he now has a
pawn on this S<:Juare!. All is not lost though.
Instead of 28 ... llk4+ Black should play
28 ... lbfd3+ 29 �fl lbxe1 30 �xe1 l:.xd8 when
he has a rook, three pawns and an ongoing
attack for the <:JUeen. Rogozenko annotated this
game in famous periodical Chess In.formator and
claimed that White has the advantage here (al­
though he stopped short of saying that White
23 .. Jlxe 1 +! was winning). I have my own ideas. Perhaps
Black must exchange rooks before giving the next time we play the game can start from this
knight check. After 23 ... lbf2+ 24 'W'x£2! White is position!
winning. 25 ...ixf2 26 :n fxg3 27 lL!c6 lieS

1 34
 The Fia n c h e t to Varia tion

Game 50
Hubner-Polzin
Bundesliga 2003

1 d4 lDf6 2 c4 g6 3 g3 �g7 4 �g2 0-0 5

lDc3 d6 6 lDf3 lDbd7 7 0-0 e5 8 e4 a6 9
'ifc2 exd4 1 0 liJxd4 :as 1 1 .l:d1 .l:bB 1 2 a4
Hubner decides to prevent ... b5 and as a
good classical player he probably thought that
...c5 would be no good if Black can't follow up
with ...b5.
12 h3 has been played more often and we
So the dust has more or less settled and can see from the course of the main game how
Black has emerged with three pawns for a useful it is to have the g4-square protected.
piece. Normally this would mean approximate However, Black does not seem to have too
equality but not here. Despite the fact that many problems after 1 2 h3 c5, e.g.
queens have been exchanged the white king is
still in grave danger of getting checkmated. The
bishop on f2 supported by the pawn on g3 is a
particularly potent combination. All Black
needs to do now is to get a rook to the h-fi.le.
White's next move was motivated by the fact
that he wanted to stop ...�g7 and ...l:h8 but he
forgot that Black has other ways to get his rook
to the h-file. The last chance, albeit a very small

one, was offered by 28 ..ixc5.

28 �b2? lDd3 29 �a1 .l:e6! 30 lDb8 .l:e5
The cleanest route to victory was offered by
30...g5! 31 tt:lxd7 l:h6+ 32 ..ih3 l:xh3+ 33 'iii>g2
tt:lf4+ 34 �£3 g2+ 35 �xf2 l:£3+ 36 �x£3
gxf1'ii'+ but by this stage I had just over a min­ a) 1 3 tt::l£3 b5 14 l:xd6 (14 cxb5 axb5 15
ute for ten moves. Any win would do! l:xd6 b4 is similar) 14...b4 (if the knight now
31 �f3 �h3 32 .l:d1 .l:e1 + 33 .l:xe1 lDxe1 moves then Black plays 1 5...tt:lxe4) 1 5 e5 bxc3
34 �e4 g2+ 35 �xg2 �xg2+ 36 �h2 �d4! 1 6 exf6 cxb2 1 7 .i.xb2 l:xb2 18 1i'xb2 ..ixf6,
37 �g3 �b7 38 lDd7 f5 39 �4 �7 40 C.Hansen-Van Wely, Groningen 1995, and
lDe5+ �xe5+ now 1 9 'ilfxf6 'ilfxf6 20 l:xf6 tt:lxf6 21 l:b1
The time scramble is over. Black should leads to a level endgame.
have done better than an ending with two extra b) 1 3 tt:lde2 tt:le5 14 b3 b5 and now we have
pawns but it is still a trivial win. two games:

41 �xeS lDf3+ 42 �4 lDd2 43 lDb5 l0xb3 b1) 1 5 f4 tt:lc6! 1 6 e5 (normally Black can't
44 �e5 c6! retreat to c6 after White plays f4 but Black has
The last tactical trick of the game. Now after spotted a tactical resource which saves the
45 tt:ld6+ �e6 46 tt:lxb7 gS+! 47 �xg5 �e5 piece) 1 6... ..if5! 17 1i'd2 (1 7 1i'b2 tt:ld4! is also
nothing can stop the black f-pawn. fme for Black) 1 7...tt:ld4 18 exf6 tt:lxe2+ 1 9
45 lDc3 lDc5 46 �b8 �a6 47 �a7 lDd7 0-1 tt:lxe2 1i'xf6 20 ..ia3 b4! (20 ..ia3 was the only
White called it a day. Not only is he three move to save the rook on a 1 but now Black
pawns down but his bishop is boxed in the regains the piece with an equal game) 21 ..ixb4
comer. l:xb4 22 l:acl l:bb8 23 g4 ..ie4 24 tt:lc3 ..ixg2

1 35
Pla y t h e King 's Indian

25 �xg2 ..d4 26 l:.c2 ..xd2+- '12-1/2 Filippov­ after 20 l:.d3?! Black can safely take the pawn,
Khalifman, Ubeda 1 997. e.g. 20 .....xb3 21 lL!b5 � (21 .....xc4 22
b2) In view of the above, in a game between lL!xd6) 22 lLlxd6 lL!xe4 23 lLlxc8 l:.xe8 24 .ixe4
the Indian stars, Harikrishna-Sasikiran, ..txa 1 with advantage to Black.
Lausanne 2001 White tried 1 5 ..te3. lbis con­ 20 :xe4 :xe4 21 �xe4 �xc3 22 l:[d1 'ii'e7
tains a nasty trap as 1 5 ... b4? is strongly met by
16 .ixc5! bxc3 17 ..txd6!. Young prodigies see
this son of thing so Black preferred 1 5 ... 'ti'c7
and after the further moves 1 6 cxb5 axb5 1 7
lLlf4 .ie6 1 8 l:.act b4 1 9 lLlxe6 l:.xe6
(19 ... fxe6!?) 20 lL!a4 the game was unclear.
1 2 c5!?
•.•

1 2...a5 to take control of the c5 and b4-

squares is another approach but Black may not
have liked this because of the reply 1 3 lL!db5.
13 tt:lde2 tt:le5 14 b3
Hubner must have been tempted by 14 f4
but eventually senled on the solid choice. After
14 ...lLlxc4 1 5 b3 (1 5 e5 ..tfS 16 'ifb3 lLla5 is not
an improvement for White) 1 5 ...lL!a5 16 e5 .ifS Black's choice has led to an endgame with a
1 7 ..a2 lL!d7 18 l:.xd6 White has a positional token advantage to White, but one where he
advantage but lacks co-ordination. Unclear is shouldn't really have any difficulty holding the
my verdict. draw.
14 ...�g4! 1 5 �e3 23 'ii'f3 �b4 24 �d5 �c5 25 �g2 �g7 26
1 5 f4 lL!£3+ 1 6 ..tx£3 ..tx£3 1 7 e5 ..c8! 1 8 'ii'c3+ 'ii'f6 27 'ii'xf6+ �xf6 28 :e1 b6 29
l:.xd6? 'ti'h3 and rather embarrassingly the h4 h6 30 f4 h5 31 �3 a5 32 :e2 :c8 33
knight on e2 has no square. l:[e1 l:[b8 34 g4 hxg4+ 35 �xg4 l:[h8 36
1 5 ...tt:lf3+ 1 6 �h1 tt:ld4! :n �g7 37 f5 gxf5+ 38 :xt5 f6 39 l:[f4
:e8 40 l:[e4 :e5 41 :xe5 % -%
As is often the case in these type of posi­
tions Black's initiative proved strong enough to
overcome the positional defects.

Game 51
Galliamova-Renet
Koszalin 1997

1 d4 tt:lf6 2 c4 g6 3 tt:lf3 �g7 4 g3 0-0 5

�g2 d6 6 0-0 tt:lbd7 7 tt:lc3 e5 8 h3
Ths is a common move order. Most of the
time White will foUow up with e4 and we have
By exploiting the pin Black is able to steer transposed into lines already considered. This
the game towards a drawish position. game deals with lines where White refrains
1 7 �xd4 &om playing e4. Why would he (or she in this
17 lL!xd4!? ..txd 1 1 8 l:.xd 1 cxd4 1 9 ..txd4 is case) want to do that? WeU, there are a couple
also about equal and may be a better winning of main reasons. Firstly, the bishop on g2 may
chance. now exert more influence on the game as its
1 7 ...cxd4 1 8 :xd4 �xe2 1 9 'ii'xe2 tt:lxe4 diagonal is not blocked by a pawn on c4. Sec­
1 9 ...'ifb6 looks like a serious alternative as ondly, White may be able to use the e4-square

136
 The Fia n c h e t to Va ria tion

for his, or her, pieces. Usually this means a b) 9 ..i.e3 exd4 to �d4 .l:te8 1 1 ..d2 (oth­
knight using the S<:JUare as a stepping stone erwise Black would be sorely tempted to sacri­
towards d6. There are also drawbacks to White fice an exchange on e3) 1 1...�5 (another
not playing e4, the principal one being that it is problem with not playing e4 is that Black may
easier for Black to advance his pawns. be able to occupy this square with his pieces; to
avoid this White now gives up his bishop for a
knight) 12 ..i.gS ..d7 13 ..i.xf6 ..i.x£6 14 .l:tad1
..i.g7 1 5 e3, Gelfand-J.Polgar, Dortmund 1 996.
The game is about level. White has more space
but could still end up regretting having parted
with the bishop pair.
c) 9 b3 is a solid move. Black could just reply
9 ...l:.e8 but Baburin-Gallagher, .Mind Sports
Olympiad 1 999 went a different way: 9 ... exd4
to .!i)xd4 l:.b8 1 1 ..i.b2 l:.e8 1 2 l:.e1 �5 1 3 e4
(I was planning to meet 1 3 f4 with 1 3. .. c5)
1 3. .. ..i.d7 (the problem with 1 3...c5 1 4 �2 bS
is that after 1 5 f4 Black can't play 1 5...�6 on
account of 1 6 eS!; there are no miraculous re­
8 . . a6
. sources this time) 1 4 f4 �6 1 5 �2 bS 16
There is no need for Black to react any dif­ cxbS axbS 1 7 b4 (With my queenside play
ferently. Remember that most of the time halted I became concerned about drifting into a
White is just going to play 9 e4. However, if passive position. What I really wanted to do
you are concerned about the dull games that was get a knight to c4 but there is no route.
can, but not necessarily, arise after 9 dxeS then Then I found one.) 1 7....!i)hS 1 8 �h2 �5! (If
you can play 8 ... exd4 9 .!i)xd4 and then either White now takes the piece then Black will get
9.. .l:te8, when White has nothing better than to attacking chances against his king, e.g. 1 9 fxeS
e4 (Games 43-46), or 9 ....!i)b6 which i s covered ..i.xeS 20 l:.e3 �g3! 21 l:.xg3 hS! is very good
in Game 52. for Black. White does better to play 20 'iti>g1 but
9 dxe5 after 20 ....!i)xg3 21 .!i)d4 cS Black has plenty of
Such an exchange is usually considered play, not only against the white king but also on
wimpish and played by those who are angling the long dark diagonal. Baburin didn't like the
for a 'Iuick draw. Here White can claim that he look of all this and played ...) 1 9 �3 (... as I felt
is trying to prove that Black's ... a6 was prema­ certain he would) but after 1 9...�4 20 .!i)xc4
ture. The point is that Black will not be able to bxc4 21 a3 cS! Black had an active game. The
get by without playing ...c6 and this will create a idea is to meet 22 'ii'xd6 with 22...cxb4 23 axb4
hole on the b6-s<:juare. White will then play cS ..i.xh3! and this fascinating variation actually
to fix the <:JUeenside pawn structure thereby occurred in the game. Unfortunately, I eventu­
creating outposts for his knights (or other ally lost through over-excitement.
pieces) on b6 and d6. In my view this is not too 9 ... dxe5 1 0 iLe3 'iVe7
serious. True, Black is likely to experience some Black tries to hold up cS and there is no
discomfort on the queenside but this is com­ need, just yet, to worry about 1 1 .!l)ds as this is
pensated for by the fact that White's exchange weU met by 1 t ....!i)xdS 12 exdS e4.
on eS gives the black pieces plenty of freedom. 1 1 'iVb3
Sometimes this freedom can be translated into Here are a couple of other examples to show
a kingside attack while on other occasions it just you how play can develop:
leads to mass exchanges and an early peace a) 1 1 ..d2 c6 1 2 l:.ad1 �5 1 3 ..i.gS aS 14
treaty. Let's take a look at the White alterna­ ...e3 l:.e8 15 .!i)clz ...£8 1 6 g4 .!i)fd7 1 7 .!i)de4 f5
tives: 18 gxfS gxfS 1 9 .!i)d6 l:.e6 with good play for
a) 9 e4 transposes to Game 47. Black, Brunner-Gabriel, Ziirich 1 995.

137
Pla y th e King 's Indian

b) 1 1 1i'ct c6 12 l:td1 it:)hS!? 13 g4 it:)f4! Black missed a chance to win a pawn. He

(there is no question of retreating once it:)hS should have played 20...it:)xc5! 21 :XeS �xd4
has been played) 14 �xf4 exf4 1 5 1i'xf4 �5 1 6 22 l:txd4 it:)e6 with a massive fork. White has
e4 �e6 1 7 e S f6 1 8 exf6 �xf6 with good com­ nothing better than 23 l:tcc4 though after
pensation for the pawn, Vaganian-Gallagher, 23 ...it:)xg5 24 l:td7 Black can just defend his 2nd
Bundesliga 1998. rank with 24. . .llf7.
1 1 ... c6 1 2 . . .'�h8
Black plans to move his knight and push his
f-pawn. That is how he creates counterplay in
this line.
1 3 lbd2 �8 1 4 lbde4 f5 1 5 ..ig5
As mentioned in the notes above White
must get his bishop out of the way of f4.
1 5 ... �f6 1 6 ..ixf6 lt:lxf6 1 7 lbd6 e4
So the white bishop still gets locked out of
the game but this time not by his own pawn on
e4 but by a black one.
1 8 e3 l:lb8 1 9 l:lfd1 ..ie6 20 1Wa3 h5

1 2 c5
It's funny how seldom White has played this
move as if Black takes the pawn, 12 ... it:)xc5, he
gets caught in a nasty pin by 1 3 1i'a3 or 1 3
1i'b4. More often White has played 1 2 l:tfd 1
with a couple of examples:
a) 1 2...l:te8 1 3 it:)gS �5 14 �xeS 1i'xc5 1 5
�e4 it:)xe4 1 6 it:)xe4 'iVaS! (Black keeps e7 free
for his rook as from there it defends the sensi­
tive points b7 and fl; in a previous game Black
had got into trouble after 16 ...1i'e7 t 7 it:)d6) 1 7
it:)d6 l:te7 1 8 .:tact 1i'c7 'h-'12 Ruck-Gallagher, Black stands well. White has one trump in
Charleville 2000. White has the choice between this position - the knight on d6. In return Black
taking on c8, which kills the game, or letting the has two good bishops, his own outpost on dS
bishop out to e6 when it will in no way be an and attacking chances against the white king.
inferior piece to the white knight on d6. The players eventually agreed to a draw in an
b) 12 ... �8 is perhaps more logical than my unclear situation, probably because of looming
choice as it frees the f-pawn and envisages a time trouble.
nice home for the knight on c7. From there it 21 h4 lbd5 22 lt:le2 b6 23 lt:ld4 bxc5 24
can go to e6, or perhaps dS after White has 1Wxc5 ..id7 25 lt:lb3 ..ixb2 26 l:lab1 .if& 27
played cS. Ribli-Beliavsky, Hungary 1 998 con­ lt:la5 lt:lc3 28 l:lxb8 l:lxb8 29 "fia7 l:lf8 30
tinued 13 cS �7 14 .:tact �h8 1 5 it:)e4 f5 16 l:ld2 lbd5 Yz - Yz
i..gS (it is much too dangerous for White to
play 16 it:)d6 on account of 16... f4; the bishop Game 52
must get out of the way of the f-pawn) 16 ...1i'e6
Vaganian-Volokitin
17 it:)d6 e4 1 8 it:)xc8 (White would prefer to not
make this exchange but after 1 8 it:)d4 �xd4! 1 9 European Ch., lstanbu/ 2003
l:txd4 1i'xb3 20 axb3 �6 Black wins material)
1 8...l:taxc8 1 9 it:)d4 1i'xb3 20 axb3 and now 1 lt:lf3 lt:lf6 2 c4 g6 3 d4 ..ig7 4 g3 0-0 5

1 38
 The Fia n c h e t t o Variation

�g2 d6 6 tt:lc3 tt:lbd7 7 0-0 e5 8 h3 exd4 9 been better to just play 21 ...'ii'e5) 22 �£3 'ii'fS
tt:lxd4 tt:lb6!? 23 1i'd2 1i'e6 24 l:fd 1 l:ad8 25 �xdS �xdS 26
1i'xd5 'ifxe2 27 l:e1 and now in Illescas­
Moreno Ruiz, Burgos 2003 Black played
27...tbe5? and lost the ending. But 27...'ifb5! is
much better. Why exchange queens when you
have the worse pawn strucrure and the oppo­
nent's king is more exposed? After 27 ...'ifb5 the
chances look about equal.
19 ...tt:lxc5 20 l:txc5 �f8 21 l:tc2 c5 22 tt:le2
1i'd7 23 tt:lf4 l:tad8 24 l:td2 d4! 25 �xb7
1i'xb7 26 exd4 'itb4!

An interesting idea to try and punish White

for delaying e4 by forcing through a quick ... dS.
10 b3 d5
10...c5 has been played more often although
White probably retains a small edge after 1 1
lt:k2 �e6 1 2 �b2 dS 1 3 itJxdS lDfxdS 1 4
�xg7 'Ot>xg7 1 5 cxdS �xdS 1 6 e4. I n such mid­
dlegame positions the advantage is often with
the side who retains the fianchettoed bishop.
His king is safer!
1 1 �a3 l:te8 1 2 c5 tt:lbd7 1 3 c6 tt:le5 14 27 l:te2
cxb7 �xb7 1 5 l:tc1 a6 16 tLla4 tt:le4 1 7 27 dS �h6 28 l:e1 �xf4 29 l:xe8t l:xe8 30
tt:lc5 tt:lxc5 1 8 �xc5 tt:ld7 gxf4 'ifxf4 31 d6 l:d8 should be a draw because
of the exposed position of the white king.
The position is also drawn after the text and
the remaining moves were 27...1i'xd4 28 'ifxd4
cxd4 29 l:xe8 :XeS 30 l:d 1 gS 31 lDhS l:e2 32
l:xd4 l:xa2 33 l:d8 h6 34 g4 l:b2 35 l:b8 aS 36
l:b5 �b4 % - % .

This appears to be a good and solid way for

Black to meet 8 h3.

Early Deviations
Now we shall look at two other variations for
White. Game 53 deals with 7 'ifc2 and Game
54 with an early b3 from White.
19 e3
The last moves have been all about the bat­ Game 53
tle for the cS-square. In a game not long after
Loginov-Ryskin
this one White preferred to keep his bishop and
played 19 .i.a3. Play continued 1 9...c5 20 lfr2 Wisla 1992
'ii'gS 21 h4 'iVhS (luring the bishop to f3 so he
can hit it later with . tDcs but it may well have
.. 1 tt:lf3 d6 2 d4 tt:lf6 3 c4 g6 4 g3 �g7 5

1 39
Pla y the King 's Indian

.ig2 0-0 6 0-0 ll'lbd7 7 1i'c2 precaution to take before embarking on the
This queen move is the main alternative to 7 queenside advance.
lDc3. White plans a quick l:td 1 to create pres­ Sec also the suggested improvement in
sure on the d-file. This variation was initially Gamc 46.
pioneered by the Ukrainian grandmaster Oleg 1 3 f3
Romanishin. White reinforces his e-pawn. An alternative
7 ...e5 8 l:l.d1 exd4 was to play 1 3 ..ib2 when Black can play
The main line runs 8 ...1i'e7 9 lbc3 c6 but 1 3. .. lt'le5 or the immediate 1 3...c5 1 4 lbde2 bS
that would force Black to play another type of 1 5 cxbS axbS 1 6 ll'lf4 ..ib7 (so as to be able to
game to the one I've been recommending keep taking white knights that arrive on dS) 1 7
throughout this chapter. There is no reason ll'lxbS ..ixc4 1 8 ..ixe4 1i'xc4 1 9 1i'xe4 lt'lxe4 20
why Black should not follow the usual recipe, ..ixg7 l:txbS! (20...'�xg7 21 ll'lxd6) 21 ..ib2 c4
i.e. exchange on d4, play ...l:te8, ...a6 and ... l:tb8 when White has a strong bishop on b2 but
and then try and force through ... cS and ... bS. Black has pressure against the queenside pawns.
9 ll'lxd4 l:l.e8 1 0 ll'lc3 a6 1 1 b3 A draw is the most likely outcome.
After 1 1 h3 l:tb8 12 e4 we have transposed Before playing a move such as 1 2 ...1i'e7
to the 1 2th move notes in Game 50. Black would also have had to take the reply 1 3
1 1 .. .l:l.b8 1 2 e4 lbd s into account. I n certain positions Black
When White plays b3 in these fianchetto would avoid 1i'e7 because of this move. Here,
lines it does not necessarily mean that he is Black seems to be doing all right after 1 3 lbds
going to follow up with ..ib2. Most White play­ ll'lxdS 1 4 cxdS lbcS 1 5 ..ib2 ..id7 (so that b4
ers believe that the bishop is more actively can be met by ..ia4; 1 5 ...lt'lxe4? 1 6 l:tet would
placed on e3. The move b3 is played with the have left Black stuck in a fatal pin). Black will
primary objective of protecting the pawn on c4. seek counterplay by either ...c6 or ... fS at an
1 2 ..ib2 is still perfectly playable in this posi­ appropriate moment.
tion. After 1 2. ..lt�e5 (the plan of ... cS and ... bS is 1 3 ...ll'le5 1 4 .ie3
not so effective when White has not played e4 The alternative 1 4 f4 lbc6 did not concern
as this pawn is not a target back on e2) there Black. The white position is very loose. After
are a couple of Van Wely games: the text Black is ready for the customary queen­
a) 1 3 l:td2 hS 1 4 l:tft h4 (as in Rogozcnko­ side offensive.
Gallagher , Black uses his h-pawn to pound the 14 . . . c5 1 5 ll'lde2 b5 1 6 cxb5 axb5
position around the white king) 1 5 lbds ll'lfd7
(rather than exchanging on dS Black plans to
drive the white knight back by playing ...c6) 1 6
e4 hxg3 1 7 hxg3 lt'lg4 1 8 l:te 1 c6 1 9 lt'le3 lbdeS
with quite a promising game for Black, Piker­
Van Wely, Dutch Ch. 1 991 .
b) 13 cS is the most critical test of ... lt'les. Of
course Black can't play the capture 1 3 ... dxc5 on
account of 14 lbc6 but 13 ... d5 1 4 e4 c6 1 5 h3
1i'c7 1 6 exdS cxdS 17 b4 lDc6 t 8 1ib3 ..ie6 led
to an unclear game in Gross-Van Wely,
Bundesliga 1997.
1 2...1i'e7!?
In Baburin-Van Wely, Leukerbad 1992
Black played the traditional 12 ...c5 13 lbdc2 bS 1 7 1i'd2?
but after 14 cxbS axbS 1 5 ..if4 lt'les 1 6 ..ixeS! This allows Black to take over the initiative.
l:txeS 17 f4 l:te7 1 8 eS found himself in trouble Although the pawn on d6 is attacked White
because his queen was badly placed opposite won't be able to find the time to take it. 17 ll'lf4
the black rook. The text, therefore, is a sensible was better. This would avoid the tactical prob-

1 40
 The Fia n c h e t to Varia tion

!ems he ran into by having too many pieces on

the e-file and also threaten to play a knight to Game 54
d5. 17 ... �e6 or 1 7.....tb7 (18 lili.b5 ..txe4 1 9
Villamayor-Gallagher
fxe4 l:.xb5) are reasonable replies while a case
could also be made out for 1 7 ... b4 1 8 lbcd5 Calcutta 200 1
tt'lxd5 1 9 tt'lxd5 'ii'a7.
1 7 . . . b4 1 8 lba4 .i.a6?! 1 ll'lf3 d6 2 d4 ll'lf6 3 g3 g6 4 .i.g2 .i.g7 5
This is a strong move which prevents White b3
capturing the d-pawn as the white knight on e2
would then be lost, but 1 8...ltJxf3+! 19 ..txf3
tt'lxe4 is more powerful. I don't see how White
can obtain a playable game, e.g. 20 ..txe4 'ii'xe4
21 �f2 .i.b7 and White can resign. 20 'ii'd3
�£5! 21 ..txe4 ..txe4 22 1i'd2 �c6 is equally
hopeless while 20 'ilfct .i.xal 21 'ilfxal tt'lxg3
22 tt'lxg3 'ilfxe3+- leaves White with a material
deficit and an exposed king. Black didn't miss
this idea when presented with a second chance
a few moves later.
1 9 %1ac1 .i.b5 20 llc2 .i.d3 21 llcc1 ll'lxf3+!
22 .i.xf3 .i.xe4 23 .i.xe4 'ii'xe4 24 .i.f4
'ii'xe2 25 'ii'xe2 :Xe2 26 .i.xd6 llbe8 27
lllxc5 lllg4 The double fianchetto can be an irritant for
the King's Indian player who is hoping for an
active game. In my younger years I had quite a
few reversals against this line as I was striving
for too much. My results improved dramatically
once I decided to settle for equality.
After 5 0-0 d6 6 b3 Black can play 6 ... c5 with
similar play to the game or he can also play
6 ... e5 7 dxe5 dxe5!, e.g.
a) 8 ..tb2 (not 8 tt'lxe5? tt'lg4) e4 and now:
at) 9 tt'le5 1i'e7 to 'ii'c t �£5 (t o ...tt'lbd7 is
also good) 1 1 lbc4 1i'e6 1 2 lbc3 tt'lc6 and Black
was already better in D'Amore-Gallagher, Is­
tanbul Olympiad 2000.
a2) White can maintain the balance by ex­
Black's little combination has led to a win­ changing queens, e.g. 9 'ii'xd8 l:.xd8 10 tt'lg5
ning endgame. Although White has won back �£5 1 1 g4 ..txg4 1 2 tt'lxe4 tt'lxe4 13 �xg7
his pawns his king is now in serious trouble. �xg7 1 4 ..txe4 lbc6 1 5 tt'lc3 with a level game.
One doesn't always need queens to conduct a b) 8 ..ta3 l:.e8 9 tt'lc3 tt'lc6 1 0 tt'lg5 �£5 1 1
mating attack. tt'lge4 tt'lxe4 1 2 tt'lxe4 'ifxd 1 1 3 l:lfxd 1 tt'ld4 1 4
28 h4 llxa2 29 ltJe4 .i.h6 30 llc4 .i.e3+ 31 l:ld2 l:lad8 1 5 �ft ! tt'lb5?! (1 5.....txe4!? 16
�1 llxe4? �xe4 tt'lb5 1 7 l:.xd8 l:lxd8 1 8 ..tb2 c6 19 c4
This wins easily but 3t...tt'lh2+ 32 �el ..tf2 tt'lc7 and even though White has the bishop
mate was slightly more accurate! pair he can claim no advantage) 16 l:lxd8 l:lxd8
32 lieS+ �g7 33 .i.f8+ �6 34 lld6+ lle6 1 7 ..tb2?! (17 �e7! gives White a clear advan­
35 llxe6+ fxe6 36 .i.xb4 ll'lh2+ 37 �e1 tage after 1 7...l:ld4 1 8 tt'lf6t 1i'h8 19 �xb7
i.d2+ 38 .i.xd2 ll'lf3+ 39 �e2 ll'lxd2 40 b4 �xc2 20 l:tct and an edge after t 7...l:le8 18
ll'le4+ 41 �3 �e5 0-1 tt'lf6t ..txf6 19 ..txf6 tt'ld6 20 c4 tt'le4 21 �xe4

141
Pla y t h e King 's Indian

.i.xe4 22 l:.d 1 .i.f5) b6 18 e3 l:[d7 19 �e 1 1/2-1/z der in Stocek-Gallagher.

Ekstrom-Gallagher, Swiss League 1 994. 9 . . ..!t:lc6 1 0 11t'd2 .!t:lc7!
5 ...c5!

The key move. This, as we shall see, is a very

Playing dS is also quite reasonable but may good square for the knight. Neither side can
not suit the style of the King's Indian player. now profitably avoid the coming exchanges.
Although many sources suggest that Black can 1 1 .l:d 1 cxd4 1 2 .!t:lxd4 .!i:lxd4 1 3 ..ixd4
equalise by playing a quick ... eS, my experience 11t'xd4 1 4 11t'xd4 ..ixd4 1 5 .l:xd4 .!t:lb5!
there is that White can easily end up with a The idea behind 10 ...tbc7 is revealed. The
nagging edge. The difference to the above note knight is heading for its ideal home on d6, clos­
is that Black will no longer be able to recapture ing the d-file and defending b7 so that the
at once on eS but will have to meet dxeS with bishop can be developed. That the move also
...lbg4 or ... .!Dd7. The text quickly brings things contains a nasty trap is merely incidental. I
to a head and promises Black an easier life didn't believe for a second that my grandmaster
6 c4 opponent was going to fall for it.
After 6 .i.g2 cxd4 7 tbxd4 dS! Black has a 1 6 .l:b4??
comfortable game. For example: 8 c4 dxc4 But he did! Instead:
(8 ... e5!?) 9 bxc4 0-0 10 0-0 1ib6 1 1 'W'ct .i.d7 a) 1 6 l:[dS has been played against me the
1 2 .!Dd2 tbc6 1 3 cS 1i'c7 1 4 tbbS 1i'c8 1 5 tbe4 two other times I have reached this position.
tbxe4 16 .i.xe4 tbes 1 7 .!Dd4 l:[b8 1 8 l:[b1 'W'c7 White has no advantage, e.g. 1 6....!Dd6 17 tbc3
19 1i'c2 l:[fc8 20 l:[fct 'W'aS 21 'W'c3 1i'xc3 22 .i.e6 1 8 l:[d2 (18 l:[dd1 l:[fc8 1 9 l:act :ab8=
:xc3 tbc6 23 tbxc6 .i.xc6 24 :c4?? (24 :e3=) Panchenko-Gallagher, Bad Worishofen 1 994)
24...b5! 0-1 Dizdarevic-Svidler, Bled Olympiad. :ac8 1 9 .!Dds .i.xdS 20 .i.xdS l:[c3. Normally
2002. White loses material after both 25 l:[b4 aS the bishop would give White a slight edge in
and 25 cxb6 .i.xe4 26 :xe4 :xb6. such a position but if anyone is better here it is
6 ...0-0 Black. The knight is beautifully placed on d6
6 ...tbe4 7 .i.g2 'W'aSt 8 tbfd2 tbxd2 9 .i.c3! and he has control of the c-file. The remaining
is a nice trick which enables White to maintain moves were 21 �g2 �g7 22 :ad1 f5 23 l:.d3
an edge l:[fc8 24 b4 �f6 25 .i.b3 :Xd3 26 :xd3 :c6 27
7 ..ig2 d5! :e3 tbc4 28 l:[d3 .!Dd6 29 :e3 tbc4 (perhaps
7...tbc6 transposes into the Yugoslav varia­ Black could play on with 29 ... e5 but in these
tion but the text is better. I worked this line out team matches a draw with Black is the usual
over the board against Panchenko 9 years ago. aim) 30 l:.d3 1/2-1/2 Stocek-Gallagher, European
8 cxd5 Team Ch., Plovdiv 2003.
8 dxcS dxc4 is fine for Black b) 1 6 l:[d2 l:[b8 (it's best to wait for White to
8 .!i:lxd5 9 0-0
•.. weaken his queenside before playing .!Dd6) 1 7
9 'it'd2 tbc7! 10 0-0 tbc6 was the move or- a4 and only now 1 7. . ..!Dd6 1 8 tbc3 l:[d8! i s level.

142
 The Fia n c h e t to Varia tion

1 6 Jld8!
.. material.
2) The key move in the two variations dis­
cussed is the little pawn move ... a6. This is
played to prepare the advance ... bS. In the old
days Black used to play with ... aS to prevent
White from advancing on the queenside. In the
Fianchetto variation the move ... a6 can be con-
sidered aggressive and the move ... aS defensive.
3) Don't play the move ... c7-cS lightly. Make
sure you have a concrete follow-up as this
move can seriously compromises the long-term
health of Black's position.
4) The queenside advance is not really a
queenside attack at all but an attack on the
white centre. Note how often Black wins the
My opponent sunk dejectedly in his chair as important pawn on e4 in the Gallagher Varia­
he came to terms with the fact that 1 7 .UxbS tion
just loses to 1 7...lldt+ 1 8 �ft �h3, and that S) Some�a �c:r Black has sacrificed the
with the knight on bS dominating its opposite pawn on d6 the manoeuvre .. ..Ub6-e6 helps to
number on b 1 he has no way to defend his increases the pressure.
back rank. 6) My current assessment of the Gallagher
1 7 lbd2 Variation is that Black gets a playable middle­
The only chance, albeit a slight one, was of­ game but that against certain set-ups (see
fered by 17 �e4 .Udt+ 1 8 �g2 �6 19 �c2 Games 4S and 46) he cannot get away with
.Ucl 20 �d3. White will now attempt to break sacrificing the pawn on d6. Positions that look
the pin with a4 and .Ua2. However, after bad for him, and which I previously thought
20... �d7 21 a4 �c6+ 22 £3 aS! the rook on b4 is were bad for him, are probably not bad for
embarrassed. It is trapped after 23 .Ub6 lbc:8 him.
while both 23 .Ud4 and 23 .Uh4 are strongly met 7) In the lines with 8...a6 when White plays
by 23 ..lt'lf5 as if White takes the knight the pin
. the move 1i'c2 (or 1i'e2) Black must ensure that
will remain permanent. the move l:td 1 is not going to embarrass his
1 7 lbd4! 0-1
. . . queen. The best square for the black queen is
Not 17 ...l:txd2 18 l:txbS when White should usually e7, but only if the reply �S is not dan­
escape with a draw. Now after prolonged gerous. In practice this often translates into
thought a disgusted Villamayor threw the towel meeting 1i'c2 with ...c6.
in. His defensive attempts were: 8) When White exchanges on eS the pawn
a) 1 8 .Ue1 lbc:2 structure (c4 and e4) means there is a hole on
b) 1 8 l:td 1 lt'lxe2t 19 �ft lbc:3 20 .Uct d4. It is usually a good idea for Black to ma­
ll'lxa2. noeuvre his knights towards this square.
c) 1 8 l:txd4 .Uxd4 19 lbc:4 .Ub8 and Black is 9) Advancing the h-pawn to soften up the
just a clear exchange up. position around the white king can be a good
idea. This advance should be considered when
Summary Black has no chance of active play on the
1) The Fianchetto Variation is a solid line queenside, or conversely, after his initial queen­
played by players who like to retain total control side play has tied White down and the opening
of the position. Therefore I am recommend an of a new front may over-stretch the opponent.
aggressive, risky approach as the most unpleas­ 1 0) If White refrains from playing e4 Black
ant for them to face. In order to arrest the ini­ can exploit this by trying to occupy the e4-
tiative from White Black is willing to make square with a knight, or by advancing ...d6-dS
positional concessions and, if need be, sacrifice or ... f7-f5, depending on circumstances..

1 43
 CHAPTER NINE I
The Four Pawns Attack

1 d4 tt::lf6 2 c4 g6 3 tt::lcl �g7 4 e4 d6 5 f4 White appears to be coming to terms with this

The Four Pawns Attack is the most overtly plan I am suggesting that Black reacts with the
aggressive line available to White in the King's more traditional ... c7-c5.
Indian. By move 5 he has already managed to 5 0-0 6 tt::lfl
...

construct a massive centre. Such a centre may Sometimes White plays 6 ..te2 first but as
look impressive but there are drawbacks to it. White nearly always develops his king's knight
Firstly, pawns don't move backwards. As the to f3 and his Icing's bishop to e2 this move
game progresses White may regret his rash start order has no independent significance. Play
to the game as key squares deep inside his camp should just transpose into one of the main lines,
become vulnerable to invasion by enemy e.g. after 6 ..te2 c5 7 d5 e6 8 l"Ll£3 exd5 9 cxd5
pieces. The point is that they can no longer be ..tg4 we have transposed back to the main line.
protected by pawns and never wiU be. How­ 6 ... c5! 7 d5
ever, this ideal scenario of the black army occu­ This is by far the most common reaction.
pying the heart of the white position is stiU a White blocks the centre and plans to foUow-up
long way off and for now, Black must recognise with e5 once his king has been evacuated to
that his situation is already verging on the criti­ safety.
cal. I f he just develops passively then before There are two other strategies available to
long White wiU hit him with e4-e5 and he will, White. He can take on c5 (7 dxc5) when Black
in all likelihood, be blown away. White is not should not recapture with 7 ... dxc5 as after 8
ready to do this yet as by spending 4 of his first 1i'xd8 l:lxd8 9 e5 he is already worse, but in­
5 moves moving pawns he is somewhat lagging stead should play 7 ...'it'a5! This move reveals
in development. Blasting open the centre with the tactical justification for Black's 6th move.
his king stiU resident on el and with Black's The point is that if White now plays 8 cxd6
nicely tucked away in the comer would be tan­ Black can take advantage of the pin on the
tamount to suicide. In fact, it is Black who knight on c3 to play 8... l"Llxe4! White should
should be looking to open the centre as quickly avoid this at all costs and instead defend his e4
as possible in order to try and take advantage of pawn with 8 ..td3. This line is examined in
his lead in development. In the King's Indian Game 59.
Black usually counters in the centre by ... e5 or The other possibility for White is to main­
... c5. In the Four Pawns Attack it is not easy for tain the tension in the centre and prepare cas­
Black to arrange ...e5. There is a system based tling with 7 ..te2. Black should then capture on
on playing l"Lla6 foUowed by ... e5 which has d4. This is also examined in Game 59.
become quite popular in recent years but as 7 e6
...

1 44
 The Fo ur Pa wns A t tack

0-0 6 ll:lf3 c5 7 d5 e6 8 �e2 exd5 9 cxd5

�g4

The most logical move. Black plans to ex­

change on d5 in order to open the c-line. His
Icing's rook will then be able to occupy the open This is the most solid and reliable way to
file or exert pressure against the White pawn on meet the main line of the I :our Pawns Attack.
e4 (depending on how White recaptures on d5). Black is going to give up his bishop for the
Certain players prefer to transpose into a knight on f3 and there are two good reasons for
type of Benko by sacrificing a pawn with ... b7- doing this. Firstly, the knight plays an important
b5. This can be done immediately with 7 ... b5!? role in supporting the advance e4-e5 and with­
8 cxb5 a6, or by starting with 7...a6 and follow­ out its presence this plan becomes less danger­
ing up with ... b7-b5. Whilst these Benko strate­ ous. Secondly, Black has less space and will,
gies are interesting I have decided to stick to the therefore, find it difficult to accommodate a full
main lines for my suggested repertoire. set off minor pieces within his camp. In gen­
8 .b2 eral, it is a good idea to aim for exchanges when
The alternative is to exchange with 8 dxe6. you have less space (and of course avoid them
This is examined in Game 58. when you have a space advantage) .
8...exd5 9 cxd5 There are two main alternatives which I do
White can also recapture with the e-pawn or not consider here, 9...b5 and 9.. l:le8. These
.

play the speculative 9 e5?!. These are both ex­ lines have their followers but I don't like them
amined in Game 57. as they allow White to play to e5. I prefer to
9 . �g4
. . frustrate my opponent's ambitions, especially
This is the approach I am recommending when playing an aggressive player who is just
against the main line and the ideas behind the lusting to play e4-e5. A good rule of thumb:
move 9 ... ..ig4 are examined in Game 55. This against sharp guys, play solid and against solid
position is very well known to chess theory as it guys, play sharp (Fianchetto Variation, for ex­
arises not just from the King's Indian but also ample). Anyway, solid and sharp is all relative.
from the Benoni. The Benoni move order is 1 It is hardly possible to avoid complications in
d4 liJf6 2 c4 e6 3 ltJc3 c5 4 d5 exd5 5 cxd5 d6 6 the Four Pawns Attack!
e4 g6 7 f4 ..ig7 8 ltJ£3 0-0 9 ..ie2 ..ig4 and we 1 0 0-0
have reached the diagram position. Black's last move did not actually prevent
White from playing to e5 - it just rendered it
hannless.
Game 55
For example 10 e5 dxe5 1 1 fxe5 ltJfd7 12 e6
Banikas-Gallagher
..ix£3! 1 3 ..ixf3 ltJe5 and now:
French League 200 1 a) A fter 1 4 exf7+ l:lxf7 1 5 0-0 ltJbd7 Black
has a good central position. The mobility of
1 d4 ll:lf6 2 c4 g6 3 llJc3 �g7 4 e4 d6 5 f4 White's minor pieces are somewhat restricted

145
Pla y th e King 's In dian

by the pawn on d5 and if he advances it any 1Vg3t 1 9 �hl �g7! he will be mated by ...l:lh8;
funher it is liable to be lost - passed pawns are 18 'Ot>g2 is a better try but Vaisser points out
very strong when they are supported by another that after 1 8 ...ll'lf6 19 �f3 ..id4! Black is win­
pawn but isolated in the opposing camp they ning as 20 l:lht can be met by 20.. 'iV£2+) .

are just a weakness. 17 ...'it'xh5. White should now settle for the bad
b) 14 0-0 fxe6 1 5 �e3 (1 5 dxe6? lfuf3+ 1 6 ending that arises after 1 8 1i'xh5 gxh5 - bad
l:lxf3 1Vxd1+ 1 7 ll'lxd1 l:lxf3 1 8 gxf3 ll'lc6 and because he is going to lose his e-pawn. Instead,
the pawn on e6 is doomed) 1 5 ... ll'lxf3+ 16 l:lxf3 if he plays 18 1Vxd7? he will be put to the
l:lxf3 17 1Vxf3 exd5 18 lilid5 ll'lc6 19 .i.xc5 sword: 18 .. .l:tad8 (forcing the queen far away
�h8 with a roughly level game. from his king) 1 9 'ifb5 'it'g4+ 20 �£2 (or 20
1 0 ...ltJbd7 �h2 �xc3 21 bxc3 �g7! with the same ...l:lh8
Although Black is going to take the knight mate we have already seen) 20 ... �d4+ 21 �el
on f3 he doesn't have to do so at once. He (21 ..ie3 'it'xf#!) �xc3t 22 bxc3 l:lxe4+ with a
might as well wait for White to spend a tempo crushing attack.
on h3. The only reason for Black to exchange a2) 1 5 g5 lt'lg4 16 �xg4 hxg4 17 l:let (17
at once is if he is worried about the knight run­ 1Vxg4 �xc3 18 bxc3 l:lxe4 19 �d2 'it'e7 20
ning away. Practice has shown, though, that 1 1 l:lael li'lb6 is fine for Black) 1 7...c4 (Black of­
ll'ld2 �xe2 1 2 'ifxe2 l:le8 1 3 'it'f3 l:lc8, prepar­ fers his g-pawn in order to activate his knight as
ing ...c4, is nothing for Black to worry about. quickly as possible) 18 .i.e3 (18 'it'xg4 ll'lc5 1 9
1 1 .:l.e1 1i' f3 ll'ld 3 20 l:le2 'it'a5 gives Black good com­
The main line but White can also play 1 1 h3 pensation for the pawn) 1 8 ... ..ixc3 1 9 bxc3
.i.xf3 12 .i.xf3. Black has tried many moves l:lxe4 20 1i'xg4 1Ve7 21 �£2 ll'lc5 with an un­
now but I am suggesting 1 2...l:te8, as very often clear game, Kouatly-Kindermann, Tmava 1 987.
White will just play 1 3 l:le 1 transposing into the It is not easy for White to get at the weak dark
main line. It is a good idea to link up your rep­ squares around the black king and meanwhile
ertoire in this fashion as it cuts down the work­ he has to watch out for his own weak pawns
load. Apart from 13 l:let White has two other and naked king.
ideas which need to be examined: b) 1 3 a4 c4 14 ..ie3 1i'a5 (there are other
a) 13 g4 h6 (this is the usual reaction to g4) plans for Black but this is similar to the one we
14 h4 h5!? and now: adopt in the main line - the main idea is to play
... ll'lc5) 1 5 id4 (15 l:lel would actually trans­
pose to the notes to move 14) 1 5...l:te7!? (now
1 5 ... ll'lc5 is not so good because of 1 6 e5 so
Black changes plan - he intends to double
rooks on the e-file and maybe sacrifice the ex­
change on e4) 16 �hl a6 (the immediate
1 6.. .l:tae8 1 7 ll'lb5 is complicated but probably
good for White) 1 7 g4 (17 1i'et is more solid
when 1 7 .. .l:tae8 1 8 'iVh4 ll'lh5 1 9 �xg7 (19
�xh5 �xd4] lt'lxg7 20 l:lae 1 is about equal)
1 7...l:lae8 1 8 g5 ll'lxe4! 1 9 ll'lxe4 (1 9 �xg7
ll'lg3t) l:lxe4 20 bxe4 l:lxe4 21 1i'f3 f5! gave
Black good compensation for the exchange in
Peev-Velimirovic, Sofia 1 972.
at) 1 5 gxh5? ll'lxh5 1 6 �xh5 'ifxh4!! (theory 1 1 ....:ea
had originally given 1 6 ...gxh5 17 'ifxh5 b5 as With his rook on el White really was threat­
offering compensation for the pawn but Vais­ ening to play e5. The text prevents this and
ser's analysis demonstrates than the text is forces White to play h3 in order to defend his
much stronger) 17 'it'g4 (White can't keep the e-pawn. There is one viable alternative to
piece with 1 7 ..ig4 as after 17 ... ..ixc3 1 8 bxc3 ...l:le8, namely t t ...ll'le8. This strange-looking

146
 The Fo ur Pa wns A t tack

retreat removes the knight out from the central often plays 14...c4 in order to clear the cS­
firing line. It is heading for a new home on c7, square for his knight. The main line then runs:
from where it will help to suppon the advance 1 S �e3 'ii'aS 16 �h1 ti)cs 17 �xeS! 'ii'xcS 18
... bS, and it also no longer blocks the king's eS dxeS 1 9 fxeS ti)fd7 20 lbe4 1i'b4 21 e6 fxe6
Indian bishop (as it did on f6). Still, this is all 22 �g4! ti)eS 23 �xe6+ �h8. White is slightly
quite time consuming and I prefer the tradi­ better according to Vaisser.
tional 1 1 ...l:.e8. 14 �e3
12 h3 �xf3 1 3 �xf3 So why isn't it such a good idea to play 14
a4. Well, the reason is that Black can follow a
similar path to the previous note but without
having to play the unnecessary ... a6. The tempo
saved on this move means that White won't
have the time to force through eS. Therefore
14... c4! and now:

It is time to take stock and discuss what each

side is trying to achieve. The most imponant
factor in the position is the unbalanced pawn
structure. White has a 2-1 majority in the centre
whilst Black has 3-2 in his favour on the queen­
side. Pawn majorities need to be pushed so
White, as we already know, will be looking to a) 1 S �e3 ti)cS 16 �xeS (White cannot be
play e4-eS while Black will be looking to ex­ very happy to give up his strong bishop but the
pand on the queenside, usually starting with the e-pawn is attacked and defensive moves such as
advance ... bS. Normally it is an advantage to 16 �£2 are just met by 1 6...ti)d3!) 16 ...1i'xc5+
have an extra central pawn but here this is (the big difference to the previous variation is
counter-balanced by the fact that it is easier for that this is now check) 1 7 �h 1 lbJ7! with a
Black to advance his majority. Black has placed pleasant game for Black. His pieces co-ordinate
his pieces in such a way that it will be very diffi­ very well.
cult for White to play eS and after dxeS to re­ b) 1 S l:.e2 (note that 1 5 ti)bs, attacking the
capture fxeS without losing a pawn. Sometimes d6 pawn, can be met by 1 S...a6 as 16 lLlxd6?
White switches w another plan - to play eS and just loses the knight to 1 6...1i'h6+) 1 S ...ti)cs 16
after dxeS not to recapture but w play 5. Al­ eS dxeS 1 7 fxeS ti)fd7 18 e6 lLleS 19 exf7+
ternatively White can gain even more space on �xf7 20 �e4 lLJed3 21 d6 lLlxe4 22 lLlxe4 1!f5!
the kingside by playing g4. But exposing one's 23 g4 �d4+ 24 �h2 'ifeS+ 25 ti)g3 1i'xd6 and
king like this is liable to backfire as we saw in Black had a winning position in Bosboom
the notes to move 1 1 . Lanchava-Gallagher, Cappelle Ia Grande 2002.
13 'ifa5!
... 1 4 b5
...

This is the most accurate way for Black to This move is self-explanatory. Black statts
prepare ...bS. 13...a6 is also played quite often the queenside attack he envisaged with his pre­
but then White can prevent bS with 14 a4!. This VIous move.
is not so good after 1 3...1i'aS - for an explana­ 1 5 a3
tion see the next note. After 13 ...a6 14 a4 Black This little move is designed to draw the sting

147
Pla y t h e King 's Indian

from b4. White will now be able to maintain White's position so he opens the e-file to enable
the knight on c3 as after 1 5 ... b4 1 6 axb4 Black his comrade to join him) 28 e5 dxeS 29 fxe5
must recapture with the queen. l:txeS (White played eS in order to get a couple
1 5 ...ltlb6 of connected passed pawns but he is under too
Dramatically raising the stakes. On d7 the much pressure for them to have much influ­
knight was performing a good defensive duty ence on the game) 30 ¢>h1 l:te4 31 i.e1 1i"c7
(stopping White from playing eS) but not doing 32 l:tab1 (32 i.xd2 cxd2 33 l:td1 i.xa1 34 l:txa1
much to help the queenside offensive. Now it is l:te 1+ wins for Black) 32...l:tde2! 33 i.xc3 l:t2e3
heading for c4 from where it will create great 34 d6 'iVxd6 35 i.b4 'iVc6 36 1i"f1 l:txh3+ 37
confusion in the white camp. But at a price. gxh3 l:te2t0-1
White can now play e5. The rime for generalisa­ b) In his notes to the above game Nunn
tions is over. Tactics and the direct confronta­ gave the following variation: 20 lbbS lbxb2 21
tion between the pieces will now decide who lbxd6 b3 22 1i"b1 lbxa4 23 lbxe8 i.xa1 24
emerges with the better game. 'iVxa1 b2 25 'iVa2 'iVxet+ 26 i.xe1 b1'iV with a
The alternative 1 5 ... b4 is considered in the winning position for Black. It is within this
next game. minefield that Vaisser suggests an improvement
1 6 e5 for White. Instead of 23 lbxe8 he gives 23 l:ta3!
White takes the bull by the horns. 'iVb4 24 l:txb3 'iVxb3 25 'iVxb3 l:txb3 26 lbxe8
16 i.£2, was shelved after a couple of crush­ as unclear.
ing defeats for White, but Anatoly Vaisser The Croatian Grandmaster Cebalo is a great
claims that all is not so clear. Let's take a look: fan of Vaisser's book on the Four Pawns At­
16 ...lbc4 1 7 'iVc2 (17 b3? 'iVxc3 1 8 bxc4 lbxe4! tack (Beating the King's Indian and Benoni, Bats ford
19 l:tct lbx£2! 20 l:txe8+ l:txe8 21 �£2 1i"e3+ 1 997) and it was no surprise to see him testing
22 ¢>g3 i.h6 0-1 was a disaster for White in this variation with White. In Cebalo-Balcerak,
Kouady-Barencilla, Doha 1 993) 1 7...lbd7 Bid 2000 Black preferred 23 ... c4!? (instead of
(threatening 18 ...lbxb2) 18 i.e2 (1 8 a4 b4 1 9 23...1i"b4) and won an amazingly complicated
lbbS a6� 1 8...l:tab8 (1 8...lbxb2 19 lbxb5) 1 9 a4 game Gust to give you an idea the next few
b4 and now: moves were 24 lbxe8 i.fB 25 d6 1i"b4 26 l:txa4
1i"xa4 27 lbc7 c3). And in Cebalo-Mohr, Croa­
tian Team Ch. 2003 Black tried 23...l:ted8 and
this worked out all right for him as well.
So, to be honest, nobody really knows
what's going on in this variation and it may well
take weeks of computer-aided analysis to find
out. I've tried a bit and still don't know what to
make of it all. All that remains for me to do is
wish you luck in the amazing event of any of
your games actually following this much theory!
Incidentally, I once met 1 6 i.£2 with
1 6...lbfd7 and after 1 7 'iVc2 b4 1 8 axb4 1i"xb4
we have transposed into the 1 7th move notes
to Regez-Gallagher. Certainly an option worth
a) Kozul-Nunn, Wijk aan Zee 1991 contin­ considering if the above complications seem
ued 20 i.xc4? bxc3 21 b3 a6! (keeping the overbearing.
bishop out of b5) 22 l:tect (planning to win the 1 6 ...ltlc4!
pawn on c3 with i.e1) 22 ... lbb6 23 i.ft c4! Such moves demonstrate the advantage of a
(Black gives up a pawn to open the b-file) 24 knowing an opening well. Without prior knowl­
i.xc4 lbxc4 25 bxc4 l:tb2 26 1i"d3 l:td2 27 1i"f3 edge it would take great effort and courage to
f5! (there is no respite for White; Black is not play such a move, especially when there is a
content with having one rook in the heart of safer-looking alternative.

148
 The Fo ur Pa wns A t tack

17 exf6 ttJxe3 18 .lbe3

At first sight it may seem that White can win
a piece by 1 8 'ii'c t as two black pieces are now
attacked. The problem, though, is that after
1 8 ... ..ixf6! 19 .l:.xe3 .l:.xe3 White can't take the
rook as 20 ... ..i.d4 wins the queen.
1 8 . . Jlxe3 1 9 fxg7 llaeB

a) The fact is, though, that the safe-looking

16 ... dxe5 is not very good. After 1 7 fxeS :XeS
18 ..ixcS l:lxet+ 1 9 1i'xe1 White has a consider­
able advantage - his passed d-pawn is acrually
quite strong and the position has opened more
for his bishop pair.
b) There is another option, however, which
has emerged relatively recently in a couple of This is the position Black had in mind when
exciting clashes involving the world's leading he embarked on the complications with
Four Pawns Attack experts. The move is 1 5 ... ltlb6. White's slight material advantage is
16 ...tbfd7!? and look at these fireworks: 1 7 e6 compensated for by the fact that the black
ltlc4 1 8 exd7 :Xe3 19 :Xe3 lLlxe3 20 1i'e2 rooks are extremely active and the mobility of
'ii'd8! 21 'ifxbS (21 'ifxe3 ..id4) 21 ...l:lb8 22 the white bishop restricted by the pawn on dS.
'ii'c6 :Xb2 and now: The pawn on g7 is of course doomed.
b1) 23 l:lct l:lb6 24 lbbS!? (wow! 24 1i'c8 20 f5 b4
.l:.b8 25 'ii'c6 .l:.b6 looks like a draw) 24 ...:Xc6 Black can 'win' the queen for two rooks with
25 dxc6 (considering that Banikas later 20....I:.et+ 21 'ifxe1 :Xet+ 22 .l:.xe1 but he cor­
switched sides he must have an improvement rectly avoids this as in the resulting position
on Black's play; I suspect it was here as there White's material advantage is more pronounced
arc certainly more useful moves than ... a6) - rook and two minor pieces against a queen is
25... a6 (25... ..if6!?) 26 lbxd6 'ifb6 27 lbe8 usually a decisive advantage. Black certainly
(nothing can stop the pawns now so it's just a doesn't want to go in for such a line when it
question of whether Black can scramble a per­ means getting rid of his main trump in this
perual) 27 ... 1i'h2 28 .l:.d1 lLlxd1 29 d81i' ..id4+ position - his active rooks.
30 �h2 lbe3 31 ll:ld6+ �g7 32 lbeS+ �g8 33 A nightmare scenario for Black would arise
h4! 'ii'f2 34 'ith3? (34 tbf6+ �g7 35 ltlh5+! gxhS if White could support his pawn on g7 with f6.
36 'ii'g5t 'itf8 37 c7 wins) 34...'ii'g1 35 ll:ld6+ '/2- The text is directed against this. Black activates
1/2, Banikas-Kotronias, Corinth 1998. his queen so that a furure f6 is just likely to lose
b2) 23 lbe4 ..id4 24 �h 1 .l:.b6 25 .l:.e 1 (here the pawn to ...'ii'h4 or ...'ii'd4. There are several
we go again) 25...:Xc6 26 dxc6 c4 27 lbxd6 other ways Black can prevent f6. The first is to
'i'h4 28 .l:.b 1 ..ib6 29 tbxc4 lLlxc4 30 .l:.xb6 play 20...�xg7 - not a bad move but not the
lLlxb6 31 c7 lbxd7 32 c81i'+ ltl£8 33 'ii'b8 '/2-1/2 most accurate. Black only wants to play this
Vaisser-Banikas, French League 2001 . when he has nothing better to do. Another
Well that's certainly an area for further possibility is to play 20 ... gxf5 but this would
investigation for the analytically minded leave Black with a seriously weakened kingside.
amongst you. The final possibility is perhaps the best of

1 49
Pla y t h e King 's Indian

all. The novelty 20...a6! was introduced in the 25 1i'f2 (25 f6 looks critical but I suppose Bani­
game Segura Ariza-Moreno Camero, Ayamonte kas was worried that both his advanced pawns
2002. This move also frees up the black queen would evenrually drop off; I'm not so sure)
as it is no longer required to defend bS, but 25 ...1i'xf2+ 26 llx£2 �xg7 27 .i.e2 I suspect the
unlike the main line, Black retains his nice posi­ position is about equal and the game evenrually
tional advantage on the queenside. The game ended in a draw after many advenrures.
continued 21 f6? (this seems to just lose) 23 ...hxg6 24 :n
21 ...11fd8 22 lDe4 ll8xe4 23 i.xe4 11fxf6! 24 It is too dangerous for White to take the a­
.i.c2 11fxb2 25 a4 b4 26 llbt 11fc3 27 llct 1i'e5 pawn. After 24 llxa7 llx£3! 25 gx£3 11fg3+ 26
28 .i.b3 aS 29 .i.c4 �xg7 and Black, who has �fl (26 11fg2? llel mate) 26...1i'x£3+ 27 �gl
total domination of the chessboard won in a (27 'if£2? 'ifht+ 28 1i'g1 llet+ 29 �xel 1i'xg1+)
few more moves. Look how useless White's 27...11fg3+ 28 �fl Black has at least a draw and
bishop is in this type of position. It's more like probably more after 28 .. .lle5. By defending his
a big pawn. bishop on £3 White rules out such sacrifices.
White can do better than 21 f6, but 20 ... a6 24 ...a6!
still looks like a good idea. A nice little move. The white bishop is al­
21 axb4 'ifxb4 22 Wd2 ready restricted - now it's time to smother the
Of course not 22 llxa7?? as then 22 ... llet+ knight. It can no longer go to bS and later we
wins the queen for just one rook. see Black playing ... f5 to funher reduce the
22 . .Wh4!
. knight's scope. Note how Black still hasn't
bothered to caprure on g7 - the pawn is not
going anywhere.
25 lDd1
This is acrually the first new move of the
game. Opening theory stretches far into the
middlegame these days. Previously 25 1i'f2 1i'd4
26 �hl ll3e5! had been played with an unclear
game according to theory. Black is certainly not
worse here and I believe that White needs to
defend well to maintain the balance after an
exchange of queens.
25 ...ll3e5 26 lLif2 f5!

Better than 22...1i'f4 as on that square the

queen can be embarrassed by a rook on fl . for
example, Dearing-Moss, Hastings 1 995/96
continued (after 22...11ff4) 23 lift ! 11fxf5? 24
i.e4! (attacking the queen twice and cutting the
lines of communication between the black
rooks) 24 ...11fg5 (the only move to avoid imme­
diate material loss) 25 ll£3! and White wins the
pinned rook on e3.
23 fxg6
After our game Banikas swore never to
touch this line again so I was somewhat sur­
prise to see later in the year the game Banikas­ As previously mentioned, this is played to
Dochev, Kavala 2001. This was the moment he limit the options of the knight - now it is de­
chose to improve by refraining from the ex­ nied access to the e4- and g4-squares. Still, I
change on g6. After 23 lift a6 24 .i.g4!? 1i'g3 was a little hesitant before playing this move as

1 50
 The Fo u r Pa wns A t tack

now Black's king position is less secure. 30 "ifxa6 .1:td2!

27 �d1 ?! Now that the white queen no longer defends
Please don't ask me to explain this move. this square, Black penetrates to the 7th rank
The only thing I can think of is that White was 31 �c6??
going to extreme lengths to lose his pawn on d5 Completely overlooking Black's idea. In fact
in order to give his bishop some open diago­ White still had good chances to defend himself
nals. The problem is that whilst White will get with 31 1i'a3!, e.g.
diagonals Black will also get more files for his a) 31 ..ltxt2? 32 l:.xt2? l:.e1+ 33 l:.fl 'ii'd4+
rooks. 34 �h2 l:.xfl wins for Black but the combina­
The move White would like to play is 27 tion is unsound because White can flick in 32
lbd3 but this is strongly met by 27 ...1i'd4+, cre­ 1i'c3+!
ating a powerful pin on the knight. At the b) 3 1 ...1i'd4 looks strong as Black now
board I was most concerned about 27 l:.a1 - threatens to take on both f2 and b2. However,
now that there are no sacrifices on f3 the rook again 32 1i'c3! saves the day as 32 .. ltxb2?! is
can return to active service. It seems to me the now met by 33 1i'xd4+ cxd4 34 ll.dt. Black
game is about equal, for example: 27...1h.g7 28 would do better to play 32..ltb8 here.
l:.xa6 l:.e1+ 29 �h2 l:.8e3 (Black plans ... f4 and c) Perhaps 31 ...1i'f6 offers Black the best
1i'g3 mate but first prevents White from playing chance of some advantage after 32 1i'c3 l:.b8
'ii'c3+; the relevance of this is seen in the varia­ but the most likely outcome is a draw.
tion 29 ... f4?! 30 1i'c3+! �g8 31 tl:le4 l:.8xe4 32 31 ....1:te1 !
i.xe4 when the queen on c3 stops ...'ii'g3 mate;
after 32.. ltxe4 33 l:.xd6 Black has to fight for a
draw) 30 lbd3 f4! 31 'ii'c3t �g8 32 l:.a8+ �f7
33 l:.a7+ �e8 34 l:.a8+ �f7 with a draw by per­
petual check. I f either side tries to avoid the
repetition they get mated.
27 ...�xg7
A good moment to take the pawn as I was
curious to see what White was going to do next.
I half expected him to return to f3 with his
bishop but instead he lashed out on the queen­
side.
28 "ifa5 .l:txd5 29 �f3 .l:td4
Better than going back to the e-file. I had a
nasty trap in mind but wasn't really expecting A simple but devastating tactic. The first
my opponent to fall into it. point is that 32 l:.xe1 1i'x£2+- 33 �h2 'i'xet just
loses a piece, but White's main problem is that
because the rook on f1 is pinned Black is
threatening ... 1i'x£2+- anyway. I thought the only
way to stave off immediate defeat was 32 tl:lh1
but my opponent said he refused to even con­
sider this sick-looking move. The most precise
way to win after 32 tl:lh1 is 32 ... 11Vd4+ 33 �h2
1i'e5+ 34 �g1 (34 tl:lg3 l:.xfl 35 1i'xf1 f4 wins
as if the knight moves there is a devastating
discovered check) 34...l:.xfl+ 35 1i'xf1 l:.e2! 36
1i'f3 (the only move to save the queen)
36... l:.e1+ 37 �f2 l:.xh1 and Black wins easily.
White chose a continuation that gave him a
couple of checks but they soon run out.

151
Pla y th e King 's In dian

32 ..,7+ �h6 33 ll:\g4+ a) In a couple of games Black has taken the

What else?. Now 33 ltJh 1 l:txfl+ 34 �xfl opportunity to exchange off dark-squared bish­
l:td 1+ is even easier. ops with 19 ... �d4 and the position after 20
33 ...fxg4 34 'ilf7 �h1 i.x£2 21 l:tx£2 c4 22 l:te2 aS 23 �g4 lL!cS!
White was placing his meagre hopes on this does seem fine for him, e.g. 24 eS ltJd3 2S ltJe4
move. ltJxf4 26 ltJf6+ rlilg7 27 ltJxeS+- l:xe8 28 l:£2
34...l:lxf1 + 35 'ilxf1 dxeS, Cebalo-Krstic, Velika Gorika 2002 and
3S �xft l:tdt+ 36 �e2 1i'e1 mate. White went on to win but it just looks lost for
35 ...�g7! 0-1 him at this point.
The black king is totaUy safe and White is an b) The aforementioned Cebalo-
exchange and a pawn down for nothing. Shchekachev, Biel 2003 steered a different path:
1 9...aS!? 20 lL!a4 l:teb8 21 eS?! (the standard
Game 56 breakthrough in the 4 pawns which White often
plays even if it costs him material, as here; in
Regez-Gallagher
my view White should have settled for some­
Ziirich 2003 thing like 21 �h1) 2 1 ...1i'xf4 22 l:e4 'ii'gS 23
exd6 lL!es (I suspect White may have already
1 d4 ll:lf6 2 c4 g6 3 ll:\c3 .i.g7 4 e4 d6 5 f4 been regrerting his decision to play eS) 24 l:a3
0-0 6 llJf3 c5 7 d5 e6 8 .i.e2 exd5 9 cxd5 lL!xa4 2S 1i'xa4 (2S l:exa4 c4! looks seriously
.i.g4 10 0-0 ll:lbd7 1 1 l:le1 Ilea 1 2 h3 .i.xf3 good for Black) 2S ...l:xb2 26 d7 (White is rely­
1 3 .i.xf3 'ila5 1 4 .i.e3 b5 1 5 a3 b4 ing on his d-pawn but has left his kingside vul­
Normally I play 1 S ...ltJb6 here but my op­ nerable) 26 ...1i'd2! 27 �e2 'ii'xdS 28 l:d3! (an
ponent seemed to be playing so quickly and excellent resource which keeps White in the
confidently that I thought it prudent to avoid game) 28 ...lLlxd3 29 l:eS+- �£8 30 �f3? (White
the long theoretical variations. After all, when had to play 30 �xd3! when the pawn on d7
you play someone a lot lower-rated it's usually a gives him plenty of tactical opportunities to
good idea to get them out of the book as soon save the game) 30...1i'd6 31 :Xa8 l:bt+ 32 �d1
as possible. I had also seen the rather impres­ lLlb2! 33 d81i' 1i'xd8 34 :Xd8 lLlxa4 3S �h2
sive Cebalo-Shchekachev, game that follows lL!b6! 36 �c2 (36 �xeS :Xd1 ! 37 .:.X£8+ rlilg7-
and thought why not give that a try. +) 36...l:b2 37 �xeS lL!d7! 38 �xfB ltJx£8 (so
1 6 axb4 'ilxb4 1 7 l:la3!? the fireworks have finally died away and Black
has emerged into a completely winning end­
game) 39 �e4 l:bS 40 l:a8 rlilg7 41 l:a7 l:cS
42 �b1 lL!e6 43 �a2 lL!c7 0-1 .
1 7 . . .l:leb8
This move took me an hour and I can't say
that I was overjoyed with it. I just didn't see a
better alternative. Black's basic idea in this posi­
tion is to play ...ltJb6 and ...lLlfd7 in order to
create pressure on the queenside but I had to
reject this immediately as after 1 7...lL!b6 1 8 l:b3
1i'aS 1 9 l:a3 White has a forced draw. That is
no good against someone nearly 400 points
lower-rated.
The next thing I wondered about was
A novelty, played instantly, by my opponent, whether I could get away with 1 7...lL!xe4 (of
who later informed me that he had worked course 1 7...1i'xb2 loses to 1 8 l:b3) but had to
hard on the move with Fritz. More common reject this because of 1 8 lL!xe4 l:xe4 1 9 i.xe4
had been 17 1i'c2 lLlb6 1 8 i.£2 lLlfd7 1 9 l:te2 1i'xe4 20 �xeS 1i'xf4 21 l:f3! when Black does
and now: not have enough for the exchange.

1 52
 Th e Fo ur Pa wns A t tack

So then it was time to look at the obvious stantly. Black must take control of the e4-
move 1 7 ... c4 but it didn't appeal to me, e.g. 1 8 syuare even at the cost of allowing White a
l:ta4 1i'xb2 1 9 .:Xc4 l:tac8 20 l:te2 1i'b7 21 protected passed pawn in the centre.
l:tc6!? (21 l:txc8 l:txc8 22 ..id4) 21...l:txc6 22 23 "ii'e 1 �f8?
dxc6 'ifxc6 23 l:tc2 and White, despite being a I finally gave up on the idea of playing
pawn down is probably better. . ..lbd7-b6 just at the moment when it was a
The only idea left was to play a rook to b8 - good move. I was nervous about leaving my
but which one? I finally decided on the text as I kingside undefended after 23 ... lbb6 24 exd6
wasn't sure what I was going to do after lbxd6 25 l:te7 but 25 ...l:txe7 26 1i'xe7 lbbc4 is
17 ...l:tab8 1 8 l:te2. The point is that if I can probably just good for Black.
never play ...lbd7-b6 and . ..lbf6-d7 (because of 24 �h4!
the draw) then, in order to get out of the way of By now I was getting seriously worried. My
e5 and activate the bishop on g7 I am going to major preoccupation's were: why is this guy
have to retreat the knight on f6 to e8. Therefore playing so well and where has all my time gone?
it had to be ...l:teb8 even though the rook was 24...l:tbb7
rather well placed where it was. Sterling defence along the second rank as
1 8 l:te2 24 ... ..ixc3 25 l:txc3 1fxf4 26 exd6 lbxd6 loses
I didn't look too long at 1 8 e5 dxe5 19 d6 to 27 ..ig3. I was waiting for e6 so that I could
exf4 as I just assumed that Black would always grab a risky pawn and see what my opponent
have sufficient compensation in such positions. was like in a tactical battle. Surely he couldn't be
1 8 ... 85 as strong as in a strategic game!
The a8 rook needs to be brought into play. 25 e6 �xc3
19 �f2 lbe8 20 �h1 l:ta7 21 �g3 The exchange sacrifice 25.. .l:.e7 26 ..ixe7
My opponent was still playing incredibly l:txe7 might be OK if it wasn't for that dreadful
quickly, following a typical plan for such posi­ knight on fB while I also thought about just
tions. It was getting quite intimidating. playing 25 .....ig7 26 e7 lbd7 but felt that my
21 ...�d4 position was too accident prone for the coming
Again after a long think. I finally decided time scramble.
that there was no good way to stop e-'k:5 so I 26 l:txc3 "ii'xf4
may as well at least get my bishop active. I actu­
ally came quite close to playing 21 ...g5?! in order
to take control of the e5-square while my mind
even toyed with the sick-looking 21...f6. It's
bad, but not that bad.
22 e5 f5!

27 e771
Short of time it would have been more un­
pleasant if White had just played 27 b3!, retain­
ing all the advantages of his position, rather
than going in for the forcing line I was ready
for.
finally it was my tum to play a move in- 27 . . . �d7 28 �g3 "ii'd4 29 �xd6

1 53
Pla y t h e King 's Indian

Whilst my opponent was thinking about this that has been fixed on h3.
move I suddenly became terrified by 29 l:.d2 51 iLb7 ll:ld7
..f6 30 ..e6+ when I saw that the planned 51...�e2 52 .i.£3+ �el was perhaps the sim­
30...'itg7 fails to 31 1Lxd6! .!Dxd6 32 'ifxd6! plest.
'ifxd6 33 e8.!D+. I relaxed a little when I saw 52 iLea
that I could get away with 30...�h8. Or 52 �ft .!Db6! 53 .i.c6 .!Dc4 54 ..ib7
29 ...ll:lxd6 30 es• + ll:lxeB 31 l:txeB+ ll:ltB .!Dd2+ 55 �el f3 56 gxf3 .!Dx£3+ 57 �fl .!Dd4
32 d6 and Black wins.
32 ..e6+ �g7 33 'ifd6 .!Dd7 34 l:.e7+ �h6 52 .....ll:\e5 53 iLb7 f3 54 gxf3 ll:lxf3+ 55
seemed all right. �1
32 ...l:txb2 33 •e6+ Or 55 �g2 .!Det+ 56 �gl .!Dd3 57 �g2
A relief as I had seen that White can draw .!Df4+ 58 �h2 �f2 and wins
with 33 ..e7! l:.xe7 34 dxe7 l:.bt+ 35 �h2 55 .ll:\d4
..

..d6+ (35.....e5+ 36 g3 l:.b2+ 37 ..ig2 l:.xg2+ 38 Now White has no time for ..ic8-g4 because
�xg2 'ife2+ 39 �gl=) 36 g3 l:.b2+ 37 �hl=. Black just plays �f3.
33 ...�g7 34 l:te7+ �h6 56 �g2 ll:le2 57 iLea ll:lt4+ 58 �h2 �2 59
34...l:.xe7 35 ..xe7+ 'itg8 36 l:.ct planning iLb7 ll:le2 60 iLe6 lLld4 61 iLd7 lLlf3+ 62
..ie2-c4 is not so easy to deal with. �h1 �g3 0-1
35 'ii'e 1 l:td71 36 'ii'e 1 +?
At last a serious error. I wasn't sure where Game 57
my queen was going after 36 l:.d3, just some­
Conquest-Mestel
where that protects h4
36 .•.•d2 37 'ii'xd2+ Hastings 1986/87
There is nothing better.
37 ...l:txd2 38 l:txe5 l:t2xd6 39 l:txd7 ll:lxd7 1 d4 ll:lt6 2 e4 d6 3 ll:\c3 g6 4 e4 Jig7 5 f4
40 l:txa5 �g5 0-0 6 ll:lt3 e5 7 d5 e6 8 iLe2 exd5 9 exd5

Finally I made it to the time control. White's A much more cautious, and less popular,
next move is a mistake as the active black king approach than the 9 cxd5 of the previous
gives him a winning position. Anyway, what­ games. White keeps the pawn structure bal­
ever White chose it was now his tum to suffer. anced and relies on his extra space to provide
41 l:td5?! l:txd5 42 ..bd5 ll:lt6 43 iLb7 �4 him with the advantage. However, his position
44 �h2 ll:le4 45 iLa6 �e3 46 �g1 f4 47 would be much healthier if his pawn were back
iLb5 g5 48 iLea ll:lt6 49 iLb5 h5 50 iLe6 on f2. On f4 it blocks in the bishop on c1 and
h4! weakens squares on the e-file, especially the
The key move as Black can win by playing important central square e4. Black will try and
... f3 and then eventually collecting the pawn control this square while at the same time mak-

1 54
 The Fo ur Pa wns A t t a c k

ing sure that White can't play f4-f5 as this It is no great surprise that we no longer see
would liberate his bishop and give him attack­ much of 9 e5.
ing chances against the black king. 9 �f5
...

Before examining 9 exd5, though, we have The most logical move, immediately taking
to take a look at another White option, the control of e4. 9 ...l:te8 10 0-0 .i.f5 leads to the
speculative 9 e5?!. This move enjoyed a period same position.
of popularity in the 1 980's but has virtually There is one other, more radical idea that is
disappeared from practice now that the worth mentioning: 9...lt:)h5!? 10 0-0 �xc3 1 1
strength of the reply 9...lt�e4! has been estab­ bxc3.
lished. Previously 9 ... dxe5 or 9 ...lt:)g4 were
played with unclear complications. After
9 ... lt:)e4 there is:
a) 10 lt:)xd5 dxe5 (the immediate 1 0 ...ttk6 is
often recommended but taking on e5 first cuts
down White's options) 1 1 fxe5 ttk6 12 �d3
(after 12 0-0 lt:)xe5 1 3 lt:)xe5 .i.xe5 I don't see a
lot of compensation for the pawn) 12 ... £5
(Black could also consider 12...1Va5+!? as after
13 �2 lt:)xd2 14 �xd2 'iVdS he will win the
pawn on e5, while 13 �ft f5 leaves the white
king badly placed) 1 3 exf6 (after 1 3 �f4 l:te8 1 4
0-0 Black can either take the pawn o n e 5 or
play 14...�e6 - both look good for him) 1 3
exf6 lt:)xf6 14 lt:)xf6+ 1Vxf6 1 5 0-0 �g4. Black It is most unusual for Black to give up his
has a slight advantage due to his lead in devel­ King's Indian bishop in this fashion but here it
opment. is justified as both White's bishops are severely
b) 10 cxd5 lili:c3 1 1 bxc3 �7! and now: restricted by the pawn structure. It goes without
bt) 1 2 0-0 dxe5 1 3 fxe5 (the original game saying that Black will now try and keep the
with 9 ...�4, Calvo-Diez del Coral, Malaga position blocked. He has two ways to go about
1981 went 1 3 lt:)g5 h6 14 �4 f5 1 5 �6 e4 this:
with a clear advantage for Black; 20 years ago a) 1 1 ...£5. This ensures that White can never
information didn't travel so well and it wasn't play f5, but it also restricts Black's pieces and
for another five years that the strength of �4 creates a hole on e6. In fact both sides now
was appreciated) 13 ...lili:e5 14 �e3 (14 �e5 appear to have a bad position - lucky they are
.i.xe5 1 5 �h6?? 'irh4 0-1 has occurred more playing against each other! 1 2 lt:)g5 (another
than once) 14... lt:)xf.3+ 1 5 �xf3 'iVd6! and White plan would be to try and activate the dark­
is just a pawn down for nothing, li Zunian­ squared bishop by means of �d2-e1 -h4)
Gheorghiu, Dubai Olympiad 1 986. Although 1 2...lt:)g7 1 3 �f3 �7 14 l:tet (the critical test
White has a passed d-pawn it is firmly block­ of Black's strategy would be 1 4 �6 �xe6 1 5
aded and Black has a good grip on the dark dxe6 lt:)f6 1 6 �d5 1i'e7 1 7 l:tet but after
squares. 17 ... �4 Black has the advantage according to
b2) After the above Gheorghiu game White Gligoric) 14...lt:)f6 1 5 l:tbt l:te8 1 6 l:txe8+ 1Vxe8
tried to revive the line with 12 e6, but this has 17 l:tb2 �d7! 1 8 l:txb7 l:tb8 1 9 l:txb8 'iVxbS
also failed to inspire: 1 2 ... fxe6 13 dxe6 lt:)b6! (the extra pawn on the c-file is irrelevant) 20
(Black should avoid the greedy 13 ... �xc3+ 14 1Vc2 h6 21 lt:)h3 1i'e8 22 �d2 .i.a4 23 'iVct
.i.d2 when he might win more material than in � 24 ft:)£2 ft:)x£2 25 �£2 'iVe7 26 �gt �8
the game but also allow White more attacking 27 'iVbt '1>-11> Forintos-Giigoric, Ljubljana
chances) 14 0-0 �xe6 1 5 lt:)g5 �d5! 16 1i'c2 1 969.
'iVf6. Not only is Black a pawn up but he also b) t t ...lt:)g7!? This time Black hopes to block
has active pieces. f5 with pieces but it does allow White to sacri-

1 55
Pla y th e King 's Indian

lice a pawn to open the position. 12 fS!? (the of his f-pawns, e.g.
quiet 12 ..te3 lbd7 13 ..tf2 lLlf6 doesn't cause a) t 2 lLlh4 lLle4 (the usual response to lLlh4)
Black any problems - he will follow up with 13 lLlxfS gxfS 14 ..txe4 fxe4 with about equal
...l:r.e8 and either bishop or knight to £5; this srill chances.
may be White's best approach as the game is b) 1 2 'W'c2 lLla6 1 3 a3 lLlc7 14 ..txfS gxfS (it
about level) 12 .....i.xf5 1 3 ..tf4 'W'e7 14 1i'd2 f6! is better for Black to keep his queen as it de­
(Black plugs up the holes on the dark squares fends fS) 1 5 ..i.d2 bS with an unclear game.
and plans to block the e-file with a knight on 1 2 �6 1 3 al
...

eS) 1 5 l:r.ael lbd7 16 ..i.dt lLleS. It is hard to A standard reaction to prevent ...lLlb4.
believe that White has enough compensation 1 3 . . .ltlc7 1 4 g4
for a pawn. After 1 7 lLlxeS Black will recapture
with the d-pawn in order to keep his king as
safe as possible.
1 0 0-0 :es 1 1 .i.dl
Black plans to meet 1 1 lLlh4 with t t ...lLle4!
not fearing the shattering of his kingside pawn
structure. In return he will increase his grip in
the centre, e.g. 1 2 lLlxfS gxfS 1 3 lLlxe4 fxe4 14
..i.e3 ..i.xb2 15 l:r.bt 'W'f6 16 'itb3 ..i.d4 1 7 ..txd4
1i'xd4+ with at least equal chances for Black,
Antoshin-Boleslavsky, Leningrad 1 956.
1 1 ...'ii'd 7!

White had prepared this advance with his

1 2th move. It srill doesn't mean that it's good.
Another game, Peng Zhao Qin-J.Polgar, Novi
Sad Olympiad 1 990 went instead 14 'W'c2 bS! (a
thematic advance - swapping the b-pawn for
the d-pawn is a good deal for Black as it allows
her to take more control in the centre and bring
the rather passively-placed knight on c7 to life)
1 5 cxbS lLlfxdS 16 tLlxdS lLJxdS 1 7 ..i.x£5 gxfS
1 8 l:r.bt lle4! (the black pieces dominate the
centre of the board) 19 l:r.dt lLlb6 20 b4 'W'xbS!
21 :Xd6 c4! (now the passed pawn decides the
Black doesn't cede any ground in the centre. game very quickly) 22 'W'£2 c3 23 'W'g3 'W'e2 24
Instead t t ...lLle4 t 2 lLlxe4 ..i.xe4 1 3 ..i.xe4 l:r.xe4 lLles c2 0-1 .
14 lLlgS! allows White good attacking chances - White resigned as, after 2 5 l:r.a 1 :XeS! 26
he is not worried about losing a pawn if he can fxeS 1i'xe5, Black will emerge a piece to the
get f5 in as then both his rook and bishop can good.
participate in an attack against the black king. 1 4. . ..i.xg4! ?
The key to this variation for Black is control An interesting though not necessarily coura­
of the e4-square and as the above variation geous decision. The point is that it is very easy
demonstrates controlling a square does not for a grandmaster to calculate that the sacrifice
necessarily mean occupying it. guarantees at least a draw. The difficult decision
1 2 hl comes later when he has to decide whether to
White plans to lift Black's blockade by play­ take the perpetual check or gamble for more. In
ing g4; risky, as we shall see. On other moves fact this position was already known to theory
Black is again not worried about the doubling at the rime of this game but no-one had sug-

1 56
 The Fo ur Pa wns A t tack

gested this obvious piece sacrifice. Perhaps I'm not convinced that Mestel's decision to
because Black has another satisfactory con­ spurn the draw was objectively correct, but as
tinuation: 14 ... .ixd3 1 5 1i'xd3 b5! 16 cxb5 the saying goes, 'Fortune favours the brave'
l:r.eb8 17 a4 a6 with good Benko style compen­ 20 . . . bxc4 21 �xc4 .l:le7!
sation for the pawn. Black prepares to double rooks on the e-file.
1 5 hxg4 'irxg4+ 1 6 �h2 'irh5+ 1 7 �g2 22 'ird3
'it'g4+ 1 8 �h2 22 llJg5 is not as powerful now as Black can
White cannot avoid the draw as 1 8 �h 1 avoid the queen exchange by 22 ... h6! as 23
'ifh3+ 19 tDh2 lbh5 lands him in trouble. The 1i'xh5 lbxh5 is now check. After 23 lDh3 1i'f5
moment of decision has arrived for Black. 24 .id3 1i'd7 25 'iV£3 l:r.ae8 Black has excellent
1 8 ... b5!? compensation for the piece but White has bet­
ter chances of organising a successful defence
than in the game.
22...�h6 23 �g2
White was worried about 23 ... .i.xf4+ 24
.i.xf4 'iVg4+ so he makes this rather sad admis­
sion of his earlier mistake.
23 ...l:lae8 24 �d2 �xf4!!

Mestel plays for the win, but with a most

surprising move on the queenside. We saw a
similar idea in the notes to move 14 - Black
wants to weaken White's protection on d5 as if
this pawn drops then more black pieces can
come flooding into the attack.
19 .l:lg1 'irh5+ 20 �g3?1
Straightaway Black is rewarded for his brav­ With the investment of a second piece the
ery. Conquest chooses the wrong square for his Black attack becomes irresistible.
king. He should have played 20 �g2 when after 25 �xf4 'irg4+ 26 �g3
20... bxc4 21 .ixc4: This is why the White king had retreated
a) Look what happens if Black tries to play from g3 - so he would have this defence. How­
the same way as with the king on g3: 21 ...l:r.e7 ever, it is still not enough to save him.
22 lbg5! (there is another interesting tactical 26 ....1:1&3 27 'irf1 ltJh5!
shot, 22 lbe5!?, but the text is probably Stronger than 27...'ifxc4.
stronger) and the queens arc forced off as 28 �2
22...1i'h4 is met by 23 l:r.ht . After 22 ... l:r.ae8 23 There is no defence, e.g.
'i'xh5 lbxh5 Black still has some compensation a) 28 �h1 lbxg3+ 29 l:r.xg3 'ifxg3 30 llJh2
for the piece but the exposed position of the l:r.e 1 and Black wins.
white king is not so crucial without queens on b) 28 'iV£2 lDf4+ 29 �h1 l:r.x£3 (there are
the board. other wins as well but this is the most spectacu­
b) It is also to late for Black to take a draw. lar) 30 .i.h2 (or 30 'ifh2 l:r.xc3! 31 bxc3 1i'f3t 32
After 21...1i'g4+ 22 �h1 'it'h3+ 23 lDh2 the l:r.g2 lbxg2 33 1i'xg2 'ifxc3 and Black wins back
move 23 ...lbh5 is not so strong as White can his piece and emerges countless pawns to the
defend with 24 11'£3. good) 30...'it'h3 31 'ifd2 l:r.e2!! (a brilliant inter-

157
Pla y th e King 's Indian

ference combination which cuts the queen's The move 9 �e2 also has to be examined. It
line of defence along the 2nd rank) 32 �xe2 (or is rarely played here, but the position can be
32 1i'xe2 lt'lxe2 33 �xe2 l:.f2) 32... 1i'xh2+!! 33 reached via the alternative move order 1 d4
'iti>xh2 l:th3 mate. lt'lf6 2 c4 g6 3 lt'lc3 .i.g7 4 e4 d6 5 f4 0-0 6 �e2
28 .. Jixf3 29 1i'h3 1Wxc4 0-1 c5 7 d5 e6 8 dxe6 fxe6 9 it'J£3. Black now has a
Black is three pawns up with a crushing at­ good solid plan which promises him a level
tack. game, namely 9 ..lt'lc6 10 0-0 e5!?. This advance
.

e5 has two advantages; it liberates the bishop

Game 58 on c8 and reinforces Black's control of d4.
These two plusses outweigh the fact that the
Christiansen-Kasparov
move concedes control of d5 - this square can
Moscou1 lnterzona/ 1982 now be occupied by a white knight. After
10...e5, the game Feletar-Jurkovic, Pula 1 995
1 d4 lt:lf6 2 c4 g6 3 lt:lc3 �g7 4 e4 d6 5 f4 continued 1 1 fxe5 dxe5 1 2 1i'xd8 l:.xd8 1 3 �g5
0-0 6 lt:lf3 c5 7 d5 e6 8 dxe6 l:.£8 1 4 �d3 �g4 with about equal chances.
In the previous games White played 8 �e2. 9 lt:lc6 10 0-0 lt:ld4!
..•

This time he prefers to exchange in the centre There is another plan based on playing
himself. A very different son of position will 1 0...a6 foUowed by ... b5 but it is most logical for
arise where White hopes that his active pawns Black to immediately occupy the imponant
on e4 and f4 will enable him to develop a king­ central outpost. The position after 1 1 lt'lxd4
side attack. The main drawback to 8 dxe6 is cxd4 1 2 lt'le2 e5 would be favourable for Black
that it allows the knight on b8 easy access to so White must come up with another plan.
White's main weakness - the hole on d4 - as 1 1 lt:lg5!?
there is no longer a pawn on d5 to prevent
...it'Jc6. First Black must decide with what to
recapture on e6.
8 ...flle6!

The grandmaster conducting the white

pieces in this game, Larry Christiansen, is weU
known for his attacking instincts. His idea is to
foUow-up the text with f5 or e5 and then attack
8 ... �xe6 is less convincing as it encourages the black king. If Black decides to chase the
White to play £5, just the move he needs to play knight with 1 1 ...h6 it will just return to £3. Al­
to obtain good attacking chances. For example, though White will have then wasted a tempo
after 8 ... �xe6 9 �d3 lt'lc6 1 0 f5 �d7 1 1 0-0 this pales into insignificance when compared
White already has a healthy initiative. By playing with the lasting damage created in front of the
8 ... fxe6 Black keeps more control in the centre black king. The pawn on g6, in particular,
and makes it harder for White to develop an would become a target for white pieces (the
attack. bishop on d3 after e5 or a knight on h4). As we
9 �d3 shall see Black has plenty of defensive and

1 58
 T h e Fo ur Pa wns A t tack

counter-attacking resources without having to �xg5 but he finally concluded that 1 4...�e6 15
resort to the immediate ... h6. Kasparov, him­ tbd5 �xd5 1 6 exd5 e4! (to open lines for the
self, has suggested the quieter 1 1 lbe2 as an black pieces) 1 7 �xe4 'ii'e7 1 8 ..id3 l:.ae8 was
alternative while a more recent game featured a good for him.
sort of deferred Christiansen plan. It is worth a 1 3 ...gxf5 1 4 exf5 b5!
look as it was quite spectacular.
Petronic-Petrovic, Yugoslavia 1 995 contin­
ued 1 1 1i'h1 �d7 1 2 ibg5 'ii'e7 13 e5!? (White
goes all in; 1 3 'ii'e 1 also comes into considera­
tion) 1 3...dxe5 14 fxe5 tbh5 1 5 g4 l:.xft+ 1 6
..ixft ? (16 'ii'xft was better) 1 6... h6! 1 7 tbge4.
White has achieved what he set out to do - trap
the knight on h5 - but the price he has paid is
too high. He has exposed his own king and
Black now punishes him for this extravagance.
1 7 ... ..ixe5! 1 8 gxh5 'ifh4 19 h3 l:.£8 20 �g2
..ic6. Just look at the black pieces. Every single
one of them (if we don't count the king and
pawns) is playing a major role in the attack.
White is a piece up but Black's extra firepower By sacrificing a pawn on the queenside Black
in the critical area of the board must have made hopes to take total control of the centre. Inci­
him feel like he was a piece down. 21 �e3 l:.f3!! dentally, this is the reason he was willing to
(Black takes aim at h3; he is not afraid to sacri­ weaken his kingside with his previous two
fice more material to strip the white king bare) moves. Now 1 5 cxb5 d5 looks terrible for
22 'ii'e 1? (the best chance was to take the rook White and 1 5 b3 is strongly met by 1 5... b4 (ac­
though extensive analysis has shown that the cording to Kasparov) so Christiansen looks for
position after 22 �xf3 'ii'xh3+ 23 �g1 tbx£3+ a tactical solution to his problems.
24 1i't2 tbd4! is very good for Black) 1 5 ..te3?!
22 ...l:.xh3+! 23 �g1 l:.hHl 24 �xh1 'ifh2+ 25 Kasparov thought White's best chance was
�ft 'ii'xh1+ 26 �t2 'ifh2+ 27 'i!?ft 'ii'h3+ 28 15 tbxb5! tbxb5 1 6 cxb5 d5 as the pawn centre
�t2 �g3+! 0-1 (29 tbxg3 'ii'g2 mate). is not quite as dangerous without the support
1 1 . . . e5! of the knight on d4. Still, this doesn't look like
We have already seen such a reaction by much fun for White.
Black in the notes to move 9. As well as liberat­ 1 5 bxc4 1 6 ..txc4+ �h8
•..

ing the bishop and increasing Black's control A player of Kasparov's class is not going to
over d4 the move also prevents White from fall for the rrap 1 6...d5? 1 7 tbxd5! tbxd5 1 8
advancing e5. �xd4 cxd4 1 9 'ii'b3 when White regains the
12 f5 pinned knight.
Now that e5 has been ruled out this is the 1 7 .i.xd4 cxd4 1 8 ltld5 ..ta6!
only way to continue the attack. With hindsight An excellent move. The point is that after 19
it would probably be better to play the feeble 1 2 �xa6 tbxd5 the black knight will invade on e3.
fxe5. White prefers to give up material for some
12 ...h6! vague attacking chances rather than to play
1bis is a good moment to attack the knight. such a position.
Black has a concrete reason for playing ... h6. 1 9 ltlxf6 ..txc4!
13 ltlh3 Now Black wins the exchange as he attacks
White would prefer to go back to f3 but this both knight and rook
interferes with the protection of his f-pawn. 20 ltlh5 .i.xf1 21 Wg4!
During the game Kasparov was slighdy con­ The best chance. Black must deal with the
cerned about the piece sacrifice 13 fxg6 hxg5 14 mate threat on g7 so White at least gets to acri-

1 59
Pla y t h e King 's Indian

vate his rook.

21 11t'd7 22 .z:l.xf1
..•

A strong manoeuvre. The rook looked well

placed on c1 but Kasparov understands that it
At first glance it appears that White has quite can do even more damage on the e-file.
good compensation for the exchange. Kas­ 33 �d3 .z:l.e5 34 lL\xg5 hxg5 35 .z:l.f2
parov, though, has seen deep into the position The power of the rooks is well illustrated by
and realised that by using his d-pawn as a decoy the following variation: 35 �xd2 lle4 36 h3
he will be able to break White's grip on the llf4 37 llgt llf3. White is almost falling off the
kingside. edge of the board.
22 . . . d3! 23 11i'f3 d2! 24 g4 35 ....z:l.a4 36 h3 .z:l.e3+! 37 �xd4
I f White drops the f-pawn he will be just los­ After 37 o!Dxe3 dxe3 Black will soon have a
ing. 24 f6 looks tempting but after 24 ... ..i.xf6 25 new queen while 37 �xd2 llxh3 is equally
o!Dxf6 'iVg7! Black wins back the knight (check it hopeless for White in the long run. Rook
for yourself; don't forget the pawn on d2). Ths against knight is usually a massive advantage in
is the son of linle tactic that strong players arc the endgame.
always using to control the game. Even when 37 . . ..z:I.Be4+ 38 �d5 .z:l.e2 39 .z:l.f3 .z:l.e1 40 f6
things look really quiet there is usually plenty .z:l.f4 0-1
bubbling just beneath the surface. The white pawn can be stopped by the black
24. . ..z:l.ac8! king; nothing will stop the pawn on d2. In fact
Black plans 25 ...llc1 the performance of the black d-pawn can be
25 11t'd3 11i'a4! 26 lLlf2 compared to that of a pacemaker in a long dis­
White must defend the pawn on g4 but now tance race who was supposed to drop out a few
Black gets the chance to exchange queens. laps before the finish, but instead continued
Don't forget that the decoy pawn is still alive and won.
and extremely dangerous. Any pawn which
reaches the 7th rank should be treated with Game 59
respect.
Soppe-Panno
26 ...11i'd4! 27 11Vxd4 exd4
Tripled pawns arc not usually very good and Buenos Ains 1999
it's not because they are tripled that they arc
good here; it's because two of them are ad­ 1 d4 lLlf6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5 f4
vanced passed pawns. 0-0 6 lLlf3 c5 7 dxc5
28 lLlf4 .z:l.feB 29 lbe6 .z:l.c1 30 lbd1 �f6! The previous games all featured White
The bishop plans to suppon the d2 pawn blocking the centre with 7 d5 but this time we
from g5. I f White wants to remove it he has to shall examine lines where the black c-pawn is
give up his powerful knight on e6. exchanged for the white d-pawn. In the main
31 �2 �g5 32 �e2 .z:l.c5! game White exchanges on c5 while in the notes

1 60
 The Fo u r Pa wns A t tack

we shall look at 7 �e2 cxd4 8 lL!xd4. The re­ Black) 1 1 ...dxe5 1 2 ..ic5 l:.e8 1 3 1i'xd8 l:.xd8 14
sulting pawn structure is the same in both cases 0-0 l:.d2 1 5 l:lad1 l:.xd1 1 6 l:.xd1 ..ie6= Uhl­
(after 7 dxc5 Black plays ...1i'a5xc5) but 7 dxc5 mann-Fischer, Varna Olympiad 1962. This
contains more venom as White gains time at­ variation alone is enough to explain why 7 ..ie2
tacking the black queen. is unpopular.
After 7 �e2 cxd4 8 lL!xd4 b) For those of you who have no desire to
exchange queens so early then 8...lL!a6 is a good
alternative. The idea is to pressurise the white e­
pawn with ... lL!c5 and very often to follow up
with the advance ...e5. White has:
b1) 9 ..ie3 lL!c5 10 �£3 ..ih6! (by pinning
the pawn on f4 Black increases the power of
the corning advance e5) 1 1 0-0 (1 1 1i'd2 e5 1 2
lL!de2 exf4 1 3 lLlxf4 l:lc8 is quite good for
Black) 1 1 ...e5 1 2 lL!db5 lL!e6 1 3 'ii'xd6 a6 14
�b6 'ii'd7 1 5 lL!d5 (1 5 'ii'e5 is interesting but
the complications after 1 5 ...'ii'c6! are not unfa­
vourable for Black) 1 5 ... lL!xd5 1 6 1i'xd7 (per­
haps it would be better to just play 16 exd5
axb5 17 'ii'xd7 ..ixd7 1 8 dxe6 ..ixe6 19 cxb5
Black has two equally valid approaches. He ..ixf4 with an unclear endgame) 16 ...lL!xb6!
can aim for exchanges with the idea of steering (hardly a sacrifice as Black gets 3 pieces for the
the game into a roughly level endgame (line 'a') queen) 17 'ii'e7 (1 7 1i'd6 lLlxc4 1 8 1i'b4 axb5)
or, if he prefers, he can direct play towards a 17 ... axb5 1 8 f5 ..ig5 1 9 1i'b4 lL!d4 20 c5 lL!c4
complex rniddlegame where his chances are no 21 b3?? (Black has the advantage but there is no
worse than White's (line 'b'). need to lose the queen in 1 move) 21.....id2!
a) 8...lL!c6 9 ..ie3 and now: 0-1 , Skotdad-Lesiege, Parthenay 1 992.
at) 9.....ig4 is played quite often but after 10 b2) 9 ..i£3 is an alternative. Boleslavsky sug­
lLlxc6! (1 0 �xg4 lL!xg4 11 'ii'xg4 lL!xd4 is the gested that Black play 9 ... ..ig4 10 �e3 1i'c8
tactical justification of Black's 9th move; in fact, while other sources offer 9 ... lL!c5 10 iLlb3
this is a very common theme in such positions) iLlfd7. Both look OK for Black. These lines
10 ... �xe2 1 1 lLlxd8 ..ixd1 1 2 l:.xdt l:lfxd8 1 3 have received very little attention in tournament
�e2 l:tdc8 the temporary pawn sacrifice 1 4 c5! practice.
ensures White of some advantage. For example, 7 'ii'a 5
...

14...dxc5 1 5 e5 lL!e8 1 6 l:.d7 l:.c7 1 7 l:.xc7 If it wasn't for this move then 6...c5 would
lLlxc7 1 8 ..ixc5 or 1 4...l:.c6 1 5 .cxd6 exd6 1 6 not be the main line against the Four Pawns
� £3 when in the first case White is more active Attack. Few Black players are willing to defend
and in the second he has the healthier pawn the passive position that arises after 7 ... dxc5 8
structure. The best that Black can hope for in 1i'xd8 l:.xd8 9 e5.
such positions is a draw and he will have to 8 .i.d3
suffer to achieve it. 8 cxd6 lL!xe4 is not good for White, but 8
This rope-a-dope strategy may be a legiti­ ..id2 is OK even though it is hardly ever
mate tactic for the titans defending the black played. After 8 ...1i'xc5 9 b4 1i'b6! (9 ... 1i'xb4 10
pieces in world championship matches but I lL!d5 is dangerous for Black) 1 0 ..id3 ..ig4 1 1
cenainly don't recommend this approach for l:lb 1 lL!c6 the chances are about level.
the rest of us. B 'ii'xcS 9 'ii'e2
...

a2) 9...e5 is a better idea. This also leads to White plans ..ie3 and 0-0. Much of the play
an endgame but one where Black has more will now revolve around the battle for the d4-
prospects. For example, 10 lLlxc6 bxc6 1 1 fxe5 square. Black will play moves such as ...lL!c6,
(1 1 0-0 exf4 1 2 �xf4 l:le8 is quite good for ...�g4xf3 and ... lL!f6-d7, all of which increase

161
Pla y the King 's Indian

his control of d4, while White will try and de­ tion is 1 3 ... lDc5 14 i.b1 lDa4 while another is
fend it with i.e3, 'if£2 and a rook to the d-file. 1 3... i.xf3 1 4 .l:lx£3 i.d4!?. Cifuentes-Roeder,
9 . lbc6 10 .b3 Wa5
. . San Sebastian 1 997 now continued 1 5 'iti>h 1
10 ... 1i'h5 is not so popular as at the end of ..i.xe3 1 6 'ifxe3 'ifb6 17 'ifd2 'ifd4 18 lDe2
the tricky line 1 1 h3 lDg4 12 i.d2 lDd4 13 'iffl 'ifg7 with a roughly level game. Note how the
(13 tilid4 1i'h4+) 13 ... lDxf3+ 14 'ifxf3 i.d4 black queen has taken over the role of the
White can force the exchange of queens with bishop on the long diagonal.
1 5 'iti>e2!. His extra space should then give him a 1 3 . . .ttJc5 1 4 .i.b1 .i.xf3 1 5 gxf3
slight plus in the ending though this line is still Recapruring with the queen would allow
perfectly playable for the unambitious Black Black to occupy the d4-square. Besides, White
player. is happy to take with the pawn as he hopes to
1 1 0-0 .i.g4 1 2 :c1 develop an attack along the semi-open g-file.
Inexperienced White players sometimes pre­ 1 5 . .ttJa4
.

fer to immediately advance their queenside

pawns. This can lead to trouble. For example,
Terzic-Cvitan, Bosnia 1 999 continued 12 a3
lDd7 13 b4 'ifhs 14 'ifd2 (14 .l:lc1 aS 1 5 bS
lDd4 16 'if£2 lDcS! 17 ..i.b1 i.xf3 18 i.xd4
..i.xd4 19 'ifxd4 lDb3! also cost White an ex­
change in another game) 1 4 ... i.xf3 1 5 .l:lx£3
lDd4 1 6 .l:lh3? (the only move that doesn't lose
material - Black is also threatening ...'!i)b3 - is
the unpalatable 1 6 ..i.xd4) 1 6 ... 'ifxh3! 1 7 gxh3
lD£3+ and Black won the exchange and the
game.
1 2 ...�7

In general it is usually a good idea for Black

to swap of the white knight on c3 in the King's
Indian. This is true here as without the knight
Black will find it easier to develop an initiative
on the queenside and neither will he have to
concern himself with White playing an annoy­
ing lDdS.
1 6 ltJxa4
White can avoid the exchange but only with
the retreat 1 6 lDd t !? This is not so srupid as the
black knight on a4 is not ideally placed. Vegh­
Petrov, Gausdal 2001 continued 1 6....1:lac8
(16...lDc5 or 1 6 ....1:lfc8 are other options) 1 7 a3
This move is an integral part of Black's strat­ 'ifhs 1 8 b4 eS!? 19 'iti>h 1 lDd4 20 fxeS dxeS 21
egy. The King's Indian bishop's diagonal is cS with chances for both sides.
uncovered while the knight heads off towards 1 6 ...1i'xa4 1 7 h4
the queenside where it may try and exchange An extremely aggressive reaction. For
itself in order to minimise the effects of Black's White's kingside attack to work he does need to
space disadvantage. play moves such as h4-h5 and f5 in order to
1 3 "ii'f2 weaken g6 and to open more flies. He usually
White unpins and plans to recaprure on f3 plays a couple of preparatory moves before
with the pawn (see below for explanation). An launching his attack. The text looks premarure
alternative is 13 'ifd2!? when one possible reac- as White still has very few pieces ready to par-

1 62
 The Fo u r Pa wns A t tack

take in an attack. A couple of examples where ensured that the white attack is over even be­
\X'hite's play was more measured: fore it got started. Note how Black hasn't
a) 1 7 b3 'it'a3 1 8 c5 dxc5 19 �xeS (19 rushed to occupy the outpost on d4. In fact the
:XeS!?) 19...'ifxc5! 20 l:xc5 �d4 21 l:dt square he really wants for his knight is e5 as this
�xt2+ 22 �x£2 l:fd8 23 l:cd5 e6 24 l:xd8+­ is nearer the kingside where the action is taking
l:xd8 25 :XdS+- �xd8 with a level endgame, place.
Topalov-Kasparov, Iinares 1 994. The bishop is 28 fxa5 l!Jxa5 29 11'f4 g5! 30 'il'd2 l:lad8 31
better than the knight but Black's superior l:lh2 11'f8 32 h5?! g4 33 'il'f4 11'f7 34 lt>g2
pawn structure compensates. l:lg5 35 h6 11'c7 36 l:ld1 b5!
b) 17 l:fd l l:ac8 1 8 b3 'it'a5 19 l:d5 'it'c7 20 \X'hite was hoping to save himself by block­
l:cdl b6 21 a3 l:fd8 22 h4 e6 23 l:g5 1i'e7 24 ing everything up but by means of a pawn sacri­
h5 'iff6 25 hxg6 fxg6 with a roughly level game, fice Black can infiltrate via the c-file.
Topalov-Dolmatov, Elinite 1995. \X'hite is now held to account for recklessly
17 � 18 b3 .i.h6!?
. . . advancing the pawns in front of his own king.
37 cxb5 11'c3 38 l:lhh1 11'f3+!
The exchange of queens doesn't mean the
end of the attack. In fact, with the open g-file
and pawn on f3 it becomes even stronger.
39 11'xf3 gxf3+ 40 lt>f2 l:lg2+ 41 lt>a3 l:ldg8
42 .i.d3 l:l8g3 43 .i.c4 f2+ 44 \t>a2 l:lg1 45
l:lh2 l!Jxc4?!
45 ...�g4! would have cost White a whole
rook. Now Black just wins a piece but it is still
good enough to win the game. The remaining
moves were: 46 �x£2 l:xdl 47 �xg3 :d3+- 48
�h4 �2 49 :gz �& so �hs l:d2 st :g3
l:h2+ 52 �g4 �St 53 �f4 :h4+ 54 �e3
l:xh6 55 �d4 l:h2 56 l:c3 l:d2+ 57 �e3 :Xa2
Often in such positions Black will play ...a5- 58 l:cS+- �g7 59 l:a8 l:a3 60 b6 :Xb3+- 61 �d4
a4 in order to create a weakness on the queen­ l:d3 mate.
side and to open the a-file for his rooks. Here,
though, with the white attack already underway Summary
such a policy was deemed too slow. The idea 1) I have no magic wand to wave here. I am
behind the text is to play ... e5 and exchange off just suggesting that Black enter the main line
the dark-s�uared bishops. If Black can then with 6 ... c5 and follow its most reliable branch
patch up the holes around his king his superior 9...�g4.
minor piece (his knight is much stronger than 2) 9...�g4 is supposed to be the solid option
\X'hite's light-squared bishop) should give him but it still leads to incredibly sharp positions.
the better game. The theory is more important here than in
1 9 �h1 aS 20 f5 .i.xa3 21 11'xa3 11'b61 many other lines as a small slip can cost one the
Black knows that \X'hite won't exchange game. So if you are trying to learn the King's
queens he is looking to transfer his own queen Indian, make this line one of your priorities.
to the kingside. White's attack will only be dan­ 3) Games 57-59 tackle variations that are not
gerous if Black ignores it. quite as popular at they used to be, but they
22 11'h6 'il'd8 23 l:lc2 'it>h8 24 l:lg2 l:lg8 25 have all enjoyed their moments in the sun and
f4 11'f8! 26 'il'g5 f6 27 11'g4 11'h6 may well do so again. 7 dxc5 of Game 59 is the
By some accurate defensive play Black has most difficult for Black to deal with.

1 63
 CHAPTER TEN I
White Plays an early h3

1 d4 lLlf6 2 c4 g6 3 lLlc3 ..ig7 4 e4 d 6 5 hl pending on whether White plays iD£3 or not.

This is a popular way of meeting the King's
Indian amongst White players looking for an White plays 6 lLlfl
uncompromising struggle. The move h3 has two 1 d4 lLlf6 2 c4 g6 3 lLlcl ..ig7 4 e4 d6 5 hl
main ideas. The first is to pave the way for .i.e3 0-0 6 lLlf3
without having to worry about been harassed by The move order 5 iD£3 0-0 6 h3 is at least as
...�. Very often the bishop will go to gS first frequent.
but if Black attacks it with ... h6 then it drops 6 ...e5
back to e3. The second idea behind h3 is to sup­ There are several move orders Black can
pon the advance g4. This is played both ro gain employ to reach the desired set-up. Most of
space on the kingside and to dissuade Black from them quickly converge. l:or example, 6...lDa6 7
playing ... fS. The intentions behind h3 are similar .i.gS eS 8 dS takes us back into Game 60.
to those behind £3 in the Samisch Variation but 7 d5
the two variations usually lead to quite different 7 dxeS dxeS 8 'ifxd8 .l:.xd8 9 .i.gS .l:.e8 is
types of games. In h3's favour is the fact that very similar to the Exchange Variation (Chapter
White can st:ill develop his knight to £3 and that 6). In the line that I am recommending for
the dark squares are not funher weakened (the Black it doesn't really make any difference that
move £3 opens the diagonal gl - a7 upon which White has played h3 instead of .i.e2.
White's king is often to be found) but, on the 7 ...lLla6
other hand, h3 does nothing to protect the all
imponant e4-square.
My main recommendations are centred
around Black playing the traditional ...eS in con­
junction with the modem ...lba6 which has, in
fact, been the main line for many years. The
struggle can be very sharp and both players can
attack on either wing (the centre is almost always
blocked). White tends to win games by taking
control over the crucial e4-square, whilst Black's
victories usually occur when he achieves the
advance ... e5-e4 or when White has neglected the
safety of his king (or a combination of both).
The material is split up into two sections de-

1 64
 White Pla ys an Early h 3

and now: to achieve ...f5 . For example, 8......e8 9 g4 lDd7

a) 8 .i.g5 is the subject of Game 60. 10 llg1 �h8 1 1 ...d2 lDdc5 1 2 liJh4 and f5 is
b) 8 .i.e3 of Game 61. not such a good idea anymore (12...c6 is still
As I recommend that Black meets 8 .i.g5 reasonable).
with 8...h6 9 .i.e3 the positions are very similar. 9 ..ie3 ltJc5
The difference between having the pawn on h6 9...llJh5, which is a popular line with the
and h7 will be explained later. The games below pawn back on h7, is not so good here as White,
are model games for Black. Life will not always at some point, will be able to gain an important
be so smooth but I have a lot of faith in the tempo with 'ifd2.
systems I'm proposing. 10 ltJd2 a5
The knight's position on c5 must be secured
Game 60 against b4.
1 1 g4
Barsov-Gallagher
A quieter line is 1 1 .i.e2 .i.d7 (on the imme­
Calcutta 2001 diate 1 1 ...lDh7 White may try 1 2 h4!?; ....i.d7 is
a useful developing move if only because the
1 d4 lt'Jf6 2 c4 g6 3 lt'Jc3 ..ig7 4 e4 d6 5 queen's rook is now more in the game) 12 0-0
lt'Jf3 0-0 6 h3 e5 7 d5 ltJa6 (12 g4 would of course transpose into the main
The actual move order of this game was game) 1 2...llJh7!? with ... t7-f5 to soon follow.
7...a5 8 .i.g5 lDa6 9 g4 llJc5 10 llJd2 h6 1 1 White has tried many moves now. Here are a
.ie3. As the knight is just en route to c5 it few examples:
would have been equally possible to get there a) 13 'ii'c2 l[Jg5!? 1 4 h4 llJh7 1 5 h5 liJf6 (of
via d7 (i.e. 7 ... liJbd7). Black's last five moves four have been with this
8 ..ig5 knight and it has ended up where it started; in
return for this scandalous waste of time he has
lured the white h-pawn forward and White is
about to take back his move ...c2) 16 ...d1
gxh5 17 .i.xh5 liJd3 18 .i.£3 ...c8 19 ...c2 lLlf4
20 c5 dxc5 21 .i.xc5 l:le8 22 liJe2 .i.b5 23 llJc4
llJ4xd5! with a good game for Black, Beliavsky­
Sher, Bern 1 995.
b) 13 l:le1 f5 1 4 exfS gxf5 15 f4 exf4 16
.ixf4 ...f6 1 7 llJ£3 l[Jg5 1 8 l:lct l:lae8 with an
active game for Black, Jokovic-Tratar, Bled
2003
c) 1 3 �h2 liJg5 1 4 f3 ...e7 1 5 liJb3 liJxb3 16
'ifxb3 b6 17 'ifc2 f5 1 8 exf5 gxf5 1 9 f4 exf4 20
.i.xf4 llJe4 21 .i.d3 and White was fractionally
8 h6
... better although he over-pressed and lost in
White played .i.g5 in order to encourage this Krasenkow-Radjabov, Wijk aan Zee 2003. I
move which is often a liability for Black in the don't see why Black didn't just play 14 ... f5.
King's Indian. Here, though, I am suggesting 1 1 c6
. . .

that we fall in with White's plans as it is possi­ A useful rule of thumb in these lines is to
ble to play a system where h6 is actually quite play ...c7-c6 only in reply to White's g2-g4.
useful. Once White has played g2-g4 his king won't be
You may well ask why not just play 8......e8 very safe on the kingside so he will usually be
as we do when White delays liJ£3 (sec Games looking to castle queenside (this game is the
62-63) Well, that is also perfectly playable but exception that proves the rule). It is therefore a
the slight difference in the position (llJ£3 in­ good idea for Black to be ready to open lines in
stead of .i.d3) means that it is harder for Black this sector of the board at a moment's notice.

1 65
Pla y th e King 's Indian

Playing ...c7-c6 is quite safe but Black must be White may pick up a pawn but his king has no
more careful about taking on dS. I f he mistimes home.
this exchange then the white pieces may be­ c) 14 gS is probably White's best move. Af­
come very active on the queenside (after White ter 14 ...hxgS t S hxgS (better than t S ..ixgS 1VaS
recaptures with cxdS a diagonal is opened for 1 6 1i'b 1 cxdS 1 7 cxdS bS 1 8 a3 .l:.tb8 with an
his light-squared bishop and the knight on d2 exceUent game for Black, Chiburdanidze-Nunn,
may be able to jump into c4). A second rule of Linares 1 988) tS ...lLlh7 we have:
thumb for Black, therefore, is to only play
... cxdS when you can foUow up with ... b7-bS.
1 2 .i.e2 .i.d7
The plan is to play ...aS-a4 and ...1i'aS to de­
velop a queenside initiative.
1 3 0-0!?
I was completely taken aback by this move, a
novelty my opponent had prepared beforehand.
StiU, it was not an unpleasant surprise. I just
assumed that my opponent had lost his mar­
bles. Who in their right mind would casde king­
side in such a position?
The main move is 1 3 h4 and after 1 3 ... a4
there is:
a) 14 hS (It is not such a good idea for White cl) t6 tLlf3 cxdS 1 7 cxdS 1VaS 1 8 1i'd2 .l:.fc8
to ignore the threat of ... 1VaS) 1 4 ... 1VaS (threat­ 1 9 �ft bS Black had active play in Akesson­
ening ltlxe4 as one of the white knights wiU be Gallagher, Istanbul Olympiad 2000.
pinned) t S f3 a3! (and this was the other threat; c2) 1 6 .l:.gt 1VaS 17 1Vbt !? (ugly but it does
ideaUy White would like to meet this move with defend against the threat of ... a4-a3 and by de­
1 6 b3 but this just loses the knight on c3) 1 6 fending e4 aUows White to recapture on dS
1i'c2 axb2 1 7 1i'xb2 (exchanging his a-pawn for with the knight; previously 1 7 .l:.ct cxdS 18
White's b-pawn has gained Black a positional cxdS bS had been played which is quite promis­
advantage on the queenside) 17 ... cxdS 1 8 cxdS ing for Black) 1 7 ... cxdS (I prefer to clarify
bS! 19 .l:.b 1 (The pawn could not be taken. 1 9 things straight away) 1 8 lLlxdS .l:.fe8 1 9 f3
lihbS .l:.tb8 20 a4 ..ixbS! and now 2 1 axbS 1i'd8?! 20 1Vct ! ..ic6 21 lLlbt! lLle6 22 lLlbc3
loses a rook and 21 ..ixbS lbd3+! a queen) lbd4 23 1i'd2 .l:.c8 24 �£2 .1:.£8 2S .l:.g3 with
1 9.. .l:.fc8 (It's not unusual for Black to sacrifice some advantage to White, Sadler-GaUagher,
the pawn on h6 in this fashion in the King's Bundesliga 2002 (1 -0, 40).
Indian. In return he hopes to take control of It was better to play 1 9 ... ..ic6! when 20 �£2
the dark squares) 20 hxg6 fxg6 21 ..ixh6 is met by 20...lLle6 and 20 1Vc2 by 20...bS! 21
lbd3+! 22 ..ixd3 .l:.xc3 23 ..ixg7 �xg7 24 �e2 cxbS ..ixdS 22 exdS e4 with a mess from which
(24 ..ie2 was better although Black has more Black is more likely to emerge triumphant.
than enough for the pawn after 24 ... l:.ac8) Before moving on a word about another
24 ...ltlxg4! 2S 1Vb4 .l:.xd3! 26 fxg4 (26 1fxaS move White has played here, 1 3 £3. It takes the
.l:.e3+! 27 �ft .l:.xaS 28 fxg4 .l:.xa2 is an even sting out of the ... a4 and ... 1VaS idea as the e­
worse endgame) 26 ...1fxb4 27 .l:.xb4 .l:.g3 28 pawn is now protected. 1 3 ...l:.b8!?, intending
.l:.b2 .1:.£8 29 �dt ..ixg4+ 30 �cl .l:.c8+ 31 1 4 ... cxdS foUowed by ... bS, has been played by
�bt ..ie2 32 .l:.et ..id3+ 33 �at .l:.g2 34 lLlb3 the strong grandmasters Ehlvest and Smirin .
.l:.xb2 3S �xb2 .l:.c4 0-1 J.Ivanov-V.Georgiev, 1 3 ...h5!
Salou 2000. White loses a second pawn. A There is no longer any point in continuing
wonderful King's Indian game. with ... a4 and ... 1i'aS as Black's main target, the
b) 14 ..ixcS dxcS l S lLlxa4 cxdS 1 6 ltlxcS white king, has unexpectedly disappeared to the
dxe4 is promising for Black according to Nunn. other side of the board. My main plan now was

1 66
 White Pla ys an Early h 3

to blast open the whole kingside as quickly as possibility of White playing cS (or even 1 9 gS).
possible. True, this could backfire as Black's That's better, got that off my chest.
king is there as weU but it is easier for Black to 1 9 ll'lxd6 b6!
feed his pieces into a kingside attack than it is It's always nice to sacrifice and then play a
for White. quiet move. This is the star move of the game.
Initially I was ttying to make the direct ap­
proach work (something with ...f4 and ...tlJgS)
but the problem was always the cS-square -
both for the bishop after ... f4, ..icS and, in
other variations, for the knight after lbxb7-cS.
It's weU worth spending a tempo to cut out
these possibilities. The shaky position of the
white king cannot be repaired so easily.
20 f3?!
After this White is in trouble. I was expect­
ing him to return the pawn by 20 cS. After
20... f4 21 ..i.d2 bxcS 22 ll:le4 I consider the
chances to be about level. Black also has a very
interesting sacrificial possibility in 21 ...ll:lg5 22
14 q;,g2 cxd5 1 5 exd5 f.3 ..ixg4!
White usually recaptures with the c-pawn in 20 . . .f4 21 �f2 �xg4 22 fxg4?
such positions but 1 5 cxdS is strongly met by The decisive mistake. White should have
15 ...b5! (remember the rule of thumb). By re­ tried 22 ll:le4 even if 22.....if5 with the idea of
capturing with the e-pawn he hopes to use the capturing on e4 and attacking the black king
e4-square for his knights. gives Black a good game.
1 5 ... hxg4 1 6 hxg4 ll'lh7 22 . . .'ifxd6 23 �f3 ll'lg5 24 �h4 ll'lxt3 25
It should come as no surprise to you that 'ifxf3
Black plans to play ... f5. The knight is also quite
handily placed on h7 as in some lines it can
jump out to gS.
1 7 ll'lde4
17 ..txcS dxcS 1 8 lbde4 is nothing for Black
to worry about. He just plays 1 8...b6 foUowed
by ... f5
17 ...ll'lxe4 1 8 ll'lxe4 f51
White had been hoping for 18 . .1te7 19 cS!
.

when his central attack crashes through before

Black's kingside attack is even out of the start­
ing blocks. To be honest I didn't consider de­
fending the d-pawn pawn for a second but it
did take me a long time to find the correct fol­
low up to the sacrifice. Barsov said after the game that he was to­
As an aside, a warning if you get ChessBase tally oblivious to the corning danger. He had
Magaifne or the MegaBase. They have a guy assumed that he would be able to blockade the
called Tsesarsky who annotates King's Indian passed pawns but Black's idea is just to get rid
games. The guy ruins many good games with off them to open up lines around the white
his forthright but erroneous annotations. Here king.
for example he awarded my fine move a '?' and 25 ... e4! 26 'ifxe4 .:Sea 27 'iff3?!
said that Black could gain the advantage with This is just hopeless and Barsov played it in
18 ...1tc7. Not true. He didn't even consider the a panic when he suddenly realised that 27 1tc2

167
Pla y th e King 's Indian

t3+! 28 :X£3 l:x£3 29 �x£3 'iV£8+! 30 �g2 'iVf4 The position is the same as in the previous
gives Black a decisive attack. games apart from the fact that the black pawn
27 ...l:le3 28 1i'd1 is on h7 instead of h6. This seemingly insignifi­
28 'iV£2 'iVd7 is crushing. cant difference is enough to make the way
28 ...f3+ 29 �h3 Black played in the previous game unattractive,
29 �h1 gS 30 ..i.xgS (30 ..i.£21i'h6+ 31 �g1 but at the same time introduces a new possibil­
'iVh3) 30...l:e2 wins. ity that wasn't really available to Black in the last
29 . . .lr.e2 30 �g3 �e5 game.
It's a real shame that I didn't find 1 1 . . .a4!
30......xg3+1 31 �xg3 ..i.e5+- 32 �h4 l:h2+- 33
�gS ..i.f4+ 34 �xg6 l:h6 mate
31 c5 1i'f6 32 �xeS 1i'xe5 33 l:lh1 0-1
\Vhite lost on time but it's mate in two any­
way.

Game 61
Nikcevic-Tkachiev
Cannes 1996

1 d4 lt:\f6 2 c4 g6 3 lt:\c3 �g7 4 e4 d6 5

lt:\f3 0-0 6 h3 e5 7 d5 lba6 8 .i.e3

11.....i.d7 is less promising because of the re­

ply 12 gS. It is better for Black to get on with
his queenside play at once, even if it means
sacrificing a pawn. The Russian grandmaster
Dolmatov pioneered the text in this particular
position, but it is a well-known idea from other
King's Indian variations. The point is that if
\Vhite wants to take the pawn he will have to
give up his important bishop for a knight. In
the resulting position Black's control of the
dark squares guarantee him long-term posi­
tional pressure. The reason that such an ap­
proach was not recommended in the previous
This time \Vhite doesn't bother trying to game is that Black needs to activate his bishop
squeeze ...h7-h6 out of Black but develops his via the h6-square - which was not available to
bishop directly to e3. him there.
8 ...lt:\c5 1 2 b4
8...lbh5 is an important alternative that can I have my doubts about this move as \Vhite
even be considered as the main line. However, I opens the queenside before he is ready to take
see no reason to complicate life. The plan we over the initiative there. Let's have a look at the
used against ..i.gS works perfectly well here as alternatives.
well (I did cover 8. ..lbh5 in SOKID for those a) Few strong players are willing to accept
of you who prefer that move). the pawn sacrifice. The one example I have,
9 ll\d2 Bagirov-Dolmatov, Lucerne 1993 wasn't very
9 ..i.xcS dxcS 10 tDxeS tbxe4 is not a very informative: 12 ..i.xcS dxcS 13 tbxa4 'iVaS 14
good idea for \Vhite. lDc3 ..i.h6 15 'iVc2 ..i.d7 16 l:g1 1/z-'lz. Black's
9 ...a5 10 g4 c6 1 1 �e2 next move would have been 16.....i.f4 with good

1 68
 White Pla ys an Early h 3

play for the pawn. does not have much cover.

b) 1 2 h4 'iVaS!, with the double threat of 1 8 f3
... tt:'!xe4 and ... a3, should be compared to the Of course 1 8 gxhS lL'lxhS 19 .i.xhS?? fails to
previous game. 19...tt:'!d3+.
c) 1 2 'it'c2 'ifaS 1 3 0-0-0 .i.d7 14 �b1 cxdS 1 8 ...�d7 1 9 �2 ltJh7 20 �g2 h4 21 ltJd2
15 cxdS bS 16 .i.xcS dxcS 17 l:ct l:ac8 1 8 gS •e7 22 ltJdb1 �f6!
tt:'!e8 19 .i.g4 f5 20 gxf6 tt:'!xf6 21 f3 c4 was
good for Black in Kalantarian-Dolmatov, 11-
nares 2000.
d) I mentioned that Black didn't play
1 t ....i.d7 because of the reply 1 2 gS. What
about 1 2 gS here? Well, nobody has played it
but the main difference is that Black can retreat
his knight to d7 instead of e8, which is impor­
tant as he retains control over cS.
1 2 ...axb3 1 3 axb3 .l:lxa1 14 'ifxa1 lba6!
Black prevents White from playing b4 and
plans to play cS in order to secure the useful
outpost on b4 for his knight. White's next
move re-introduces the threat of b4 but it
forces Black to play what he was going to do Note how Tkachiev, a very strong grand­
anyway. master, calmly manoeuvres his pieces to their
1 5 'ifa3 c5 1 6 ltJf1 best squares before launching his attack. He can
White should probably play 16 tt:'!a2 to pre­ do this because White has no counterplay and
vent Black's next move and to force through is reduced to fiddling about with his knight on
b4. This is the only way for White to generate the back couple of ranks (he's trying to get it to
some play in this position and is certainly pref­ c2 to exchange of Black's knight on b4 he -

erable to doing nothing as he did in the game. doesn't quite make it).
Black would reply with either 16 ... tt:'!e8 or 23 �f2?!
16 ... tt:'!d7 and then ... fS. White is trying to avoid the exchange of
1 6 ...ltJb4 bishops that Black had planned with .i.f6. This
Threatening 17...tt:'!c2+- . is because his bishop is, theoretically speaking,
1 7 •c1 h5! the better piece (his pawns are on light squares
while Black's are on dark squares) But which
bishop would you prefer in the rniddlegame?
White's passive one on f2 or Black's active one
on gS. The lesser evil was to acquiesce to the
bishop exchange.
23 ...�g5 24 ...b2 f5
At last.
25 lba3 fxg4! 26 hxg4
On 26 fxg4, 26...l:xf2+!? 27 �xf2 'iff6t 28
�e1 'iff4 would be very strong but there is an
even more decisive line for Black: 26 ...'iff6! 27
l:ft (27 .i.e1 tt:'!d3!) 27 ...tt:'!d3! (we'll see this
combination again in a minute) 28 .i.xd3 'iVBt
29 �g1 'ifxd3 with a completely winning posi­
With the queenside more than secure It IS tion.
time for Black to turn his attention to the king­ 26...'iff6 27 ltJd1
side. As usual in this variation the white king White defends his bishop on f2 to stop

1 69
Pla y the King 's Indian

Black from playing ..txg4. But...

27 . . ..bg4!
...Black plays it anyway.

7 . . .e5 8 d5 WeB
This is the modem, dynamic approach to the
King's Indian. The alternative, 8 ... c6, can be
28 'ifc3 considered the traditional approach. The idea
And White can't take as after 28 fxg4 he is behind 8 ..."ife8 is, I hope, obvious to you. Black
blown away by 28 ...lbd3!. The text, however is unpins the knight on f6 so that he can move it
no better. and play ... £7-5. Black could also have un·
28 ...ltJ&21 0-1 pinned with 8 ... h6 when the bishop normally
White resigned as his only move to stay de­ drops back to c3. Although that is how I rec­
fending f3 is 29 'ifd3 but then 29. ..luct ! 1s a ommended Black play against 6 ltJf3 I am not
killer. so keen on that against the ..td3, lt:lgc2 set-up.
A fter 8 ...'ife8 one of Black's hopes is that the
White delays, or omits, ll'lf3 bishop on gS may actually tum out to be mis­
placed. It could end up just pointing into thin
1 d4 ll'lf6 2 c4 g6 3 ll'lc3 �g7 4 e4 d6 5 h3 air, or it may block the g-file for the white rooks
0-0 6 �g5 while it will also have to watch that it doesn't
In this section White delays the development get stranded by 13lack playing f4.
of his king's knight as he usually prefers a set­ 9 g4
up with ..td3 and lt:lge2 although he has been
known to change his mind and develop his
knight to f3 later. I am recommending that
Black plays the very modem system 6 ... lt:la6 7
..td3 eS 8 dS "ife8!?
If you can learn how to handle such posi­
tions then you will be well on your way to
King's Indian mastery.

Game 62
Zotnikov-Gallagher
Arosa 1996

1 d4 ll'lf6 2 c4 g6 3 ll'lc3 �g7 4 e4 d6 5 h3 This double-edh:.cd move is the consistent

0-0 6 �g5 ltJ&6 7 �d3 follow-up to an early h3. The idea is not to start
The main line. For early deviations, sec a kingside pawn storm but to make Black think
Game 63. twice about playing ... f5. I f Black goes ahead

1 70
 White Pla ys an Early h 3

with the ... £7-fS idea, the g-file, where the black 1i'xe4 'irhS!? (1 8...1i'f7) 1 9 l:.cl (with the inten­
king is currently in residence, will open. How­ tion of swinging his queen's rook to the king­
ever, the piece which has the least reason to be side) 1 9...�d7 (an attacking player like Ku­
satisfied with g2-g4 is the white king. His pre­ preichik would be considering ...l:tf4 at each
ferred home on g1 is now out of bounds (un­ tum - here it was rejected because of 20 �xf4

less your name is Barsov) so he will have to stay 1i'xh4 21 1i'g2 - but it is clear that it, and other
in the centre or risk the queenside. Playing attacking ideas, will be more effective once the
Black in the King's Indian I am always happy to queenside is developed) 20 l:tc3 l:tae8 21 l:tcg3
see White playing g4 as even if things go wrong bS! (21...llf4 22 1i'g2! is less good as Black can't
there should still be swindling chances. play ...e4) 22 1i'g2 e4! 23 �e7 (after something
A couple of other tries for White: like 23 �e3 �xb2 White has no killer blow)
a) 9 lbge2 is a more solid alternative, when 23 ... e3! 24 f3 (completely hopeless, but so is
after 9 ... lbd7 we have a couple of examples: everything else) 24...l:txe7 0-1 Paunovic­
a1) 10 a3 is interesting. The idea is to follow Kupreichik, Yugoslavia 1992. After 2S :Xg7
up with b4 in order to make the knight on a6 1i'xh4+ there is no good square for his king.
look silly. I suppose l O ...lbacS 1 1 �c2 aS is I f a game is lost so quickly without any clear
possible but Black can also just play lO ... fS 1 1 error then it must mean that the whole strategy
b4 when 1 1 ...h6 1 2 �d2 cS! 1 3 dxc6 bxc6 1 4 is wrong. I have serious doubts about White's
0-0 lbc7 led t o a tense game i n Atalik­ combination of g4 and lbf3 in this game.
l vanesevic, Halkidiki 1 998. Also possible is b) 10 a3 (danger!) and now:
1 1 ...f4 12 f3 �f6 1 3 �xf6 lbxf6 1 4 1i'd2 1i'e7
with reasonable prospects for Black. Such a
position is only playable because he has ex­
changed off his passive bishop. I f White now
plays 1 S 0-0-0 then 1 S ... cS looks like a decent
reply.
a2) 10 0-0 f5 1 1 exfS gxf4 1 2 f4 e4 1 3 �c2
'ifhS! (threatening h6) 1 4 lbd4 1i'xd1 1 S l:taxd1
lbb4 16 �b1 lbcS with an active game for
Black, Cramling-Gallagher, Bern 1 992.
b) 9 a3 lbcS 10 �c2 aS 1 1 1i'f3?! lDfd7 12
lbbS lba6 1 3 g4 lbdcS 1 4 lbe2 (14 b4 axb4 1S
axb4 lbxb4! is good for Black) f5 1S 1i'g2?! fxe4
1 6 �xe4 lbxe4 1 7 1i'xe4 �d7! 1 8 a4 lbcs 1 9
'it'e3 �xbS 20 axbS e4! an d the rest was car­
nage, Chemin-Gallagher, Basle (rapid) 1 99S. b1) 10 ... lbacS 1 1 �c2 aS and now Agrest­
What I like about these h3 systems is that when Milov, European Ch., Ohrid 2001 continued 12
things go wrong for White they go really wrong. lbge2 lbb6! (a rare exception to the good
9 . . .ll'ld7 King's Indian rule of not putting one's knights
Black is not bluffed out by White's g4. He is on b6) 1 3 b3 f5 1 4 l:tg1 fxe4 1 S lbxe4 lbxe4 16
still going to play ... fS even if it means exposing �xe4 a4! (White's queenside strucrure will now
his own king. Better that than a passive position be spoiled as he can't play 1 7 b4) 17 lbc3 axb3
with nothing to do. 1 8 1i'xb3 lbd7 1 9 �e3 b6 20 l:ta2 1i'e7 21 l:tg3
10 lbge2 lbcS (this is such a good square for the knight
Or: that now White feels obliged to give up his
a) 10 lbf3 f5 1 1 gxfS gxfS 12 l:tg1 (12 exfS is strong bishop for it) 22 �xeS bxcS 23 a4 �h6
very strongly met by 12 ...e4!) 12 .. .'it>h8 13 lbh4 24 lbe2 �d7 2S 1i'c2 �f4 26 l:tf3 :as 27 lbc3
lbdcS 14 �c2 fxe4 1 S lbxe4 lbxe4 16 �xe4 l:tfa8 1/z-1/z. Black could certainly have contin­
lbcS 17 'ii'c2 (the bishop must be protected as ued the game.
retreating it would allow 1 7 ... e4!) 1 7 ... lbxe4 1 8 b2) Black can also play the other knight to

171
Pla y t h e King 's Indian

c5, e.g. t o...lDdc5 1 1 �c2 f5 but after 1 2 b4 over the initiative. White should have played 13
lihe4 1 3 �xe4 fxe4 14 �xe4 he must immedi­ b4 lDd3+- 14 .i.xd3 exd3 1 5 'ifxd3 in order to
ately bring the knight on a6 back into play with try and keep the knight on a6 out of the game. I
14...�b8. had envisaged a neat way to get the beast back
10 ...lbdc5 into play: 1 5 ... e4 1 6 'ifd2 c5! 1 7 b5 �b4! al­
10 ... f5 at once has also been played but the though the position after 1 8 axb4 cxb4 19 lDd4
text is more accurate. Black wants to be in a bxc3 20 1i'xc3 is just unclear.
position to recaprure with the bishop if White 1 3 ...lbxe4 14 .be4 ltJc5 1 5 lbc3
takes twice on f5. It is essential for White to keep a piece on e4
1 1 .i.c2 to prevent Black from playing ... e4. Black must
1 1 �g3!? has been suggested as an im­ take <Juick action as if White has the time to
provement but it seems to me like an admission consolidate then his blockade of e4 will give
of defeat to give up the bishop so easily when him a positional advantage. The first move
Black is about to open the position with ... f5. King's Indian players look at in such positions
1 1 ...f5 1 2 a3 is the exchange sacrifice ...llf4. Here there is an
even more dynamic solution.
1 5 ...b5!
Not deep, but a visually surprising move. Of
course 16 �xb5 is impossible but Black's main
point is that after 1 6 cxb5 �xe4 1 7 �xe4
1i'xb5 the position has opened and White's king
is very exposed. Rather than go for this White
prefers to give up a pawn to retain his blockade.
1 6 .i.e3 lbxe4 1 7 lbxe4 bxc4 1 8 ..-c2 ft5
1 9 0-0-0

It may look like a <Juiet little move but a3

spells danger for Black in this system. White is
threatening to lock the knight on a6 out of the
game. Black is forced to take action immedi­
ately and that means exchanging in the centre.
Alternatively:
a) In Eliet-Gallagher, Charleville 2000, White
tried the over-ambitious 12 exf5 gxf5 13 a3 but
overlooked that after 13. ..e4 1 4 �e3 lDd3+- 1 5
..txd3 exd3 that he cannot play 1 6 1i'xd3 as
16 ... f4! wins a piece. If the bishop moves there White was well aware that his king was not
follows 1 7 ... £3. Instead he was reduced to 16 going to lead a <Juiet life on the <JUeenside but
�f4 but after 1 6... fxg4 1 7 �xd3 g3 Black had a castling kingside is at least as risky and staying
strong attack. In passing, I should just mention in the centre is out of the <JUestion.
that 12 gxf5 gxf5 1 3 �g3 should be met by 19 ....i.d7 20 .l:lhe1
13... f4. White plans to use his rooks on the 2nd rank
b) Hazai awards the move 12 £3 an '!' but af­ to defend b2.
ter 1 2...�b4 13 ..tb1 fxg4 14 fxg4 he fails to 20 ......86
consider Black's best move, 1 4...1i'f7! Threatening 21 ....i.a4
1 2 ...fxe4 1 3 lbxe4? 21 ltJc3 .l:laeS?! 22 .l:ld2 .l:lbB
This narural recaprure allows Black to take Correcting my previous error which had

1 72
 White Pla ys an Early h 3

been prompted by some imaginary concerns a position.

about White playing f4. Of course the rooks b) The best chance was offered by 27 'irfS
belong on the open b-file. �g6 28 1i'e6+ as after 28 ... �h8 29 ..tf6 �xc2
23 •e4 llb3 24 llc2 llfbS 25 llee2 30 ..txg7+ �xg7 31 1i'e7+ White scrambles a
White has successfully defended his main draw by perpetual check. Black shouldn't be so
weakness on b2. It is hard to see how Black can greedy. Instead, he can play 28 ... ..tf7 29 'irfS c6!
make further progress on the queenside. How when his big centre gives him a clear advantage
can he increase the pressure then? WeU, ever in a stiU complex position.
since White played 'ire4 the old adage that a 27 ...exf4 28 ••f4 �xc3!
queen is a bad blockader (the black e-pawn lbis is not the sort of move one should play
must be blockaded and the queen is too strong lightly and I didn't. I used up most of my re­
a piece to be reduced to the role of blockader) maining time making sure there would be no
would not leave my head. If only the queen nasty surprises on the weakened dark squares
could be attacked then the white position is around my king. 28.....tg6 looks good but I
sure to fall apart. didn't want to give him the chance to play 29
25 ...g5!! ..tf6.
29 bxc3 ••83+ 30 �d2 .:Xc3!?

Oh, what pleasure I got &om that move! I

particularly enjoyed the fact that the two key This is actually the safest move in the posi­
moves in the game were sacrifices of the tion as with best play it leads to a very favour­
knight's pawn (1 5 ... b5!). Black's plan is simple - able endgame.
to transfer his bishop to g6. The move g5 had 31 llxe8+?
to be played at once as otherwise White could The main line was 31 .l:txc3 .l:tb2+- 32 .l:tc2 (32
prevent the manoeuvre with h4-h5. �d 1 ..ta4+- 33 .l:tcc2 .l:tb1+) 32 ... l:txc2+- 33 �xc2
26 .bg5 �aS 27 f4 'ii'd3t- 34 'it>ct 'ii'xe2 35 ..th6 'ii'e7 which
White allows the opening of the long diago­ should be winning for Black.
nal in order to get some vague attacking 31 . . ..:Xe8 32 llxc3 -.82+!
chances of his own against the black king. A Not 32...'ii'b2+- 33 .l:tc2 1i'b4+- 34 .l:tc3 which
couple of other moves are also worth examin­ is just a draw.
ing: 33 llc2
a) 27 'irg2 ..tg6 28 ttk4 is a fragile attempt 33 �d1 1i'e2+- 34 �ct 'ii'e l+ 35 �c2 .l:te2+­
to keep e4 blockaded and the black bishops out and Black wins.
of the game. However, 28 ... c3! 29 lLlxc3 .l:txc3! 33 ...-.85+! 34 llc3 ••d5+
30 bxc3 .l:tbt+ 31 �xb1 (31 'it>d2 'ird3 mate) Now the king is forced to the open files on
3t ...'ii'xe2! and Black wins the other rook as the queenside.
weU. A nice combination but there are also 35 �c1 lle1 + 36 �b2 'ifb5+ 37 �82 lle2+
other ways to win. It is simply inevitable in such 0-1

1 73
Pla y t h e King 's Indian

tain his grip on e4. 14 it.gS should be met by

Game 63 1 4...-tfS.
To sum up, Black has a slightly easier time
Partos-Gallagher
after 6 it.e3 than 6 it.gS. He can foUow the
Swiss uague 1997 same plan but doesn't need to spend a tempo
unpinning on ... 'iVe8.
1 d4 liJf6 2 c4 g6 3 ltJc3 .i.g7 4 e4 d6 5 h3 6 ...ltJa6
0-0 6 .i.g5 Black prepares the traditional King's Indian
I ncidentally, 6 �e3 is played from time to central strike, ...e7-eS.
time when 6.)Da6 foUowed by eS is again the If he plays this at once then 6... e5 7 dxeS
right plan. For example, dxeS 8 'ifxd8 :txd8 9 lbds wins a pawn al­
After 6 .ie3 ltJa6 there is though Black may have some slight compensa­
a) 7 itlf3 eS 8 dS transposing to Game 6 1 . tion after 9...itlbd7 1 0 ltlxc7 :tb8. StiU, I cer­
b) 7 g4 e S 8 dS ltlcs ( 8. . .c 6 i s an alternative) tainly don't recommend that you foUow this
9 f3 aS 1 0 'ii'd2 c6 1 1 dxc6!? bxc6 12 0-0-0 path. I can't work out whether the few grand­
itlb7 13 cS dS 14 exdS (14 gS d4) itlxdS 1 5 masters who have played like this did so on
.ic4 (1 5 ltlxdS cxdS 1 6 'ii'xdS 'ii'xdS 1 7 :txdS purpose or because they were drunk or some­
�e6) ltlxe3 16 'ii'xe3 'ii'e7 with a roughly level thing like that.
game, Gomez-Topalov, Seville 1 992. Another way to prepare ... eS is the immedi­
c) 7 it.d3 (more common) 7 ... e5 8 dS lbd7 ate 6...'iVe8 which has been gaining in popular­
(8 ...ltlh5 is also interesting) and now White has ity recently. It is fine for Black to use this move
several possibilities: order if he is planning to transpose into our
main line, but some of the independent lines
there do not look so trustworthy to me.
6 ... c5 is also quite popular, but after 7 dS e6
8 it.d3 exdS 9 cxdS! we are in the Modem Be­
noni which is outside the scope of this book.
7 liJf3
a) 7 it.d3 was covered in the previous game.
b) 7 g4 is not so common although I have
had to face it a couple of times.

c1) 9 ltlge2 itldcS 1 0 it.c2 f5 1 1 exfS (1 1 f3

.ih6) gxfS 1 2 0-0 f4 1 3 it.xcS ltlxcS 14 f3 is
unclear according to Kuzmin.
c2) 9 g4 lbdc5 10 .ib 1 ?! (10 .ic2) f5 1 1 exfS
gxfS 12 ltlge2 'ii'h4 13 a3 e4 14 gxfS �xfS 1 5
itld4 it.g6 1 6 �c2 itld3+ 1 7 it.xd3 ..txd4! was
exceUent for Black in Avshalumov-Kupreichik,
Blagoveshchensk 1 988.
c3) 9 a3 ltlacS 10 it.c2 f5 1 1 b4 (1 1 f3 .ih6)
ltlxe4 12 .ixe4 fxe4 1 3 ltlxe4 1i'h4 1 4 g4!, Bar­ Black can either continue with his ... eS plan
lov-Kir.Georgiev, Yugoslavia 1 99 1 is given as or switch to a Benoni now that \X!hite has
better for White by several commentators, but played the weakening g4 move. Here is an ex·
the simple 1 3 ...ltlf6! promises Black a good ample of each strategy:
1.,>ame. White is far too underdeveloped to main- b1) 7 ... e5 8 dS 'ii'e8 (8 ... c6 is an alternative) 9

1 74
 White Pla ys an Early h 3

tbge2 �h8 (I wanted to be ready to play ...tt'lg8 1 3 0-0-0 a6 1 4 .i.c2 b5 1 5 tt'ld4 bxc4 16 'ifxf4
and ... f5 at a moment's notice) 10 'ifd2 c6 1 1 ..td7 17 'ife3 'ifb8 1 8 f4 'iVb6 19 l:.d2 tt'ld3+- 0-1
lbg3 cxd5 1 2 cxd5 ..td7 1 3 ..te2 tt'lcs 1 4 f3 Kunte-Zhang Zhong, Shenyang 1999.
tLlg8! 1 5 h4 f5 16 gxf5 gxf5 1 7 exf5 ..txf5 1 8 a12) 10 g3 f5 1 1 ..td2 tt'lcs 1 2 ..tc2 fxe4 1 3
tt'lce4 (18 tt'lxf5) ..txe4 1 9 tt'lxe4 tt'lxe4 20 fxe4 tt'lxe4 tt'lxe4 14 ..txe4 tt'l f6 1 5 ..tc2 e4 with a
'iVa4! 21 ..td3 ti'lf6 22 'ife2 tt'ld7 with an un­ decisive advantage for Black, Bronstein­
clear position, Suba-Gallagher, Kuala Lumpur Nijboer, Wijk aan Zee 1 992.
1 992. White has the bishop pair but his king is a2) Therefore a number of White players
exposed. have tried 9 0-0 when 9...exd4 10 tt'lxd4 tt'lcS
b2) 7 ... c5 8 d5 e6 9 ..td3 exd5 (9 ...1i'e8!?) 10 seems fine for Black as White has no time for
exd5 (the white position would be too loose 1 1 ..tc2 ti'lfxe4! and must play something like
after 10 cxd5) 10 ... 'ife8+l (unpinning) 1 1 'ife2 1 1 l:.et . But more Black players have continued
(1 1 tt'lge2 tt'ld7 looks all wrong for White) with 9...tt'lh5, e.g.:
1 t ...tt'lc7 (Black is going to play ...b7-b5) 1 2
0-0-0 b5 1 3 cxb5, Gyimesi-I.Botvinnik, Tel
Aviv 2001 , and as Hazai points out Black
should just play 13 ...tt'lfxd5! as 14 ti'lxd5 ti'lxdS
15 ..te4 is refuted by 1 5.....txb2+ 16 'ifxb2 (16
�xb2 1i'e5+ 17 �c2 tt'lc3 is very good for
Black) 16 ...'ifxe4 1 7 .i.h6 'ifc4t 18 �d2 f6 with
a decisive advantage for Black.
Obviously White can do better than this but
this shows you the sort of thing to aim for.
7 .. Ji'e8!?

a21) t o l:.et exd4 1 1 tt'lds f6 1 2 .tct c6 1 3

ti'lf4 tt'lxf4 14 ..txf4 1i'd8 1 5 lbxd4 f 5 with
advantage to Black, Shabalov-Edelman, New
York 1 993.
a22) 1 0 'ifd2 was played in Akesson­
Shulrnan, Stockholm 1 999. Black played 10 ... f5
and got in some trouble after 1 1 exf5 gxf5 1 2
dxe5 dxc5 1 3 l:.fc1 e4 14 tt'lxe4! but t o... f6 1 1
..th6 ..txh6 1 2 1i'xh6 ti'lf4 1 3 l:tad1 'iff7 14 b3
'iVg7 seems like a relatively easy way to equalise.
a23) 1 0 dxe5 dxe5 1 1 l:.e 1 led to victory for
Of course Black could just play 7...e5 as after White in Barsov-Sutovsky, York 1 999, but only
8 uS we have transposed into Game 60 and 8 because Black went in for a speculative ex­
dxe5 dxe5 9 'ifxd8 l:.xd8 10 tt'lds l:.d6 is fine change sacrifice. Such a continuation cannot be
for Black, but the text is an interesting alterna­ dangerous for Black. 1 Lc6 is a good solid
tive. The idea is to play ... e7 -e5 next move, but choice while even after 1 t ...tt'lc5 1 2 tt'lds 'iVd7
since Black will already be unpinned he can (Sutovsky played 1 2...c6?! 1 3 tt'lc7 'ifd7 14
then meet d4-d5 with ...tt'lf6-h5. tt'lxa8 ti'lxd3 1 5 l:.e3 but it didn't work out) 13
8 �e2 tDc7+ �h8 14 ..tc2 tt'lc6! Black has no prob­
Let's take a look at a few other moves: lems.
a) 8 ..tu3 e5 and now: b) 8 e5 may look quite dangerous but is ac­
at) 9 d5 tt'lh5 falls in with Black's plans, e.g. tually quite hannless. After 8... dxe5 9 dxeS tt'ld7
al l) 10 g4 ti'lf4 1 1 ..txf4 exf4 12 'ifd2 tt'lc5 we have a couple of quick draws:

1 75
Pla y th e King 's Indian

b1) to 'ife2 f6 1 1 exf6 exf6 12 .i.e3 1/z- 1/z, his posltlon. White wants to suppon his e4-
Bykhovsky-Istratescu, Tel Aviv 1994. square with lL!f3-d2 but is worried about his
b2) 10 ..i.f4 b6 (to. ..c!frlc5 cenainly deserves bishop stuck out on gS. A retreat to e3 will only
consideration) 1 1 .i.d3 .i.b7 1 2 be4 .i.xe4 13 encourage Black to play ... f4, d2 is required for
lL!xe4 lDdc5 1 4 lL!xc5 lL!xc5 1 5 0-0 lldB 16 the knight so the bishop went all the way home.
'ifc2 'ifc6 1 7 l:lad1 lDe6 1/z- 1/z, San Segundo­ A more testing continuation would be 1 1 exf5
Spassov, Moscow Olympiad 1 994. gxf5 12 lLJh4 (and not 12 lL!xeS? 1i'xe5 when
c) 8 g4 can be met by 8...e5 9 d5 lDd7 trans­ the bishop on e2 is pinned) 1 2 ...lL!f6 with a
posing to Zotnikov-Gallagher, or by 8 c5 9 d5
... double-edged game.
e6 10 dxe6 'ifxe6 (1 0 ... .i.xe6 1 1 'ifxd6 h6! is a 1 1 ...lbc5 1 2 lbd2 lbf6
playable pawn sacrifice) 1 1 'ife2 lleB 12 0-0-0 I was quite tempted by the continuation
lL!xe4 13 lL!xe4 'ifxe4 14 'ifxe4 llxe4 1 5 llxd6 1 2 ... a5 1 3 exf5 e4! but in the end opted for the
l:leB was comfortable for Black in Hear-Wood, solid text. White is vinually forced to give up a
London 1993. pawn now as if Black can play ...aS next move
Recently White has played 9 ..i.g2 here a few he will have achieved a perfect King's Indian
times and I think Black should just head for the position.
strange Mar6czy with 9 ... cxd4 rather than get­ 1 3 b4 lbcxe4 14 llldxe4 lbxe4 1 5 lbxe4
ting involved with 9 ... h5 to .i.xf6 exf6 1 1 gxh5 fxe4 1 6 ..ie3 ..id7 1 7 a4 'fie7 1 8 'flb3 :tf7
f5 12 hxg6 fxe4 13 lL!g5 fxg6 14 lL!cxe4 .i.xd4 1 9 c5 'flfB!
1 5 0-0, as he did unsuccessfully in the game
Nielsen-Michelakis, Copenhagen 2003. I believe
this messy position to be somewhat in White's
favour.
8 ...e5 9 d5
9 dxe5 dxe5 to lDd5 lDxe4! 1 1 .i.e7 c6 1 2
..i.xf8 'ifxfB is the son of exchange sacrifice
King's Indian players dream about.
9 ...lDh5 10 g3
Not a desirable move, but otherwise Black
will play ... lL!h5-f4 with an easy game.
1 0 ...15
Shirov has played to ... f6 here but consider­
ing the move my opponent played after to .. f5 I .

am reluctant to recommend this. My original plan had been to double rooks

1 1 ..ic1 on the f-file but now that White can't stop the
threatened ....i.g7-h6 with 1i'd2 the text is a
good choice which forces White to weaken his
kingside. He doesn't want to allow ....i.g7-h6 as
the bishop on e3 is holding his position to­
gether.
20 g4 'fle7!? 21 :tc1 :tafB 22 b5?!
22 cxd6 cxd6 23 .i.xa7 .i.h6 is very good for
Black but 22 gS, trying to keep the black queen
out of h4 makes more sense. I was planning the
typical rejoinder 22...l:lf4 23 h4 h6 but 23.....i.g4
is interesting.
22 . . .1Fh4 23 :th2 ..ih6 24 ..ixh6 :txf2!?
A rather speculative continuation. I consid­
ered 24...'ifxh6 25 1i'e3 1i'f4 to be a bit better
The move of a man who is not happy with for Black but couldn't resist the text.

1 76
 Whi t e Pla ys an Early h 3

25 l:txf2 l:txf2 26 �e3 lth2+? the bishop on g4 plays an imponant defensive

26 ...l:tg2+ 27 �d2 'irxh3 was more accurate, role by preventing 'irc8+-
preventing the defence with l:tg1 given in the 32 . . .b6 33 �g5
next note. White has no move, e.g. 33 �xb6 axb6 34
27 �d2 'lfxh3 a7 1i'a3.
33 . . .h4 34 '1Fc7
Or 34 ..i.xh4 'irf4+ 35 �d1 l:txh4.
34...l:txe2+ 35 :Xe2 'lfxe2+ 36 �c1 'lff1 +
3 7 �d2?!
37 �c2 poses slightly more problems but
Black is stiU winning easily after 37 ... h3!
37 . . .e3+! 38 �c2 'lff5+ 39 �c3 'lfxg5 40
'lfb8+ �g7 41 'lfxa7+ �h6 42 '1Fxb6 e2 43
'1Fg1 'lff4 44 a7 'lff1 0-1

Summary
1) The systems with an early h2-h3 by White
usually lead to a tense struggle where Black
must play actively to avoid getting squashed.
28 l:te1 ? 2) White usually plays an early g2-g4, gaining
Missing his last chance. After 28 l:tg1 ! ..i.xg4 space on the kingside and making it more diffi­
29 l:txg4 'irxg4 30 'irc4 hS 31 cxd6 cxd6 32 cult for Black to play ..f7-f5. Black must either
.

�dl ! Black has a draw with 32...l:.ht+ 33 �d2 go ahead and play . f7-f5 anyway or open lines
..

l:th2. He could play on with something like on the queenside (... c7-c6) so that the white
32...1i'h3 but not without risk. After the text king won't be safe on this side of the board.
White gets caught in a pin and has no good way 3) Black must take care that the knight he
to prevent the advance of the black kingside often has on a6 doesn't get shut out of the
pawns. game by White playing a2-a3 and b2-b4.
28 . . .�xg4 29 cxd6 cxd6 30 'iVc4 'iVf3! 31 4) The c4-squarc can be of paramount im­
a5 h5 32 a6 ponance in the lines where Black plays ... f7-f5.
White was reluctant to play 32 ..i.xa7 as this If White can maintain a blockade of this square,
allows the black queen to penetrate to the preferably with a knight, then he usually has a
queenside. For example, after 32 ..i.xa7 'ira3! 33 pleasant game whilst if Black can advance ...e5-
a6 bxa6 34 bxa6 (34 b6 e3+) 34...l:.h3! Black has c4 then everything changes and it is White who
a decisive attack. Note how in many variations is usually struggling.

1 77
 CHAPTER ElEVEN I
The Averbakh Variation

1 d4 lLif6 2 c4 g6 3 l2Jc3 �g7 4 e4 d6 5 difficult for Black to play ... cS but playing a
�e2 0-0 6 �g5 quick ...c6 and ... lbc7 introduces many interest­
This line bears the name of the famous Rus­ ing ideas into the position. Such a plan would
sian grandmaster Yuri Averbakh. It is a solid be too slow in the real Four Pawns Attack (sec
restraining system which by preventing the Chapter 9) but is quite acceptable here as
immediate ... e7-e5 makes it more difficult for White's bishop is offside on gS. Against 7 h4
Black to obtain his traditional kingside counter­ Black can react in traditional fashion with 7 ... e5.
play. I t is quirt: reasonable for Black to switch The systt:m recommended involvt:s a very the­
plans and play for a quick ... c7-c5 but there is matic pawn sacrifice. In the 67 game we look at
no need for him to abandon the idea of playing the solid 7 lbf3 whilst in the Game 68 we take a
for ... eS. I t just needs further preparation. There look at Uhlmann's speciality 7 'ii'c2.
are several ways this can be done. The most
obvious is to attack the bishop with 6 ... h6 and Game 64
after it retreats (1 �e3) to play 7 ... e5. The prob­
Korniushin-Ozolin
lem here is that is not always a good idea for
Black to play ... h6 so early in the King's Indian To!tJSk 1997
as White may be able to gain a tempo with
'ifd2. Another way to prepare ... eS is with 1 d4 lLif6 2 c4 g6 3 l2Jc3 �g7 4 e4 d6 5
6 . ..lbbd7. This is quite playable but I'm rec­ �e2 0-0 6 �g5 lL\a6 7 "ii'd2 e5 8 d5
ommt:nding a third, more flt:xiblt: way, 6 ...lba6. After 8 dxeS dxeS 9 'ifxd8 l:r.xd8 neither 10
This prepares ... e7-e5 by defending the c7-pawn lbdS l:r.d6 nor 10 lbxeS lbxc4 cause Hlack any
and is more flexible than 6 ... lbbd7 because it problems.
doesn't hem in the bishop on c8. Although a. _ _ l2:\c5
Black's main idea is to play ... e7-e5 sometimes 8. ..c6 and 8...1i'e8 arc commonly played al­
White prevents this (1 f4 for example) but, as ternatives. I covered the former in SOKID and
wt: shall see, Black has other ways of develop­ it can easily lead to similar positions to the text
ing counterplay. where Hlack is merely delaying ...c7 -c6 until his
White has a wide variety of replies to 6 ...lba6 knight has been established on cS.
and each of these is examined in their own 9 f3
game. The most popular response is 7 'ii'd 2 and An important point is that Black can meet 9
th is is examined in the first .L,>ame below. The b4 with 9 ... lbcxe4! 1 0 lbxe4 lbxc4 1 1 ..ixd8
sharpest moves arc 7 f4 and 7 h4. These arc lbxd2. The position is very complicated but
examined in ( ;ames 65 and 66. 7 f4 makes it practice has shown that after 1 2 i. xc7 both

1 78
 The A verb a k h Varia tion

12 .. .'�Je4 and 12 ...e4 give Black good play. An may mean him having to worry about White
example of the latter: 1 3 l:tcl e3! 1 4 fxe3 (14 recapturing with the knight. 1 2 ltJxd5 is weU
�d3 ex£2+- 15 �xf2 ..tfS and 14 ..txd6 ltJe4!) met by 1 2 ... ltJe6.
14 ...ltJe4 1 5 �d3 l:te8 and Black has exceUent 1 2 cxd5 �d7 1 3 tt:lhl 'iFeB
compensation for his pawn. A slightly awkward move but it does unpin
the knight and prepare the desirable <jueenside
advance ... b5. How brave are you? I f you are
willing to sacrifice a piece to create chaos on
the board then the immediate 1 3...b5!? (pat­
ented by the Cuban master A.Perez) is the
move for you. The guiet manoeuvring game
that one is likely to get after 1 3...1fe8 is now
likely to be replaced by a roUer-coaster ride.
Take a look at the foUowing variations before
you make your mind up. After 1 3... b5:
a) firstly, White can politely decline the offer
with 1 4 ltJf2 but then Black can develop his
queen actively with 1 4...1fb6. A logical foUow
up for Black would then be to double rooks on
9 a5
... the c-flle.
Now that White's e-pawn is protected b2-b4 b) 14 �xb5? is a mistake. The point is that
must be prevented. after 1 4... ..txb5 1 5 ltJxb5 l:tb8 White cannot
10 g4 retreat with 1 6 ltJc3 on account of 16 ...l:txb2!
White's standard plan in this position is to when 1 7 1i'xb2? loses to 1 7...ltJd3+. 1 6 1fe2
advance his kingside pawns. There is not much 1Vd7 1 7 a4 is better but Black still has a good
chance of Black getting mated but White hopes game after 17 ...ltJxa4!
that the space he will gain will futther restrict c) 14 ltJxb5 is the critical test. The amazing
Black's active options. Sometimes White starts response is 14 ...ltJfxe4!!
with 10 h4 but it usually just comes to the same
thing.
1 0 c6
...

and White has two moves:

cl) 1 5 ..txd8 ltJxd2 16 ..te7 (16 �xd2 l:tfxd8
17 ltJxd6 ..txg4 is good for Black) 16 ...ltJxf3+
Black needs to seek counterplay on the 17 ..tx£3 ..txb5 1 8 ..txfB l:txf8, Keipo-Perez,
gueenside as ... fS is not really on the cards. At Cuba 1 996. The dust has settled and in return
least not yet! for the exchange Black has a pawn, active
1 1 h4 cxd5 pieces and two potential passed pawns in the
Black takes immediately on d5 as delaying centre. Quite adequate compensation

1 79
Pla y th e King 's Indian

c2) 15 fxe4 f6 when White has a couple of It is a good idea for Black to exchange some
possible bishop moves (returning the piece pieces as he has less space. Without the bishop
offers no advantage): on d7 this square will become available to either
c21) 16 �h6 tt'lxe4 17 1i'e3 �xh6 and the queen or king's knight.
whichever recapture White makes Black will 1 7 tt:le3
have compensation for the piece. I'll just men­ Black is not worried about 17 h6 as if White
tion that 18 1i'xe4 can be met by 18 ... f5 19 gxfS closes the h-file he will no longer have any at­
�xfS and 18 1i'xh6 by either 1 8...tt'lg3 or tacking chances on the kingside. Although the
18 ...1i'b6. Of course the reason Black has com­ bishop on h8 will be locked out of the game
pensation is the terribly exposed position of the this is not irreversible.
white king. 1 7 . . .�xe2 1 8 'ifxe2 b3 1 9 a3 :cs 20 0-0
c22) 16 �e3 tt'lxe4 17 'ii'c2 fS. At first sight a4 21 :Sc1 'ii'd 7
it seems that Black doesn't have much but on
closer inspection we can see that White is fac­
ing serious problems. For example, 1 8 gxfS
lbg3! 19 0-0-0 (19 :g1 'ii'xh4 is good for Black
after both 20 tt'lt2 tt'lxfS 21 �g5 1i'xf2+! 22
�xt2 tbd4+ and 20 �g5 1i'xh3 21 1i'c3 tt'lxe2!
22 1i'xh3 tt'lxg1) :cs 20 tt'lc3 �xfS 21 �d3
�xd3 22 'ii'xd3 e4! 23 'ii'd2 �xc3 24 bxc3
:xc3+! (this is why Black was in no tush to take
the rook in the comer; after 25 'ii'xc3 tbe2+ 26
�b2 tt'lxc3 27 �xc3 'ii'c7+ 28 �d2 'ii'c4 he has
a winning position) 25 �b2 :d3 and Black is
winning. Instead of taking on f5 White can try
18 tt'lg5 but here too Black has good play after
18 ... tt'lg3, e.g. 1 9 :g1 tt'lxe2 20 'ii'xe2 f4 21 �t2 White's position has one main drawback -
e4! or 1 9 0-0-0 tt'lxh 1 20 :xh 1 :cs. the exposed position of his king. For the mo­
To conclude, 13 ...b5 looks worth a try if you ment the black pieces cannot get anywhere near
have a gambling nature. The above variations it but this may change as the game progresses.
are only the tip of the iceberg, given to demon­ For the time being both sides concentrate on
strate the sort of possibilities open to Black. doubling their rooks on the only open ftle.
There are no guarantees that it won't end in 22 :c4 :c7 23 :tc1 :feB 24 :1c3 1i'e8
tears against a strong defensive player but I , 25 :b4 tt:lfd7 26 tt:lc4 1i'f8 27 �e3 1i'e7 28
myself, would be inclined to take the risk. tt:la5?!
14 tt:lf2 b5 1 5 h5 b4 1 6 tt:lcd1 �b5 It would have been better for White to ex­
change on g6 at some point as now Black will
be able to open lines advantageously on the
kingside.
28 . . .11Vh4 29 �g2 gxh5 30 g5 tt:lfB
The knight is heading for f4.
31 tt:lh3?
Now that the black queen can't defend d6
White should have brought his knight back
with 31 tt'lc4! That would have left the situation
far from clear whereas after the text Black
should be winning.
31 ...tt:lg6 32 'iff2 tt:lf4+ 33 tt:lxf4 1i'xf2+?
33 ... 'ii'xg5+! 34 �h 1 exf4 would have been
even stronger.

1 80
 The A verb a k h Varia tion

34 �xf2 exf4 35 .i.d4 lba6 36 .txg7? �h4 on account of 10 ...lL'Ixf4 so he would have
White could have offered more resistance to part with his important dark-squared bishop.
with 36 Axc7 Axc7 37 Axa4. White has three ways to deal with the threat of
36 . . .llJxb4 37 .td4 :.Xc3 38 .txc3 1bc3 9 ... lL'Ie6. He can prevent it with dS, retreat the
0-1 bishop from gS, or defend his f-pawn.

Game 65
S.Mohr-Uhlmann
Bundesliga 1994

1 d4 llJf6 2 c4 g6 3 llJc3 .tg7 4 e4 d6 5

.te2 0-0 6 .tg5 lba6 7 f4! ?

9 'ifd2
Defending the f-pawn, so that 9...lL'le6 can
be met by 10 ..ih4, is considered the main line.
The alternatives:
a) 9 �h4. This prophylactic retreat is
becoming increasing popular as White has
failed to prove any advantage in the main line.
The bishop is less vulnerable to attack on h4
An aggressive move which leads to a strange than gS, from where it could be hit by a knight
version of the Four Pawns Attack. One of the on e6 or e4 or a pawn on f6. By retreating the
main points behind 7 f4 is to prevent Black bishop now White reasons that he won't have
from playing ... eS. In this way White hopes to to waste a tempo at a more important moment.
show that the knight on a6 is misplaced. Don't As in the main game Black has a choice
forget that the main point of 6 .. .'�)a6 was to berween 9 .. b5 or 9 ... d5.
. We arc going to

prepare the advance ... eS. This approach will concentrate on the latter. After 9...d5 1 0 e5
not suit aU White players, though, as with the lL'le4
bishop on gS stuck outside the pawn centre
there is an element of risk involved in White's
strategy. Black should not react with the tradi­
tional response to the 4 Pawns A ttack, 7 ...c5, as
after 8 dS the knight on a6 would indeed be
condemned to passivity. Instead, he should
prefer the modest-looking 7 ... c6 followed by
8 ... lL'Ic7 and then decide berween establishing a
foothold in the centre with ... dS or seeking
active play on the queenside with ... bS. The
potentially problem piece on a6 often settles
down on the healthy central square e6.
7 . . .c6 8 llJf3 llJc7
9 ... lL'Ie6 can now be considered a threat. For we can already see the benefit of 9 �h4 -
example, after 9 0-0 lL'le6 White can't play 10 the bishop is not attacked. A number of j,'lliTIC S

181
Pla y t h e King 's Indian

have gone 1 1 lbxe4 dxe4 1 2 lbg5 f6 1 3 exf6 the queenside. Yakovich-Atalik, Beijing 1997
exf6 14 lbxe4 but 1 4...lbe6 gives Black good continued 9...'ife8 (if Black is willing to sacrifice
play. In Korchnoi-Xie Jun, Schuhplattler Veter­ a pawn he can also play the immediate 9 ... b5!?)
ans-Ladies 2000 White introduced 1 1 0-0! and 1 0 'ifd2 b5 1 1 e5 b4! (an important counter­
quickly built up strong pressure after 1 t ...lbe6 attack) 12 exf6 exf6 1 3 ..ih4 bxc3 14 'ifxc3?!
12 g3 f6 1 3 cxd5 cxd5 14 'itb3. Black needs to (1 4 bxc3 is better when the game is about level
improve on this. Annotations to the game sug­ after 14 ... ..if5) 14 ... 'ifxe2! 1 5 llfe1 (it looks as
gest that White is doing very well after 1 t .....i£5 though Black has fallen into a childish trap -
12 lbxe4 ..ixe4 13 lbg5 ..i£5 14 g4 ..ic8 15 f5 how docs the queen escape?) 1 5...lbb5!! Bril­
as Black has wasted so much time moving his liant. The point is that after 1 6 cxb5 Black has
bishop. But what has White done with this 1 6 ...'ifxb5 while 16 'ifb3 lbxd4! 17 lbxd4 'ifg4
time? Pushed all the pawns in front of his king is also good for Black. White decided to settle
which may give him attacking chances but also for the slightly worse endgame that arises after
leaves his own monarch quite exposed. I be­ 16 llxe2 1 6 ...lbxc3 17 bxc3 ..ia6 and managed
lieve 1 5...f6 leads to an unclear game. This is to hold the game with some careful defence.
just an untested suggestion while the latest de­ 1 0 .bf6
velopment in the variation is the 1 1 ... f6 that was t o e5 is of course met by t o ... lbe4.
introduced in IUescas-Comas, Spanish Ch. 1 0 exf61
.. .

2002. After 12 lDxe4 dxe4 1 3 lbe1 lbe6 14 lbc2

g5 15 fxg5 fxe5 16 llxfB 'WxfB White could
have obtained an edge with 17 d5! lbf4 1 8 lbe3.
A possible improvement for Black is to play
16 ... ..ixf8 with the idea of meeting 1 7 dxe5 with
a spot of pawn grabbing - 1 7 ... 'Wb6t followed
by ...'ifxb2.
b) 9 d5 is a radical way to prevent ... lbe6. I t
i s not very popular although it i s probably no
worse or better than any of the other lines. The
recommended antidote is 9 ... cxd5 10 cxd5
lbh5! (threatening ... f6).
b1) Now 11 f5 lDf6 12 fxg6 hxg6 13 'ifd2
(13 0-0) 13. ..lba6! is good for Black according
to Seirawan. This is all deep stuff where you Rules are made to be broken, though you
may be forgiven for thinking that the black have to know when to break them! The natural
knights are aimlessly wandering backwards and recapture, t o.....ixf6, leaves Black with a diffi­
forwards. In fact the purpose of their meander­ cult game after 1 1 cxd5 cxd5 12 e5 ..ig7 13 h4!.
ings was to clarify the central situation and to White has a space advantage and attacking
carve out some reasonably safe squares for chances and Black has no counterplay. Black's
themselves. The f6-square is now secure for the choice of recapture obviously compromises his
king's knight (there is no longer e5 to worry pawn structure but there are certain compensa­
about) while the queen's knight can go to c5. tions. For example, it is not so bad to have
b2) 1 1 'ifd2 looks better when 1 l ...f6 12 doubled f-pawns in front of the castled king.
..ih4 ..ih6 1 3 g3 e5!? 1 4 dxc6 ..ixe6 leads to a The back one can stay at home on defensive
very double-edged position where Black has duty while the front one can happily advance.
weaknesses but the white bishop on h4 is in With just one f-pawn you can't do both. Recap­
trouble. turing with the pawn should also give Black
9 d5
... first use of the open e-file.
Finally Black reacts in the centre. As previ­ 1 1 exd5
ously mentioned there is an alternative plan 1 1 0-0 is well met by 1 1 ...dxe4 1 2 lbxe4
which involves not striking in the centre but on ..ig4! Black's strateb'Y is based on pressurising

1 82
 The A verbakh Varia tion

the d4 pawn. By eliminating the knight on f3, In other games White has preferred 15 l:tad1
playing . . fS and a rook to the d-file Black
. f5 16 lt:'le5 to tempt Black into playing ... f6, the
should be able to force the pawn to advance to reasoning being that the time is well spent as
d5. He will then look to blockade the pawn the long diagonal of the bishop on g7 will now
with a knight on d6, which in conjunction with be closed. Black is advised to fall in with
his better bishop and safer king should, at the White's plans as he has another way to activate
very least, compensate for White's central his bishop. After 1 6 ... f6 1 7 lt:'lf3 .i.h6 1 8 g3 g5
passed pawn. All this was demonstrated in the Black has good play. This is the same strategy
first important game in this line, Yakovich­ Black uses in the main game.
Smirin, Munich 1 993. Play continued: 1 3 l:tad 1 1 5 f5 1 6 lLlb5?!
•..

'ife7 1 4 lt:'l£2 .i.xf3 1 5 .i.xf3 f5 (this must be Rather optimistically sending his knight on a
played before White can play fS) 1 6 d5 cxd5 1 7 journey to d6. True, this is a nice outpost but
cxd5 1i'd6 1 8 g3 lt:'lb5 1 9 .i.g2 and now Black now has the rime to develop a kingside
19 ... 'ifb6!, followed by ... lt:'ld6 would have given attack. Note that White no longer has the pos­
Black a good game (20 d6 would just push the sibiliry to remove the masterful bishop on e4.
pawn to its doom). Many years ago 1 6 . . . i&.h6! 1 7 g3 g5!
Nimzowitsch taught us that the knight is an
ideal blockader, especially of an advanced cen­
tral pawn. Here Black originally blockaded the
pawn with his CJUeen but this was never in­
tended to be a permanent state of affairs. After
the Yakovich-Smirin game White searched for
other ways of injecting life into this system and
came up with the plan in the main game. Ini­
tially, Black struggled but his recent results have
improved.
1 1 . . . cxd5 1 2 c5 ..11Lf 5!

There has been a wonderful economy about

Black's play in this game. No useless moves,
everything done with a clear purpose. Uhlmann
used to be one of the world's best players and
even now, well into his eighth decade, he still
knows how to pack a punch. I can testify from
personal experience.
1 8 lLle5?
White's best chance was to sacrifice the ex­
change with 1 8 fxg5 lt:'lxg5 19 lt:'lxg5 .i.xg5 20
1i'c3 .i.xct 2 1 'irxct when he has some com­
White has a potentially dangerous lJUeenside pensation though Black's game is still to be
pawn majority but also weaknesses in the centre preferred after 21 ...a5! (opening lines for the
created by the move £2-f4. If this pawn were rooks). The text leads to complications that are
back on f2 White would have a clear positional only ever going to end in Black's favour.
advantage but now Black is able to park his 1 8 . . .gxf4 1 9 c6 bxc6 20 l:l.xc6 "ifg5
bishop on the dominant e4-square. White will Black was unable to play ... fxg3 on the pre­
be extremely reluctant to take the bishop as it vious move as the f4 pawn was pinned to his
will then be replaced by a strong passed pawn. bishop on h6. Now with Black threatening this
1 3 0-0 i&.e4 1 4 b4 lLle6 1 5 l:.ac1 capture White can find nothing better than a

1 83
Pla y the King 's Indian

desperate piece sacrifice. Another aggressive move. When playing

21 ltJxf7 �xf7 Black in the King's Indian and your opponent
Of course Black takes this way to defend the starts an attack on the h-file, don't panic. It's
knight on e6. never mate. The text was quite popular in the
22 ttJcl6+ �e7 23 ltJxe4 early 1 990's when the world class Russian
Now White has no pieces left to attack with Grandmaster Bareev lost a lot of games with it.
but the reality was that if he didn't make this Nowadays it is not seen so much but the way
exchange it was Black who had the stronger Black deals with it is very instructive and shows
attack. In fact, he still docs. White may as well how the King's I ndian should be played - with
have resigned here. great energy!
7 . . .e5 8 d5 h6 9 .i.e3 ltJc5 1 0 'it'c2
1 0 f3 doesn't fit with an early h4 on account
of l O ...lDhS but 1 0 ..tf3 is played sometimes. A
recent example is Kogan-V.Georgicv, Salou
2000. Play continued 10 ... a5 1 1 g4 1i'd7! 12 gS
llJg4 13 ..txg4 1i'xg4 1 4 gxh6 ..t f6 15 'ifxg4
..txg4 1 6 f3 ..thS. White has won a pawn but he
will be unable to hold onto it for long. He now
felt obliged to part with his strong bishop as he
was quite fearful of the damage that Black's
knight might inflict. After 1 7 �xeS dxcS 1 8
lDce2 �h7 1 9 llJg3 �xh6 20 lbxh s �xhS 2 1
lDe2 ..txh4+ 22 �e2 l:.h8 23 f4 exf4 24 lDxf4+
�g4 25 lDd3 ..tg3 26 lDxcS f5 Black had the
23 . . .fxe4 24 b5 f3 25 'ifb4+ �7 26 'it'd& better ending (active king, superior minor piece,
.Uae8 27 .i.xf3 �g8! 28 .i.g2 'ife3+ 29 �h1 dangerous kingsidc pawns) and went on to win .
.Uxf1 + 30 .i.xf1 'it'f3+ 0-1 1 0 . . .c6
It's mate in two.

Game 66
Vakovich-Smirin
European Ch., Saint Vincent 2000

1 d4 ltJt6 2 c4 g6 3 ltJc3 .i.g7 4 e4 d6 5

.b2 0-0 6 .i.g5 ltJa6 7 h4

As we also saw in Chapter 10 when White

starts advancing on the kingside Black should
open lines on the quecnside so that the white
king won't be safe there. With the white queen
on c2 this becomes even more desirable.
1 1 h5
There was already a trap for White to fall
into: 1 1 b4? lDcxe4! 1 2 lDxe4 lDxe4 13 1i'xe4
..tf5 1 4 'W'f3 e4 and Black wins the rook on al .

1 84
 The A ve rb a k h Varia tion

If it weren't for this tactic then Black would a3 29 b3 ltJcxd3 30 ltJxd3 .:tc3 31 ltJxf4 .:txe3
have played 1 1 ...a5 in order to prevent b4. 32 ltJd3 b4 33 �c2 hS 34 �d 1 i.a4 35 bxa4 b3
1 1 . . .cxd5 36 �d2 i.h6 37 f4 bxa2 38 i.e2 exf4 39 i.£3
1 l ...gS!? has been played by Kasparov. It is i.g7 40 eS dxeS 41 d6 e4 42 ltJxf4 a11i' 43
not a bad move in this particular position but in .:txa 1 i.xa 1 44 �xe3 ex£3 45 ltJds fxg2 46 'Ott2
general Black should be very wary about block­ i.eS 47 d7 i.g3+ 48 �xg2 i.h4 49 ltJb4 i.d8
ing the kingside in this fashion. I prefer the 0-1 .
more dynamic approach used by Smirin (and 1 5 . . .l:ac8!
others).
1 2 cxd5 'ii'a 5
Black threatens to take on e4 so White has
no rime to take on g6. 1 3 fx!,>6 ltJcxe4 is good
for Black.

An exceUent move with a cunning idea be­

hind it. Smirin's preparation had no doubt re­
vealed the game Yakovich-Bekker Jensen, Ex­
celsior Cup 2000 where Black played the the­
matic 1 5 ... ltJh5?! but landed in trouble for tacti­
1 3 �d2 �d7 cal reasons after 1 6 b4! 1i'xb4 1 7 ltJbS 1i'a4 1 8
Discovered attacks can be devastating but 1i'xa4 ltJxa4 1 9 ltJxd6. The next note shows
Black doesn't have to worry here (at least not how Smirin planned to deal with this continua­
yet) as there is no good square for the knight on tion against his move. But can't White just take
c3 to move to. on h6 now? Yes, but after 1 6 i.xh6 i.xh6 1 7
1 4 hxg6 fxg6 1 5 f3 .:txh6 �g7 Black has good compensation for
By solidly defending the e-pawn White fi­ the pawn. This is similar to the game continua­
nally 'threatens' to take on h6. The move £3, tion.
though, can be considered a positional conces­ 1 6 llb1
sion from White as it allows Black to play 16 b4 1i'xb4 17 ltJbS is now met by the
... ltJhS and take over the initiative on the king­ powerful CJUeen sacrifice 1 7 ...1i'xb5!! 1 8 i.xbS
side. In Hauchard-Hebden, CappeUe Ia Grande i.xbS. Although Black has just two pieces and
1 998 White tried to live for a while without a pawn for the queen, his threats against the
playing f3 but soon had to admit defeat. Play white king and queen (now we see why
continued 1 5 .:tb1 (threatening b4) 1 5...'ifb6 1 6 1 5 ....:tac8 was better than 1 S ...ltJhS) give him a
i.e3 aS 1 7 ltJh3 .:tac8 1 8 1i'd2 a4 1 9 f3 ltJhS! decisive attack. For example after 19 'iVb 1 i.d3
20 ltJt2 (20 i.xh6 ltJg3! 21 i.xg7 'Otxg7 22 .:th2 20 1i'd1 ltJfxe4! 21 i.xh6 (what else? 21 fxe4?
.:th8 is good for Black) 20...ltJf4 21 i.ft 'iVb4 allows mate in one) 21 ...i.a6! 22 lbe2 ltJd3+ 23
22 ltJcd 1 ltJa6 23 ltJd3 1i'xd2+ 24 'it>xd2 bS and �ft ltJ<.It2 24 1i'a4 ltJxh1 Black even has a
Black's energetic play has given him a good material advantage to go with his attack (25
game. In fact he continued to play with great 1i'xe4 ltJg3+!).
energy. The rest of the game is worth giving: 25 1 6 . . .'ii'd8
.:tel .:txcl 26 �xct ltJcS 27 ltJ1 t2 .:tc8 28 �b1 1 6 ...1i'b6 may look more active but White

1 85
Pla y th e King 's Indian

replies 17 �e3. place on the d4-square. He will then obtain an

1 7 �xh6 annoying passed pawn which also prevents the
As White is going to suffer anyway - Black knight on a4 from returning to play.
would play ... tLlhS against neutral continuations 29 'ii'xd4 exd4 30 'it>b3 �xg2 31 �xg2 b5!
- he may as weU have a pawn for his troubles.
1 7 ...�xh6 1 8 J:txh6 'itg7 1 9 l:th1 lLlh5

The previous intermediate move was quite

simple - this is pure class. Black forces an end­
This is not an original pawn sacrifice - the ing with good knight against bad bishop. Even
idea is known from other variations of the with all Black's pawns isolated the chances to

King's Indian. So what does Black have for the hold such an ending are slim. I f nothing else
pawn? WeU, firstly in order to win it White had being under permanent pressure is likely to lead
to exchange off his active dark-squared bishop to an eventual collapse.
for Black's passive one. Because of the pawn 32 lLlb6 axb6 33 �f1 d3 34 'it>c3 l:th1 35
structure this means his bishop's protection of l:td1 'it>f6 36 'it>d4 b4 37 l:tb1 lLle2+ 38 'it>e3
the dark squares will be sorely missed - the g5
subsequent course of the game should drive Maybe 38 ...'it>e5 is more accurate. For exam­
this point home. Secondly, White's king doesn't ple 39 �xc2 .:r.xb1 40 f4+ 'it>f6 41 �xd3 l:hb2
have a safe home and thirdly his minor pieces 42 �c4 gS! is hopeless for White.
are awkwardly developed - just lcx>k at his 2 1 st 39 'it>f2 lLlf4 40 'it>e3 lLle2 41 l:td1 lLlf4 42
move, influenced by the fact that he wants a l:tb1 d2 43 'it>xd2 l:th2+ 44 �e3 �e5 45
knight on e2, not a bishop. I don't know if �a6 lLlg2+ 46 q.,d3 lLlf4+ 47 'it>e3 l:th3 48
Black can claim a theoretical advantage here but :n ?
all King's Indian players should delight in such I'm not sure that Black has made the most
positions. of his position but until here White had de­
20 1Wd2 l:th8 21 �f1 1Wb6 22 lLlge2 fended very weU. He should have played 48
22 'W'gS is not dangerous as Black just replies .:r.g1 to prevent g4. Black now gets a passed
22...lLlf6. pawn that quickly decides the game.
22 . . .lLlf4 23 J:txh8 lLlcd3+! 48 . . g4! 49 �e2 lLlg2+ 50 'it>d2 g3 51 l:tg1
.

Games between strong players contain many l:th2 52 �a6 'it>f4 53 e5 q.,xe5 0-1
more intermediate moves than games between
average players. There are a few more to come!
Game 67
24 'it>d1 J:txh8 25 lLlxf4 lLlxf4 26 'it>c2 l:th2
Burnier-Gallagher
While White is trying to improve his king
position Black piles the pressure on his Achilles
Neuchdtel 2002
heel - g2.
27 l:.c1 �h3 28 lLla4 'ii'd4! 1 d4 lLlf6 2 c4 g6 3 lLlc3 �g7 4 e4 d6 5
Black is happy to exchange queens if it takes �e2 0-0 6 �g5 lLla6 7 lLlf3

1 86
 The A verb a k h Varia tion

A modest move. White has no attacking advantage. If White had played 10 0-0 then
ambitions. He just wants to casde kingside and Black would have continued in the same fash­
steer the game towards a Classical King's In­ ion, though meeting 1 1 'ifxd4 with 1 1 ...l:le8)
dian where Black has committed himself to an 1 1 ...g5 1 2 ..tg3 fbxe4! 1 3 'ifxe4 f5 1 4 'ifd3 f4
early ...lba6. Of course there is nothing wrong 1 5 'ifxd6 'ifxd6 1 6 l:lxd6 fxg3 1 7 hxg3 lbc5.
with an early ...lba6 there; it is just that White Although White is a pawn up Black's powerful
avoids all the complex lines with ...fbc6. bishops give him good compensation. In Alex­
7 h6 8 .i.el
... androv-Bologan, Kstovo 1 998 Black emerged
This is Milov's favourite move with which victorious on the 37th move.
he has won a number of games including one b3) 9 0-0 can be met by the interesting
against your author. However, I didn't play the 9...g5!?. Black forces the exchange of White's
opening so weU in that game. There are two bishop for a knight but in doing so weakens his
other bishop moves: kingsidc. As a rule such play is risky if White
a) 8 ..tf4 e5! 9 dxe5 lbh5 10 ..te3 dxe5 gives hasn't casded but quite acceptable once he has
Black an easy game. White has been wasting committed his king to the kingside. It wiU then
time with his bishop whilst Black has been be difficult for him to attack on this side of the
making useful move. Neither 1 1 0-0 c6 1 2 board. The game Zakharevich-Kokorev, Mos­
'ifxd8 l:lxd8 1 3 l:tfd1 l:le8 1 4 g3 lbf6 1 5 lbd2 cow 1 999 now continued 1 0 ..tg3 lbh5 1 1 'ifd2
lbg4 16 ..txg4 ..txg4 17 f3 ..te6, Milov-Smirin, lbc7 (or 1 1 ...e6 1 2 lbe1 lbxg3 1 3 hxg3 c5 'lz - 'lz
Haifa 1 995, nor 1 1 'ifcl 'it>h7 12 0-0 c6 1 3 l:ld1 Kallai- C.al.lagher, French League 2002 - Black
'ife7 14 c5 lbc7 15 l:ld6 lbe8 16 l:ld2 fS, Seri­ has easy equality) 1 2 d5 lbxg3 1 3 hxg3 e6 1 4
genko-Navrotescu, Decin 1 996, gave Black l:lad1 (after 1 4 dxc6 bxc6 1 5 l:lad1 d 5 1 6 cxd5
cause for complaint. cxd5 1 7 exd5 fbxd5 1 8 fbxd5 'ifxd5 19 'ifxd5
b) 8 ..th4 (the most common choice) 8...c6 exd5 20 l:lxd5 ..te6 Black regains the lost pawn
(a useful waiting move: 8 ... e5 transposes direcdy with the better game) 1 4 ... exd5 1 5 exd5 c5 (the
into a line of the lba6 Classical) and now: game has taken on the character of a Benoni
where the bishop on g7 is a particularly strong
piece now that its opposite number has been
eliminated) 1 6 lbh2 a6 1 7 g4 b5 1 8 b3 bxc4 1 9
bxc4 l:lb8 2 0 ..td3 lbe8 21 l:lb1 l:lxb1 22 l:lxb1
fbf6 23 ..tfS 1/z-'lz.
Black could have continued the struggle.
8 .!Dg4 9 .i.c1
...

b1) 9 eS can be met by 9 ... lbe8 but also

9 ... dxe5 1 0 dxe5 lbg4 1 1 'ifxd8 (on 1 1 ..tg3
Black can hit e5 again with 1 1 ...'ifa5) 1 1 ...l:lxd8
12 ..tg3 and now Black should play 12 ... h5 in
order to give his knight on g4 a retreat square.
The game looks about equal.
b2) 9 'ifd2 e5 10 0-0-0 exd4! 1 1 'ifxd4
(White would prefer to recapture with the Don't worry, I also struggle to get my head
knight but this runs into a little tactical trick: 1 1 around this variation. One way to look at it is
fbxd4 fbxe4! 1 2 ..txd8 fbxd2 gives Black the that White after playing his sixth move in the

187
Pla y th e King 's Indian

Classical offers his opponent a pact. You can Up until here my opponent had been playing
have the pretty useful move ... lba6 for free and very quickly as he had been following some
you can even move again. But in rerum you opening theory which said White was better
must also play the possibly harmful and the here. But they didn't take the text into account.
positively ugly moves ... h7-h6 and lbf6-g4. I This is a standard ploy in such positions (see
think I would take him up on his offer. Game 21) and makes the set-up with the
9 ...e5 10 0-0 c6 bishop on f3 look rather suspicious. It's not
In the game I lost to Milov I played 10...'ife8 obvious how White can deal with ...lLJgS and in
here and ended up worse after 1 1 dxeS dxeS 12 my opinion he is already fighting for equality
b3 'it'e6?! 1 3 .i.a3 cS 1 4 lbds lbf6 1 S 'ii'c2, 1 5 �f4 tt:\g5 1 6 tt:lb3?!
Milov-Gallagher, Pula 2000. Simply overlooking the reply. Probably
1 1 h3 White should settle for 16 h4 lbxf3+- 17 lbxf3
Probably a slight inaccuracy. 1 1 dS is also .i.e6 with an edge for Black.
doubtful as Black just plays ... f7-f5. The knight 1 6 ...�xh31 1 7 1i'xd6?!
sortie to g4 will then have been given real Obviously 17 gxh3 loses to 1 7...l2Jxh3+ but I
meaning. That leaves 1 1 dxeS and now: assumed he was going to try 17 �xgS 'ifxgS 1 8
a) 1 l ...dxeS 1 2 h3 lbf6 13 'ifxd8 llxd8 14 eS with the idea of lbe4 if Black caprures on eS.
.i.e3 lbe8 (1 4...�e6!?) 1 S llfd1 llxd1+ 1 6 llxd1 I had planned to play 1 8... .i.f5 as even after 19
.i.e6 17 a3 and Black had to suffer a little in exd6 Black has the advantage.
Milov-Ponomariov, Biel 2000 before obtaining 1 7 ...tt:lxf3+ 1 8 gxf3 1i'h4
a draw. Of course Black keeps the <:JUeens on.
b) 1 t ...lbxeS seems like the better recaprure, 1 9 �h2
e.g. 1 2 lbd4 lbcs 13 .i.e3 aS 1 4 f4 lbed7 1 S White is keeping g3 free for his <:JUeen.
.i.f3 a4 (aS is a good square for the black <:JUeen 1 9 . . .Wh5 20 lle3
in such positions) 16 llb1 lle8 17 .i.f2 'ii'aS 1 8 20 'it'g3 �e6 is excellent for Black.
'ii'c2 'iVb4 (Black has taken over the initiative) 20 . . .llad8 21 1i'g3 tt:lb4!
19 lbce2 lbf6 20 lbg3 hS 21 h4 l2Jg4 22 .i.xg4
.i.xg4 23 f5 lbd7 24 a3 112-112 Milov-Motylev,
linares 2001 . Knowing 1\filov rather well, I can
assure you that he must have hated his position
to offer a draw here.
1 1 ...exd4! 1 2 tt:\xd4 tt:lf6 1 3 lle1 lieS 1 4
�f3
14 f3 seriously weakens the dark squares
while 14 .i.fl is met by 14 ... lbcs.
14...tt:lh7!

21 ... �eS is also strong but after the text all

Black's pieces get to join in the attack.
22 f4 tt:\c2! 23 1i'xh3 1i'xh3 24 llxh3 tt:\xa1
25 tt:\a5
The point was of course that 2S lbxa1 is met
by 2S ... �xc3 and ...lld1+. Black has a decisive
material advantage. The remaining moves
were...
25 . . .lld2 26 tt:\xb7 llbB 27 tt:\a5 llbxb2 28
e5 llxf2 29 tt:\xc6 �fB 30 tt:\e4 llfe2 31

1 88
 Th e A verb a k h Varia tion

tiJf6+ �g7 32 tlJeB+ �hB 33 �h1 l:le1 + 34 King's Indian player to face. I can't say it both­
�g1 l:lbb1 35 tiJdB l:lxg 1 + 36 �h2 llh1 + ered me. It's just time to take the game into
0-1 Benoni territory.
8 . . . c5 9 d5 e5 1 0 dxe6 �xe6 1 1 l:ld1
Game 68 Both players were probably quite content
with their position. I was happy because I had a
Bonsch-Gallagher
lead in development and was expecting to solve
Bundesliga 2003 the position tactically. Better than suffering with
no counterplay! Bonsch was probably quite
1 d4 tiJf6 2 c4 g6 3 ttJc3 �g7 4 e4 d6 5 happy as well because he is a positional player
�e2 0-0 6 �g5 lt:Ja6 7 -.c2 and Black has a lot of weaknesses.
This position had been reached once before
and there too White immediately attacked the
d-pawn but with 1 1 0-0-0. Then, after 1 1 ...'it'a5
White decided that he had to prevent ...tt:\b4 so
he played 1 2 a3 but Black played it an}way!
1 2 ... tt:\b4! 1 3 axb4 cxb4 1 4 'ifa4 (there is noth­
ing better) 1 4 ... 'ii'xa4 1 5 tt:\xa4 tt:\xe4 1 6 �e3 b5
and Black has the better game despite the ex­
change of queens, Bagirov-Smirin, Batumi
1 999.
1 1 ...ttJb4
The only thing that perturbed me about the
black position was the knight on a6. It makes
sense, therefore, to throw this move in and find
A speciality of Uhlmann so It IS no great out where the white queen is going.
surprise to sec Uwe Bonsch, a former colleague 1 2 .b1 !?
in the East German team, trying out one of the A surprise. I had assumed he would play 1 2
ideas of his old mentor. 'ii'd2 when I was seriously toying with the idea
7 . . . h6 of playing the speculative piece sacrifice
One of the main points behind 7 1i'c2 is to 12 ...tt:\xa2!? 1 3 tt:\xa2 tt:\xe4 when Black has
render the immediate 7 ... e5 unattractive, e.g. decent compensation for the piece. The more
7...e5 8 dxe5! dxe5 9 l:.dt 1i'c8 (9...�d7 1 0 sensible continuation, however, is 1 2 ...'ii'a5
ttJd5±) 1 0 l2Jd5 1i'c6 1 1 ..i.xf6 �xf6 1 2 �g4 when 1 3 a3 lbc6 14 �xh6 is nothing to worry
'ii'c6 1 3 �xeS l:.axc8 14 b4!? with an edge for about on account of 1 4...tt:\xe4! 1 5 tt:\xe4 �xh6.
White, Uhlmann-Reschke, Dresden 1 995. 1 2 ....:es
So before playing ...e7-e5 Black forces the 1 2 ... 'ifa5 is considered a more accurate move
bishop back to avoid this line. The main reason order by Hazai (he doesn't like Black's position
he is usually reluctant to play ... h7-h6 in such after 1 2...l:.e8 1 3 tt:\£3) meeting 13 ..i.xd6 with
positions is that White sometimes gains a 1 3 .. .l:.fd8 and 1 3 tt:\£3 with 1 3...tt:\g4 14 0-0
tempo with 'ii'd 1 -d2, but this is no longer so tt:\ge5.
relevant here as White has already played 'ii'd 1 - 1 3 �xd6?!
c2. Another surprise. This rime I thought he
7 ... c6, however, is a perfectly reasonable al­ was finally going to start developing his king­
ternative to the game continuation. The idea, as side but Bonsch was probably scared off by
we saw in Game 65 is to play ... lbc7-e6 variations such as 1 3 tt:\£3 tt:\xe4!? 14 tt:\xe4!
8 �f4!? ..i.fS. I really wanted to play this but I'm not
Now 8 �c3 e5 is fine for Black so White sure Black has sufficient compensation here
decides to prevent ... e7-e5, probably assuming after 1 5 tlJfd2. There arc plenty of other 13th
that this would be the most unpleasant for the move options for Black such as 13 ...'it'a5,

1 89
Pla y th e King 's Indian

13 ...ltxl7 or 13 ...g5 and I don't share Hazai's but the game is far from over. The text just
opinion that Black is worse here. drops a whole piece.
1 3 ...'iFa5 14 a3?! 22 . . .'iFc6! 23 �1
It is a mistake to drive the knight back to the 23 tt'ld2 1i'xf3 24 li::lx f3 d3, 23 l:tg4 f5 and 23
centre. He had to play 14 li::lf3 when I would ..td3 f5 are the reasons White can't save his
still have sacrificed a piece with 14...li::lxe4!?. knight. The remaining moves were ...
The key line is 15 1i'xe4 li::lxa2 16 0-0 ..txc3 17 23 ...:xe4 24 'iFd3 'iFe6 25 :g3 :e8 26
bxc3 li::lxc3 which looks at least equal for Black. .i.f3 f5! 27 .i.d2 q.,h7 28 c5 h5 29 b4 h4
1 4. . .lZ\c6 1 5 lZif3 lZ\xe4! 30 :h3 .i.f6 31 .i.d1 'iFd5 32 .i.f3 32 . . .'iFa2
33 �g2 :g4+! 34 .i.xg4 fxg4 0-1

Summary
1) Don't play 6 ... e5?? It loses.
2) Be careful about the knight on a6 ending
up out of play. In many lines where Black plays
e5 it quickly comes to c5. In others it remains
on a6. That may be all right if the piece is per­
forming an important defensive task but not if
it is redundant and locked out of the game.
Note how in Game 65 Black quickly brings the
piece into play via c7. Remember the old adage:
one badly placed piece makes the whole posi­
tion bad.
Finally T get to play the sacrifice I had been 3) Don't panic when White flings his king­
dreaming of. side pawns down the board. There may be a
1 6 'iFxe4 .i.h3! variety of reasons for this but he is unlikely to
A great follow-up. be making a whole-hearted attempt to mate
1 7 'iFd3 you. Black doesn't get mated very often in the
After 1 7 'ifxe8+ l:txe8 1 8 gxh3 White has King's Indian. Possible motives behind White's
enough material for the queen but a dreadful advance on the kingside include gaining space
position. I believe Black is winning, e.g. in this sector of the board, scaring you into
18 .....ixc3+ (18 ...tt'ld4 is also not bad) 19 bxc3 exchanging queens or simply by advancing his
'ifxc3+ 20 �fl l:txe2 21 <ot>xe2 1i'xc4+ 22 l:td3 own pawns White hopes to prevent Black from
(or 22 'it>d2 'ifdS+ 23 �e3 li::ld4) 22...1i'e4+ 23 advancing his own - a sort of prophylactic
l:te3 (23 �d2 c4) 23...'ifc2+ 24 'it>ft 'ifd1+ 25 measure against being attacked himself.
'it>g2 1i'xd6 with a decisive advantage. 4) As in most variations of the King's Indian
1 7 ....i.xg2 1 8 :g1 .i.xf3 1 9 'iFxf3 lZ\d4 20 Black must play with great energy. I f you just
:Xd4 want to get your pieces out then the King's
The point of the combination is that this Indian is not the opening for you. Black must
move is forced because the knight on c3 is play with purpose or risk ending up in an infe­
pinned and therefore not defending the bishop rior position.
on e2. 5) Don't be afraid to sacrifice material to ob­
20 ...cxd4 21 .i.b4 'iFc7 22 lbe4?? tain the initiative. In the notes to Game 64
22 tt'ldt! was the only way to resist. I was there is a very interesting piece sacrifice while
unsure how to continue as I can pick up a sec­ Game 66 features a very thematic sacrifice of
ond pawn on c4 or h2, or play for an attack by the pawn on h6 in order to take over control of
sacrificing the d-pawn. Black is clearly better the dark squares.

190
 CHAPTER TWEL VE I
Other Systems

1 d4 ltJf6 2 c4 g6 3 ltJc3 .l11.g7 4 e4 d6 with cxf5 and f4 (or occasionally lLlc2-g3). This
l11c final chapter features the remaining is a solid line in which Black's chances of a
'important' variations in the King's Indian. successful kingside attack are slim. The main
These lines don't Guite merit a chapter of their drawback to an early i.d3 is that it slackens
own but are still relatively mainstream. White's already rather shaky !>Ti
, p on d4. There­
fore, it is no great surprise that Black's most
5 il.d3 popular defences are based on a Guick assault
against this point.
1 d4 lDf6 2 c4 g6 3 ltJc3 �g7 4 e4 d6 5 5 0-0 6 ltJge2 ltJc6
. . .

�d3 A fter 6 ... c5 7 dS Black has to be ready to

play a well known line in the Modem Benoni
whilst 6 . e5 is played surprisingly rarely.
..

7 0-0
Now:
1) In SOKID I concentrated mainly on
7 ... ltJhS, the variation I have played myself for a
number of years. It has also become the main
line of 5 �d3 and Game 70 reveals the latest
state of affairs there.
2) There have also been some interesting
developments after 7 e5 so that is covered in
...

Game 69.

Game 69
Some strong grandmasters play this system
Bareev-Tkachiev
Guitc rq.,TUlarly. I ts main supporter has been the
American !.,>randmaster Yasser Seirawan whilst Cap d'Agde 2002
in the last couple of years Ivan Sokolov has
used it as his main weapon a1.,>ainst the King's 1 d4 lDf6 2 c4 g6 3 ltJc3 �g7 4 e4 d6 5
Indian. �d3 0-0 6 ltJge2 ltJc6 7 0-0 e5 8 d5 ltJd4 9
White has a simple plan. He will develop his ltJxd4
kin!,osidc Guickly with i.d3, lbge2 ami 0-0 and 9 �gS!? is a favourite of Vasser Seirawan but
be ready to meet any subscGuent ... f5 by Black what works well for him docs not always work

19 1
Pla y th e King 's Indian

for other players. Yasser tends to operate on his 9 exd4 1 0 lt:lb5

...

own, admittedly fairly advanced, positional The alternative is 1 0 lbe2 .:.e8 1 1 f3 cS with
planet and trying to emulate his style is not an unclear game. White will try and use his
advisable. After 9 ... h6 to ..i.h4 cS we have: central majority whilst Black will try and open
lines so that his passed d-pawn can be come a
force. After 12 ..i.gS 'iVc7 1 3 'iVd2 lbd7 we
have:
a) 1 4 .:act bS 1 5 cxbS a6 1 6 b6 1i'xb6 1 7 f4
h6 1 8 ..i.h4 aS 1 9 ..i.f2 ..i.a6 with good play for
Black, Gonzalez Velez-Arizmendi, Spanish Ch.,
Palencia 1 999. Another example of a 'Gon­
zalez' playing p<x>rly against Arizmendi ..
b) 14 f4 bS 1 5 b3 bxc4 16 bxc4 ltJf6 17
..i.xf6 ..i.xf6 1 8 ltJg3 with chances for both
sides.

a) 1 1 llhd4 exd4 1 2 lbe2 1i'e8 13 ..i.xf6

...i.xf6 14 1i'd2 ..i.g7 1 5 b4 b6 16 bxcS bxcS 1 7
l:.ab1 1i'd8 1 8 .:.b3 h S 1 9 � h 1 �h7 20 ltJg1 =
Seirawan-Nunn, Cannes 1 992.
b) 1 1 dxc6 bxc6 1 2 b4 .:.e8 13 bS cxbS 1 4
cxbS gS 1 5 ..i.g3 ..i.b7 and Black had a n active
game in Dzhindzihasvili-Benjamin, Los Angeles
1 991. However, White's plan should not be
dismissed out of hand as he got into trouble in
this game by playing b4-b5 prematurely.
c) 1 1 .:.b1 , preparing b2-b4, is the most
popular choice. Now we have: 1 0 lt:le8!?
...

cl) 1 1 ...g5 1 2 ..i.g3 hS 1 3 f3 h4 1 4 ..i.e1 ltJhS An interesting move. It had already occurred
has occurred a couple of times and is worthy of a few times before but theory had dismissed it,
attention.. rather unfairly, as just good for White. The
c2) 1 t .....i.d7 12 f3 (12 b4 ltJxe2+ cxb4 1 4 standard line has been t o....:.e8 1 1 .:.e 1 and
.:.xb4 1i'c7 looks all right for Black) 1 2...g5 1 3 now:
..i. f2 ltJhS 1 4 b4 b6 1 5 bxcS dxcS, Graf­ a) 1 1 ...ltJg4 1 2 h3 a6 1 3 hxg4 axbS 1 4 cxbS
Kotronias, Aeroflot Open Moscow 2004. 1i'h4 1 5 g3 1i'h3 16 ..i.ft 1i'xg4 17 1i'xg4 ..i.xg4
I see no reason to be unhappy with Black's 1 8 �g2 is well known to give White an end­
position in lines 'cl ' or 'c2'. game plus whilst Milov's 1 5 ..i.f4!? may be even
c3) 1 1 ...'iVd7!? (Black unpins - he wants to stronger.
play ... t7-f5) 1 2 b4 b6 1 3 ltJxd4 exd4 1 4 lbe2 b) 1 t .. . ..i.d7! is a much better move and, to
lbg4 1 5 bxcS bxcS 1 6 ..i.c2? lbes 1 7 ..i.a4 'iVc7 be honest, I hadn't realised that Black could get
1 8 1i'b3 ..tg4 1 9 f3 .:.fb8 20 1i'd1 ..i.c8 21 ..i.b3 such an easy game in this fashion. I think I may
..i.a6 22 f4 ltJxc4 23 1i'd3 lba3 0-1 , Gonzalez de have been fooled by assessments of + = in vari­
Ia Torre-Arizmendi, Mislata 2003. White's per­ ous places which are not really += at all but
formance in this game was very poor and I'm more like =. Now:
sure his play can be improved upon. Even his b1) 1 2 ltJxd4 ltJxd5 1 3 cxdS (moves such as
resignation was a bit premature as after 24 1 3 lbc2 or 1 3 ..i.fl offer White no advantage
'ii'xa6 c4? he has 25 .:ct. Black should still win after 1 3 ...ltJb6; these are easy Mar6czy positions
however with 24...d3! 25 1i'xd3 c4. for Black as he has exchanged a minor piece

1 92
 O ther S ys tems

and has good development) 1 3. ..�xd4 1 4 1i'c2 when Hazai gives 1S ...l:la8 16 �b2 cS 1 7 i.ct
(or 14 1i'b3 c6=) 1 4... cS 1 S dxc6 �xc6 1 6 �e3 as best play for White. Black has an easy game
�xe3 17 llxe3 dS with equality, Mastrovasilis­ after 1 7 ... f5. The reason Black didn't play
Kotronias, Athens 2003. 1 S...tt:Ja6 is because of the awkward reply 1 6
b2) 1 2 �gS h6 1 3 �h4 gS 1 4 �g3 tbg4 1 S �d2 when after 1 6...tt:Jcs 1 7 �xaS 'iiVxaS 1 8
h3 tt:Jcs 1 6 �f1 d3 1 7 �xeS �xeS 1 8 tt:Jc3 1i'f6 1i'b4! 1i'd8 1 9 llad1 f5 h e may have quite rea­
19 1i'xd3 hS and Black had good dark square sonable compensation for the exchange but it's
control in return for his pawn, Arduman­ not the sort of thing you really want to play.
Kotronias, Zouberi 1 993. 1 5 . . .:as 1 6 lLic2 lLia6! 1 7 'ifa3?! �g4!
1 1 b4 Now White is in serious trouble as his solid
This doesn't work out so well. Perhaps blockade of d3, the bedrock of his game, is now
Bareev was unaware of the theoretical recom­ crumbling.
mendation, 1 1 �c2 1i'f6 1 2 f4 with advantage 1 8 �b4
to White. Or perhaps, like me, and presumably So he decides his best chance is to give up
Tkachiev, he just didn't believe it. This theoreti­ his queen to try and hold the square. But it's
cal assessment is based upon Sagalchik-Yuneev, pretty hopeless. 1 8 1i'b2 looks forced when
Miedzybrodzie 1 991 and the continuation after 1 8 ...tt:JcS 1 9 tbe1 Black can choose be­
1 2 ... c6 1 3 tba3 1i'e7 14 f5 tiJf6 1 S �gS h6 1 6 tween 1 9 ... f5 and the more spectacular
�h4 gS 1 7 �g3 tt:Jd7 1 8 c S which is supposed 1 9...�e2!? 20 �gS (20 �xe2 d3) 20 ...1i'xgS 21
to be good for White. Well, maybe, maybe not, 1i'xc2 l:la3 with a large positional advantage.
but Black can just play 1 4... �eS. This is surely 1 8 . . . lLicS 1 9 �xeS :xa3 20 �xa3 c5
fine for him . The knight on a3 is a rather for­
lorn-looking creature while Black may even be
able to consider taking on f5 and playing ...�h8
and l:lg8 in some lines.
1 1 . . .a5!
Carving out a potential outpost on cS which
plays an important role in the game.
1 2 bxa5 c6 1 3 t0&3 :xa5 1 4 '1Vb3
I'm not sure that Black should have been of­
fered any encouragement to attempt to ma­
noeuvre his knight to cS. The immediate 1 4
tbc2 i s probably better.
1 4 . . .lLic7!

Black is just winning. The remaining moves

were...
21 :ab1 'ii'a 5 22 �c1 f5! 23 exf5 �xf5 24
�xf5 :xt5 25 :xb7 1fxa2 26 t0&3 d3 27
lLibS d2 28 �xd2 1fxd2 29 lLixd6 :ta 30
lLibS 'ifd3 31 :e7 'ifxc4 32 lLic7 �d4 33 g3
'ifc2 34 �h1 'ii'd3 0-1

Game 70
I .Sokolov-Golubev
Bundesliga 2003

1 5 �d2?! 1 d4 lLif6 2 c4 g6 3 lLic3 �g7 4 e4 d6 5

Now he really had to play the move 1S tbc2 �d3 0-0 6 lLige2 lLic6 7 0-0 lLihSI?

193
Pla y th e King 's Indian

narural but 10 ..ic2!? forces Black to exchange

his knight and after 1 0 ...lL!xc2 1 1 1i'xc2 fS 1 2
exfS gx£5 1 3 f4 lL!f6 1 4 h3 White probably has
a slight advantage) 1 0 ...d2 fS 1 1 exfS gxfS 1 2
f4 ( 1 2 ..ig5 f4! led to a most complicated game
in Christiansen-Gallagher, Bern 1 996) 1 2...lL!g6!
1 3 g3 ..id7 1 4 l:tae1 a6 1 5 b3 with equal
chances.
8 e5 9 d5 li:Je7
. ..

Forcing White to deal with the threat to d4.

It is most uncommon for Black to move this
knight before playing ... e7-e5 in the King's In­
dian.
8 �c2
This has become the main line. I suppose
the original idea was to prevent a black knight
coming into d4 (now after 8 ... e5 9 d5 lLld4? 1 0
lLlxd4 exd4 1 1 lL!b5 White just win s a pawn)
but as we see in variation 'b' below it's not even 1 0 a4
clear that it wants to go into d4. The real reason White's whole set-up is based on neutralising
for the text then is that the bishops are simply the advance ... f7-f5. The only problem is that
better placed on c2 and c1 than they arc on d3 Black hasn't even played the move yet and so
and e3 (bishops work well from a distance), White needs to find some good waiting moves
There are two alternatives: until Black does play it. The text is the most
a) 8 d5!? lL!e5 9 f4 lL!xd3 10 1i'xd3 £5!? (I popular as not only does it gain space on the
wasn't sure if I liked this move or not but I qucenside but it also prepares to deploy the
didn't want to let White play f4-f5) 1 1 ltJJ4 c5 queen's rook actively on the third rank. The
12 dxc6 bxc6 13 ..ie3 ..id7 14 exfS brxfS 1 5 c5 other waiting moves include:
e5 16 lLlde2 d5 17 fxe5 ..ixe5 1 8 ..id4 ..ixd4+ a) t o �h 1 fS 1 1 exfS lL!xfS (White's idea is
1 9 lL!xd4 1i'g5 20. l:tae1 (the Hungarian master to meet 1 l ...gxfS with 12 lL!g1 ! followed by 1 3
Hazai, commenting in ChessBase Maga�ne, f4 which A talik claims is good for White) 1 2
stopped his analysis here and said that White is lL!e4 (1 2 g4 lLld 4 1 3 gxh5 lL!t1) 1 2...lL!f6 1 3
better because he has a strong blockading ..ig5 1i'e8 1 4 1i'd2 lLlxe4 1 5 ..ixe4 ..id7 1 6
knight on d4 against a passive bishop on d7; I l:tae1 1i' f7 an d Black i s close to equality, Soko­
disagree) 20...l:tae8! 21 :Xe8 l:txc8 22 lL!x£5 lov-Spassov, Istanbul 2003.
l::t£8! 23 lL!g3 :Xfl� 24 lLlxfl ? (overlooking a b) 10 l:tb1 fS 1 1 exfS gxfS (1 1 ...iLixf5 is the
little tactic; 24 1i'xfl! lLlxg3 25 hxg3 1i'xg3 26 solid option but the text is also fine as Black
...f2 ...e5 27 1i'e2 should be a draw) doesn't have to worry about 1 2 iLig1 here) 1 2 f4
24 ...1i'xg2+! 25 �xg2 lLJf4+ 26 �t1 lL!xd3 27 lL!g6 1 3 fxe5 dxc5 (I've seen this assessed as
lL!a4 �f7 and Black has a clear advantage as the better for White but why arc we playing the
white knights are all over the place, he has a King's Indian - for positions like this as far as I
passed pawn and the so called passive bishop is am concerned) 1 4 c5 'iti>h8 1 5 b4 f4! (Black
not looking so passive anymore (0-1 , 42), Van concedes c4 in order to bring his light-squared
der Werf-Gallagher, Cannes 1 997. bishop into play; he also gains the useful fS­
b) R ..ic3 e5 9 d5 lLlc7!? (9...lLJd4 looks more square) 16 lL!e4 ..ig4 17 l:tb3 lL!h4 1 8 l:tf2?

1 94
 O th e r S ys tems

(White seems oblivious of the danger; he had to Or 14 tL!ce2 cS!? 15 lDxf4 exf4 16 iDhS �eS
play 1 8 h3, or perhaps 17 h3 with unclear play) 17 l:tf3 'ifh4 1 8 l:th3 'iVgS 1 9 �hl l:te8 20 l:gl
1 8 ... £3! 1 9 gxf3 tL!xf3+ 20 l:tbxf3 l:txf3 21 l:tg2 l:te7 21 l:tg3 'ifh4 22 l:th3 'iVgS 23 l:tg3 'ifh4 24
�h3 22 lDgS �xg2 23 �xg2 iDf4+! 24 �xf4 l:th3 '12-'lz Pinter-Nataf, Barumi 1 999.
l:txf4 25 h4 l%g4+ 26 tLig3 l:txgS 27 hxgS 'ifxgS 1 4. . .Wh4 1 5 l:tg1
and Black soon won, Feletar-Gallagher, Pula White was, presumably, worried about Black
2000. sacrificing on g2. I suspect something has al­
1 0 ...f5 ready gone wrong with his preparation and I
If Black is going to play ... f7-f5 then he may will be most surprised if we see Sokolov repeat­
as well do so at once. Many players delay for a ing the super prophylactic plan of �hl and
move or so with lO ... aS or 10 ...�h8 and then l:tgl .
after 1 1 l:ta3 play 1 1 ...f5. This doesn't really 1 5 . .e4 1 6 f3
.

make sense as White gains more from l:ta3 than White must play this before Black can play
from whatever move Black plays. However, it is iDeS.
perfectly feasible to play lO ...aS, to block 1 6 ...�d7 1 7 lt'lf1
White's queenside play, and to then srubbomly 17 fxe4 tLlcS! 1 8 �xf4 tLig4! is the point!
refuse to play ... f7-f5 at all. For example, 1 7 . . . lt'ld3! 1 8 �xd3 exd3 1 9 'iVxd3 lt'le5 20
Dorfman-Nataf, Marseille 2001 continued (af­ Wd1 lt'lxc4 21 l:tb3
ter lO...aS) 1 1 l:ta3 �g4 (to provoke f2-f3 After 21 l:tal �d4 22 g3 1i'f6 23 l:tg2 White
which closes the third rank for the rook and the saves the exchange but is left in a miserable
dl -hS diagonal for the queen) 1 2 f3 �d7 1 3 position.
iDbS �h8 14 'iti>hl tLig8 1 5 'ifel �c8 1 6 g4 21 . .lt'las 22 :a3 :aes!?
.

iDhf6 17 'ifh4 c6 with a roughly level game.

1 1 exf5 gxf5 1 2 lbg3!?
White is hoping for 12 ...tL!xg3 1 3 fxg3! with
pressure against f5. Black does best to avoid
this.
1 2 ...lt'lf4

It may have been more prudent to repeat

moves against a man of Sokolov's calibre but
Golubev is not someone who shirks his re­
sponsibilities when he has the better of it.
23 lt'lg3 l:tf6!
Black is going to try and prove that White's
1 3 l:ta3 14th and 1 5th moves were part of a helpmate
In Korchnoi-Radjabov, Buenos Aires 2001 combination
White preferred 1 3 iDhS and after 1 3 ... iDxhS 14 24 lt'lce2 lt'lc4!?
'ifxhS iDg6 1 5 f4 exf4 16 �xf4 (16 tL!c2 has How was Sokolov planning to respond to
also been played) 1 6...'iff6 17 l:tadl iDxf4 1 8 24... f4? Obviously not with 25 tLlc4 (25 iDfl
l:txf4 'ifh6 1 9 l:th4 ifxhS 20 l:txhS �xc3 2 1 1i'f2! is also a killer as if the knight on e2 moves
bxc3 �d7 Black comfortably held the draw. Black plays l:tel) 25...1i'xh2+ 26 �xh2 l:th6
1 3 ...ltJeg6 1 4 �h1 mate. Therefore, I imagine 25 iDxf4 .:Xf4 26

1 95
Pla y th e King 's Indian

i.xf4 1i'xf4 27 lt:lhS 1i'e5 28 lt:lxg7 'itxg7 with peripheral variation of the King's Indian it can
some chances of survival. be quite a tricky system to face for the unpre­
25 %1.c3 pared. To spend a couple of tempi manoeu­
25 l:tb3 loses to 25 ... i.xa4 and after 25 l:tal vring the knight to g3 (that's where it is going)
f4 White is forced to take on f4 again. may seem like rather strange behaviour but
25 ...%1.g61 26 'ii'c2 from there it exerts influence over the e4 and
White gives up the exchange as variations fS-squares so it is not easy for Black to play on
such as 26 l:td3 f4 27 lt:lxf4 (27 i.xf4 lt:lxb2) the kingside. That is, perhaps the sophisticated
27.. ..J:I.xg3 are extremely unpleasant. part of White's strategy. The crude part is actu­
26 ....i.xc3 27 'ii'xc3 ally what most White players arc interested in.
They are going to play h4, hS and try and anack
on the kinhJSide. I f Black blocks the h-pawn
with hS, they play i.gS and try and make i.xhS
work.

Black is just winning here but Golubev now

played like a man who had about a minute left
to reach move 40. Which he probably did! He
should have just returned his knight to the cen­
tre with 27...lLle5 as he is winning comfortably Here is an example of what I mean. This is
after 28 1i'xc7 1i'xa4. Instead he played the game Iiardet-Peng Xiaomin, Geneva 1 997.
27 ...lt:lb6?! and the rest was a disaster. The re­ 1 d4 lt:lf6 2 c4 g6 3 lt:lc3 i.g7 4 c4 d6 5
maining moves were: 28 1i'xc7 l:tc8? 29 1i'xb7 lt:lgc2 0-0 6 lt:lg3 eS 7 dS aS 8 h4 hS 9 i.gS c6?!
l:tg7 30 1i'xa7 lt:lxd5 31 1i'a5 1i'c4 32 b3? 1i'xb3 (9...1i'e8! is a better move) 10 i.e2 cxdS 1 1
33 lt:ld4 1i'a2 34 i.h6 l:tf7 35 1i'e1 l:te8 36 1i'c1 lLlxdS! i.e6 1 2 i.xhS! i.xdS 1 3 cxdS gxhS 14
f4 37 lZ:Je4 1i'xa4 38 1i'd2 l:txe4? 39 fxe4 lt:le3 lLlxhS lt:lbd7 1 5 1i'£3 (White now has an un­
(39...lt:lf6� 40 i.xf4! lt:lg4 41 h3 lt:lf6 42 l:tel breakable pin and it just remains for him to
1i'a8 43 lt:lfS i.xfS 44 exfS lLle4 45 1i'd4 dS 46 bring up the reinforcements and deliver mate)
l:tb 1 l:tfB 47 i.h6 l:.f6 48 1i'e5 1 -0 1 s.. ..J:I.c8 1 6 0-0 l:tc2?! 1 7 l:tfcl l:txb2 1 8 l:tc3! a4
1 9 l:tacl (the queenside is irrelevant) 1 9 .. ..J:I.xa2
5 lt)ge2 20 'iVfS! (now the rook on c3 will swing to the
kingside) 20 ... a3 21 lt:lxg7 'itxg7 22 l:tg3 �h8
Game 71 23 i.xf6+ lLlxf6 (23 ...1i'xf6 24 1i'h5+ and mate
next move) 24 l:tc8! 1 -0
Liardet-Gallagher
Quite frightening. Because of the potential
unk 1998 dangers on the kingside quite a few players,
myself included, prefer to delay castling and
1 d4 lDf6 2 c4 g6 3 lDc3 .i.g7 4 e4 d6 5 immediately start operations on the qucenside
lDge2 with a6, c6 and bS. I f Black also refrains from
This system was devised by Hungarian play­ playing ...eS, which he usually does, he must be
ers in the 1 960's and whilst it remains a fairly careful that White can't breakthrough in the

196
 O th e r Sys tems

centre with e4-e5 himself. ers, including Korchnoi, have played 6 g3

To be honest, it is not bad to play the nor­ against me. Here, though we can just play
mal King's Indian moves (...0-0, ...e5), it just 6 ... 0-0 7 .i.g2 lbbd7 8 0-0 e5 and the only dif­
requires accurate play. However, if you have a ference to the proposed repertoire in the Fi­
choice between two roughly equivalent lines, anchetto Variation is that the white knight is on
one of which thwarts the opponent's plans and e2 instead of £3. Its unclear who this favours
the other which doesn't then one should obvi­ but if Black captures on d4 in the near future
ously choose the former. we will transpose directly to that chapter.
5 . 86
. . 6 . . .c6
Forewarned is forearmed as the saying goes.
During my preparation for this game I had
noticed what liardet did to his illustrious op­
ponent in the above game so I was more con­
vinced than ever that it would be prudent to
delay castling.

7 .i.e2
Alternatively:
a) White can also play 7 a4 to prevent ... b5
but then Black replies 7...a5!. Playing ... a6 and
... aS may seem like a criminal waste of time but
the point is that Black has secured a couple of
6 ll'lg3 useful outposts for himself on the queenside.
And what about if they play 6 £3, you may b4 is obvious but after he plays ... e5 the c5-
ask, transposing into the Samisch? Didn't you square will also fall into his hands. A possible
recommend that we play 6...c5? How are we continuation is 8 .i.e2 0-0 and then:
going to get back into that? at) 9 h4 e5 1 0 d5 lba6 1 1 .i.g5 lbc5 1 2 'ifc2
WeU, I'm afraid you are not going to get 'ifh6 1 3 .:a3 (to prevent lbb3-d4) 1 3. .. �d7
back directly into that. But don't lose any sleep with a comfortable game for Black.
over it, and no, I'm not going to write a whole a2) 9 f4 e5! 10 dxe5 (10 fxe5 dxe5 1 1 d5
new chapter because of it. No-one has ever 'ifb6! looks a little awkward for White)
played the Samisch via this move order against 1 0... dxe5 1 1 'ifxd8 .:xd8 1 2 f5 (White tries to
me. Just play percentage chess and forget about make it difficult for Black to develop his queen­
it. And if it does happen, then don't panic. I side and doesn't allow him use of the e5-square
suggest you play 6 ... 0-0 when you can meet 7 but he is, nevertheless, stiU balancing on the
�e3 with 7...lbbd7 foUowed by 8 ... c5, and 7 edge of a positional precipice) 12 ... lba6 1 3 0-0
lbg3 or 7 .i.g5 with 7 ... c5 (8 dxc5 'ifa5!? 9 lbd7 1 4 .i.e3 lbdc5 1 5 .:adt �d7 and White is
�xf6 exf6! 1 0 cxd6 f5 was the fascinating con­ fighting for equality, !·lear-Gallagher, San Ber­
tinuation in a game I had with Dorfman - very nardino 1 991
unclear). Play shouldn't be too different from b) This was not the first game I had against
what you are used to. Uardet in this line. Some years previously he
But whilst no-one has ever played 6 £3 played 7 h4 against me and after 7... h5 8 �e2
against me, strangely enough, a couple of play- b5 9 cxb5 axb5 10 b4 0-0 (often when Black

197
Pla y th e King 's Indian

has played ...h5 in response to h4 he has to ... bxc4 and ...tt:\xe5, exploiting the undefended
worry about piece sacrifices on h5, but delaying state of the bishop on d2) 0-0 1 6 b3 l:.ad8 1 7
castling until White has played a move such as tt:le4 bxc4 1 8 tt:lxc4 tt:lxc4 1 9 .ixc4 ..tfS! 20
b4 renders it extremely unlikely that White can tt:lg3 tt:lxe5 21 lLlxfS gxfS 22 ..txa6 l:td4! (White
conduct a kingside attack without allowing is allowed no peace) 23 l:.c2 tt:lg4 24 h3 tt:lf6 25
serious counterplay on the queenside) 1 1 ..tg5 ..tc3 l:.e4+ 26 l:.e2 l:.a8 27 ..txf6 l:.xa6 28 ..txe7
tt:lbd7 1 2 1i'd2 tt:lb6 1 3 0-0 (13 l:.ct) tt:lh7! 1 4 l:.xa2! 29 l:.xe4 fxe4 with a favourable endgame
.ie3 e6 (it turns out that it's White who has for Black.
problems on the kinbrside) 1 5 d5 cxd5 1 6 ..txb5 b) 8 0-0 0-0 9 e5! is probably better for
(16 exd5 bxc3 17 ...xc3 tt:lxd5 is good for White. Black should therefore play either
Black) dxe4 17 tt:lgxe4? (1 7 a4 d5 is less clear) 8 ...tt:lbd7, preparing ... e7-e5, or 8 ... bxc4 9 .ixc4
1 8 tt:lg5? d4 1 9 l:.fd 1 e5 and Black won mate­ d5 10 ..tb3 dxe4 1 1 tt:lgxe4 when 1 t ...tt:ld5 1 2
rial, Liardet-Gallagher, Geneva 1 993. ...£3 i s better for White and 1 1 ...tt:lxe4 1 2 lLlxe4
7 ... b5 8 cxb5 1i'xd4 too risky, or so I wrote back in the mid-
Or: 1 990's. AU I can say is that it doesn't feel quite
a) 8 e5!? is not considered very dangerous, so risky these days with Fritz 8 assisting the
e.g. 8...dxe5 (8 ... tt:lfd7? 9 exd6 exd6 10 .if4 defence.
tt:lf6 1 1 tt:\ge4) 9 dxe5 ...xdl+ 10 tt:lxdl (10 8 axb5 9 b4
...

.ixdl tt:\g4 11 f4 bxc4 1 2 ..te2 ..te6) 1 0...tt:lfd7 I'm not too convinced by White's plan. It
1 1 f4 with a couple of examples: seems to me he plays half-heartedly on both
wings.
The move b2-b4 can be an effective counter
to an early ... b5 by Black in the King's Indian -
first blockade and later seize the initiative with
a4 - but here White has a slightly inferior ver­
sion as he has already committed his knight to
g3. In similar positions arising from the
Samisch (not part of our repertoire) the knight
usually settles on the more relevant b3-square.
9 .0-0 1 0 l1b1 lt:'lbd7 1 1 0-0 �b7 1 2 �g5
..

h6 1 3 �e3 e5 14 •c2 •e7 1 5 l1fd1 exd4

1 6 �xd4 h5 1 7 f3 lt:'le5

at) In Tyrtania-Gallagher, Bad Worishofen

1 993 Black attacked the centre at once with
1 1 ...f6. After 1 2 exf6 exf6 1 3 tt:le4 f5 (on
1 3...�e7 I was afraid of 14 ..td2!) 1 4 tt:ld6+ q;e7
15 tt:lxcS+- l:.xc8 16 ..td2 (1 6 ..te3!? tt:lf6) tt:\c5!
1 7 .ib4 tt:lbd7 1 8 lL\c3 (1 8 l:.c1 is well met by
18 ...a5! as although 1 9 ..txc5 tt:lxc5 20 cxb5
cxb5 21 ..txb5 may appear to win a pawn it
loses the game after 2 1 ...tt:ld3+!) 18 ...�£7 1 9
l:td1 ..t f8 the game was about level.
a2) Perhaps 1 1 ...tt:lb6 offers more chances of
a complicated middlegame. Goormachtigh­
W.Watson, Brussels 1 986 continued 1 2 tt:le3 I prefer Black. White is kept occupied by the
.ie6 1 3 .id2 tt:l8d7 1 4 l:.ct ..th6 1 5 l:.ft (1 5 hole on c4 as the last thing he would want is a
0-0 looks more natural, although White may black knight establishing itself there. The move
have been worried about some combination of £3, which was required to protect e4, has also

1 98
 O th e r S ys tems

weakened the White kingside. foUows up with e2-e3. This can be quite a diffi­
1 8 tt'lf1 h4 19 "it'd2!? .:tfe8 20 �f2 tt'lh5! 21 cult line for the unprepared Black player. In my
g4 early days with the King's Indian I struggled a
21 i6'xd6 i6'g5 with . ..lbf4 to foUow is much bit playing with ... e7-e5. The turning point for
too dangerous for White. I wasn't too unhappy me carne when I suddenly noticed a game
to see the text either as the white king position where Srnyslov was playing Black against his
is beginning to open up. own system. He played with ...c7-c5 and won
21 . . . hxg3 22 hxg3 .:tad8 23 g4 tt'lf6 24 really easily. I copied him and my results im­
tt'lg3 "it'e6 25 �g2? proved drarnaticaUy.
White should have played 25 g5. Now he is 2) In Game 74 White plays �g5 with his
losing. pawn on e4. This arises after the initial moves 1
25 ...d5! 26 g5 tt'lxf3! d4 tt:lf6 2 c4 g6 3 tt:lc3 i.g7 4 e4 d6 5 i.g5.
This line is not entirely respectable and if Black
knows what he is doing he should be able to
emerge from the opening with a good position.

Game 72
C . Horvath-Rajlich
Budapest 2002

1 d4 tt'lf6 2 c4 g6 3 tt'lc3 �g7 4 tt'lf3 0-0 5

�g5 d6 6 e3 c5

White had seen that 26...dxe4 27 i6'xd8

wasn't so clear but completely overlooked this
shot. Of course 27 bx£3 just loses to 27 ... dxe4.
27 "it'f4 tt'lxe4!
I f White takes on £3 Black takes on c3 and if
White takes on e4 Black recaptures defending
his knight on £3.
28 tt'lcxe4 dxe4 29 tt'lxe4 tt'ld4 30 �xd4
:xd4 31 �f3 "it'xa2+ 32 �g1 c5 33 tt'lf6+
�xf6 34 "it'xf6 .:tg4+! 35 �1
35 �xg4 i6'g2 mate and 35 �h1 i6'f2! were
alternative ways for the game to end. 7 �e2
35 . . ."it'c4+ 36 �2 :e2+! 0-1 Alternatively:
White resigned because uf 37 �xe2 l:lg2t 38 a) 7 d5 is gaining in popularity as White
'it>e3 l:lxe2. players are beginning to accept that the main
line is just a dead end. It is covered in the next
Systems with an early �g5 game.
We shaU now look at a couple of lines where b) 7 h3 provides a haven for the bishop on
White plays an early .i.g5. h2, thereby preventing lines with ...h7-h6, ...g6-
1) Garnes 72 and 73 feature the Srnyslov g5 and tt:lf6-h5. Black has:
System, a quiet restraining line named after the b1) 7... i.f5!? 8 g4 i.e4 9 i.g2 cxd4 10 exd4
ex World Champion Vassily Srnyslov. It is tt:lc6 1 1 i.e3 i.x£3 1 2 i.x£3 lt:kl7 1 3 0-0 e5 14
characterised by the moves 1 d4 tt:lf6 2 c4 1,>6 3 dxe5 lt:klxe5 1 5 i.d5 tt:le7 16 �g5 '1>-11>, Hort­
�3 i.g7 4 tt:l£3 d6 (0-0) 5 i.g5 and White Kinderrnann, Munich 1 99 1 .

1 99
Pla y th e King 's Indian

b2) 7..li1c6 8 d5 lt::la5 9 lt::ld2 a6 10 a3 b5 1 1 eCJualise with a move such as this.

cxb5 axb5 1 2 ..ixb5 ..id7 1 3 ..ie2 h6 1 4 ..if4 9 .ltJe4
. .

l:lb8 1 5 l:lbt 1i'c7 t6 0-0 c4 (Kuligowski­ 9...lt::lbd7 is an imponant alternative which I

Hawelko, Polanica Zdroj t 984), with compen­ have recommended in previous works (and still
sation according to ECO. recommend it). The idea is to prepare ...1i'd8-
c) 7 dxc5 is anti-positional. After 7 ...dxc5 8 b6, which is not playable at once in view of tO
1i'xd8 l:lxd8 9 ..ixf6 ..ixf6 10 lt::ld 5, 10 ...lt::lc6 is ..ixf6 foUowed by t t lt::ld 5. After 9...lt::lbd7 we
very comfortable for Black whilst 10 .....ixb2!? have:
t t l:lbt ..if6 t 2 lt::lc7 ..ic3+ 1 3 �e2 ..ifS t 4
l:ldt lt::lc6 t 5 lt::lxa8 l:lxa8 looks like good value
for an exchange.
7 h6 8 ..i.h4 ..i.f51
...

a) t O d5 1i'b6 t 1 lt::la4 (ugly, but 1 t 1i'd2 g5

t 2 ..ig3 lt::le4 t 3 lt::lxe4 ..ixe4 and t t 1i'h3 g5 t 2
..ig3 lt::lc4 13 lt::lxe4 ..ixe4 t 4 lt::ld2 ..ig6 both
seem to lose a pawn, whilst t t b3 g5 t2 ..ig3
Black has tried numerous moves in this posi­ lt::lc4 1 3 lt::lxe4 ..ixe4 t 4 l:lct ..ib2 costs White
tion but my attention was originally drawn to an exchange for uncenain compensation)
the modest-looking text when I spotted that it t t ...1i'a5 t 2 lt::ld2 lt::lb6 1 3 lt::lc3 (13 it::lxb6 1i'xb6
was the choice of Smyslov when somebody had again leaves the b-pawn in difficulties)
the temerity to play his own system against him. 1 3 ...1i'b4! t 4 1i'h3 (t4 e4 1i'xb2! t 5 l:lct ..id7
There are at least a couple of good reasons for looks like a relatively safe pawn) t 4... lt::lbxd5!
putting the bishop on fS. The first is to intro­ (the exposed position of the bishop on h4 is the
duce the possibility of ...lt::le4. Exchanging his key point in this simple, but pleasing combina­
knight on f6 for White's on c3 is almost always tion) 1 5 cxd5 (or t 5 lt::lxd5 lt::lxd5 t 6 1i'xb4
a good deal for Black as with all the excess lt::lxb4 t 7 ..ixe7 l:lfe8 t 8 ..ixd6 l:lad8)
baggage removed from the long diagonal the t 5 ... 1i'xh4 t 6 1i'xb7 1i'b4! t 7 1i'xb4 (there is no
full force of the King's Indian bishop is likely to choice for White as t 7 1i'xe7 1i'xb2 loses mate­
be felt. The second reason for putting the rial) 1 7 ... cxb4 t 8 lt::lb5 lt::lxd5! t 9 ..if3 ..id3! 20
bishop on f5 is that it covers bt and in a sur­ ..ixd5 ..ixb5 2t ..ixa8 ..ixft 22 ..ie4 ..ia6 0-t,
prising number of variations this allows Black Pachmann-Smyslov, Amsterdam t 994.
to mount a decisive assault against the b2 pawn, The moral of the story: don't play the Smys­
which in this variation is more likely to have a lov System against Smyslov
coating of sugar than the usual arsenic. b) 10 l:lct 1i'b6 (10 ...lt::le4 1 t lt::lxe4 ..ixe4 t2
9 0-0 lt::ld2 g5! has also been suggested as roughly
Or: level) t t b3 l:lfe8 t 2 h3 g5 1 3 ..ig3 lt::le4 t 4
a) 9 h3 has been played but after 9 ... lt::lc4 lt::lxe4 ..ixe4 t 5 dxc5 lt::lxc5 t 6 lt::ld4 l:lad8! with
White is simply a tempo down on the game. good play for Black, Law-Gallagher, British
b) I haven't seen any examples of 9 ..id3 but Ch., Hove t 997.
perhaps White should already be looking to c) to 1i'd2 lt::le4 is level while 10...1i'b6!? t t

200
 O ther S ys tems

lbd5?! tt:Jxd5 1 2 exd5 l:fe8, with ideas of ...�e4 1 2 ...�c8! 13 d5! �xb2 14 l:l.b1 �g7
also deserves consideration. 14 ... �f6 1 5 �g3 e5 has been played which
1 0 lLlxe4 turned out to be I.JUite promising for White
10 �d3 tt:Jxc3 1 1 bxc3 �xd3 1 2 'ifxd3 tlJc6, after 1 6 dxe6. However, Black could play
Poluliakhov-Kengis, Podolsk 1 990, just leaves 1 4 ... �f6 just as a means of driving the bishop
Black with the better pawn structure. back to the less active g3-square and continue
1 0 ...�xe4 1 1 lLld2 with, weU I don't know; 1 5 ... a6, 1 5 ...1i'a5 or
I like this line because one small slip can even 1 5... �g7 come to mind. The moves ...e7-
suddenly leave White in a lost position. For e6 or ...e7-e5 remain an option for later.
example, 1 1 'ii'd2 g5 1 2 �g3 'iib6 1 3 l:fd l ? (1 3 1 5 f4 a6
l:adl is better) 1 3...tt:Jc6 14 l:acl (or 14 d5 Black prepares to open lines on the queen­
lbd8 and White can kiss goodbye to his b­ side with ... b7-b5.
pawn) 14 ...l:ad8?! (perhaps a touch too sadistic; 1 6 l:l.b3 b5 1 7 l:l.g3
Black could have cashed in at once with the
same mini-combination that he played on his
next move) 1 5 b3 (1 5 d5 would have saved the
pawn although Black would stiU be better)
1 5 ... �xf3 16 �x£3 cxd4 1 7 �xc6 dxe3! and
Black soon won, Skare-Westerinen, Gausdal
1 992.
1 1 . . .�f5 1 2 e4!?

White doesn't have anything concrete for

the pawn, just space and vague attacking
chances. But it will also be a long time before
the extra pawn can have any real effect. I con­
sider the chances to be about equal.
1 7 . . .lLld7 1 8 �h1 lLlb6
Maybe Black should keep this knight nearer
to the kingside. 1 8...b4 is an idea foUowed by
White decides to give up a pawn for a space pushing the a-pawn at every available opportu­
advantage and some vague attacking chances nity.
sacrifices. I suspect this is the best practical try 1 9 �d3 bxc4 20 �c2 c3 21 lLlb1 a5 22
as the alternatives are far from awe-inspiring, lLlxc3 �a6 23 l:l.e1 lLlc4 24 e5! dxe5! 25 f5
e.g. This attacking ploy has occurred hundreds
a) 1 2 tiJb3 tiJc6 1 3 d5 tt:Jb4 14 a3 tt:Ja6 1 5 of times in similar positions but that doesn't
'ifct 'iib6 with good play for Black, Del Rey­ make it any less dangerous!
Gofshtein, Lisbon 1 999. 25 ...g5?
b) 1 2 �£3 tlJc6 1 3 �xc6 (1 3 tiJb3 cxd4 14 25...tt:Jd6! 26 fxg6 f5 is winning for Black ac­
exd4 g5 1 5 �g3 'ifb6 is very good for Black) cording to Hazai but I would stiU be pretty
bxc6 1 4 e4 �e6 (14 ... �c8!? 1 5 dxc5 .i.xb2 16 scared after 27 1i'h5!. I t is certainly better than
l:bt �g7 is at least an edge for Black) 1 5 d5 the text however, which just loses out of hand.
�d7 1 6 'ifc2 l:b8 1 7 l:abl cxd5 18 cxd5 f5 26 �xg5! hxg5 27 l:l.xg5
with some advantage to Black, Lebel-Sharif, Surprisingly, there is absolutely no defence
French League 1 992. here.

20 1
Pla y the King 's Indian

a) 8...e5, the typical King's Indian response,

and the jury is still out on the position after 9
ttJd2! Most sources say White is better but I
have my doubts. Black should probably start
with 9 ...1i'e8 or even 9...1i'd7 as Golubev did
recently in a blitz game.
b) 8...c6 9 ttJd2 exd5 t o cxd5 leads to a
Modern Benoni which seems quite difficult for
Black, e.g. t o...1i'e7 1 1 ..ie2 lL!a6 12 0-0 lL!c7 1 3
e4 g5 1 4 ..ig3 ttJd7 1 5 ltJc4 lL!e5 1 6 lL!e3 ..id7
17 a4 and White's position was preferable in
Yusupov-Markowski, Netherlands 1 998.
c) 8... b5 is the Benko option. It is probably
slightly dubious. l •'or example, after 9 cxb5 a6
t o ttJd2 there is:
27 ...86 c1) t o ... axb5 1 1 ..ixb5 ..ia6 12 ..ixa6 l:lxa6
Or 27 ...1i'd6 28 l:lxg7+l �xg7 29 1i'g4+ �h8 1 3 1i'e2 1i'b6 14 ltJc4 1i'b7 1 5 e4 ltJbd7 16 0-0
30 f6! and Black gets mated. with advantage to White.
28 lbg7+! �xg7 29 -.g4+ �h6 30 f6 1 -0 c2) to ... e6 1 1 dxe6 ..ixe6 12 ..ie2 axb5 13
..ixb5 d5 1 4 0-0 1i'b6 1 5 a4 l:ld8 1 6 ..ixf6 ..ixf6
Game 73 17 'iV£3 ..ig7 1 8 l:lad1 , Agrest-Cicak, Swedish
Team Ch. 2003, which doesn't look like quite
Vusupov-Gallagher
enough for the pawn.
Dresden 1998 d) I have suggested in the past that Black can
try 8...1i'b6 to try and reach similar positions to
1 d4 iDf6 2 c4 g6 3 lDc3 .i.g7 4 iDf3 d6 5 those discussed in Game 72, but even I forgot
.i.g5 0-0 6 e3 c5 7 d5 to play it when given the chance.
In the past I was quite puzzled as to why c) 8...g5 9 ..ig3 ltJh5 is the most principled
nobody ever played this move. In my opinion it reaction to White closing the centre and proba­
is the only test of Black's opening strategy. bly the reason most White players avoided 7 d5
7 ... h6 8 .i.h4 in the past. Black does !>,et a powerful King's
Indian bishop to compensate for his kingside
weaknesses. Here are a couple of examples:
e 1) to ..id3 f5 1 1 ttJd2 (1 1 e4? f4 1 2 lL!xg5
fails to 12...1i'e8! whilst 1 1 ltJxg5 leads to a
draw after 1 t ...ltJxg3 12 hxg3 hxg5 13 'ii'h5 l:lf6
14 'ii'h7+ �£7! 1 5 'iVhS+ �g8 16 'ii'h7+ �£7 17
'iVhS+ �g8 1 8 'ii'h7+ 1/2-1/2, Pert-Smirnov, Av­
iles 2000) 1 t ...ltJxg3 1 2 hxg3 and now 12 ...e5?
1 3 g4! aUowed White to take control of the light
squares in I .Sokolov-Radjabov, Sarajevo 2002,
but 1 2...ttJd7 1 3 ltJ£3 lLif6 14 1i'c2 ltJg4 seemed
perfectly satisfactory for Black in Fiorito­
l lanos, Rosario 2000.
e2) 10 1i'c2 l:le8 (to... f5 is an alternative as
8 a6
... long as Black doesn't meet 1 1 ..ie2 with
I'm not at aU sure what the best move is but 1 t ...ttJd7?? 12 lL!xg5! ltJxg3 13 lLic6 when
l'm presenting my game with Yusupov as the White was winning in Speelman-Polzin, Bun­
main line as it contained some interesting open­ dcsliga 2002) 1 1 ..ie2 lL!xg3 12 hxg3 e6 1 3 ttJd2
ing play. The numerous alternatives include: exd5 14 ltJxdS ltJc6 1 5 l:ld 1 l:lb8 1 6 ltJe4 a6 17

202
 O th e r S ys tems

a6 17 a3 ..te6 1 8 lLlxd6 lLld4 19 .:f.xd4 cxd4 20 �g5 h6

lLlxe8 'ii'xe8 21 lLlc7 'ii'c6 22 lLlxe6 'ii'xe6 23 e4 I t makes sense to put the yuestion to the
(23 exd4 .:f.e8) 23 ....:f.e8 24 0-0 'ii'xe4 25 'ii'xe4 bishop before committing oneself in the centre.
1l2- 1l2 Agrest-Kozul, Plovdiv 2003. 6 �h4
9 tt:ld2 6 ..te3 invites Black to play 6 ...lLlg4. Several
Ruling out any ... g6-g5 foUowed by ... lLlf6-h5 games have continued 7 .tel eS 8 dS f5 9 ..te2
ideas that Black may have still entertained. lLlf6 1 0 gxfS gxfS 1 1 ..thS+ lLlxhS 1 2 'ii'xhS+
9 ...tt:lbd7 'it>ffi with about eyual chances. Black has lost
9 ... b5 1 0 cxbS transposes to line 'c' above. I the right to castle, but he has gained the bishop
didn't want to play this sacrifice just yet but it is pair and has a strong central position.
certainly a useful option to have available. 6 . . . g5
1 0 �e2 tt:lh7 1 1 e4 tt:le5! This is more accurate than 6 ...c5 as that gives
Wbite the extra possibility of 7 eS!?.
7 �g3 c5 8 d5
Colin Crouch, who for some reason kept re­
peating this line, has also tried 8 dxcS. After
8 ... 'ii'a 5 9 'ii'd2 dxcS 10 h4 lLlhS he suffered one
painful defeat when he allowed Black to play
...ltJxg3 so in the next game he played 1 1 ..txb8!
and managed to draw. Black could also have
played 9...1i'xc5.
8 ...'iFa5 9 �d3?!
9 'ii'd2 is the lesser evil, but Wbite players
were reluctant to play this as Black can take the
bishop pair with 9 ... lLlh5.
9 ...tt:lxe4!
1 2 �g3
12 f4 would have been met by 1 2 ...g5! with
good play for Black.
12 . . .f5 1 3 f4 tt:lg4 14 �xg4 fxg4 1 5 0-0
tt:lf6 1 6 e5 tt:lh5 1 7 'iFe2
Wbite would like to play 1 7 e6 but he can't
as it loses ro 1 7 ... ltJxg3 1 8 hxg3 ..td4+ 19 �h2
l:.fS and lthS mate.
1 7 ...tt:lxg3 1 8 hxg3 b5!
Now is a good moment to play ...b7-b5.
believe Black has the better game here.
1 9 l:lae1 �f5 20 ttld1 bxc4 21 tt:lxc4 dxe5
22 tt:lxe5 -.xd5 23 tt:le3 -.e6
The game came to a sad conclusion now. I
was extremely short of time but 23 ...'ii'xa2! This combination IS the justification of
looks like a serious advanta).,>e for Black. Black's play.
24 ttl3xg4 l:tad8 % - % 1 0 �xe4 �xc3+ 1 1 bxc3 -.xc3+ 1 2 �1
f5!
Black now wins back one of the white bish­
Game 74
ops and should remain a pawn up. In such
Spassky-Fischer
sharp positions, though, material is not of
Sveti Stifan/ Belgrade (m16) 1992 paramount importance.
1 3 l:tc1
1 d4 tt:lf6 2 c4 g6 3 tt:lc3 �g7 4 e4 d6 5 A supposed improvement over Stein-GeUer,

203
Pla y th e King 's Indian

Moscow 1 966 which went 1 3 lbc2 1i'f6 14 �c2 Black has completed his development and
f4 1 5 h4 .l:t£8! with an excellent game for Black. his forces co-ordinate beautifully. White ... well,
1 3 ...1Wf6 14 h4 let's just mention his rook on h1 and leave it at
After 14 1i'h5t Black happily plays 1 4...�d8 that.
as his king has better prospects on the queen­
side than White's on the kingside.
14...g4!

23 f4 tt:\xc4! 24 tt:\h5 irf7 25 irxc4 'ii'xh5

26 :b2 :g3!
Black prepares the fatal doubling of his
Geller only considered 14 ... fxe4, when 1 5 rooks on the g-file, having calculated 27 1i'a6
1i'h5t is good for White. Many people criticised 1i'xd5! 28 1i'xa7 �c6.
fischer for his antiquated openings in this 27 .b2 irf7
match, but if you had a whole stack of novelties Matanovic has pointed out that 27...1i'g6!
gathering dust on the shelf after a twenty year was possible as after 28 1i'a6 l:txg2t 29 �ft
lay-off, I'm sure you would also be trying to get l:tg8 White has no mate.
them in when you made your comeback. 28 .i.f3 :dg8 29 'ii'b3 b6 30 ire3 irf6 31
1 5 .i.d3 :e2 .i.b5!
In Milov-Gallagher, Las Vegas 2002 I was Not falling for 31 ...e5? 32 dxe6 �c6 33
rather surprised to see my strong opponent �xc6! l:txc3 34 l:txe3 when White would be
srumbling into this position by accident. Here very much back in the game.
he played 15 llk2 and after 1 5 ... fxe4 1 6 �g1 32 :d2 e5! 33 dxe6 .i.c6 34 <M1 .i.xf3 0-1
�£5 17 lbc:3 lbd7 1 8 1i'e2 I would accept that
White has one pawn's worth of compensation, Summary
but not two. I finally won in 67 moves. 1) 5 �d3 tends to lead to a tough struggle
1 5 ...f4 1 6 lDe2 where Black's chances do not seem inferior.
1 6 �h2 doesn't help as after 1 6...g3 1 7 fxg3 2) 5 lbge2 and 6 lbg3 is surely not the an­
fxg3 it is check. 1 6 �xf4 was probably a slight swer to the King's Indian. I am recommending
improvement on the text. the same system I have for years, based on
1 6 ...fxg3 1 7 tt:\xg3 :t8 1 8 :c2 tt:\d7! delaying ... 0-0 and instigating immediate queen­
Black is more than happy to give back his side play.
extra pawn if it involves the rapid development 3) Jjnes with an early �g5 are best met by a
of his quecnside and increased attacking quick ... c7-c5 as White can experience trouble
chances on the kingsidc via the open g-ftle. on his quecnside dark squares which arc not so
1 9 irxg4 lDe5 20 'ii'e4 .i.d7 21 �g1 0-0-0 secure, especially after the moves ... h7-h6 and
22 .i.t1 :98 �g5-h4 have been included.

204
 INDEX OF VARIA TIONS I

The Classical Variation

1 d4 lLlffi 2 c4 g6 3 lt'lc3 il.g7 4 e4 d6 5 iLe2 0-0 6 lt'lf3 eS 7 0-0
7 Jl.e] Game }()
7 dxc5 Jxe5 H 'ifxdH l:txdH 9 il.g5 l:tcH
10 lt'ld5 Game 3 1 ; 10 0-0-0 Game 32
7 d5 lt'la6
H il.g5 Ganll' 28; H 0-0 Game 29
7 .. .lLlc6
7 ... lt'la6
H .!:tel c6 9 il.fl
9 ... Jl.g4 Game 26; 9 ... cxd4 Game 27
H Jl.e] lt'lg4 9 il.g5 'ifeH
10 h.1 (;ame 2 1
1 0 dxe5 dxc5
I I lLld2 Came 22
I I hJ
l l . . . f6 c;al//1' 2 5
l l . ..h6 1 2 il.d2 lt'lf6 13 Jl.e] 'ife7
1 4 a) Ctlllle 2 J; 1 4 lLldS Ca111e 24
8 d5 lt'le7 9 lt'le1
9 il.g5 Came 20
9 lLld2 aS 10 aJ lt'ld7 I I l:tb I f5 1 2 b4 �hB
13 'ifc2 Came 18; 13 f3 Came 19
') b4 lt'lh5
10 cS Came 1 7
10 g 3 f5 I I lt'lg5 lLlf6 1 2 f3 f4
1 3 b5 Callll" 1 5; 1 3 �g2 Game 16
10 l:tc t fS I I lL\gS lLlf6
1 2 iLf3 c6
1 3 Jl.e] Came· 1 1 ; 1 3 il.b2 Came 12
12 f3
1 2 ... lt'lh5 Game· 1 J; 12 ... �hH Ga111r I.J
9...lLld7 10 il.e3
10 f3 Game• I 0
I 0 lLldJ f5 I I il.d2 lLlf6 1 2 f3 f4
1 3 g4 Came 9
D c5 g5 14 cxd6 cxd(>
1 5 lt'lf2 Ga111r 8
1 5 l:k l lt'lg6 1 6 lt'lb5 .l:!.t7 17 'ifc2 lLlcH I H a4 h5 I 9 lLlf2
I <J il.fH c;ame 6; I <J il.d7 c;ame 7
••. •••

205
Pla y th e King 's In dian

10 f5 11 f3 f4 12 .i.a g5 13 l:tct
•..

1 3 a4 Game 1; 1 3 b4 Game 2; 1 3 .!iJd3 Game 5

lJ l:tfti 14 b4 Game 3; 14 c5 Game 4
...

The Slmisch
1 d4 .!t:\fti 2 c4 g6 3 .!t:\c3 .i.g7 4 e4 d6 5 f3 0-0 6 .i.e3
6 .!t:\ge2 c5 7 dS e6 8 tt:lg3 exdS 9 cxd5
9... .!t:\h5 Game 40
9... h5
10 .i.g5 Game 38; 1 0 .i.e2 Game 39
6 .i.g5 c5 7 d5 e6 8 1i'd2 exd5
9 .!t:\xd5 Game 4 1 ; 9 cxdS Game 42
6...c5 7 dxc5
7 d5 Game 37
7 .!t:\ge2 1i'a5
8 dS Game 35; 8 1i'd2 Game 36
7 dxc5 8 1i'xd8 l:txd8 9 .i.xc5 .!t:\c6 10 .i.a3 Game 33; 1 0 .!t:\d5 Game 34
.••

The Fianchetto Variation

1 d4 .!t:\fti 2 c4
2 .!t:\£3 g6 3 g3 .i.g7 4 .i.g2 d6 5 b3 c5 Game 54
2...g6 3 .!t:\o .l.g7 4 g3 o-o 5 �g2 d6 6 o-o .!t:\bd7 7 .!tJc3
7 1i'c2 Game 53
7 ... e5 8 e4
8 h3
8 ... a6 Game 51; 8...exd4 Game 52
8 exd4
.••

8 ...a6
9 l:tct Game 48
9 h3 Game 41
9 1i'c2 exd4 10 .!t:\xd4 l:te8
1 1 .!t:\de2 Game 49; 1 1 l:td 1 Game 50
9 .!t:\xd4 l:te8 10 h3 a6 11 l:te1
1 1 .i.e3 Game 46
11 ...l:tb8 12 l:tb1 Game 45; 1 2 1i'c2 Game 44; 1 2 .i.e3 Game 43

The Four Pawns Attack

1 d4 .!t:\fti 2 c4 g6 3 .!t:\c3 .i.g7 4 e4 d6 5 f4 0-0 6 .!t:\13 c5 7 d5
7 dxc5 Game 59
7 e6 8 .i.e2
•••

8 dxe6 Game 58
8 exd5 9 cxd5
•..

9 exd5 Game 57
9 .i.g4 10 0-0 .!t:\bd7 11 l:te1 l:te8 12 h3 .i.xf3 13 .i.xf3 1i'a5 14 .i.e3 b5 15 a3
. ..

15 ... .!t:\b6 Game 55; 1 5 ... b4 Game 56

Other Variations
1 d4 .!t:\fti 2 c4 g6 3 .!t:\c3 .i.g7 4 e4
4 .!t:\£3 0-0 5 .i.g5 d6 6 e3 c5
7 .i.e2 Game 72; 7 d5 Game 73
4... d6 5 .i.e2
5 .i.g5 Game 74; 5 .!t:\ge2 Game 71;
5 .i.d3 0-0 6 .!t:\ge2 �c6 7 0-0
7 ... e5 Game 69; 7 ... .!t:\h5 Game 70
5 h3 0-0
6 .i.g5 .!t:\a6
7 .i.d3 Game 62; 7 .!t:\£3 Game 63
6 .!t:\0 eS 7 d5 .!t:\a6
8 .i.g5 Game 60; 8 .i.e 3 Game 6 1
5. . .0-0 6 .i.g5 .!t:\a6 7 1i'd2 Game 64
7 f4 Game 65; 7 h4 Game 66; 7 .!t:\£3 Game 61; 7 1i'c2 Game 68

206
 INDEX OF coMPLm GAMES I

Andruet-Spassov, .f'!Jia 1990 ............................. . . . . . . . . . ................. .................................. 32

Banikas-Gallagher, French I .£ague 200 1 .................................................................... 146
Bareev-Tkachiev, Cap d'Agde 2002....................................................................... 1 92
Barsov-Gallagher, Calcutta 200 1 ................................................................................. 166
Berkes-Jobava, European Team Ch., Plovdiv 2003 ....................................................... 88
Bogdanovski-Kempinski, Halkidiki 2002 .............................................................. 102
Bonsch-Gallagher, Rundesliy,a 2003............................................................................ 190
Bumier-Gallagher, Neuchdtel 2002............................................................................. 187
Christiansen-Kasparov, Mosc01r Interzonal 1982...................................................... 159
Conquest-Mestel, Hastinl,S 1986/87.......................................................................... 155
Drasko-Gallagher, Aosta 1990 .............................................................................. 123
Dreev-Bologan, Shanghai 200 1 .................................................................................... 109
Dreev-Gallagher, Gibraltar 2004 ..................... ...................................................... 1 1 1
Fioramonti-Cvitan, Geneva 1995 .................................................................................. 34
Fritz 6-Har Zvi, Israeli Team Cb. 2000 ......................................................................... 97
Galliamova-Renet, Koszalin 1997 .............................................................................. 136
Gelfand-Markowski, Polanica ZdiT!) 1998 ................................................................... 8 1
Golod-Gallagher, ZiindJ 2003 ....... ............................. ....................... ....................... 83
Graf-Guseinov, /Jubai 2003 ......................................................................................... 100
Grooten-Motylev, Hooy,ol'ens 2003 ................................................................................ 75
Hauchard-Krakops, Cappelle Ia Grande 1997 ........................................................... 1 15
Hohler-Gllagher, Ren1 1994 ....................................................................................... 122
Horvath.C-Rajlich, Budapest 2002 ............................................................................. 200
Hubner-Polzin, Hundesliga 2003 ............................................................................. 1 3 5
Jackelen-Gallagher, Rundesli,J!,a 2002............................................................................ 7 1
Kaganskiy-Golod, l.rrael 1999 ....................................................................................... 1 7
Kamsky-Kasparov, New ) "ork (rapid) 1994 ................................................................. 54
Knott-Gallagher, British Cb., Torquqy 2002 ................................................................ 74
Korchnoi-Xie Jun, Amsterdam 200 1 ............................................................................. 22

207
Pla y th e King 's Indian

Komiushin-Ozolin, Tomsk 1997................................................................................ 1 79

Koutsin-Frolov, Kiev 1995 .......................................................................................... 9
Legky-Gallagher, French League 200 1 ........... ............................................ ............... 85
Lehtivaara-Gallagher, Neuchdtel 2004 .................................................................. 104
Liardet-Gallagher, unk 1998 .................................................................................... 19 7
Ljubojevic-Kasparov, l.inare.r 1993 ............................................................................. 61
Loginov-Ryskin, Wish 1992 ....................................................................................... 139
Lputian-Dohnatov, RostotJ 1993 ................................................................................... 58
Lukacs-Rajlich, Budapest 200 1 ...................................................................................... 64
Malakhatko-Jenni, Istanbul Ofympiad 2000 ................................................................. 4 1
Mihajlovic-Kovacevic, Yugoslav Ch. 1996............................................................ 107
Miles-Sax, l__ondon 1980 ............. . . . . . . . . . . ..................... ...................................... ................. 26
Mohr.S-Uhhnann, Bundesliga 1994 ............................................................................. 182
Moor.O-Ekstroem, Ziirich 2002................................................................................... 92
Nikcevic-Tkachiev, Cannes 1996 ............................................................................... 169
Partos-Gallagher, Swiss uague 1997 .......................................................................... 1 75
Piket-Kasparov, Tllbu'l!, 1989 ........................................................................................ 14
Pinter-Nunn, Thessaloniki Ofympiad 1988 .................................................................... 3 7
Ponomariov-Radjabov, Wijk aan Zee 2003 ............................................................... 48
Regez-Gallagher, Zurich 2003 .................................................................................... 153
Renet-Gallagher, French uague 2003..................................................................... 127
Roeder-Hebden, Bern 1992........................................................................................... 29
Rogers-Gallagher, Bundesliga 1997............................................................................... 69
Rogozenko-Gallagher, Bundesliga 1999 .................................................................... 1 32
Rufener-Gallagher, Bern 1997 ....................................................... ........................ 1 1 7
Salgado-Gallagher, L 'Hospitalet 1992 ........................................................................ 90
Shirov-Radjabov, Unares 2004 ..................................................................................... 4 3
Soffer-Mittehnan, Israeli Team Ch. 2003 ................................................................. 78
Sokolov.I-Golubev, Bundesliga 2003 ................ ..................................................... 194
Soppe-Panno, Buenos Aires 1999 ................................................................................. 161
Spassky-Fischer, Sveti Stifan/Belgrade (m16) 1992 ................................................... 204
Svetushkin-Romero Hohnes, Bled Ofympiad 2002 ................................................ 1 1 3
Tregubov-Gallagher, French ua._{l,ue 2002 ............................................................ 1 30
Vaganian-Volokitin, European Ch., Istanbul 2003 .................................................... 138
Van Wely-Degraeve, Mondariz 2000 ........................................................................... 50
Van Wely-Fedorov, European Team Cb., Batumi 1999............................................... 52
Van Wely-Gallagher, Biel 2000 .................................................................................... 73
Vera-Nataf, Montreal 2003 .............................................................................................. 21
Villamayor-Gallagher, Calcutta 200 1 ........................................................................ 14 1
Williams.S-Gallagher, British Ch., Scarborough 200 1 ............................................... 126
Yakovich-Smirin, European Ch., Saint Vincent 2000 ............................................... 185
Yin Hao-Ye Jiangchuan, Shanghai 2000 .................................................................. 124
Yusupov-Gallagher, Dresden 1998 ............................................................................. 203
Zhu Chen-Ye Jiangchuan, Three Anvws Cup, Ji Nan 200 3 .................... . .
... ....... 46
Zotnikov-Gallagher, Arosa 1996 ............................................................................... 1 71

208