Professional Documents
Culture Documents
Betting
Betting
Betting To Win
A Professional Guide
To Profitable Betting
Professor Leighton Vaughan Williams
HIGH STAKES
This edition published in May 2004 by
High Stakes Publishing
21 Great Ormond St
London WC1N 3JB
T: 020 7430 1021
www.highstakes.co.uk
www.gamblingbooks.co.uk/publishing
The right of Professor Leighton Vaughan Williams to be identified as author of this work has
been asserted by him in accordance with the Copyright, Designs & Patents Act 1988.
All rights reserved. No part of this book may be reproduced, stored in or introduced into a
retrieval system, or transmitted, in any form or by any means (electronic, mechanical,
photocopying, recording or otherwise) without the written permission of the publishers.
Any person who does any unauthorised act in relation to this publication may be liable to
criminal prosecution and civil claims for damages.
Conclusion 217
Appendix 1: Some Interesting And Useful Internet Sites 218
Bookmakers .................................................................................. 218
Spread Betting Companies ........................................................... 219
Betting Exchanges ........................................................................ 219
Sports And Betting Information And Advice ................................ 219
Bibliography 234
Introduction
IF EVER THERE WAS A GOLDEN AGE of betting, this is it. The
choice is extraordinary, the opportunities are bountiful, and at last
the betting man or woman has a real chance of earning money at
the same time as enjoying the thrill of the chase.
In such an environment, there is a clear need for a book which is
able to act as a guide through this choice, and through the
opportunities on offer.
My research career, which has focused on the study of betting
and financial markets, pre-dates this Golden Age, to a time when
the identification of profitable betting strategies was so much more
difficult. This served to ensure, of course, that the cutting edge of
the analysis was sharpened all the more effectively. The tools that
have resulted are based on an extensive study of the most rigorously
reviewed literature from across the world, and through original work
of mine that has been published at the forefront of the international
academic literature. These tools are all the more valuable when
applied to today’s punter-friendly betting environment.
I was approached to write this book to mark this Golden Age of
betting. My remit was to provide a clear and accessible guide, based
on the most rigorous analysis available anywhere.
Betting has always been fun, but it has not always been profitable.
Times have changed, and this book is my contribution to the new
times. I hope you enjoy reading it as much as I have enjoyed writing
it – and, most of all, I hope that it helps turn a pleasant hobby into a
profitable investment. Let battle commence!
Part One: Betting With The
Bookmaker
Chapter 1: Fair And Unfair Odds
THE ESSENTIAL IDEA behind bookmaking is a very simple one,
and the example of setting odds about the outcome of a coin toss
will serve to explain it.
Clearly the outcome of a normal coin toss is restricted to two
possibilities, i.e. ‘heads’ or ‘tails’. If it is a fair coin, there is a 50 per
cent chance of heads and a 50 per cent chance of tails. This means
about a half of all coins will end up heads and a half tails. Moreover,
the more coins we toss, the closer we get to the even split predicted
by theory.
If you are a bookmaker, and you are only interested in breaking
even, you will offer to match the punter ’s stake if he calls correctly,
but keep his stake if he calls wrongly.
For example, if the punter calls ‘heads’, and stakes £10, you will
take the £10 and wait for the coin toss. If it lands on ‘heads’, the
punter wins, and so you match the £10 stake, and return the original
stake. This is known as giving odds of ‘even money’ or ‘evens’. If it
lands on tails, however, the punter loses, and you keep the £10 stake.
This way, by offering ‘evens’, the bookmaker wins £10 every time
the punter is wrong, and loses £10 every time the punter is right.
On average the punter is right half the time and wrong half the
time, so the bookmaker can expect to win £10 and lose £10 on an
equal number of occasions.
The offer of ‘evens’ about a coin toss, or any other event with a
50-50 chance of occurring, is known as an offer of ‘fair odds’.
Take now the case of dice. What is the chance that a single die, if
thrown, will show a six? Clearly the answer is 1 in 6. What is the
chance that it will show a four? Again, the answer is 1 in 6. The
reason is that there are six possible outcomes, a 1,2,3,4,5 or 6, and
they all have an equal chance of coming up.
Before throwing the die, therefore, the fair odds about calling the
correct number is 5 to 1. This means that the bookmaker should pay
the punter 5 times his stake (as well as returning the stake) if the call
is correct, but keep the stake if the call is wrong.
This way, the bookmaker wins £10 every time the punter is wrong,
and loses £50 every time the punter is right. On average, the punter
BETTING TO WIN
is right one time in six, and so out of six throws, the bookmaker can
expect to win £10 on five occasions (a total of £50) and to lose £50 on
one occasion. In other words, in the long run, the bookmaker can
expect to break even, as can the punter.
So much for fair odds! The bookmaker ’s purpose, however, is to
make a profit, and that means offering less than fair odds.
Take again the case of the coin toss. The bookmaker who wishes
to make a long-term profit might offer, say, 4 to 5 about the coin
turning up ‘heads’ and 4 to 5 about it turning up ‘tails’.
This means that the bookmaker will pay out £4 if the punter calls
correctly, but keep £5 if the punter calls wrongly. Since the punter
will call correctly only half the time, on average the bookmaker wins
£5 half the time, and loses £4 half the time. Clearly, at these odds,
the bookmaker wins in the long run.
To summarise, if the odds set by the bookmaker are fair, the
bookmaker and the punter will break even in the long run. If the
odds are less than fair, the bookmaker will win, and the punter will
lose, in the long run.
Now it would seem foolish to bet on the toss of a coin at less than
fair odds (evens), unless you enjoy the game, or unless you know
something you shouldn’t about the coin. The same applies to the
dice game.
Of course, it is possible to win in the short run, even at less than
fair odds, by sheer luck. There is nothing, for example, to prevent a
call of heads winning ten or twenty times in a row. This way of
thinking is not recommended, however, for those looking to a long-
term betting strategy that will win money.
What’s Fair?
In the case of the coin toss, or the dice game, it is easy (unless the
game is bent) to calculate what the fair odds should be. On this
basis, it is straightforward to decide whether a bet at any given odds
is likely to yield a long-term profit, loss, or a break-even position.
What constitutes fair odds about a horse winning, or a dog, or a
football team, is altogether more difficult to identify, and two different
judges may come up with totally different answers. It is this
difference of opinion that has always allowed the betting public to
live in the hope that they can beat the bookie. The problem, of course,
14
FAIR AND UNFAIR ODDS
is that the bookmakers tend to get the odds right more often than
the punter, and, besides, bookmakers are able to offer a set of odds
which taken as a whole are designed in their favour. This margin in
their favour, or ‘over-round’, tends to be particularly large where
the bettor has a relatively large number of alternatives to choose
from, such as a 20-runner field in a horse race. In addition there has
until recently been the crippling effect of betting tax deductions.
The Golden Age of betting that we are presently experiencing
has brought with it changes in all these factors. First, the quantity
and quality of information available to bettors is now so plentiful
that, if used properly, it can be used to significant effect in countering
the advantage that the bookmaker has historically possessed in the
setting of odds, and particularly so where information is changing
rapidly. Second, the number of different bookmakers is now so
enormous that the bettor is able to shop around for odds in such a
way that at best odds the effective margin is wiped out, or even
turned to the advantage of the bettor. Finally, the switch from a
turnover tax to a tax on the gross profits of bookmakers, and the
simultaneous reduction in the burden of taxation on bookmakers,
has led to the present era of deduction-free betting.
The battle between bettor and bookmaker is at last set on
something approaching a level playing field. Indeed, in many ways
the pitch is tilted to the advantage of the bettor. It is a battle that
really can be won.
15
BETTING TO WIN
16
FAIR AND UNFAIR ODDS
17
BETTING TO WIN
first three. In horse racing, this is strictly defined. In a race of five, six
or seven runners, a place means finishing in the first two; in races of
eight or more runners it usually means to finish in the first three, or
the first four in handicaps of sixteen or more runners. Sometime
bookmakers offer more generous terms, including a place for the first
five home in some high-profile handicaps. For the record, handicap
races are races where the horses carry different weights, depending
on their ability, in order to make the race more competitive.
The usual place terms are shown below:
18
FAIR AND UNFAIR ODDS
Assume now that instead of betting £20 on each horse, you bet
£10 each way on each (i.e. a total of £20 per horse). This is made up
of £10 to win and £10 to place on each horse. Again your total stake
is £120.
However, only one horse will win, say Horse A, and another will
finish second, say Horse B.
The return to the winning horse is calculated as follows:
To summarise
19
BETTING TO WIN
20
FAIR AND UNFAIR ODDS
fraction they actually are. Win bets only are accepted if there are
fewer than five runners.
Handicaps
Runners Fair place fraction Actual place fraction No. placed
5 0.38 0.25 2
6 0.40 0.25 2
7 0.42 0.25 2
8 0.24 0.20 3
10 0.26 0.20 3
12 0.27 0.25 3
14 0.28 0.25 3
16 0.20 0.25 4
20 0.21 0.25 4
Non-handicaps
Runners Fair place fraction Actual place fraction No. placed
5 0.38 0.25 2
6 0.40 0.25 2
7 0.42 0.25 2
8 0.24 0.20 3
10 0.26 0.20 3
12 0.27 0.20 3
14 0.28 0.20 3
16 0.29 0.20 3
20 0.30 0.20 3
21
BETTING TO WIN
Multiple Bets
There exist a range of multiple bets, i.e. bets that yield a return if
more than one part of the bet comes up. The simplest is a win double.
In a win double, the bettor chooses two selections. To win, both of
the selections must be successful. If both horses win, the odds are
multiplied. If either loses, the bet is lost.
Example
You select Horse A to win the first race on the card at 2 to 1. You
select Horse B to win the second race on the card, at 10 to 1. You
now ask for a £10 win double on these two horses.
The double is calculated like this. Your £10 stake goes first on to
Horse A. If it wins, your return is £30 (2 times £10 plus your stake
returned). That £30 is then placed on Horse B. If Horse B wins, your
return is £330 (£30 at 10 to 1 yields £300, plus your £30 stake returned).
All multiple bets follow this principle.
A common alternative to the win double is the win treble. These
operate exactly like doubles, except that there are three selections
instead of two. Accumulators can be constructed in the same manner,
built up of four or more selections.
To calculate your return, convert the odds on offer into decimal
notation, if they are not shown that way to start with. To do this,
add one to the traditional (or fractional) odds. For instance, odds of
2 to 1 are represented in decimal notation as odds of 3.0. Odds of 7
to 2 (3.5 to 1) are represented, in decimal notation, as odds of 4.5.
Now multiply these decimal odds together.
So, if you have three selections at 2 to 1, 3 to 1 and 4 to 1, these are
converted into their decimal equivalents of 3.0, 4.0 and 5.0. If they
all win the return is calculated by multiplying these odds together,
i.e. 3 x 4 x 5 = 60.0.
On this basis, a £20 win treble, for example, at the odds shown
above would yield a return of 60 × £20.00, i.e. £1200.
There are numerous multiple bets, usually with strange names,
which simply combine a number of different multiple bets. For
example, if you make five selections, these can be combined into
22
FAIR AND UNFAIR ODDS
23
BETTING TO WIN
Forecast Bets
Forecast bets are similar to multiple bets, except that the bet involves
more than one horse in the same race rather than different races.
The best-known example of a forecast bet in the UK is the
bookmakers’ Computer Straight Forecast. This involves selection of
the first two horses to pass the post in the correct order.
Unfortunately, there is no simple way of working out in advance
how much you will win if you are right. Instead, the payout is
calculated according to a complex formula devised by the
bookmakers. Need one say more?
The Tote’s competitor to the Computer Straight Forecast used to
take the form of the Dual Forecast. In the Dual Forecast punters
would specify the first two home in either order (rather than the
correct order), and the pool of winning bets was shared (after a 24%
deduction).
Most independent analysis suggested that the Dual Forecast offered
better value, in most circumstances, than the Computer Straight
Forecast. The Dual Forecast has now been replaced by the Tote Exacta.
This is the Tote’s direct competitor to the Computer Straight Forecast.
24
FAIR AND UNFAIR ODDS
Choice
The rise of telephone betting, and latterly betting on the Internet,
has led to an explosion of choice for those seeking value.
Perhaps the best-known forum for highlighting the abundance
of choice, and for pinpointing value, is the Racing Post’s ‘Pricewise’
column. This has been hosted by a succession of tipsters, including
most recently Henry Rix and Melvyn Collier, both of whom have
gone on to tip on a private subscription basis. At the time of writing
the incumbent Pricewise is Tom Segal.
The Pricewise column, which appears every Saturday, and
periodically on other days of the week, highlights a selection of
bookmakers who are offering early prices about each of the horses
contesting selected races. These races are usually, though not always,
broadcast live on terrestrial TV.
Pricewise by no means highlights all the offers available, but it is
a good start, and it is complemented by advice and analysis.
Latterly, those seeking to compare the odds on offer for best value
have had the benefit of a number of Internet sites which have been
set up to cater for this very need.
A good example is Oddschecker. Oddschecker is available at
www.oddschecker.com and at www.oddschecker.co.uk
Easyodds is another odds comparison service, available at
www.easyodds.com
Bestbetting can be found at www.bestbetting.com
Other useful odds comparison sites include:
25
BETTING TO WIN
• www.smartbet.com
• www.bookiesindex.com
• www.betbrain.com
• www.crastinum.com
• www.bookiebusters.net
26
FAIR AND UNFAIR ODDS
27
BETTING TO WIN
half of your bet is a winner (the handicap bet at + ½) and the other
half of your bet is effectively void (the bet at level handicap) and
that part of your stake is returned. If you bet on Liverpool, and the
game is drawn, again half of the bet is effectively void (that portion
of the stake is returned) and the remaining half of the total stake (on
the handicap bet at - ½) is lost.
28
BETTING ON THE TOTE
29
BETTING TO WIN
evidence which shows a tendency for the odds available about such
so-called longshots with the Tote to be rather more generous than
with the bookmakers. However one sees it, the choice that the two
betting mediums offer the UK punter is clearly of significant benefit,
if used properly.
Most countries do not offer bettors the luxury of betting with
bookmakers or with the Tote. In the USA, for example, horse racing
and greyhound racing are operated by a Tote (so-called parimutuel)
monopoly. Bookmakers at the track are outlawed. The same applies
in many other countries, such as France, Canada, Hong Kong, Japan
and New Zealand.
Australia has an in-between system, whereby the Tote and
bookmakers co-exist at the racetrack, but where the Tote has a
monopoly off-course.
To summarise, the co-existence of bookmakers and the Tote offers
a clear benefit to the betting public, who can choose the better odds
on offer. For less fancied runners, in particular, this is often available
with the Tote. Either way, it is worth comparing the odds before
placing a bet, while taking account of the fact that it is not possible
to know with certainty what odds you are getting unless you take a
price with a bookmaker.
30
BETTING ON THE TOTE
Place: this is a bet on the horse to finish in the first three (usually),
although the number of finishers which qualify for a
winning place bet can vary depending on the number of
runners and the type of race.
Each way: this is a win and a place bet on the same horse. It is
available on all races of five or more runners.
Exacta: this bet requires selection of the first and second-placed
horse in the correct order.
Jackpot: this bet requires selection of the winners of all six specified
Jackpot races (usually races 1 to 6). It operates daily at a
selected meeting.
Placepot: This bet requires selection of a placed horse in all of six
specified Placepot races (usually races 1 to 6). It operates
at every meeting.
Quadpot: This bet requires selection of a placed horse in each of four
selected Quadpot races (usually races 3 to 6). It operates at
almost every meeting.
Trifecta: This requires selection of 1st, 2nd and 3rd in the correct order
in nominated races of eight or more runners.
Scoop 6: This requires the selection of the winners of each of six
nominated TV races, and operates on a Saturday only. The
races may be selected from more than one meeting.
31
BETTING TO WIN
32
SPREAD BETTING
33
BETTING TO WIN
34
SPREAD BETTING
Answer
Profit = Stake (Outcome - Buy price)
= £20 (15 - 11) equals £20 (4)
= £80
Answer
Profit = Stake x (Outcome - Buy price)
= £20 (9 - 11)
= - £40
LOSS = £40
35
BETTING TO WIN
Example
At the start of the match, the quote about the number of corners
is 10 – 11.
You buy corners at 11, for £20 per corner.
If you now change your mind and decide to close the bet entirely,
you must sell at 10 for £20.
You lose: (buy price - sell price) stake
= (11 - 10) £20
= £20
36
SPREAD BETTING
Take the same example as above: You buy corners at 11 before the
match.
Now examine the following scenarios.
Scenario 1
After ten minutes of play, four corners have been taken.
The quote has risen to 13 – 14.
If you wish to close the bet totally, you can sell at 13, for the same
stake (£20).
You now earn a sure profit of (new sell price - old buy price)
stake = (13 - 11) £20 = £40.
Scenario 2
After 30 minutes of play, no corners have been taken.
The quote has now fallen to 7 – 8.
You decide to close the bet, by a sell at 7 for £20.
You incur a sure loss of (old buy price - new sell price) stake =
(11 - 7) £20 = £80.
In each case, you have decided to close the bet to guarantee a
given profit or to cap a given loss.
37
BETTING TO WIN
Example
Opening of market: FTSE 100 is trading at 5015
The market-makers quote a spread on the day’s closing price of,
say, 5010 – 5020
One hour after the opening of the market: FTSE 100 has risen to
5025
The market-makers are now offering a spread of 5020 – 5030.
You have two choices if you wish to trade.
Choice A: buy at 5030.
Choice B: sell at 5020.
You must then choose your unit stake, say £10 per point.
Now consider the following scenarios.
Question
Scenario 1
At the close of the day’s trading, the FTSE stands at 5080.
What is the profit/loss if you adopted Choice A?
What is the profit/loss if you adopted Choice B?
Outcome A:
Having bought at 5030, you correctly predicted that the spread
was too low.
For every point over 5030 you win your unit stake, i.e. £10.
Therefore, you win (5080 - 5030) £10 = £500.
Outcome B:
As a seller at 5020, you incorrectly predicted that the spread was
too high.
For every point over 5020 you lose your unit stake, i.e. £10.
Therefore, you lose (5080 - 5020) £10 = £600.
38
SPREAD BETTING
Risk Management
Whether you are selling or buying, you can indicate to traders a
level at which you would like your bet to be closed without further
reference. This is known as a stop loss, and it is designed to mitigate
risk.
Example
You ask, in May, for a price on the June FTSE 100, i.e. the closing
price on a stipulated day in June.
You are given a June FTSE 100 quote of 5310 – 5320.
Your view is that the market will rise, so you buy at 5320, for £10
a point.
As it is an unpredictable market you decide to take out some
‘insurance’. To do this you place a stop loss order with the traders
at, say, 5270.
This means that as soon as the quote on June FTSE 100 falls to, or
below 5270 – 5280, your bet will be closed at the first trading
opportunity, for a loss of (5320 - 5270) £10 = £500.
This is a protection against further losses if the market continues
to fall. However, the stop loss cannot be guaranteed at the precise
level requested, particularly in fast-moving markets.
Example
You ask, in May, for a price on the June FTSE 100, i.e. the closing
price on a stipulated day in June.
The quote is 5350 – 5360.
You think that the level is too high, and therefore you sell at 5350,
for £10 per point.
However, you decide to leave a guaranteed stop loss order at 5450.
For this there is a premium charged of, say, 3 points. Hence, the
selling level is now at 5347 (5350 - 3).
Now, assume that between the date of the agreement and its expiry
39
BETTING TO WIN
Limit Orders
You can place a limit order at the same time as you open a bet, or
anytime thereafter, in order to close the bet if it reaches a pre-
determined profit level.
Example
You buy, in May, the June FTSE 100, at 5270, for £3 per point, in
the belief that the market will rise.
However, you think that if the market goes up by 100 points it
might well then fall back, so you place a limit order to sell at 5370 for
£3 if the quote reaches 5370 – 5380 to close the bet.
Assume now the spread for the June FTSE reaches 5370 – 5380 a
few days later in May.
Your limit order is activated and you sell at 5370 for £3.
You win: (5370 - 5270) £3 = £300.
N.B. You can also place forward orders to open a bet (buy or sell) if
a market rises or falls to a level you specify.
40
SPREAD BETTING
The spread is quoted at 5520 – 5530. You think that the spread is
too high, and therefore you SELL at 5520 for £5 a point.
Over the next few days, and before expiry of the contract, the
FTSE has gradually fallen, so you ask for a new quote. On calling,
you are offered a spread of 5460 – 5470.
You buy the June FTSE at 5470 for £5.
You have closed out with a sure profit of:
(5520 - 5470) £5 = £250
Options
This section is intended for advanced study only. Please feel free to
skip it if you wish.
Options allow you to choose whether to buy or sell a particular
market.
There are two types of option – calls (backing the market to rise)
and puts (backing the market to fall).
Options offer the security of knowing your exact downside when
buying a call or a put.
41
BETTING TO WIN
Call Options
Example
In April, with the FTSE 100 standing at 5000, you believe
that the market is set to rise substantially. However, rather
than bet on the June FTSE futures market, you decide that it
is less risky to buy a call option, because your maximum
downside is known in advance.
Say that the June FTSE 5200 call stands at 40 – 50.
You decide to buy at 50 (this is the premium) for £10 a point.
42
SPREAD BETTING
Put Options
Example
In May, with the FTSE 100 standing at 5050, you are offered a
quote for a June FTSE 4900 put at 84 – 94. This gives you the right to
sell the index at 4900 on or before the expiry date, having paid a
premium of 94.
Say, now, that the index falls to 4710. Since your put strike price
was 4900, you exercise your Option.
The difference between the strike price and the market price =
4900 - 4710 = 190.
However, the premium is 94.
So your profit = 190 - 94 = 96 x £5 = £480.
Let us now consider the opposite scenario. You buy a 4900 put
with a stake of £5 per point. The index then rises. You do not exercise
your option to buy, since your put is worthless.
Irrespective of how far it rises, your loss is your stake multiplied
by the Option price (in this case 94):
Loss = 94 x £5 = £470.
43
BETTING TO WIN
44
BETTING ON THE ‘EXCHANGES’
45
BETTING TO WIN
will win, you may decide to lay Leeds at, say, 3 to 1, instead of backing
them. Perhaps you are willing to lay Leeds at 3 to 1 up to a stake of
£20. This means that you are willing to accommodate a backer (or
backers) up to a total of £20 at 3 to 1. If your bet is fully matched, you
will be liable to pay out £60 (3 x £20) if Leeds go on to win, but you
keep the £20 if they lose or draw.
Betting exchanges include Betfair (www.betfair.com), Betdaq
(www.betdaq.co.uk, www.betdaq.com), Sporting Options
(www.sportingoptions.co.uk), GGBet (www.ggbet.com) and Betsson
(www.betsson.com).
Intrade As An Exchange
Intrade is a betting exchange which works in a similar way to futures
markets. In particular, standardised contracts are bought and sold
between exchange members. Members of the exchange trade with
each other, with the operator charging a transaction fee on each
trade.
All contracts can be bought or sold, and there are two types of contract
– short-term contracts and long-term contracts. Short-term contracts
usually cover individual events, such as a particular football match.
Long-term contracts usually cover seasonal events, such as the total
number of points obtained by a team over the course of the whole
season.
There are two basic methods of trading. The first involves PIX
(Percentage IndeX) contracts, the second totals contracts. Let us
consider these in turn.
PIX Contracts
PIX contracts are percentage representations of potential outcomes
of sporting events.
The outcomes range from a notional 0 to 100. 100 represents a
positive outcome, i.e. a win. Zero (0) represents a negative outcome,
i.e. a non-win. The contracts can, in principle, trade anywhere
between 0 and 100. When the outcome is determined, however, the
contract will expire at either 0 or 100.
Multiple PIX contracts may be listed on the same event. The
simplest case is where there are three possible outcomes , i.e a win,
a draw and a loss.
46
BETTING ON THE ‘EXCHANGES’
Example
Arsenal v. Liverpool
There are three possible outcomes, and therefore three linked PIX
contracts :
1. Arsenal win
2. Liverpool win
3. Draw
Totals Contract
Totals contracts are priced to represent some likely quantity, such as
the total number of goals scored in a football match, or number of
runs in a cricket match. Each totals contract will have a given range,
as well as a prescribed tick size and tick value. If an event results in
more or less (e.g. runs, points) than the pre-set range, the contract
will expire at the specified minimum or maximum price.
Example
Man United Total Season Points:
Tick Size 0.1, Tick Value 10p, Min Price Value 50, Max Price Value
100.
In this example the range is specified as 50 points to 100 points.
So if a client decides to BUY 10 Man United contracts at a price of
47
BETTING TO WIN
80, and the final result is 82, the trader has a profit of 20 points (10
contracts of 2 points) or 200 ticks (tenths of a point). Since each tick
in the example has a specified value of 10p, the profit is £20 (200
ticks at 10p per tick). The maximum profit is 200 points (10 contracts
of 20 points), which would be realised if Man Utd amassed 100 or
more points.
The Tradindex Sports betting exchange also works on a 0/100
principle. Bets are placed with other clients of the exchange, in the
form of shares which trade between 0 and 100. Winning team shares
expire at 100. Losing team shares expire at 0. In a draw both teams’
shares expire at 50.
Perhaps the most interesting manifestation of this sort of exchange
is that offered by Tradesports (www.tradesports.com) where traders
come together to exchange contracts based on their opinions about
the likely outcomes of a diversity of topical issues and news stories.
48
BETTING ON THE ‘EXCHANGES’
49
BETTING TO WIN
50
BOOKMAKERS AND THE EXCHANGES
51
BETTING TO WIN
52
BOOKMAKERS AND THE EXCHANGES
Betfair Betdaq
Valiramix
To back:
2.4 to 1 (£2759) 2.3 to 1 (£2263)
2.3 to 1 (£2262) 2.2 to 1 (£1182)
2.2 to 1 (£2759) 2.1 to 1 (£476)
To lay:
2.5 to 1 (£1335) 2.4 to 1 (£1003)
2.6 to 1 (£1265) 2.5 to 1 (£3750)
2.8 to 1 (£8) 2.6 to 1 (£1706)
Istabraq
To back:
3 to 1 (£655) 2.9 to 1 (£928)
2.9 to 1 (£958) 2.8 to 1 (£1429)
2.8 to 1 (£777) 2.5 to 1 (£400)
To lay:
3.1 to 1 (£695) 3 to 1 (£300)
3.2 to 1 (£1731) 3.1 to 1 (£4141)
53
BETTING TO WIN
Landing Light
To back:
3.5 to 1 (£189) 3.5 to 1 (£190)
3.4 to 1 (£851) 3.4 to 1 (£921)
3.3 to 1 (£1030) 3.3 to 1 (£574)
To lay
3.6 to 1 (£226) 3.6 to 1 (£604)
3.7 to 1 (£1232) 3.7 to 1 (£2960)
3.8 to 1 (£45) 4 to 1 (£2260)
54
BOOKMAKERS AND THE EXCHANGES
55
BETTING TO WIN
56
BOOKMAKERS AND THE EXCHANGES
Punters Beware!
The idea of betting with the bookmakers and hedging in the betting
exchanges, highlighted above, exposes one potential trap into which
57
BETTING TO WIN
58
BOOKMAKERS AND THE EXCHANGES
Now the danger is all too clear. Hedging the mythical 50 to 1 at that
17 to 1 in the exchanges is a potential road to the poorhouse.
Not wishing to place £50 on No Goalscorer at 10 to 1 (I could get
it at 12 to 1 elsewhere if I was interested), I sent off the following e-
mail to Customer Services at UKBetting (I could find no telephone
contact number), timed at 2.17 pm.
‘Dear Sir,
All bets on No Goalscorer at the incorrect price of 50/1 have
been voided (at 14.30) as this was a palpable error. The price
should have been 10/1 as reflected in our Correct Score market.
Your stake will be returned when the First Goalscorer market
has been settled, at the conclusion of the game.’
59
Part Two: Beating The Bookmakers
Chapter 1: Theory And Evidence
The key factor to consider when considering a profitable approach
to betting with a bookmaker is that the bookie offers prices about
many more eventualities than you are willing to bet upon. In other
words, you have the luxury of picking and choosing between all
the prices on offer. Similarly, you have the choice of picking among
a significant number of bookmakers, all of whom are competing for
your custom. Sometimes these prices are so disparate that it is
possible, indeed, to bet on every possible outcome with different
bookmakers and win whatever the result. Such circumstances do
not last long, though, as there are plenty of people willing to swoop
on such prices, with the effect of forcing them down to a less
generous level.
There are, however, many less fleeting ways of tackling the
bookmaker which can, if handled carefully, offer the opportunity of
turning the odds in your favour.
Economists and statisticians have worked for over half a century
seeking to identify these methods, through an examination of what
are known as betting market ‘inefficiencies’ or ‘anomalies’. What is
meant by an ‘inefficiency’ in this sense is an opportunity to make a
superior return by the use of a defined system or approach. There
are a number of these inefficiencies, and we shall consider them
each in turn.
First, though, we shall turn to perhaps the most well known, and
certainly the most rigorously tested of them all. This is the so-called
‘favourite – longshot’ bias.
horse started the race, the better on average the value. In other
words, those who systematically bet on the favourite (the horse with
the shortest odds) would over the long-term win more, or at least
lose less, than those backing any other horse or horses in the field.
This was a startling discovery, because it suggested that it was
possible to earn above-average returns by following a simple betting
system, which required no knowledge of anything other than the
available odds.
This discovery was significant not only for horse race bettors
(obviously) but also for economists. After all, how could this loophole
exist?
To the betting public, and that included quite a few economists
and statisticians, the most important thing to find out was whether
Griffith was right, and this meant collecting lots more data, from
different racetracks and at different time periods. Amazingly, study
after study came up with the same findings. The favourite – longshot
bias is indeed real.
In the subsequent fifty years and more, only one significant US
investigation has indicated otherwise, and that solely for the case of
one atypical, relatively small US racetrack.
In a way, this is not surprising, since laboratory experiments dating
back to a classic study by Preston and Baratta in 1948 all point in the
same direction. These experiments all found evidence of a systematic
tendency by subjects (under controlled conditions) to relatively
underbet or undervalue events characterised by high probability
(short odds), and to relatively overbet or overvalue those with low
probability (long odds).
Wayne Snyder, writing in 1978, surveyed all the work to that date
which had looked at US racetrack betting, and concluded that there
was indeed a strong bias which made bets on favourites much better
value than bets on longshots. The problem was that the bias was
not big enough to cover the deductions from bets levied by the
operators of the American Tote-only (‘parimutuel’) system.
A subsequent classic US study, undertaken by Richard Thaler and
William Ziemba in 1988 was a little more optimistic (from the punter ’s
point of view). While they confirmed that track deductions were
too large to make bets on the aggregate of short-priced favourites
profitable, they found that such a strategy was profitable at odds of
64
THEORY AND EVIDENCE
65
BETTING TO WIN
66
THEORY AND EVIDENCE
Before you start trotting around the globe armed with this
knowledge of how to blunt the odds against you, beware of one
place where all is not what one might expect. I refer to the Happy
Valley and Sha Tin racetracks of Hong Kong where betting on the
horses seems to be not so much a hobby as a mad passion.
Studies, covering 5343 races, published by Kelly Busche and
Christopher Hall (1988), have shown that the traditional favourite –
longshot bias just doesn’t exist over there. The same also seems to
go for Japan. Why so? Do all the shrewdies live in the Far East, or
make their way there on the first available flight? Nobody knows
for sure, although theories abound. One of the most popular
explanations is that the Tote pools are so big that it pays professional
gamblers to set up shop with the most sophisticated data processing
models. They then use these models to mop up the money placed
in the pool by mug punters so silly as to overbet the longshots. This
brings the odds into line with the true probabilities, and eliminates
the bias. Unlike bookmakers, the pool operator is happy to pay them
and pay them again, using the money put into the pool by less astute
race fans.
So much for Hong Kong. Over here, things are very different,
and no doubt the majority of punters will continue for some time to
67
BETTING TO WIN
68
THEORY AND EVIDENCE
the return, on average, to a bet on that team, and that this pattern
persists at all odds level. In other words, when betting on football,
shorter odds tend to represent better value.
That sums up the evidence for horse racing, greyhound racing
and football, but what about all the other sports out there? Well,
fortunately for us, Cain, Law and Peel have yet again come to the
rescue with an article in the Bulletin of Economic Research.
Their data set consisted of the odds and results for 50 boxing
matches, 132 Sunday League cricket matches, 647 snooker matches
and 91 tennis matches at the Wimbledon championships. For good
measure, they also looked at 24,603 US baseball games.
Their results confirmed the existence of the usual bias in favour
of backing favourites for all the other sports, including baseball.
Indeed, a system based on backing strong favourites (roughly
defined as 1 to 2 or shorter) would have generated, in the case of
boxing and cricket, pre-tax profits ranging from 12% to 16%.
There you have it then – the evidence from a series of studies
spanning half a century. If you know nothing else, a blind strategy
of betting on favourites would have yielded a much better return
than betting on any other outcome, and the shorter-priced the
favourite the better. Indeed, in some studies, such a strategy, at very
short odds, would have produced a significant profit.
Be careful, though, before sprinting off, armed with all this
information, to back a series of short-priced favourites. Statistics,
after all, work only in the long-run, or on the average. In the long
run, of course, we are all dead. The art of turning the advantage
you may have gleaned from the evidence you are starting to
accumulate will be the subject of our next chapter.
69
BETTING TO WIN
70
HOW TO BET WHEN THE ODDS ARE IN YOUR FAVOUR
twists in the air at 11 to 10, you are free from the worry of imminent
doom, but you are hardly likely to get rich in the process.
If, on the other hand, you wager half your gross income because
the odds are in your favour, you are likely to win rather more, but
you’re also more likely to have that run-in with those unforgiving
bailiffs. Just try explaining to them how that 11 to 10 really was very
good value about the flip of a very small coin.
So what should you do when you spot value? Fortunately, John
Kelly has provided us with just the answer. His so-called Kelly
criterion tells you exactly how much to stake on favourable bets in
order to maximise your fortune without sacrificing your peace of
mind.
The mathematical details are a little complex, but the essential
conclusion of his analysis is clear enough. The Kelly strategy for
maximising long-term average capital growth is to wager a
proportion of your assets equivalent to your advantage at the
available odds.
For example, if you are offered evens about a bent coin coming
up ‘heads’, and you happen to know that it has an exactly 60%
chance of landing just that way up, what is your advantage? Well,
it’s 20%, i.e. a 60% chance of heads (you win) minus a 40% chance
of tails (you lose). The Kelly formula in these circumstances advises
a stake of 20% of your ‘fortune’ on heads. By the way, for practical
purposes it may be more prudent to substitute ‘betting bank’ for
‘fortune.’ The same, of course, applies to any event where you
calculate the advantage to be 20% in your favour.
Now, what would Kelly say about the outcome of a toss of that
two-headed coin you’ve kept for a rainy day? Well, with a 100 per
cent chance of a head, and absolutely no chance of a tail, your
advantage in betting heads, however short the odds, is a whole 100
per cent. It’s a clear case of betting your ‘bottom dollar ’; it’s mortgage
the house time. After all, you can’t lose, can you? Can you? Yes, you
can! In the real world, even a double-headed coin can turn up tails.
Stake accordingly!
71
BETTING TO WIN
72
SHOULD YOU FOLLOW THE MARKET?
73
BETTING TO WIN
74
SHOULD YOU FOLLOW THE MARKET?
75
BETTING TO WIN
76
WAS THE STEAMER VALUE AT THE OFF?
77
BETTING TO WIN
The reality was the same in both cases. By the off, the price will
have reflected more or less the genuine chance of the beast. At a
succession of prices on the way down the price was probably, on
the basis of known information, on the generous side.
Win or lose, that’s the lesson to be learned for those seeking long-
term profitability.
78
WAS THE STEAMER VALUE AT THE OFF?
79
BETTING TO WIN
win £8000, and three of £2000 to win £4000. All of these were on
Bilboa. Not a drop of big money was seen on Azertyuiop until he
had lengthened to 5 to 2. Meanwhile, even at 7 to 4 the new favourite
continued to attract two bets of £2000 apiece.
At the time, of course, those sitting at home had no way of
knowing the exact nature of the trades, but the market told its own
tale. By the time that Bilboa had contracted to 7 to 4, the signal to act
had arrived. It would never represent better value, given the available
information. Perhaps those cross-checking with the spreads could
have done even better. Either way, the nod was given Bilboa’s way.
In the event, the French-trained filly never looked in doubt, as
she jumped accurately and effortlessly, before leaving her rivals
trailing in her wake for a 21-length victory.
So there we have it. Two standout favourites in a decent-sized
race, one of whom eventually attracts the smart money.
In the world of markets and odds, life doesn’t come much more
shapely than that!
80
THE GAMBLER’S FALLACY
81
BETTING TO WIN
period of 60 days. Now, the New Jersey Lottery pays out to winners
a share of the pool of all bets, so a higher payout implies that fewer
people chose that number. If the gambler’s fallacy applied to the New
Jersey Lottery, we would expect the payout to increase if a number
appears twice in a short period of time. This is because people would
be less likely, on average, to choose a particular number after it has
just been drawn. This means a bigger payout if these numbers do
crop up because there are fewer winners to share the pool.
Terrell’s results confirmed what advocates of the gambler ’s fallacy
had always thought. Numbers which repeated within a 60-day
period had a higher payout than when they won previously on 80
of the 97 occasions when this happened. Moreover, the more quickly
the numbers repeated, the greater was the effect. To be precise, Terrell
found that the expected payout on a number that repeated increased
by 28% one day after winning. Therafter it decreased from this level
by about half a per cent each day after the number won, returning
to the original payout level (as if nothing had happened) after 60
days or so. In other words, the effects of the gambler’s fallacy virus
lasted a couple of months in New Jersey.
Interestingly, Charles Clotfelter and Philip Cook, writing in the
journal Management Science in 1993, identified a similar, albeit stronger
effect, in the fixed-odds Maryland Lottery.
The fallacy really does seem to get everywhere, at least in the
lottery world. How about horse race betting?
Well, just for a moment let your imagination lead you to the track.
It is a balmy summer’s afternoon, and you are prowling the paddock
in search of the winner of the fourth on the card? The fourth? Yes,
you already have your head down. After all, you just knew the
favourite was going to win the first race, but at the last moment that
healthy 8 to 1 about your second choice got the better of you.
Still, how on earth could you have been so stupid as to miss that
good thing in the second as well? Once – well, accidents happen.
But twice – now that’s carelessness. Never mind, you tell yourself,
you wouldn’t have won much anyway. Still, that doesn’t explain
why you missed the too-good-to-be-true 9 to 2 on offer about the
favourite in the third. What now? Well, you really do fancy that
short-priced animal in the next, and have come ready to turn over
the men with the satchels with one big thump. Hold on a minute,
82
THE GAMBLER’S FALLACY
83
BETTING TO WIN
84
DO MULTIPLE BETS MAKE SENSE?
85
BETTING TO WIN
the Dual Forecast. This bet, like the CSF, required punters to select
the first two horses past the post. However, unlike the CSF, the Dual
Forecast did not require the correct finishing order to be specified.
There is also one other aspect of the bookmakers’ pricing strategy
in the UK that reduces the ability of substitute products to provide
effective competition for CSF bets. The pricing formula used to
calculate the CSF returns is not publicised and is difficult to obtain,
although technically it is not withheld. Furthermore the likely payout
to each CSF combination is not made available to consumers during
the course of betting. Indeed, the only information to which
consumers have ready access is the actual CSF payout to the winning
selection. It would still be possible (though not without cost) for
consumers to estimate the expected price of a CSF bet by using
historic data on winning payouts. However, suppliers hinder even
this by changing the pricing formula every so often.
In summary, there are plenty of reasons why we would expect
the CSF bet to offer worse value to the punter than win bets. Is the
value actually as bad as we would expect?
Well, as it happens, this very subject has been addressed in a
paper I co-authored, with David Paton, for the journal, the Review
of Industrial Organization.
In that paper, we compared the return to CSF bets with what
would be expected on the basis of the starting prices of the horses
involved. Unsurprisingly perhaps, we found that the average CSF
payout was significantly worse, even allowing for the lower average
return generated by longshots. The worst value of all, we found,
was concentrated in pairs of longer-priced horses.
So much for the Computer Straight Forecast, but does this mean
that all multiple bets represent poor value compared to their single
equivalents?
The way in which bookmakers frame their bets certainly suggests
so. Take the Lucky 15, for example. This comprises 15 bets, made up
of all the possible combinations (singles, doubles, trebles,
accumulator). The Lucky 31 bet is the five-runner equivalent of the
Lucky 15, while the six-runner equivalent is called the Lucky 63. So
keen have the bookmakers been to encourage this sort of bet that
there are a variety of special bonuses available to punters who are
keen to dabble at this end of the market.
86
DO MULTIPLE BETS MAKE SENSE?
87
BETTING TO WIN
88
THE REAL LESSON OF DETTORI DAY
back seat when millions of pounds are at stake. In the event, for the
princely sum of one pound, a seven-horse Dettori accumulator
yielded, at starting price, £25000. At early prices, the same stake
would have knocked back the bookmaker almost a quarter of a
million!
For those who never did fancy Frankie’s mounts on the day, isn’t
all this spilt milk? Not at all, for there is a lesson to be learned. Just
look at the details. Even by the fourth race, the odds had started to
tumble on Dettori’s mounts, for no other reason that that they were
the lag end of multiple bets. Very few of these had, in fact, gone the
full monty, most of them being doubles, trebles and four-horse
Yankees. Thus, Decorated Hero in the 3.55 had opened in the
morning at 14 to 1, but was returned at half that. In the fifth, Fatefully
had opened in the morning at 100 to 30 but was returned at 7 to 4.
This tells a valuable tale, and one which in smaller measure applies
each and every day.
The point of the tale is this. If you follow particular patterns or a
particular system in placing multiple bets, beware if these patterns
or your system are likely to be followed by others. For instance, you
may back all the favourites at the televised meeting, or the horses of
a top trainer, or all the odds-on shots. If so, be wary of taking these
at starting price. By the time the liabilities have mounted up, that
healthy 6 to 1 about the last leg of your Henry Cecil Newmarket
accumulator may just have turned to 3 to 1 (if you’re lucky). The
result is that your potential winnings are halved, or worse. For such
bets, seriously consider taking the early price.
It’s a tale which the one-time odds-on offer about Marion Jones
also tells, in an alternative but complementary fashion, in the context
of the Sydney Olympic Games. True, a very different context to
racing at Ascot, but very similar market processes at work. It was
the day when a much-hyped five-timer for Jones (Golds in 100 and
200 metres, long jump and two relays) was punctured by the superior
technique of long jump specialists Fiona May and Heike Drechsler.
Again, the lesson was the same. If you follow particular patterns
or a particular system in placing multiple bets, beware if your strategy
is likely to be followed by others. Whatever the context, you can be
sure that the bookmakers will be alert to it and will take defensive
action. In the case of a series of horse races, the defence, as on Dettori
89
BETTING TO WIN
Day, was to shorten up the starting prices of the tail end of potential
winning multiples. In the case of well-hyped accumulators, of the
Marion Jones variety, the prices are shortened up well in advance.
The price of 2 to 1 with William Hill to beat the likes of long jump
specialists of the class of May, Drechsler and Tatyana Kotova was
miserly, and that’s not said with the benefit of hindsight. It was
flagged up on betting advice pages across the nation. Yet still the
punters came in droves. So it matters not whether we are looking to
Ascot or the Millennium Games. If your accumulator bet is an
obvious one, be careful before placing it. The critical question you
need to ask yourself is whether value exists at the prices available
about each part of the multiple. Of course, you might still be grateful
for the £25000 to a pound SP bettors won on Dettori day. Better,
though, to have placed the bet early and to go on earning a good
proportion of that that in annual interest.
90
THE REAL LESSON OF DETTORI DAY
91
BETTING TO WIN
92
DRAW BIASES
93
BETTING TO WIN
94
WHAT GOES UP, MUST COME DOWN!
95
BETTING TO WIN
96
BETTING IN RUNNING
97
BETTING TO WIN
wavering at the margin who pick up their telephones are more than
likely doing so to oppose him.
Now is your opportunity. Now is the time to side with the new
underdog. In truth, you may never get a better chance.
98
BETTING IN RUNNING
99
BETTING TO WIN
100
THE DIFFERENCE BETWEEN RISK AND UNCERTAINTY
make traditional ante-post betting look rather tame. Take the well-
known bookmaker who is offering odds about the year that ‘man’
will first walk on Mars. Still, I think I’ll leave alone their 50 to 1 about
it happening this year.
William Hill were even more far-sighted, offering 5 to 1 at one
time that Senator Hillary Clinton would become President Hillary
Clinton before she shuffles off this mortal coil. Betting to close on 11
November, 2040. Do they know something we don’t?
Ideally though ‘super-ante post’ bets should have a payout date
this side of an age at which you are still likely to care. Any takers on
the date of the end of the world?
101
BETTING TO WIN
102
STRIKE WHILE THE IRON’S HOT
price goes up; if it’s bad news they should likewise sprout wings to
sell, before they sit facing a loss. Unfortunately for the vast majority,
by the time they have acted it’s too late. The fleetest of foot, with the
quickest access to new information, have already reaped the spoils.
The same applies to betting markets. By the time the market has
settled down, the real gains to be made have for the most part
disappeared into the mists of time. The on-course racetrack betting
market is a clear case in point. The market may be open for several
minutes before a race, but by the time it closes the prices available
will more than likely incorporate all the information that is available
about the horses or dogs springing into action. That’s a pity for
those who always take the starting price, but the beauty of the British
system is that you don’t have to wait. If you have genuine new
information, it is vital to strike before the information filters out to
the madding crowd. Fortunately, because you are able to take a price,
what happens after you have plunged in doesn’t matter a jot. That’s
all very well if the information you obtain is the genuine article, and
it’s why in those circumstances the bookies’ telephone lines are
jammed with those all at once seeking to get on before the price
goes. If, on the other hand, your new information is simply a rumour
or a whisper from an unreliable source, take the advice of numerous
studies. On average, the starting price is the best price ever available.
On average, of course, you won’t win at starting price. That’s why
genuine information, acted upon very quickly, is a recipe for long-
term betting success. At least that’s what science tells us, and it’s
been doing so ever since it was used to track down that drunk in a
field.
103
BETTING TO WIN
104
PRESCRIBING THE RIGHT DOSAGE
105
BETTING TO WIN
106
THE HOME GROUND FACTOR
107
BETTING TO WIN
watched the silent footage. The implication of the study was that in
the absence of crowd noise the officials were more equal-handed
between the home and away sides. The original referees’ decisions,
however, more accurately mirrored the behaviour of those armchair
referees who had access to sound. Alan Nevill has his own
explanation: ‘To get the crowd off their back they wave play on.’
If true, this explanation of home advantage is fascinating, but it is
only of use to a professional gambling strategy if it is not already
fully factored into the bookies’ odds. Insofar as it is not, bettors who
back teams where the ratio of home to away fans is particularly
high will, in the long run, be on to a winning strategy.
If it were only so easy, we would all be millionaires, of course. On
the other hand, perhaps we should be.
108
THE HOME GROUND FACTOR
109
BETTING TO WIN
110
DRAW BIASES
111
BETTING TO WIN
112
SMALL IS BEAUTIFUL
113
BETTING TO WIN
114
SMALL IS BEAUTIFUL
115
BETTING TO WIN
Unless, of course, you have access to that model which is just a little
bit better than the others.
116
DRAW BIASES
117
BETTING TO WIN
118
DRAW BIASES
119
BETTING TO WIN
3. How much does the tipping service cost? The more expensive
the service, the greater the stakes required to show a net profit,
even if the tips produce a theoretical profit. Of course, the
larger the stake required, the greater the risk involved, and
any returns must be weighed against this additional risk.
Blatant dishonesty is one thing, but there are more subtle methods
of generating a notional profit. Let me point out one of the more
sophisticated ploys.
Take a service which works on the following basis:
The client is provided with a tip about a horse, along with a price
(a so-called ‘minimum price stipulation’).
The client is told to bet on the horse only if the minimum price
becomes available. If it does not become available, the advice is void.
Say, for example, that the tip for the day is Mr. Nippy, with a
minimum price stipulation of 4 to 1.
120
THE VALUE OF TIPPING AGENCIES
121
BETTING TO WIN
122
BETTING ON THE OSCARS
123
BETTING TO WIN
124
BETTING ON THE OSCARS
2001
Best Actress: Julia Roberts (11 to 10) – Result:WON
Best Actor: Russell Crowe (5 to 4) – Result: WON
Best Picture: Gladiator (4 to 5) – Result: WON
2002
Best Actress: Sissy Spacek (4 to 5) – Result: LOST
Best Actor: Russell Crowe (5 to 4) – Result: LOST
Best Picture: A Beautiful Mind (5 to 4) – Result: WON
125
BETTING TO WIN
126
SECOND IMPRESSIONS ARE SOMETIMES BETTER
127
BETTING TO WIN
very much in the eye of the beholder, and in this case the bettor is at
a clear advantage over the market-maker in being able to review the
opinions of a diverse spectrum of those doing the beholding.
The conventional tools of betting analysis can and should still be
applied, however, and in particular the well-established ‘favourite –
longshot bias’. This is the regularity, explained earlier, which indicates
that, in the absence of other information, a blind strategy of betting
at shorter odds outperforms in the long run a strategy of betting at
longer odds.
Almost everywhere that it’s been tested this phenomenon applies,
and there is no reason to expect that the Miss World spectacular
should be any exception.
The last time I studied the form of Miss World was at that 2000
pageant, when Miss India was the general favourite at best odds of
7 to 1. On this basis, she should have been my choice. The problem
for me was that another Miss India had won the year before, and
surely a representative of the same country wouldn’t win again (yes,
I was temporarily affected by that dreaded affliction which I have
diagnosed in all too many others – the ‘gambler’s fallacy’). And so,
bowing to the fallacy, but cognisant of the bias against longshots, I
plumped for an alternative strategy based around the elimination
from my calculations of any entrant rated by a mainstream market-
maker as longer than 20 to 1.
On the basis of a wide selection of all odds on offer, this left at the
time of selection only the following: Miss Colombia, Miss Costa Rica,
Miss South Africa, Miss USA, Miss Venezuela, and Miss India herself.
To distinguish between the relative merits of these contenders I
decided, in the absence of superior information, to calculate the
average probabilities as implied in the odds made available by the
different bookmakers.
This gives the following guide to the approximate average odds
(best odds in brackets):
Miss India 6 to 1 (7 to 1); Miss Venezuela 6 to 1 (7 to 1); Miss South
Africa 13 to 1 (16 to 1); Miss USA 14 to 1 (16 to 1); Miss Costa Rica 15
to 1 (20 to 1); Miss Colombia 18 to 1 (20 to 1).
The solution followed directly. Take the relatively ‘generous’ 20
to 1 (average price 15 to 1) available about Miss Costa Rica, 22-year-
old Cristine de Mezerville Ferreto. She might not have the best
128
THE PROBLEM WITH MISS WORLD
129
BETTING TO WIN
130
LOOKING FOR PATTERNS
131
BETTING TO WIN
132
THE MOTIVATION THING
133
BETTING TO WIN
League Final or that opening Test Match. Who will be most up for
it, who will be seeking to even the score? Will it be this; will it be
that?
In general, you needn’t worry too much about any of those
considerations. Instead, just follow the form. And if you get it wrong,
don’t get emotional and don’t seek revenge. Keep cool, keep calm
and get even.
134
THE MOTIVATION THING
135
BETTING TO WIN
136
DRAW BIASES
137
BETTING TO WIN
shade of odds-on to win the election outright before the poll of MPs,
how come that within moments of the declaration, all firms (bar
Coral – who adjusted fast) were now quoting Clarke as the hot
favourite. After all, the odds shortly before the declaration, in the
fixed-odds and spread markets, were generally of a mind that
Duncan Smith and Clarke would make it through, and yet Mr DS at
that time was still being quoted at odds as short as 1 to 2.
Perhaps the logic lies in the belief by some that Duncan Smith
would easily beat Portillo, but not Clarke, and that the chance of
Portillo coming through to the final play-off skewed matters.
Perhaps the fact that Clarke scored more votes than expected
among MPs had an impact. The problem with that analysis is that
Conservative MPs make up only one in 2000 of the party
membership. That apart, it was always expected that all three
candidates would poll somewhere between 50 and 60 votes. So the
difference between the actual and expected vote was marginal at
best.
All we can say with confidence is that the odds-setters themselves
had very little confidence in the odds they set about internal
Conservative Party matters. Indeed, this inability of pundits and
market-makers to fathom the mind of the Conservative Party has
echoes dating back to the accession of Alec Douglas-Home to the
job in the early 1960s. They were no better informed in 1997, when
the Clarke-Redwood pact led to a drastic shortening of the former
Chancellor ’s odds, just at the time as the Tory MPs were recoiling
so heavily from the deal as to give the prize to the Member for
Richmond.
I decided, therefore, that this was one event where following the
odds was for mugs. Instead I used my so-called ‘common sense’ in
guessing that a bigger name would always appeal more to the rank
and file. In other words, I was banking on a glorified version of
what American political pundits tend to call the ‘name recognition’
factor.
The rest, as they say, is history!
138
DRAW BIASES
139
BETTING TO WIN
140
DRAW BIASES
141
BETTING TO WIN
142
Part Three: Beating The Tote
Chapter 1: When Betting Longshots,
Getting What You’re Given Really
Can Be The Best Option
It’s a balmy summer evening at Newmarket’s July course, as you
study the horses parading before the next race. That mount of
Dettori’s looks particularly well. So you walk the few yards to the
row of betting booths representing the Horserace Totalisator Board,
and stake your last fiver that the favourite will indeed frank the
form. Or do you? Well, it’s definitely a lot more convenient than
trekking hastily back to the bookies’ boards lining the long straight,
and you’re certain to get on in time. You can’t be sure of the price
you’re getting, of course, but there’s no reason to expect a bad deal,
is there? According to standard economic theory, you would be right.
The so-called ‘Efficient Markets Hypothesis’, to be precise, suggests
that the returns to investments of similar riskiness placed in any of
a number of complementary markets should even out over time.
Whether you invest with the bookies or the Tote, therefore, you
might expect the same return. Otherwise clued-up investors should
step in and mop up any value until it disappears. But do they? In
most markets, the weight of evidence suggests that they do. Not so,
it seems, at the racetrack. There are systematic differences in what
you can expect to earn for every pound staked, depending on
whether it ends up in the Tote till or the bookie’s satchel, and this
difference is especially pronounced at long odds. This so-called
‘anomaly’, as finance specialists are apt to call it, has been guessed
at for years. The first large-scale academic study to prove what was
always expected, however, can be traced to the work of two
American academics, at the time operating out of Universities in
Chicago and Kentucky.
Paul Gabriel and James Marsden examined 1427 flat races from
the 1978 British flat racing season. Using a range of statistical tests,
they confirmed that the Tote returned on average a higher amount
of money to a unit stake than the bookmakers, and that this was
especially so early in the season.
BETTING TO WIN
146
DRAW BIASES
147
BETTING TO WIN
148
DRAW BIASES
149
BETTING TO WIN
150
DRAW BIASES
151
BETTING TO WIN
to shorten from the morning line were no. 6, the bizarrely named
Hunka Hunka Lori Z (from 7 to 2 to 3 to 1), and no. 4, Color Me
Happy (from 3 to 1 to 5 to 2). Neither of these showed any signs of
the late money effect. The odds about all the other horses had
lengthened since the morning forecast.
Turn now to the findings of Peter Asch and Richard Quandt, of
Rutgers and Princeton Universities respectively. Employing data
from 729 races at the Atlantic City racetrack, Asch and Quandt
reported that the final parimutuel odds tended to be lower than the
predicted (or morning line) odds for winners. Moreover, in the case
of winners, the later in the betting period the money was laid, the
stronger was this effect. Asch and Quandt proposed an explanation
of this finding as the withholding of smart money (money bet by
people with superior information) until late in the betting period,
in order to avoid giving out market signals which could depress the
final payout odds. As in any tote system, it is not possible to take a
price.
In other words, the best horses to bet on are those which have
shortened since being published in the morning line, and particularly
those which have shortened, or shortened further, late on. Add in
to this the well-established favourite – longshot bias, which applies
in the USA as well as the UK, and which dictates that the expected
return to bets at shorter odds tends to outweigh those at longer
odds.
Applying all this theory and evidence at Laurel Park, the clear
value lay in Holly Jolly, the 4 to 5 favourite, with some signs of
support for numbers 4 and 6. Of the latter entrants, there was little
to distinguish between them, each of which had shortened a little
in the market with no late surge of support.
These figures simply drive the bet: two Trifectas, one on horses 1,
4 and 6 to finish in that order past the post. The other bet, of course,
is the Trifecta on 1, 6 and 4.
In the event, it was no. 1, Holly Jolly who took the honours,
followed home by nos. 6 and 4. The Trifecta payout was $31 to a $2
stake, or in this case $31 dollars for minimal risk. Of course, this was
too good to be true, in the UK at least, with the bookies unwilling to
accommodate further than the simple Exacta. So it’s two Exactas,
coupling 1 with 6 and 1 with 4. Ah well, at least it’s a profit.
152
US HORSERACING ON THE TV
153
Part Four: Beating The Spreads
Chapter 1 : Profit Margins And
Profitability In The Spread Markets
The bookmaker offers you evens that the coin will land on its head,
and evens that it will land on its tail. For a fair coin, those are fair
odds, and in the long run both you and the bookmaker will come
out level. Of course, the bookmaker has to earn an honest crust,
and so a notional profit margin is built into the odds, known as an
‘over-round.’ We’ve illustrated how this works elsewhere in this
volume, but it bears repeating. Imagine the bookmaker shades the
odds about the head and the tail to 4 to 5 each. At these odds, you
have to bet £10 (£5 on heads and £5 on tails) to ensure a return of £9
whichever comes up, i.e. £4 plus your £5 stake returned on the
winning bet. The over-round can be calculated as 10 divided by 9,
i.e. 1.111 (or, more usually, 111.1%). You can apply the same logic to
any number of options. The smaller the over-round the less in-built
advantage the bookmaker has in the odds, and if the total is less
than 100% the tables are turned. For fixed odds, this is a relatively
straightforward calculation. What about spread betting markets?
Take the example of a Test Match between England and Australia,
and a performance index for the teams. Assume for a moment that
the game has to end with one side winning, so that in the event of
a draw there would be a sudden-death shootout of some kind. Now
award 25 points for the win, and none for the defeat. What would
the market-makers offer about each team? The favourite would be
expected to earn more than 12.5 points, and the stronger the degree
of favouritism the higher the expected points make-up. Say the quote
is 16 – 18 about Australia and 6 – 8 about England. The spread in
each case is two points, out of 25 points available. The over-round is
quite simple to calculate in this sort of case. Whether you buy or sell
England or Australia, it is given by calculating the share of the 2
point spread in the total 25 on offer. This is 2/25, or 8%.
In fact, of course, the draw is a very real possibility, and the market-
makers usually quote a spread based on 25 for a win, 10 for a draw,
and zero for the loss. The problem is that because we don’t know
the likelihood of a draw in advance, it makes the calculation of the
BETTING TO WIN
158
DRAW BIASES
159
BETTING TO WIN
In other words, wait till the crowd gets hot. Now is the time to
play it cool.
160
DRAW BIASES
161
BETTING TO WIN
162
DRAW BIASES
163
BETTING TO WIN
164
DRAW BIASES
165
BETTING TO WIN
166
DRAW BIASES
167
BETTING TO WIN
168
QUARB STRATEGY: USING THE ODDS TO BEAT THE ODDS
seasons. Data were collected about 207 matches for the 1999/2000
season and 240 matches for the 2000/2001 season.
In the 1999/2000 season, the Quarb strategy suggested 60 trades.
On the basis of £1 being placed on each Quarb trade, a bettor would
have won £973 on 36 winning bets and lost £384 on 22 losing bets, a
net profit of £589 over the season. The remaining two bets yielded a
return of zero.
The 2000/2001 season, 80 trades were suggested by the Quarb
strategy. The average return to these bets was lower than in 1999/
2000 but would still have yielded winnings of £836 on 50 winning
bets and losses of £450 on 28 losing bets, a net profit of £386.
The conclusion of the study was that the mid-point of all the
quotes on offer is a better forecast of the actual outcome in the
bookings market than is the mid-point of the spread offered by the
market outlier (the ‘maverick’). This casts doubt on a hypothesis
that market-makers who set quotes out of line with the prevailing
view do so because they possess better (even privileged) information,
or that they are able to process a given set of information more
effectively than the market as a whole.
Moreover, using the notion of quasi-arbitrages or Quarbs, it was
possible to devise a trading strategy on the basis of the outlying
spread that yielded significantly positive profits.
169
BETTING TO WIN
170
TOMORROW’S ANOTHER DAY
171
BETTING TO WIN
172
BETTING ON THE ELECTION SPREADS
173
BETTING TO WIN
174
CONTROL THE CONSEQUENCES
175
BETTING TO WIN
176
Part Five: Beating The Exchanges
Chapter 1: Backing And Laying On
The Exchanges
Betfair operates by allowing clients to either back or lay a bet. In
order to open an account as a client, a deposit must be made. Credit
card deposits carry a 1.5% charge, but there is no charge for UK
debit card deposits or withdrawals to any card. Deposits and
withdrawals can also be made by cheque and bank transfer, though
there is a charge for bank transfer withdrawals.
When you bet on an event in the conventional sense, this is termed
backing the event. If you wish to act as a bookmaker and offer odds
about an event happening this is known as laying the bet.
Betfair provide the best three offers on each side of the equation.
Thus the best three odds at which you can back an event are
displayed on the left-hand side of the screen, together with the
amounts that you can bet at those odds. On the right-hand side of
the screen are the best three odds at which you can lay an event.
Betdaq, GGBet and Sporting Options also operate by allowing
clients to either back or lay a bet. In the case of GGBet, however, the
terminology is different. If you wish to back an event happening
(i.e. to bet on something in the conventional sense) their term is
that you take the odds.
Betfair, Betdaq and GGBet all provide the best three offers on each
side of the equation. Thus the best three odds at which you can
take the odds (back) are displayed on the left-hand side of the screen,
together with the amounts that you can bet at those odds. On the
right-hand side of the screen are the best three odds at which you
can lay a bet. In addition to odds betting Betfair also offer line betting
and range betting. Let’s consider each of the three types of betting
in turn.
BETTING TO WIN
Introduction
With odds betting the bettor predicts whether an outcome will occur
or not, and places a bet on that outcome at given odds. In
conventional betting markets, bettors place their bet at odds set by
bookmakers, i.e. bettors take odds and bookmakers make (or offer)
odds.
Betting exchanges permit clients to both take and to offer odds.
In this way, you can bet both for and against an outcome occurring.
When you take odds you back the outcome and when you offer
odds you lay the outcome.
Clearly, if you back an outcome at given odds, you will gain if
that outcome occurs. Say, for example, you back England to beat
Australia at cricket at odds of 3 to 1, for a stake of £10. This means
that you make a profit of £30 if England win (3 times your £10 stake),
plus your £10 stake returned. Your loss if England do not win is
limited to your stake, i.e. £10.
Laying an outcome, as explained earlier, is the exact opposite of
backing it. Essentially, it is betting that the outcome will not occur.
This is what bookmakers traditionally do. For example, if you lay
Aston Villa at odds of 4.0 for £100, you think it is sufficiently unlikely
that Villa will win that you are willing to offer the equivalent of
traditional odds of 3 to 1 against it. Your total return, if Aston Villa
lose or draw, is £100 (i.e. the stake), and your maximum loss is £300.
In effect you are accepting the bets of other clients who believe that
3 to 1 is worth taking about Villa’s chances of winning.
180
ODDS BET TING
currently available, you are able to place an order for a better price.
In doing so, you are hoping that somebody will see your offer and
be willing to lay you the bet, partially or in total, at the odds you
request.
Your offer to back appears on the right-hand side as a request for
someone to lay you a bet, at the odds requested and up to the amount
requested.
Example
For a few minutes after opening on Sunday 26 May 2002, it was
possible to obtain 7 to 1 in a string of betting offices, about Quarter
Moon winning the Irish 1000 Guineas. Your stake size would have
been restricted, but say you placed £100 at that 7 to 1:
Returning to your computer screen you would have noted the
following Betfair odds, available at about 12.30 pm that day:
You could now lay Quarter Moon at 7.0 (or 6 to 1) for £100.
If Quarter Moon wins, your profit at the betting office is £700, but
you only pay out £600 to your betting exchange colleagues.
If Quarter Moon loses, you lose £100 at the betting office, but you
win £100 (minus commission) in the betting exchange. In other
words, you stand to win a net profit of £100, for the maximum
downside of 5% of £100, i.e. £5.
By laying the horse at 6 to 1 for £110, say, you would make a profit
of £90 if Quarter Moon wins, but make a small profit, even net of
commission, if it loses.
Now let us see how exactly you would go about laying the horse
at 6 to 1 on Betfair. First, you select Today’s Horse Racing from the
vertical list of sports/special bets available on the left side of the
screen.
You then select 4.00 Curragh from the list of options presented,
and you are led to a list of the best three prices at which it is possible
to back and lay the horse, together with the amounts at which you
can back or lay at these odds (see above).
As noted in the example, it was possible to lay Quarter Moon at
181
BETTING TO WIN
7.0 up to a total of £873. Note that this means that you are offering
odds of 7.0 (6 to 1) to other clients of the exchange who wish to back
the horse at that price. If you agree to accommodate all this money,
your net liability is, therefore, more than £5000. To be precise, it is 6
× £873, or £5238, which means that you would need to have that
amount of money deposited in your account, over and above any
outstanding potential liabilities on other events you may already
have laid.
Let us say that you decide to lay Quarter Moon at 7.0 for £100. To
do this you select the LAY button, and click on it.
If you wish to lay at 7.0 for £100, simply enter £100 at this point.
You are immediately warned of the implications of what you are
intending to do, i.e. that you stand to win £100 (minus commission)
if the horse does not win, but to lose a net amount of £600 if the
horse does win.
Press Enter again to confirm.
An alternative approach is to offer to lay the horse at less than 7.0,
say at 6.8 when the display prompts you with the possibility of
choosing your own odds. You now select 6.8 for £100, say, and press
Enter. The implications of your proposed offer are again displayed
before you finally confirm.
If Quarter Moon does not win, you keep your stake of £100. In
total your gross profit is £100 (minus commission). If Quarter Moon
wins, your gross payout is £680. This includes your stake of £100. In
total, your loss is £580.
If you confirm, an extra £100 is added to the BACK side of the
screen, which means that those wishing to back the horse at 6.8 (i.e.
5.8 to 1) will have the opportunity of staking £100 with you, in
addition to however much is being offered by other members of the
exchange at those odds.
In the example above, £25 was available at 6.8. With your offer,
£125 is now available. Of course, this is only an offer, and there is no
guarantee that anyone will be willing to take your offer at those
odds. In the meantime, you may have missed the opportunity to
LAY the horse, for a sure profit, at 7.0 when people were willing to
back it at that price.
182
ODDS BET TING
1. Identify cases where the back and lay prices are quite close
to each other. This way, the price you are getting will be
quite close to what the market expects. For example, if you
can back a horse at 6.8, or lay it at 7.0, whichever you do will
probably be close to the true probability. Moreover, there is
a reasonable chance that the market will move enough to
allow you to back and lay for a sure profit. After all, it doesn’t
have to move far.
183
BETTING TO WIN
184
LINE AND RANGE BETTING
Cancelling A Bet
Any offer (whether to back or lay) that has not been matched can be
cancelled. It is not possible however, to cancel a bet once it has been
matched, i.e. the offer has been accepted by another client.
185
BETTING TO WIN
186
ASIAN HANDICAP BETTING ON THE EXCHANGES
To summarise:
England 0 v. Sweden 0
Back England: You win if England win. If the match is a draw,
your stake is returned. You lose if Sweden win.
Lay England: You win if Sweden win. If the match is a draw,
your stake is returned. You lose if England win.
Back Sweden: You win if Sweden win. If the match is a draw,
187
BETTING TO WIN
188
Part Six: World Cup Fever
Chapter 1: Different Betting
Strategies
In the last chapter, we looked at Asian handicap betting on the
exchanges, using England’s opening match against Sweden as an
illustration of how it works.
In this final part of the book, I will use the same match to highlight
some of the different ways in which a betting strategy can be
implemented.
Fixed Odds
We begin on the Saturday evening preceding England’s opener. The
average punter would go to his local bookie and perhaps take the
odds on offer about one or other outcomes. The reality is that the
choices on offer were simply wondrous.
Take the fixed odds first, and consider the case of the typical
Ladbrokes punter (I choose Ladbrokes simply as an example of one
of the big established chains). The odds about the three basic
outcomes were as follows (decimal odds in brackets):
192
DIFFERENT BETTING STRATEGIES
193
BETTING TO WIN
As with Betfair, it’s only the price about a Sweden win that should
tempt backers away from the traditional bookmakers.
The 4.3 was commission-free with Betdaq, compared to a
commission of up to 5% with Betfair on the 4.4. Not a bad offer,
albeit to limited stakes, and certainly to be weighed closely against
the SportsTab offer.
Finally, to the betting exchange located at GGBET.com.
194
DIFFERENT BETTING STRATEGIES
Let’s see how this compares with what the professional layers
(the bookmakers) were willing to offer.
Just for the record, I compiled the odds again with just 15 minutes
to kick-off.
Ladbrokes
England to win: Evens (2.0)
Sweden to win: 12 to 5 (3.4)
Draw: 2 to 1 (3.0)
195
BETTING TO WIN
Betbrain
England to win: 2.3 – with Lion Bets
Sweden to win: 4.50 – with SportsTab
Draw: 3.25 – with SportsTab, SportingbetUSA
2.3 is the equivalent of a 43.5% chance (100/2.3)
4.50 is the equivalent of a 22.2% chance (100/4.50)
3.25 is the equivalent of a 30.8% chance (100/3.25)
196
DIFFERENT BETTING STRATEGIES
Betfair
Since the prices posted the evening before, the only change (apart
from the stake availabilities) was that layers were slightly better off
(it was now possible to lay Sweden at 4.4 compared to 4.5 the
previous evening), but those seeking to back Sweden to win could
only avail themselves of a price of 4.3 instead of the 4.4 previously
available.
Betdaq
Again, the odds to back Sweden to win had contracted since the
previous evening, although the odds available about an England
win had lengthened a little.
197
BETTING TO WIN
GGBet
198
DIFFERENT BETTING STRATEGIES
Spread Betting
Immediately before the England v. Sweden game, Sporting Index,
IG Index Sport, Cantor Sport, Spreadex and Sportsspread were
offering the following quotes about the major markets:
England supremacy:
0.2 – 0.5 (Sporting)
0.3 – 0.6 (IG)
0.4 (Cantor) (i.e. zero spread special offer)
0.2 – 0.5 (Spreadex)
0.3 – 0.6 (Sportsspread)
199
BETTING TO WIN
Total Goals:
2.1 – 2.4 (Sporting)
2.2 – 2.5 (IG)
2.1 – 2.4 (Cantor)
2.1 – 2.4 (Spreadex)
2.4 – 2.7 (Sportsspread)
Shirt numbers:
26 – 29 (Sporting)
25 – 28 (IG)
24 – 27 (Cantor)
24 – 27 (Spreadex)
24 – 27 (Sportsspread)
Bookings:
36 – 40 (Sporting)
36 – 40 (IG)
34 – 38 (Cantor)
33 – 37 (Spreadex)
36 – 40 (Sportsspread)
Corners:
10.25 – 11.25 (Sporting)
10 – 11 (IG)
9.5 – 10.5 (Cantor)
9.5 – 10.5 (Spreadex)
10 – 11 (Sportsspread)
200
DIFFERENT BETTING STRATEGIES
25 – 10 – 0 Index (England)
13.5 – 15 (Sporting)
14 – 15.5 (IG)
13 – 14.5 (Cantor)
13 – 14.5 (Spreadex)
13 – 14.5 (Sportsspread)
25 – 10 – 0 Index (Sweden)
8.5 – 10 (Sporting)
8.5 – 10 (IG)
9 – 10.5 (Cantor)
9 – 10.5 (Spreadex)
9.5 – 11 (Sportsspread)
201
BETTING TO WIN
Simple Quarbs exist when the top end of one company’s spread coincides
with the bottom end of another company’s spread. In the England v. Sweden
game, no such situation existed in the supremacy market.
The second thing to look for is a full Quarb. This exists when one
of the spreads lies wholly outside the average of the mid-points of
all the spreads.
Applying this to the supremacy market produces the following:
England supremacy:
202
DIFFERENT BETTING STRATEGIES
Total goals:
2.1 – 2.4 (Sporting)
2.2 – 2.5 (IG)
2.1 – 2.4 (Cantor)
2.2 – 2.5 (Spreadex)
2.4 – 2.7 (Sportsspread)
Again, the first thing to look for is whether there exist what I
have termed ‘Quarbs’ (quasi-arbitrages).
You may remember that simple Quarbs exist when the top end of
one company’s spread coincides with the bottom end of another
company’s spread. In the England v. Sweden game, this existed
between Sporting and Cantor (2.1 – 2.4) and Sportsspread (2.4 –
2.7). This means that you could buy with one company and sell
with another, and make a certain profit (loss) of zero.
This is indicative, at least, that one of the companies is a little out
in its judgement, and it may be possible to take advantage. To see if
this is the case, we need to determine whether a full Quarb exists.
This exists when one of the spreads lies wholly outside the average
of the mid-points of all the spreads.
Applying this analysis to the total goals market produces the
following:
Total Goals
2.1 – 2.4 (Sporting) MID-POINT = 2.25
2.2 – 2.5 (IG) MID-POINT = 2.35
2.1 – 2.4 (Cantor) MID-POINT = 2.25
2.2 – 2.5 (Spreadex) MID-POINT = 2.35
203
BETTING TO WIN
Shirt numbers:
26 – 29 (Sporting)
25 – 28 (IG)
24 – 27 (Cantor)
24 – 27 (Spreadex)
24 – 27 (Sportsspread)
The average of the mid-points is:
(27.5 + 26.5 + 25.5 + 25.5 + 25.5)/5 = 26.1
There is almost a full Quarb with Sporting, but not quite.
Bookings:
36 – 40 (Sporting)
36 – 40 (IG)
34 – 38 (Cantor)
33 – 37 (Spreadex)
36 – 40 (Sportsspread)
The average of the mid-points is: (38 + 38 + 36 + 35 + 38)/5 = 37
As near as it is possible to get to a full Quarb (courtesy of Spreadex),
but not quite.
204
DIFFERENT BETTING STRATEGIES
Corners:
10.25 – 11.25 (Sporting)
10 – 11 (IG)
9.5 – 10.5 (Cantor)
9.5 – 10.5 (Spreadex)
10 – 11 (Sportsspread)
The average of the mid-points is:
(10.75 + 10.5 + 10 + 10 + 10.5)/5 = 51.75/5
= 10.35
No Quarb.
There is clearly a simple Quarb here, in that the top end of the
Sportsspread quote (40) coincides with the bottom end of the Sporting
quote (40). This means that you could buy with Sportsspread and sell
with Sporting, and make a certain profit (loss) of zero.
Again, this is indicative that one of the companies may be a little
out in its judgement. To see if this is the case, we need to determine
whether a full Quarb exists.
Applying this analysis to the first match goal market produces
the following:
205
BETTING TO WIN
of the first match goal will be 41 minutes. If this is truly the best
estimate, then the value exists in a buy of minutes to the first match
goal at 40, with Sportsspread.
The most sophisticated bettors, however, will be aware of what
was discussed in an earlier chapter, i.e. that the time of first goal
markets are often set artificially high, to accommodate the
preponderance of buyers in the market. Allowing for this, any trade
would offer marginal value at best.
206
DIFFERENT BETTING STRATEGIES
61 – 64 (IG)
58 – 61 (Spreadex)
56 – 59 (Sportsspread)
25 – 10 – 0 Index (England)
13.5 – 15 (Sporting)
14 – 15.5 (IG)
13 – 14.5 (Cantor)
13 – 14.5 (Spreadex)
13 – 14.5 (Sportsspread)
The mid-points average is:
(14.25 + 14.75 + 13.75 + 13.75+ 13.75)/5 = 14.05
Almost a full Quarb, given a sell of England at 14 with IG, but not
quite.
25 – 10 – 0 Index (Sweden)
8.5 – 10 (Sporting)
8.5 – 10 (IG)
9 – 10.5 (Cantor)
9 – 10.5 (Spreadex)
9.5 – 11 (Sportsspread)
The average of the mid-points is:
(9.25 + 9.25 + 9.75 + 9.75 + 10.25)/5 = 9.65
No Quarb.
207
BETTING TO WIN
a. No bookings points
b. 20 bookings points
c. 40 bookings points
d. 100 bookings points
e. 200 bookings points
208
DIFFERENT BETTING STRATEGIES
No bookings points
Loss of £360 (Betfair)
Loss of £370 (Spreadex)
20 bookings points
Loss of £160 (Betfair)
Loss of £170 (Spreadex)
40 bookings points
Profit of £40 on Betfair (£38 after deducting 5% commission)
Profit of £30 on Spreadex
209
BETTING TO WIN
a. No goals
b. 3 goals
c. 6 goals
d. 7 goals
No goals
Loss of £230 on Betfair
Loss of £240 on Spreadex
3 goals
Profit of £70 on Betfair (£66.50 after deducting 5% commission)
Profit of £60 on Spreadex
210
DIFFERENT BETTING STRATEGIES
6 goals
Profit of £370 on Betfair (£351.50 after deducting 5% commission)
Profit of £360 on Spreadex
7 goals
Profit of £370 on Betfair (£351 - 50 after deducting 5% commission)
Profit of £460 on Spreadex
211
BETTING TO WIN
Match 1 – 30 minutes
Match 2 – 8 minutes
Match 3 – 84 minutes
Match 4 – 52 minutes
Match 5 – 70 minutes
Match 6 – 32 minutes
Match 7 – 18 minutes
Match 8 – 10 minutes
Match 9 – 29 minutes
The median time of the first goal is 30 minutes, since there are
an equal number of goals scored before and after this time.
8, 10, 18 and 29 are less than 30; 32, 52, 70 and 84 are greater
than 30.
212
IF IT LOOKS TOO GOOD TO BE TRUE, THEN IT IS TOO GOOD TO BE TRUE!
213
BETTING TO WIN
Trade accordingly.
I have also collated (Appendix 2) the times of all the disciplinary
cards issued in World Cup 2002. For the particularly assiduous
reader, a useful exercise would be to use these data similarly to
calculate the difference between mean and median times in the
bookings market.
Finally, an interesting and potentially useful perspective of goal
times can be gleaned courtesy of what is known in statistical
circles as the Poisson distribution. Originally, it was formulated to
model the rate at which Prussian officers were kicked to death
by horses. Nowadays it has a more prosaic use, i.e. to calculate
the likelihood of a given number of goals being scored by each
team. It works like this.
First, we need an estimate of the number of goals that a team
will score in a given match. From this estimate, the Poisson
distribution allows us to estimate the chance of any game finishing
with any particular scoreline.
Get this right, and there are potentially rich pickings available
across a number of markets.
The Poisson distribution is included in a table (see Appendix
3) which estimates the percentage chance of a team scoring a
given number of goals based on our estimate of the number they
are likely to score. The value of the table lies in the fact that the
predictions of the Poisson distribution tend to conform to actual
results over time.
For example, if we estimate that Southampton will, on average,
score one goal against Arsenal, then the Poisson distribution tells us
that there is a 36.8% chance that they will not score any goals on
this particular occasion. It tells us that there is a 73.6% chance of one
goal or less, a 92.0% chance of 2 goals or less, a 98.1% chance of 3
goals or less, and so on. This can be read off directly from the table.
Look down the vertical column on the left of the table to 1.0, and
read across the horizontal row to find the probability of any given
number of goals (or less) being scored. From this table we can even
work out the probability of exactly 3 goals being scored, say. To do
this, we read off the likelihood of 2 goals or less being scored (92.0%),
and then the likelihood of 3 goals or less being scored (98.1%). The
difference between these numbers (6.1%) is the likelihood of exactly
214
IF IT LOOKS TOO GOOD TO BE TRUE, THEN IT IS TOO GOOD TO BE TRUE!
215
BETTING TO WIN
216
DRAW BIASES
Conclusion
In this book, I have drawn upon the experience I have built up over
many years of professional study of betting markets.
In particular, I have investigated and sought to make accessible to
a wider audience the results of an extensive body of work
contributed by experts around the world, in order to study the way
in which bettors really can turn the odds in their favour. Much of
this work has, until known, been relatively unknown outside of
academic circles.
In doing so, I have sought to distinguish what is valid and useful
from what is merely speculative or just plain wrong.
I have tried to do this in an entertaining and topical, as well as an
informative, manner. In this quest, I hope I have succeeded.
The Golden Age of betting is now. We are living in it. I hope that
this book will help you to enjoy and profit from our good fortune to
the full.
217
BETTING TO WIN
Bookmakers
www.tote.co.uk
www.bluesq.com
www.paddypower.com
www.sportingbet.com
www.eurobet.com
www.gamebookers.com
www.ukbetting.com
www.willhill.com
www.sportingodds.com
www.ladbrokes.com
www.skybet.com
www.bet365.co.uk
www.betinternet.com
www.intertops.com
www.sportsinteraction.com
www.sportstab.com.au
www.centrebet.com
www.expekt.com
www.betandwin.com
www.stanjames.com
www.canbet.com.au
www.victorchandler.com
www.thegreek.com
www.planetpinnacle.com
www.betabet.com
www.yabet.com
218
INTERESTING AND USEFUL INTERNET SITES
www.totalbet.com
www.xodds.com
www.attheraces.co.uk
www.betonmarkets.com
www.betdirect.net
Betting Exchanges
www.betfair.com
www.betdaq.com
www.betdaq.co.uk
www.ggbet.com (see also www.iwageru.co.uk)
www.sportingoptions.co.uk
www.tradindex.com (sports)
www.intrade.com
www.newsfutures.com
www.biz.uiowa.edu/iem
www.tradesports.com
www.tradingsports.com
219
BETTING TO WIN
www.bbc.co.uk/sport
www.football365.com
www.readabet.com
www.oddschecker.com
www.oddschecker.co.uk
www.betbrain.com
www.easyodds.com
www.bestbetting.com
www.sportsbettingindex.com
www.soccerbase.com
www.sportal.com
www.smartbet.co.uk
www.attheraces.co.uk
www.gamebookers.com
www.spreadtrades.com
www.bookiesindex.com
www.bettingadvice.com
www.crastinum.com
www.bookiebusters.net
www.soccernet.com
www.onewaybet.com
www.opta.co.uk
www.bethelp.com
220
WORLD CUP 2002 STATISTICS
Group A
FRANCE v. SENEGAL
Score: 0 – 1
Goal times: 30
Yellow card times: 47+ (i.e. 45), 51
Red card times: –
URUGUAY v. DENMARK
Score: 1 – 2
Goal times: 45, 47, 83
Yellow card times: 25, 34, 51
Red card times: –
FRANCE v. URUGUAY
Score: 0 – 0
Goal times: –
Yellow card times: 11, 47+ (i.e. 45), 47+ (i.e. 45), 48+ (i.e. 45), 47
Red card times: 25
221
BETTING TO WIN
DENMARK v. SENEGAL
Score: 1 – 1
Goal times: 16, 52
Yellow card times: 7, 10, 20, 62, 82, 84
Red card times: 80
DENMARK v. FRANCE
Score: 2 – 0
Goal times: 22, 67
Yellow card times: 27, 71, 81
Red card times: –
SENEGAL v. URUGUAY
Score: 3 – 3
Goal times: 20, 26, 38, 46, 69, 88
Yellow card times: 2, 4, 8, 19, 35, 39, 40, 69, 82, 82, 87, 87
Red card times: –
Group B
PARAGUAY v. SOUTH AFRICA
Score: 2 – 2
Goal times: 39, 55, 63, 91(i.e. 90)
Yellow card times: 3, 9, 35, 38, 47+ (i.e. 45), 65, 90, 93 (i.e. 90)
Red card times: –
SPAIN v. SLOVENIA
Score: 3 – 1
Goal times: 44, 74, 82, 87
Yellow card times: 36, 46+ (i.e. 45), 65
Red card times: –
SPAIN v. PARAGUAY
Score: 3 – 1
Goal times: 10, 53, 69, 83
Yellow card times: 9, 44, 60, 80
Red card times: –
222
WORLD CUP 2002 STATISTICS
SLOVENIA v. PARAGUAY
Score: 1 – 3
Goal times: 46+ (i.e. 45), 65, 73, 84
Yellow card times: 4, 15, 68, 69, 79
Red card times: 22 (2nd yellow), 81
Group C
BRAZIL v. TURKEY
Score: 2 – 1
Goal times: 47+ (i.e. 45), 50, 87
Yellow card times: 21, 24, 44, 73,
Red card times: 86, 94 (i.e. 90) (2nd yellow)
BRAZIL v. CHINA
Score: 4 – 0
Goal times: 15, 32, 45, 55
Yellow card times: 25, 69
Red card times: –
223
BETTING TO WIN
TURKEY v. CHINA
Score: 3 – 0
Goal times: 6, 9, 85
Yellow card times: 19, 30, 46+ (i.e. 45), 62, 81
Red card times: 58
Group D
SOUTH KOREA v POLAND
Score: 2 – 0
Goal times: 26, 53
Yellow card times: 31, 70, 79, 84, 90
Red card times: –
USA v. PORTUGAL
Score: 3 – 2
Goal times: 4, 29, 36, 39, 71
Yellow card times: 34, 52, 92 (i.e. 90)
Red card times: –
224
WORLD CUP 2002 STATISTICS
PORTUGAL v. POLAND
Score: 4 – 0
Goal times: 14, 65, 77, 88
Yellow card times: 21, 25, 27, 31, 39
Red card times: –
POLAND v. USA
Score: 3 – 1
Goal times: 3, 5, 66, 83
Yellow card times: 44, 46, 63, 72, 86
Red card times: –
Group E
IRELAND v. CAMEROON
Score: 1 – 1
Goal times: 39, 52
Yellow card times: 30, 51, 82, 89
Red card times: –
GERMANY v. IRELAND
Score: 1 – 1
Goal times: 19, 92 (i.e. 90)
Yellow card times: –
Red card times: –
225
BETTING TO WIN
CAMEROON v. GERMANY
Score: 0 – 2
Goal times: 50, 79
Yellow card times: 8, 9, 29, 31, 37, 42, 42, 44, 56, 58, 60, 72, 74, 81
Red card times: 40 (2nd yellow), 77 (2nd yellow)
Group F
ENGLAND v. SWEDEN
Score: 1 – 1
Goal times: 24, 59
Yellow card times: 12, 47+ (i.e. 45), 73
Red card times: –
ARGENTINA v. NIGERIA
Score: 1 – 0
Goal times: 63
Yellow card times: 51, 73, 90
Red card times: –
SWEDEN v. NIGERIA
Score: 2 – 1
Goal times: 27, 35, 63
Yellow card times: 31, 69, 80
Red card times: –
226
WORLD CUP 2002 STATISTICS
ARGENTINA v. ENGLAND
Score: 0 – 1
Goal times: 44
Yellow card times: 13, 29, 50
Red card times: –
SWEDEN v. ARGENTINA
Score: 1 – 1
Goal times: 59, 88
Yellow card times: 55, 58, 65, 75, 78
Red card times: 47+ (i.e. 45)
NIGERIA v. ENGLAND
Score: 0 – 0
Goal times: –
Yellow card times: –
Red card times: –
Group G
CROATIA v. MEXICO
Score: 0 – 1
Goal times: 60
Yellow card times: –
Red card times: 60
ITALY v. ECUADOR
Score: 2 – 0
Goal times: 7, 27
Yellow card times: 14, 49, 54
Red card times: –
ITALY v. CROATIA
Score: 1 – 2
Goal times: 55, 73, 76
Yellow card times: 39, 51
Red card times: –
227
BETTING TO WIN
MEXICO v. ECUADOR
Score: 2 – 1
Goal times: 5, 28, 57
Yellow card times: 15, 27, 49, 61, 65, 87
Red card times: –
MEXICO v. ITALY
Score: 1 – 1
Goal times: 34, 85
Yellow card times: 2, 5, 10, 43, 55, 57, 84
Red card times: –
ECUADOR v. CROATIA
Score: 1 – 0
Goal times: 48
Yellow card times: 72, 86, 92 (i.e. 90)
Red card times: –
Group H
JAPAN v. BELGIUM
Score: 2 – 2
Goal times: 57, 59, 67, 75
Yellow card times: 21, 31, 54, 62, 82
Red card times: –
RUSSIA v. TUNISIA
Score: 2 – 0
Goal times: 59, 64
Yellow card times: 27, 50, 75, 88
Red card times: –
JAPAN v. RUSSIA
Score: 1 – 0
Goal times: 51
Yellow card times: 13, 15, 38, 42, 60, 91 (i.e. 90)
Red card times: –
228
WORLD CUP 2002 STATISTICS
TUNISIA v. BELGIUM
Score: 1 – 1
Goal times: 13, 17
Yellow card times: 22, 40, 43, 68, 69
Red card times: –
TUNISIA v. JAPAN
Score: 0 – 2
Goal times: 48, 75
Yellow card times: 21, 81
Red card times: –
BELGIUM v. RUSSIA
Score: 3 – 2
Goal times: 7, 52, 78, 82, 88
Yellow card times: 12, 14, 39, 64, 84
Red card times: –
Round of 16
GERMANY v. PARAGUAY
Score: 1 – 0
Goal times: 88
Yellow card times: 26, 35, 50, 71, 92 (i.e. 90)
Red card times: 92 (i.e. 90)
DENMARK v. ENGLAND
Score: 0 – 3
Goal times: 5, 22, 44
Yellow card times: 24, 50
Red card times: –
SWEDEN v. SENEGAL
Score: 1 – 1 (1 – 2 after extra time)
Goal times: 11, 37, 104 (excluded)
Yellow card times: 73, 94 (excluded)
Red card times: –
229
BETTING TO WIN
SPAIN v. IRELAND
Score: 1 – 1 (Match eventually decided on penalties)
Goal times: 8, 90
Yellow card times: 62, 87, 89
Red card times: –
MEXICO v. USA
Score: 0 – 2
Goal times: 8, 65
Yellow card times: 26, 37, 47, 50, 53, 67, 70, 81, 83, 84
Red card times: 88
BRAZIL v. BELGIUM
Score: 2 – 0
Goal times: 67, 87
Yellow card times: 24, 28
Red card times: –
JAPAN v. TURKEY
Score: 0 – 1
Goal times: 12
Yellow card times: 21, 44, 45, 91(i.e. 90)
Red card times: –
S. KOREA v. ITALY
Score: 1 – 1 (2 – 1 after extra time)
Goal times: 18, 88, 117 (excluded)
Yellow card times: 4, 17, 22, 55, 59, 80, 99 (excluded), 115 (excluded)
Red card times: 103 (2nd yellow) (excluded)
230
WORLD CUP 2002 STATISTICS
Quarter-finals
ENGLAND v BRAZIL
Score: 1 – 2
Goal times: 23, 47+ (i.e. 45), 50
Yellow card times: 75, 86
Red card times: 57
GERMANY v. USA
Score: 1 – 0
Goal times: 39
Yellow card times: 40, 41, 66, 68, 68, 69, 70
Red card times: –
SENEGAL v. TURKEY
Score: 0 – 1
Goal times: 94 (i.e. 90)
Yellow card times: 12, 22, 63, 87
Red card times: –
Semi-finals
GERMANY v SOUTH KOREA
Score: 1 – 0
Goal times: 75
Yellow card times: 71, 85, 94 (i.e. 90)
Red card times: –
231
BETTING TO WIN
BRAZIL v. TURKEY
Score: 1 – 0
Goal times: 49
Yellow card times: 41, 59, 90
Red card times: –
232
WORLD CUP 2002 STATISTICS
233
BETTING TO WIN
Bibliography
Ali, M.M. (1977) Probability and utility estimates for racetrack bettors, Journal
of Political Economy, 83, 803 – 815.
Ali, M.M. (1979) Some evidence on the efficiency of a speculative market,
Econometrica, 47, 387 – 392.
Asch, P. and Quandt, R.E. (1986) The Professor’s Guide to Strategies. Auburn
House, Dover, MA.
Asch, P. and Quandt, R.E. (1987) Efficiency and profitability in exotic bets,
Economica, 59, 278 – 298.
Asch, P. and Quandt, R.E. (1988) Betting bias in exotic bets, Economic Letters,
28, 215 – 219.
Avery, C. and Chevalier, J. (1999), ‘Identifying investor sentiment from price
paths: the case of football betting, Journal of Business, 72 (4), 493 –
521.
Benter, W. (1994) Computer based horse race handicapping and wagering
systems: a report. In: Efficiency of Racetrack Betting Markets (1994),
eds Hausch, D.B., Lo, S.Y. and Ziemba, W.T., pp. 183 – 198, Academic
Press, London.
Betton, S. (1994) Post position bias: an econometric analysis of the 1987
season at Exhibition Park. In: Efficiency of Racetrack Betting Markets
(1994), eds Hausch, D.B., Lo, S.Y. and Ziemba, W.T., pp. 511 – 526,
Academic Press, London.
Bird, R. and McCrae, M. (1987) Tests of the efficiency of racetrack
betting using bookmakers’ odds, Management Science, 33, 1552 –
1562.
Bird, R., McCrae, M. and Beggs, J. (1987) Are gamblers really risk-takers?
Australian Economic Papers, December, 237 – 253.
Bird, R. and McCrae, M. (1994) Efficiency of racetrack betting markets:
Australian evidence. In: Efficiency of Racetrack Betting Markets (1994),
eds Hausch, D.B., Lo, S.Y. and Ziemba, W.T., pp. 575 – 582, Academic
Press, London.
Blackburn, P. and Peirson, J. (1995) Betting at British racecourses: an analysis
of semi-strong form efficiency between bookmaker and tote odds,
Studies in Economics, University of Kent, 95/4.
Bolton, R.N. and Chapman, R.G. (1986), Searching for positive returns at
the track: a multinomial logit model for handicapping horse races,
Management Science, 32, 1040 – 1059.
234
BIBLIOGRAPHY
Brown, W.O. and Sauer, R.D. (1993), Does the basketball market believe in
the ‘hot hand’? Comment, American Economic Review, 83 (Dec.), 1377
– 86.
Bruce, A.C. and Johnson, J.E.V. (1992) Towards an explanation of betting as
a leisure pursuit, Leisure Studies, 11, 201 – 218.
Busche, K. (1994) Efficient market results in an Asian setting. In: Efficiency
of Racetrack Betting Markets, eds Hausch, D.B., Lo, S.Y. and Ziemba,
W.T., pp. 615 – 616, Academic Press, London.
Busche, K. and Hall, C. D. (1988) An exception to the risk preference
anomaly, Journal of Business, 61, 337 – 346.
Cain, M., Law, D. and Peel, D.A. (1996) Insider Trading in the Greyhound Betting
Market, Salford Papers in Gambling Studies, Vol. 1, University of
Salford, Salford.
Cain, M., Law, D. and Peel, D. (2000) The favourite – longshot bias and
market efficiency in UK football betting, Scottish Journal of Political
Economy, 47 (1), 25 – 36.
Cain, M., Law, D. and Peel, D., Bias and Insider Trading in a Variety of
Betting Markets, Bulletin of Economic Research (forhtcoming).
Camerer, C. (1989), Does the basketball market believe in the ‘hot hand’?
American Economic Review, 79, 1257 – 61.
Canfield, B., Fauman, B.C. and Ziemba, W.T. (1987) Efficient market
adjustment of odds prices to reflect track biases, Management Science,
33, 1428 – 1439.
Chadha, S. and Quandt, R.E. (1996), Betting bias and market equilibrium
in racetrack betting, Applied Financial Economics, 6 (3), 287 – 292.
Chapman, R.G. (1994) Still searching for positive returns at the track:
empirical results from 2000 Hong Kong races. In: Efficiency of Racetrack
Betting Markets, eds Hausch, D.B., Lo, S.Y. and Ziemba, W.T., pp. 173
– 181, Academic Press, London.
Clotfelter, C. and Cook, P.J. (1991) Lotteries in the real world, Journal of Risk
and Uncertainty, 4, 227 – 232.
Clotfelter, C. and Cook, P.J. (1993) The ‘Gambler ’s Fallacy’ in lottery play,
Management Science, 39 (12), 1521 – 1525.
Crafts, N.F.R. (1985) Some evidence of insider knowledge in horse race
betting in Britain, Economica, 52, 295 – 304.
Crafts (1994) Winning systems? Some further evidence on insiders and outsiders in
British horse race betting. In: Efficiency of Racetrack Betting Markets, eds Hausch,
D.B., Lo, S.Y. and Ziemba, W.T., pp. 545 – 549. Academic Press, London .
235
BETTING TO WIN
De Bondt, W. and Thaler, R. (1985), Does the stock market overreact?, Journal
of Finance, 40, (3), 793 – 805.
Dowie, J. (1976) On the efficiency and equity of betting markets, Economica,
43, (170), May, 139 – 150.
Dowie, J. (1982) Why Spain should win the World Cup, New Scientist, June
10, 693 – 695.
Figgis, E.L. (1951) Focus on Gambling, Barker, London
Fingleton, J. and Waldron, P. (1996) Optimal determination of bookmakers’
betting odds: theory and tests, Trinity Economic Papers, Technical
Paper No. 9, December, Trinity College, Dublin.
Gabriel, P.E. and Marsden, J.R. (1990) An examination of market efficiency
in British racetrack betting, Journal of Political Economy, 98, 874 – 885.
Gabriel, P.E. and Marsden, J.R. (1991) An examination of market efficiency
in British racetrack betting: errata and corrections, Journal of Political
Economy, 99, 657 – 659.
Gandar, J.M., Zuber, R.A., O’Brien, T and Russo, B. (1988), Testing rationality
in the point spread betting market, Journal of Finance, 43 (Sept.), 995
– 1008.
Gandar, J.M., Dare, W.H., Brown, C.R. and Zuber, R.A. (1998), Informed
traders and price variations in the betting market for professional
basketball games, Journal of Finance, 53 (Feb.), 385 – 401.
Gilovich, T., Vallone, R., & Tversky, A. (1985). The hot hand in basketball:
On the misperception of random sequences. Cognitive Psychology, 17,
295 – 314
Golec, J. and Tamarkin, M. (1995) Do bettors prefer longshots because they
are risk-lovers, or are they just over-confident? Journal of Risk and
Uncertainty, 11, 51 – 64.
Golec, J. and Tamarkin, M. (1998), Bettors love skewness, not risk, at the horse
track, Journal of Political Economy, 106, 1 (March), 205 – 25.
Gray, P.K. and Gray, S.F. (1997), Testing market efficiency. Evidence from
the NFL sports betting market, Journal of Finance, 52 (Sept.), 1725 –
37.
Griffith, R.M. (1949) Odds adjustment by American horse-race bettors,
American Journal of Psychology, 62, 290 – 294.
Haigh, J. (1999), (Performance) index betting and fixed odds, The Statistician,
48, 3, 425 – 434.
Haigh, J. (1999), Taking Chances: Winning with Probability Oxford University
Press, Oxford
236
BIBLIOGRAPHY
Hamid, S.S., Prakash, A.J. and Smyser, M.W. (1996) Marginal risk aversion
and preferences in a betting market, Applied Economics, 28, 371 –
376.
Hausch, D.B., Lo, S.Y. and Ziemba, W.T. (1994) Efficiency of Racetrack Betting
Markets, Academic Press, London
Hausch, D.B., Ziemba, W.T. and Rubinstein, M. (1981) Efficiency of the
market for racetrack betting, Management Science, 27, 1435 – 1452.
Henery, R.J. (1985) On the average probability of losing bets on horses with
given starting price odds, Journal of the Royal Statistical Society, 148
(4), 342 – 349.
Hurley, W. and McDonough, L. (1995), A note on the Hayek hypothesis
and the favorite – longshot bias in parimutuel betting, American
Economic Review, 85, 949 – 55.
Isaacs, R. and Jackson, D.A. (1994), Focus on sport: index betting on sports,
The Statistician, 43, 2, 309 – 315.
Johnson, J.E.V. and Bruce, A.C. (1993) Gluck’s second law: an Empirical
investigation of horserace betting in early and late races, Psychological
Reports, 72, 1251 – 1258.
Kahneman, D. and Tversky, A. (1973) On the psychology of prediction,
Psychological Review, 80, 237 – 251.
Kahneman, D. and Tversky, A. (1982) Intuitive prediction: biases and
corrective procedures. In: Judgment Under Uncertainty: Heuristics and
Biases, eds Kahneman, D., Slovic, P. and Tversky, A., pp. 414 – 421,.
Cambridge University Press, Cambridge.
Kahneman, D. and Tversky (1984) Choices, values and frames, American
Psychologist, 39, 341 – 350.
Kelly, J. (1956) A new interpretation of information rate, Bell System
Technology Journal, 35, 917 – 926.
Kopelman, R.E. and Minkin, B.L. (1991) Toward a psychology of parimutuel
behaviour: test of Gluck’s laws, Psychological Reports, 68, 701 – 702.
Krueger, T.M. and Kennedy, W.F. (1990) An Examination of the Super Bowl
stock market predictor, The Journal of Finance 45 , 691 – 698.
Ladbrokes Pocket Companion, 1990 Flat Edition, Aesculus Press, Oswestry.
Lo, V.S.Y. and Busche, K. (1994) How accurately do bettors bet in doubles?
In: Efficiency of Racetrack Betting Markets (1994), eds Hausch, D.B., Lo,
S.Y. and Ziemba, W.T., pp. 615 – 616. Academic Press, London.
Losey, L. and Talbott, J.C. (1980) Back on track with the efficient markets
hypothesis, Journal of Finance, 35, 1039 – 1043.
237
BETTING TO WIN
238
BIBLIOGRAPHY
239
BETTING TO WIN
240