Gallagher Joe - Play The King's Indian, 2004-OCR, Everyman, 211p PDF
Gallagher Joe - Play The King's Indian, 2004-OCR, Everyman, 211p PDF
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
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.
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.
Bibliography 4
Introduction 5
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
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
\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
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
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
.
11
Play tne King's Indian
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
.
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:
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
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.
17
Play the King's Indian
18
The CIBSSICBI Vana tton : M O Oern MBtn L ine
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
Game S
Korchnoi-Xie Jun
Amsterdam 2001
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
24
Th e Classi cal Vari a ti o n : Mo dern Main Line
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
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
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
28
The Classical Varia tion: 9 lbe 1 0Jd 7 with o u t 1 0 � e 3
29
Pla y th e King 's Indian
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!?
31
Pia y t h e King 's Indian
GameS
Andruet-Spassov
Sofia 1990
32
Th e Classical Varia tio n : 9 ltJe 1 ljjd 7 with o u t 1 0 i&.e3
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
.
36
The Classical Varia tio n : 9 lt:'le 1 li:J d 7 with o u t 1 0 � e 3
37
Pla y the King 's Indian
38
Th e Classical Varia tio n : 9 li:Je 1 li:J d 7 with o u t 1 0 fl.. e 3
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
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!.
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!
43
Pia y t h e King 's Indian
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
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 !?
. ..
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
49
Pla y th e King 's Indian
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
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
(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
...
54
Th e Classical Varia tio n : Th e Ba y o n e t A t tack 9 b 4
55
Pla y th e King 's Indian
56
CHAPTER FOUR I
The Classical Variation :
9th Move Alternatives
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.
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
65
Pla y t h e King 's Indian
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
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
71
Pla y t h e King 's Indian
72
The Classical Varia tio n : 7 . . .li:J a 6
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
76
The Cla ssical Varia tion : 7 . . .Ci:J a 6
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
.
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
.
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
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
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
.
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
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:
96
Th e Siimisch
97
Pla y th e King 's Indian
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
99
Pla y th e King 's Indian
Game 34
Graf-Guseinov
Dubai 2003
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
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!
. . .
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
1 04
The Samisch
1 05
Pia y th e King 's Indian
1 06
The Siimisch
Game 37
Mihajlovic-Kovacevic
Yugoslav Ch. 1996
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
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!
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 !
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
1 14
The Samisch
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
1 16
The Samisch
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!?
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.
.
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!?
.
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
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
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 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
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 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
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
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
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
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!
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
1 34
The Fia n c h e t to Varia tion
Game 50
Hubner-Polzin
Bundesliga 2003
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!?
•.•
Game 51
Galliamova-Renet
Koszalin 1997
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
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!
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
141
Pla y t h e King 's Indian
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
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
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:
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
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!
1 50
The Fo u r Pa wns A t tack
151
Pla y th e King 's In dian
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!
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!!
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
.
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!
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
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
.
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 64
White Pla ys an Early h 3
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!
.
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 68
White Pla ys an Early h 3
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
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 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
1 72
White Pla ys an Early h 3
1 73
Pla y t h e King 's Indian
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!?
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 .
1 76
Whi t e Pla ys an Early h 3
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
...
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
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
.. .
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!
1 83
Pla y the King 's Indian
Game 66
Vakovich-Smirin
European Ch., Saint Vincent 2000
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.
1 85
Pla y th e King 's Indian
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
...
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!
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
. . .
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
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.
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!
Game 70
I .Sokolov-Golubev
Bundesliga 2003
193
Pla y th e King 's Indian
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!?
.
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.
196
O th e r Sys tems
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
..
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 99
Pla y th e King 's Indian
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!?
20 1
Pla y the King 's Indian
202
O th e r S ys tems
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!
204
INDEX OF VARIA TIONS I
205
Pla y th e King 's In dian
10 f5 11 f3 f4 12 .i.a g5 13 l:tct
•..
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
.••
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
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
. ..
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
207
Pla y th e King 's Indian
208