Secrets of Practical Chess

John Nunn
First published in the UK by Gambit Publications Ltd 1 998
Copyright© John Nunn 1 998
The right of John Nunn to be identified as the author of this work has been as­
serted in accordance with the Copyright, Designs and Patents Act 1 988.

All rights reserved. This book is sold subject to the condition that it shall not, by
way of trade or otherwise, be lent, re-sold, hired out or otherwise circulated in
any form of binding or cover other than that in which it is published and without a
similar condition including this condition being imposed on the subsequent pur­
chaser.

A copy of the British Library Cataloguing in Publication data is available from

the British Library.

ISBN 1 90 1 983 0 1 3

DISTRIBUTION:
Worldwide (except USA) : Biblios Distribution Services, Star Rd, Partridge
Green, West Sussex, RH 1 3 8LD, England.
USA: BHB International, Inc, 994 Riverview Drive, Totowa, New Jersey 075 1 1 ,
USA.
For all other enquiries (including a full list of all Gambit Chess titles) please
contact the publishers, Gambit Publications Ltd, 69 Masbro Rd, Kensington,
London W 1 4 OLS.
Fax +44 (0) 1 7 1 37 1 1477. E-mail 10056 1 .3 1 2 1 @compuserve.com.

Edited by Graham B urgess

Typeset by John Nunn
Printed in Great Britain by Redwood Books, Trowbridge, Wilts.

10 9 8 7 6 5 4 3 2 1

Gambit Publications Ltd

Managing Director: GM Murray Chandler
Chess Director: GM John Nunn
Editorial Director: FM Graham Burgess
Assistant Editor: GM John Emms
German Editor: WFM Petra Nunn
Contents

Introduction 5

1 At the Board 7
Decision-making 7
The Tree of Analysis revisited 7
Evaluation functions 14
When to analyse 18
DAUT 21
Safety-nets 26
When the tactics have to work 28
Implicit commitments 30
Positional thinking 34
The method of comparison 44
Making your opponent think 46
Oversights and blunders 49
Warning signals 51
'Hard-to-see' moves 55
Time-trouble 59
Laziness 62
Determination 63

2 The Opening 67
Building a repertoire 67
Using opening books 69
Books on offbeat openings 70

3 The Middlegame 82
Good positions 82
Bad positions 88
Attack 98
'Inviting everyone to the party' 98
Over-sacrificing 1 03
Defence 1 05
4 SECRETS OF PRACTICAL CHESS

4 The Endgame 1 12
King and Pawn endings 1 12
Opposition 112
The Reti manoeuvre 117
Triangulation 1 18
Expect the unexpected 120
Chess is more than counting 1 22
Rook endings 1 24
Rook and Pawn vs Rook 125
The extra Pawn 1 28
Positional advantage 135
Minor-piece endings 138
Knight endings 139
Bishop vs Knight endings 141
Bishop endings 143
Queen endings 1 47
Queen and Pawn vs Queen 1 47
The extra Pawn 152
Common endings without Pawns 155
Rook vs minor piece 155
Rook and minor piece vs Rook 158
Quick-play finishes 1 64

5 Using a Computer 1 66
Game databases 166
Playing programs 1 69

Index of Names 1 74
Index of Openings 176
Introduction

This book i s aimed at players who are primarily interested in improving their re­
sults. If you are prepared to lose nine games in order to score one brilliant victory,
then it is probably not for you. However, most players are motivated at least
partly by over-the-board success - the thrill of winning is one of the attractions of
chess, and most players feel very satisfied when their rating improves.
The level at which one plays is governed by a number of vague and poorly un­
derstood factors. The first is what one might term ' natural talent' . By this I mean
that combination of factors which sets an upper bound to the level one can
achieve by training and practice. One cannot list precisely which factors are rele­
vant, but one may divide the possible factors into two classes. The first class con­
sists of non-chess-specific elements such as general intelligence and memory.
The second class involves a mesh of inter-related chess factors such as the age at
which one learnt the game, early chess education and so on.
By the time anyone gets around to reading this book, the 'natural talent' factor
will probably be immutable, which brings us to other factors which are more un­
der one 's control. These determine how closely one approaches the ceiling im­
posed by one's ' natural talent' . It is my belief that most players never get
anywhere near their natural ceiling, and that considerable improvement is possi­
ble with appropriate education, training and practice. Most chess books aim to
help readers improve their chess. An opening book, for example, will give gen­
eral plans and concrete analysis, both to help the reader prepare his chosen open­
ings and, after a game, to compare the course of the game with established
theory. Clearly, concrete knowledge is an important factor in establishing chess
strength; someone who has a detailed knowledge of rook and pawn endings will
have an advantage over someone who does not. An assiduous program of self­
training is bound to have a positive effect. In 1 977, Jon Tisdall explained to me
his plan for becoming a grandmaster. He had estimated how many hours of study
were required to advance by one rating point. Multiplying this by the difference
between his current rating and the grandmaster level gave the total number of
hours of study required. I laughed, and pointed out that with each advance, the
number of hours required to gain the next point would probably increase, and so
he might never make it. However, his plan proved justified, because in 1995 he
did indeed gain the grandmaster title.
There are few players who can conduct a training program stretching over
decades, and indeed time limitations apply to virtually all players. In practice this
6 SECRETS OF PRACTICAL CHESS

restricts the amount of improvement possible on the 'chess knowledge' front. In

this book I will give advice on how to use the time available for chess study most
efficiently, for example by distinguishing essential knowledge from optional
knowledge, and advising on the construction of an opening repertoire.
The third factor, which is the main focus of this book, is the efficiency with
which one applies the first two factors while actually sitting at the board. A de­
tailed knowledge of rook and pawn endings won't help a bit if one has an attack
of blind panic; an encyclopaedic memory is valueless if one is regularly seized
by an uncontrollable impulse to sacrifice a piece unsoundly. Chaotic and mud­
dled calculation; misjudgements; oversights; lack of confidence (or overconfi­
dence!); lack of determination - these and many other negative influences all
serve to whittle away one's playing strength. Such problems are not at all easy to
solve, firstly because players very often do not realize what they are doing wrong
and secondly because they imagine that there is nothing they can do to improve
matters.
This book includes a description of various common failings at the board. I
think that many readers will reach a particular section and suddenly think "Yes,
that's exactly the mistake I always make." Recognizing the problem is already
the first step towards solving it. An awareness of when one is most likely to go
wrong enables one to take special care in these 'danger situations' . Eventually,
by concentrating on a particular weakness, it is often possible to eradicate it com­
pletely.
Since many of the matters dealt with in this book are psychological in nature,
there will be quite a few examples from my own games - I can personally testify
that muddled thinking occurs at grandmaster level ! Where I have covered a fa­
miliar topic, I have made an effort to replace, whenever possible, the standard
time-worn examples with excerpts from contemporary play.
Of course, this book, while containing much useful advice and information,
cannot hope to go into detail about every aspect of the game. My aim has been
merely to start the reader along the upward path of self-improvement. I hope that
Secrets of Practical Chess will help readers to improve their results and produce
more satisfying games.

John Nunn
Chertsey, 1 997
1 At the Boa rd

Decision-maki ng one constructs some sort of analytical

tree in thinking about almost any posi­
The Tree of Analysis revisited tion; in less tactical situations, where
the opponent's replies are much less
The so-called 'Tree of Analysis' was predictable, one would not rely solely
popularized in Kotov's famous book on the 'tree' but also take other factors
Think Like a Grandmaster. The analy­ into consideration.
sis of any chess position has a tree-like It follows that the 'Tree of Analy­
structure. There are various alterna­ sis' is a very important method of
tives in the current position, which chess thinking; computers have shown
form the main branches. Each alterna­ that it is possible to play very strongly
tive permits a range of replies, which using practically nothing else. Kotov's
form slightly smaller branches, and so book described in detail the mental
on. Since there are, typically, dozens processes involved in concrete analy­
of legal moves in an average chess sis. If we assume that White is to
position, a tree including every legal move, then Kotov recommended that
move rapidly becomes too dense for White form a list of 'candidate moves'
human beings to handle. Of course, it which he is trying to decide between,
isn't necessary to consider every legal making sure that the list is complete.
move, since a considerable percentage For each of these, White creates a list
of these legal moves are nonsensical, of possible replies by Black and so on,
and this tends to thin the tree some­ following each branch in turn until a
what. Nevertheless, even if there are definite evaluation can be given. Ko­
only five reasonable possibilities at tov specifically warned against jump­
each ply (we will borrow a term from ing from branch to branch; he thought
the computer chess world and call half that you should analyse each candi­
a move a ply), after three whole moves date move until it can be definitely
there are 1 5 ,625 'leaves ' . It follows evaluated, and only then move on to
that analysing solely by means of an the next one.
analytical tree is only possible when The following example, which I
the number of reasonable possibilities have taken from Colin Crouch's inter­
for the two players is limited - in prac­ esting book Attacking Technique, is
tice this means tactical positions and given as an exercise. It can be solved
certain types of endgame. However, by Kotov's method, although this does
8 SECRETS OF PRACTICAL CHESS

not mean to say that you will find it

simple.

L. Psakhis - D. King (analysis)

L. Psakhis D. King
-

London (Lloyds Bank) 1 994

Black actually blundered a piece

away with 31 .. lLixeS?? 32 'ii'x d5+
. Fig 1
'ii'e6 33 lt xe5, but 3 1 . .. l: xeS is better.
Crouch then remarks ''The attempt at a many branches removed for the sake
tactical refutation . . . with 32 lL!e7+ of clarity. The computer analyses by a
does not quite succeed: 32 ... ltxe7 33 process of 'iterative deepening' . Let
ltxe7 :xn 34 'ii'xdS+ � h7 ! 35 'ii'e4+ us suppose it is analysing a position
lt f5+ ! 36 �h l lLif2+ 37 lt xf2 'ii'xf2 38 in which it has to choose between 40
ltf7 � g6 39 lt xf5 'ii'xf5 40 'ii'xb4 with legal moves. It will analyse all legal
a draw." Your task is to choose a move sequences of moves to a particular
after 34 'ii'xd5+ (D) in this line. The depth, taking some branches deeper,
diagram is given at the top of the next especially those with forcing se­
column and the solution may be found quences involving checks or captures.
on page 14. This will allow it to attach a numerical
It is interesting to see how Kotov's evaluation to all the 40 possibilities.
method for, as he put it, teaching Based on its evaluation, it will re­
'human beings to analyse with the ac­ order the 40 moves to put the most
curacy of a machine' compares with promising ones (i.e. those with the
the way machines actually calculate. highest evaluation) first. Then it will
In Figure 1 we see a stage in the go one ply deeper, again analysing all
computer's analysis of a position, with possible legal sequences, and devoting
 A T THE BOARD 9

more time to those moves early in the Your Chess Now, have discussed the
list. pros and cons of Kotov's recommen­
dations, but my concern is to give
practical advice and not to get in­
volved in an academic discussion.
There are several problems which
can arise as a result of Kotov's method.
The most obvious is that it can be ex­
tremely inefficient. Let us suppose
that you are analysing a possible com­
bination l .t xh7+ � xh7 2 lll g 5+.
Fig 2 There are two defences, 2 . . <it> g6 and
.

Figure 2 shows a human being ana­ 2 . � g8. You start analysing 2 . <it> g6;
. . ..

lysing according to Kotov's recipe. He it is very complicated, but after twenty

has listed three candidate moves and is minutes you decide that White has the
in the process of analysing the first of advantage. Then you start looking at
these. He has not started analysing the 2 . cat g8. After a couple of minutes it
..

other two. becomes obvious that this refutes the

sacrifice. The upshot of Kotov's in­
flexible approach is that you have lost
twenty minutes' thinking time, solely
on account of the bad luck of having
chosen the wrong move to analyse
first. A more rational approach would
be to spend a couple of minutes look­
A ing at each of the two alternatives. It is
Fig 3 possible that this will reveal one of
Figure 3 shows how a human actu­ them to refute the sacrifice, at which
ally thinks. He has started analysing point i.. xh7+ can be abandoned. It is
move A and discovered line A l . He also possible that the preliminary
wasn't especially impressed by it, so analysis will show one of them to be a
switched to analysing move B. He clear-cut loss, in which case you can
didn't like that move either, so he re­ switch attention to the other one in the
turned to his analysis of A, adding line confidence that it is definitely the
A2 to his earlier efforts. He has either critical line. If both are unresolvable
not got around to move C, or has sim­ within a short time, then this is in it­
ply forgotten about it. self useful information. It shows that
More recent authors, such as Tis­ the task of determining whether the
dall (see the Introduction ! ) in Improve sacrifice is sound or not will require a
10 SECRETS OF PRACTICAL CHESS

substantial time investment. Then the i.. b 7, but this leads to the same result
decision is really whether it is worth­ after 1 . . . cat c7. Finally, there is 1 i.. a6,
while putting in the effort to analyse which avoids the king fork, but after
deeper. We will revisit this question of l . . . cat c7 White nevertheless must lose
whether or not to analyse in the next a piece, as the knight is now trapped.
section. This is all in accordance with Kotov's
A second problem with Kotov's method. While looking at this last line,
method is that it fails to take into ac­ however, you notice a surprising point:
count the synergistic effect of analys­ after 1 i.. a6 <l; c? White can try 2
ing several lines. The analysis of move cat c5 ! ?, with the idea that 2 . . . cat xb8 3
A and that of move B are very often catd6 <l; a8 4 rt; c? leads to mate by 5
not independent of one another. Sup­ i. b7#. Having spotted this, it is cer­
pose you have rejected move A, but tainly worth noting that it might be
when analysing move B you suddenly possible to utilize it in the previously
notice a tactical possibility. It makes analysed lines. This jumping back is
sense to return to move A to see if the not in accordance with Kotov's method.
same possibility is applicable there. Sure enough, after 1 i.. b 7 catc7 White
Here is a simple example: can play 2 i.. a 6, and again Black can­
not take the knight.
So which line is correct? The an­
swer is that 1 i.. a6 rt;c? 2 rt;c5 fails to
2 d6+ 3 <l; d5 <l;xb8, and now 4 rt; xd6
...

rt; a8 5 cat c7 is stalemate. White must

prevent Black giving up his d-pawn,
and so 1 i. b7 ! rt; c? 2 i. a6 ! is the right
line (2 d6 allows the knight out).
...

Spotting a new idea may even cause

you to change your list of candidate
moves. Here is an example (D):
White has just played 1 9 .:r. g l -g5.
Black was threatening 1 9 . . tlld 3+ fol­
.

J. Gunst lowed by 20 tllc 5+, but now this

...

Das Illustrierte Blatt, 1922 threat is nullified, because on the dis­

covered check White can take on f5.
Clearly White has to move his Moreover, 20 l:r. xf5 gxf5 21 .:r. g l is a
bishop, so there are three possibilities. serious threat.
You look at 1 i.. x d7, and note that My initial thought was the safety­
1 . . . cat c7 wins one of the two minor first 19 . . .e6, but after 20 hxg6 fxg6 21
pieces by a fork. The next move is 1 .:r.dg l the position is just unclear, so I
 A r THE BOARD 11

However, having noticed the possi­

bility of . . . lL!xa2+ in the variation with
. . ..i. h6, I should have gone back to see
if this affected my list of candidate
moves. The key point is that the cap­
ture on a2 becomes possible once the
white queen no longer guards b2, and
so the move 1 9 . . . .i. c2 ! ! (D) springs to
mind:

A. Kuligowski - J. Nunn
Wijk aan Zee 1 983

started looking for alternatives. My

eye was suddenly caught by a tactical
idea. After 1 9 . . . .i. h6 the most obvious
reply is 20 .l:t xf5, but then I wondered
if some combination of . . . .i. xe3 and
. . . lL!xa2+ might force mate. After some
thought I established that 20 . . . lL! xa2+ !
2 1 lL! xa2 (2 1 .i. xa2 .i. xe3) 2 1 . . . 'i!Vxb3 White is now completely helpless,
22 .i. xh6 'ikxa2 wins for Black, and for example:
concluded that after 19 . . . .i. h6 White 1 ) 20 lL!xd5 lL!xa2+ 2 1 � xc2
would have to sacrifice the exchange 'i!Vxb3+ 22 <it>bl .l:tb7 23 hxg6 f5 wins
on g5. The resulting positions ap­ for Black.
peared to give Black some advantage. 2) 20 .l:tdg l lL!xa2+ 21 lL!xa2 'i!Vxb3
The game continued: 22 'i!Vxc2 'i!Vxa2 is also decisive.
19... .i.h6! 20 .l:tdgl .i.xg5 21 .l:txg5 3) 20 hxg6 lL! xa2+ 21 <it> xc2 (2 1
e6 22 hxg6 fxg6 23 h5?! lL!xa2 .i. xb3) 2 1 . . . °ikxb3+ 22 � d3
After 23 'Ii g2 .l:t f6 24 h5 White lL! xc3 23 bxc3 'i!Vc4+ 24 <it>c2 'i!Va2+ 25
would have had better chances of <it>d3 .l:tb2 wins the queen.
counterplay, although Black retains a
large advantage with 24 . . . lL!c2. There are also some types of chess
23 ....l:tb7 24 °ikg2 .l:tg7 25 h6 .l:tb7 calculation which have nothing to do
26 .l:txf5 exf5 27 'ikxg6+ �h8 28 °ikg2 with the 'Tree of Analysis' . One of
After 28 .i. g5 'i\i'xd4 Black defends. these is the ' goal-seeking ' approach.
28 f4 29 .i.gl .l:tg8 0-1
.•. Here is an example:
12 SECRETS OF PRACTICAL CHESS

17 .i.xh7+ �xh7 18 'ii' bS+ <Jitg8 19

.i.xg7 �xg7 20 ..g4+ <Jitb7 21 l:lf3
1-0

You can never find a move like 1 6

lbb6 ! by using the tree of analysis (un­
less you are a computer) because it
makes absolutely no sense except as
part of the whole tactical operation,
and therefore will not make it onto
your list of 'candidate moves' .
Here is a more sophisticated exam­
G. Kuzmin E. Sveshnikov
- ple.
USSR Ch, Moscow 1 973

The position seems tailor-made for

a double bishop sacrifice on h7 and
g7, the only problem being that it
doesn't work: 16 .i. xh7+ � xh7 1 7
'ii'h 5+ � g8 1 8 .i. xg7 <lt xg7 1 9 'ii'g4+
<Jith7 20 .:r.f3 (of course White can give
perpetual check) 20 . . . lb xf4 2 1 l:r. xf4
f5 defends. In order to win, White has
to isolate the element that makes the
sacrifice fail (i.e. the possibility of
. . . lb xf4) and ask a 'What if. . .' ques­
tion, in this case 'What if I could de­ V. Anand J. Lantier
-

flect the knight away from d5?' . Then Biel 1 997

the solution becomes obvious:
16 lbb6! lbxb6 There are so many pieces hanging
It does not help if Black moves his that it takes a few moments to grasp
rook, for example 16 . . . l:r. b8 17 lbxd5 what is going on in this position !
cxd5 18 .i. xh7+ <Jitxh7 19 'ii'h5+ <lt g8 White is set to lose the exchange on
20 .i. xg7 'ii'xc5+ 21 �hl � xg7 22 e3, but the bishop on g2 is trapped, so
'ii'g 4+ <Jith8 23 l:r. f3 'ii'c 2 24 f5 ! 'ii'xf5 it looks as if the most likely outcome is
25 .:r. xf5 exf5 26 'ii'h3+ (had Black a position with, for example, two bish­
played 16 . . . .l:r. a?, White would now ops against a rook and two pawns.
have 26 'ii'd4+) followed by 'ii'g3+ and However, Anand had an imaginative
the rook on b8 falls. tactical idea, namely to sacrifice his
 A T THE BOARD 13

queen with the astonishing move 20 least you know where the tricky areas
J. g6. Unfortunately, after 20 J. g6 lie.
Wxd l 21 .:l xe6+ Black can just run The most common errors in 'tree'
with his king by 2 1 . . . � f8 22 i.. a 3+ analysis are:
(22 .:lxf7+ � g8) 22 . . . li::Je? 23 J. xe7+ I) Forgetting to analyse a move
� g8 and White is losing. Now White completely. This is surprisingly easy
can ask 'What if I had a pawn on h6?' . to do. After spending twenty minutes
Then the line with l::lxf7+ would lead analysing defences A and B to your
to mate as .:lxg7+ followed by J. a3+ intended sacrifice, you decide that it
would be possible. is sound and play it. The instant your
20 h6! ! gxh6? hand has left the piece you remember
Black apparently hadn' t seen An­ defence C, which you had noticed but
and's idea at all, or he would have tried not analysed.
20 . . . li::Jxe3, although White retains the 2) Confusing similar lines. If you
advantage after 2 1 i.. xe3 'ii'e5 22 hxg7 are jumping about a lot between rather
l:r. g8 23 'ii'c l ! . Of course, the variation similar variations, then it is easy to get
20 . . . g6 21 J. xg6 represents the fulfil­ mixed up as to which position came
ment of White's idea. from which line. In other words, al­
21 J.g6!! li::Je7 though you have the positions clear in
Now that h6 is available for White's your mind, the links between them
bishop, the line 2 1 . . . 'ii'xd l 22 l:r. xe6+ which form the branches of the 'tree'
�f8 23 J. xh6+ � g8 24 i.. xf7# ends in have become muddled. You may need
mate. to start rebuilding the 'tree ' in your
22 'ii'xd4 .:lxd4 23 .:ld3 ! l:r.d8 24 mind from scratch when this hap­
.:lxd8+ <itixd8 25 i.. d 3! 1-0 pens.

The tree of analysis is certainly a While a certain amount of mental

useful technique in tactical positions, discipline can greatly increase the effi­
but it should be used flexibly and sup­ ciency of your tactical analysis, a great
plemented by other types of chess deal still depends on ' natural talent' . It
thinking. One has to strike a balance would be quite easy to overlook the
between the rigid application of Ko­ whole idea of J. g6 in the Anand­
tov's principles and jumping from one Lautier position above, and in this
variation to another too much. In par­ case a wonderfully organized tree of
ticular, I favour the 'quick scan' ap­ variations will not help you. But then,
proach, looking briefly at all the major Lautier missed it too, so at least you
lines to see if any can be quickly re­ would not be alone.
solved. With any luck, this will be Now we return to the exercise on
enough to settle the analysis; if not, at page 8. Here is the diagram again:
14 SECRETS OF PRACTICAL CHESS

39 . . . � f5 40 'ii'f7 + (40 'ii'd 7+ � g5)

40 ... �g5 41 'ii'xg7+ <ith5 winning.

Evaluation functions

When computers analyse a position,

they create a tree of analysis, and then
use a rather crude evaluation function
to assess the position at the end of
every branch. Then, by working back­
wards, they can evaluate the current
position and establish what they con­
L. Psakhis - D. King (analysis) sider to be the optimal line of play. If a
computer had a perfectly accurate
Crouch's 34 . . . 'ii? h7 is actually the evaluation function, then it would not
weakest of the three king moves, and need to analyse the position more than
leads to a draw, as in his analysis. one ply ahead - it could simply evalu­
Initially I thought that 34 . . . <ithS ate the position after each legal move
was winning, since any check on the and choose the one with the highest
eighth rank is met by . . . l:. f8+, and oth­ evaluation. The effect of analysis is to
erwise White has no defence to the increase the accuracy of the fairly
threats against his king. However, primitive evaluation function.
Fritz found an incredible defence: 35 Humans tend to analyse in a differ­
l:. e8+ l: f8+ 36 �b l l:r. xe8? 37 l:r. f8+ ! ent way, but they still use an evalua­
l:. xf8 (after 37 . . . <itih7 38 'ii'f5+ Black tion function. While they may not
is worse) 38 'ii'g 8+ ! forcing stalemate. think they have an advantage of 0.32
Therefore, Black would have to play pawns, it is quite normal to think
36 . . . ll'i f2+ 37 l:r. xf2 l:. xe8, with a large 'Well, at the end of this line I have a
advantage although no forced win. slight/fair/clear/crushing advantage.'
The final move, 34 . . . <it f8, is actu­ After looking at various lines the hu­
ally the strongest and leads to a forced man decides on the best. This process
win for Black, although it appears the is rather similar to that performed by
least likely move because it self-pins the computer.
the f2-rook. White can only try 35 l:.e2 However, the human also typically
l:r. f6+ 36 l:r.ff2 (36 l:r. ef'2 l:r. xt'2 37 'ii'a8+ uses an evaluation function in a differ­
� e7 38 'ii'e4+ <iti f6 wins) 36 . . . l:r. xf2 37 ent way. If you think that you have an
'ii'a 8+, but there is no perpetual check: advantage in the current position, then
37 ... � f7 38 'ii'e8+ (38 'ii'd 5+ <iti f6) you automatically reject moves which
38 . . . <it f6 3 9 'ii'e7+ (39 'ii'f8+ � g6) allow your opponent total equality.
 A T THE BOARD 15

This type of reasoning is very com­ Y. Seirawan - J. Nunn

mon and is extremely useful in rapidly Brussels World Cup 1 988
cutting down the number of moves King's Indian Defence
that you have to consider, but it also
has its dangers. It is almost the reverse 1 d4 ll:if6 2 c4 g6 3 ll:ic3 i.g7 4 e4 d6 5
of the computer's logic; instead of us­ f3 0-0 6 i.e3 ll:ic6 7 ll:ige2 a6 8 d5
ing the analysis to evaluate the posi­ ll:ie5 9 ll:ig3 c6 10 a4 cxd5 ll cxd5 e6
tion, the 'evaluation ' is used to prune 12 i.e2 exd5 13 exd5 l:r.e8 (D)
the analysis tree. However, because
the initial 'evaluation' is not based on
concrete analysis of the current posi­
tion, it is inherently unreliable. Typi­
cally, players simply carry over their
evaluation from the preceding move
and use it as a starting point for the
next one. If the evaluation is inaccu­
rate, then all kinds of odd things can
start to happen. Suppose you have
three possible moves, A, B and C, and
believe that you have the advantage.
You analyse A and decide that it leads
to equality; B the same. You then con­ I was very happy with the result of
clude that C must be the correct move the opening. It seemed to me that 8 d5
with very little analysis. If the position was premature, giving Black a target
is in fact equal, then C may actually to aim at with . . . c6 and . . . e6. In the dia­
lead to a disadvantage, although you gram position, I felt that the move f3
do not realize this immediately. Then, was not only a loss of time, but even
because you tend to carry over evalua­ served to weaken White's dark squares,
tions from one move to the next, you especially e3. It is thanks to this that
repeat this exercise in defective logic White cannot castle (allowing the e3-
the next move. The upshot is a whole bishop to be exchanged by ... ll:ic4 and
string of inaccurate moves; this is, in . . . ll:ie3 would really be bad for White).
fact, what is commonly known as 'lo­ Therefore I assessed the position as
sing the thread of the game' . If you slightly better for Black, but this was
experience 'losing the thread' , try af­ just wrong. Perhaps White has not
terwards to trace the problem back to played the opening in optimal style,
its root cause; surprisingly often, the but a small inaccuracy by White is not
source of the error was a faulty evalua­ enough to hand Black the advantage;
tion. he has equalized, but no more.
16 SECRETS O F PRACTICAL CHESS

14 'ii'd2
Not unexpected, since 14 0-0 ll:ic4
is bad (as mentioned above) and 1 4
�f2 'ilc7 1 5 'ii'b3 (preventing . . . lLic4)
1 5 . . . 'ile7 ! creates unpleasant pressure
along the e-file.
14 'ii'e7
•••

My first thought was to continue

14 . . . 'ii'c 7 15 0-0 lLic4 16 i. xc4 'ilxc4,
but then I saw that White could more
or less force a draw by 17 ll:ige4 ll:ixe4
( 1 7 . . . 'ilb4 1 8 ll:i xf6+ i. xf6 19 ll:ie4
'ilxb2 20 l: abl 'ile5 2 1 i. f4 'ild4+ 22 The rook emerges so that it will not
�h l even favours White) 1 8 ll:i xe4 be blocked in when the knight has to
'ilb3 ! 19 i. d4 ! 'ii'xd5 20 :fd l 'ilxd4+ retreat.
2 1 'ilxd4 i. xd4+ 22 : xd4 l: e7 23 16... h4 17 ll:ifi
ll:i xd6. After 17 ll:i ge4 ll:ih7 ( 1 7 . . . ll:ieg4+ ! ?
I could also have simply developed 1 8 fxg4 ll:i xe4+ 1 9 ll:ixe4 'ilxe4 20
by 14 ... i.d7 1 5 0-0 :cs, but once again i. f3 'ii'c4 is also possible) 1 8 'iii> g l f5
this hardly promises Black more than White has to retreat with loss of time.
equality. 17...ll:ihS? (D)
1S �f2 Up to this point Black has not made
Of course 15 0-0? ll:ic4 ! and 1 5 any real error, but now his desire to
i. d4? ! ll:i c4 are fine for Black, but play for the advantage leads him badly
the fact that White had to play this astray. Originally I intended 1 7 . . . h3,
artificial-looking king move seemed thinking that after 18 g4 there would
to justify my earlier assessment. I was be some combination based on taking
now quite excited about various tacti­ the g4-pawn. Now I discovered that
cal possibilities involving a bishop or there was no such combination. I be­
knight moving to g4, but at the mo­ came frustrated by my inability to find
ment no such idea is effective. a continuation consistent with my ear­
15 ... hS lier active play. The result was the very
Continuing my plan to 'punish' weak text-move, based on a trap which
White for his opening play. Since doesn't even work!
there was no immediate tactical blow, The right plan was 17 ll:ih7 (pre­
.. .

the only way to keep the momentum venting i. g5) 1 8 f4 ll:i g4+ 1 9 i. xg4
going seemed to be by pushing the h­ i. xg4 20 i. d4 i. xd4+ (20 . . . 'ii'f8 21
pawn. ll:ie3 favours White) 21 'ii'xd4 'ii'f6 22
16 l:hel (D) 'ii'xf6 ll:i xf6 leading to a roughly equal
 A T THE BOARD 17

ending, but o f course this was not ac­ 20 i.d4

ceptable to me. Threatening 2 1 f4.
20 ... 'iifS 21 �g3 �h7 22 gS
Sealing the h7-knight out of play.
22 ... f6 23 f4 �g4 24 �ce4 i.d7
(D)

18 �gl
After 1 8 i.. g 5, I intended the pawn
sacrifice 18 . . . 'iic7 19 i.. xh4 'ii'c 5+ 20
� 3 i..h6 21 lLJe4 �g4+ 22 fxg4 :xe4,
but after 23 i. f3 :d4 24 'ii'c3 Black is White now made the very weak
struggling to find any compensation move 25 gxf6?, which allowed the
for the pawn. Perhaps 19 . . . 'ii'b 6+ 20 dead knight on h7 back into the game;
�e3 i. h6 is better, but even so 2 1 after many complications the game
�cd l i s unclear. In fact the mundane ended in a draw.
18 . . . i. f6 19 i.. xf6 'ii'xf6 20 �e4 'ii'e7 Had Seirawan continued 25 'ii'd l f5
is probably best. 26 i.. x g7 'ii'xg7 27 i.. x g4 fxg4 28 'iid2
18... h3 then Black would have been in trou­
Now 18 . . . i.. d7 ? ! is bad after 1 9 ble. The d6-pawn is under fire and the
i. g5 ! i.. f6 2 0 f4 i.. x g5 2 1 fxg5 � g7 h7-knight can only emerge in the dis­
22 �e4, heading for f6 so, to avoid tant future. Even the solid 25 :n
losing the h4-pawn to i.. g 5, the pawn would have given White a clear advan­
has to advance. tage.
19 g4 In this example the chain of events
Cutting off the h3-pawn. was: a bad assessment of the position,
19... �f6 resulting in over-ambitious play by
The sacrifice 1 9 . . . � xg4 20 fxg4 Black; then avoidance of drawing
i..xc3 2 1 bxc3 'iie 4 is refuted by 22 lines, leading to the awful 17 . .. �h5?.
i. f3 ! 'iFxf3 23 gxh5, so the knight has Black only got back on track after his
to retreat with loss of time. position had severely deteriorated.
18 SECRETS OF PRACTICAL CHESS

When to analyse If you have thought about a position

for some time and are still unsure what
One question which Kotov did not to play, then it is essential to be ruth­
really cover is whether to analyse at lessly pragmatic. You have to ask
all, or how much to analyse. Yet this is yourself whether further thought is
of great practical importance. One really going to help you make a better
might end up playing better moves if decision. Suppose your lengthy cogi­
one could analyse for an unlimited tation is the result of indecision; for
time, but games played before the in­ example, say there are two moves, both
troduction of chess clocks (and some of which have roughly equal merit. If
postal games) show that this is not you have not been able to decide be­
necessarily the case. Too much analy­ tween them up to now, it is reasonable
sis can easily lead to fatigue and con­ to suppose that there is in fact little to
fusion. These days the tendency is choose between them. Considerable
towards faster and faster time-limits, further thought might eventually re­
and this means that apportioning one's veal some tiny difference, but it is rare
thinking time becomes ever more cru­ that this expenditure of time is worth­
cial. while. First of all, it is easy to be wrong
When analysing a given position, it when dealing with such fine distinc­
is fair to say that one almost always tions and secondly, gaining an infini­
sees more in the first five minutes than tesimal advantage is of little value if
in the next five minutes. The five min­ the result is that you run into time­
utes after that is even less productive, trouble and blunder away a piece.
and so on. I have observed that if a Games are decided by very small ad­
player spends more than 20 minutes vantages far less often than is usually
over a move, the result is almost al­ supposed. There are players, such as
ways a mistake. The normal decision­ Capablanca and Karpov, with the abil­
making process should not take longer ity to convert a small advantage into
than this, even in fairly complex situa­ a win on a regular basis, but this talent
tions. There will, of course, always be is rare even amongst grandmasters.
exceptions to any such empirical law, Games in which the players make er­
but in this case they are fairly rare. If a rors and the advantage swings back
player takes a long time over a move, and forth are far more common; in the
the reason is usually either indecision end the decision comes about as a re­
or inability to find a satisfactory con­ sult of a serious mistake. My advice is
tinuation. It is very unusual for the po­ to obey your instinctive feeling as to
sition to be so complicated that it which of the two moves is better or, if
really demands more than 20 minutes' you don't have any preference, just
thought. choose at random. I have occasionally
 A T THE BOARD 19

been tempted to toss a coin at the play it in the end anyway and further
board, but despite the undoubted psy­ thought is only going to waste time.
chological impact on the opponent, this The situation which represents the
does seem rather unsporting. worst of all worlds is when you want
If you are unable to find a satisfac­ to play X, but can't find any objective
tory continuation, then once again it reason for doing so; then you analyse
usu ally does not help to think on and on and analyse until you find some half­
hoping for a miracle. Of course, if it is baked and dubious reason why X is a
a question of coming up with some­ good move. Then you end up playing a
thing or resigning, then you may as dubious move and spending a lot of
well continue, but this is uncommon. time. However, it is usually worth­
Usually, you are better off playing while spending a few minutes trying
what appears to be the least unfavour­ to find something concrete wrong with
able continuation and saving the time your intuitive choice, because posi­
for accurate defence and possible tional intuition doesn' t insure against
counterplay later on. It is worth noting the possibility that it might be bad
that if you are generally unhappy with for tactical reasons. If, during these
your position, then this might well be few minutes you don't find anything
colouring your assessments. Before wrong, then you should just play the
choosing a move, it is probably worth move.
having a quick review of the alterna­ The following game is a successful
tives and the reasons why you deemed example of playing (almost ! ) without
them unsatisfactory, to see if they are calculation.
really as bad as you imagine. It is not
unusual suddenly to find that your in­ J. Nunn P. van der Sterren
-

tended equal continuation is slightly Bundesliga 1 99516

worse for you ; you then look at more Ruy Lopez
and more unfavourable possibilities,
become deeply depressed about your 1 e4 eS 2 lLJf3 lLJc6 3 i.bS a6 4 i.a4
position, and forget that the first line lLJC6 S 0-0 i.e7 6 :%.el bS 7 i.b3 d6 8
you looked at wasn' t actually all that c3 0-0 9 d3 lLJaS 10 i.c2 cS 1 1 lLJbd2
bad. :es 12 tLJn h6 13 lLJg3 i.f8 14 d4
Another common time-waster is exd4 IS cxd4 cxd4 16 lLJxd4 i.b7 17
insti nctively to want to play a certain b3 dS 18 eS lLJe4 (D)
move and then to spend a long time Black's inferior handling of the
trying to back up this intuitive feeling opening has left him with an offside
with concrete analysis. Don't do this. knight on a5. The idea of . . . d5 and
If your intuition is telling you strongly . . . lLJe4 is to offer a pawn in order to
to play move X, then you will probably free his position.
20 SECRETS OF PRACTICAL CHESS

d l-h5 diagonal, along which White

might like to move his queen.
19 'iWb8
...

After 19 . . . l:r. xe5 20 f3 J. b4 21 fxe4

i. xe l 22 •xe l 'ii'b6 23 �bl White
has a large advantage since an assault
on g7 is not far away.
The text-move was unexpected, but
the point is clear: if White takes three
times on e4, then Black just plays
. . . l:r. xe5.
20 e6
19 i.b2 A key moment. I thought for a few
I did consider talcing the pawn by minutes about the direct 20 'ii'g4,
19 ll:ixe4 dxe4 20 J. xe4 J. xe4 21 :xe4 when Black has almost no choice but
'iid 5 22 'iWe2 (but not 22 'ii'g4 :xe5 23 to play 20 . . . 'ii'xe5 . Then it looks as
l::lxe5 'iixe5 24 J. xh6 f5 25 'iWf4 i. d6 though there must be a tactical possi­
26 'i!Yxe5 i. xe5 27 J. e3 f4 and Black bility using the long diagonal, but I
wins) but I instinctively didn't like it. could only see one idea: 2 1 lL!df5
White 's pieces are badly tangled up, 'ii'xb2 22 lLi xh6+ <ith7 23 ll:ixf7 •xc2
while Black's development problems 24 : xe4 dxe4 25 ll:i g5+, followed by
are solved. It seemed to me that it 'ii'h5 . However, even a quick glance is
would be very hard to put the extra enough to show that the analysis of
pawn to good use in such a position. this line will be very complicated: in
Looking back on it, I think this line the middle Black can try 24 . . . : xe4, or
would also have been somewhat better he may decline the second piece and
for White, since Black still has trouble play, for example, 23 ... 'ii'f6.
using his knight effectively. The more Here is a case in which a quick scan
dynamic game continuation is at least of the other possibilities is much bet­
as good, and presents Black with diffi­ ter than the twenty minutes or so
cult decisions. which would have been required to as­
In general, White would be quite sess this double piece sacrifice. I first
happy for Black to exchange knights looked at 20 ll:i gf5 l:r. xe5 2 1 'ii'g4 g6,
on g3, as then he would have an auto­ but there seemed to be nothing clear.
matic attack by 'ii'd3 (meeting . . . g6 by Then the text-move occurred to me
e6). However, I was not prepared to and I was at once attracted to it. After
spend a tempo on forcing it by playing 20 . . . fxe6 2 1 ll:i xe4 dxe4 22 i. xe4
1 9 f3. Moreover, this move both weak­ J. xe4 23 : xe4 Black has a miserable
ens the a7-gl diagonal and block!? the position; the e6-pawn is weak, the
 A T THE BOARD 21

aS-knight remains offside and White i.. xe4 i.. xe4 25 :%. xe4 'iWf6 26 :%.f4 !
can step up his pressure with natural 'ifg5 27 'ii'xg5 hxg5 28 :%.e4 is similar
moves such as 'ii'g4 and :%. ael . Best of to the game, but with Black's kingside
all, it is simple and risk-free. pawns even weaker.
Home analysis shows that the dou­ 24 'ii'e2 fxe6 25 i.. xe4 'ifxe4 26
ble piece sacrifice leads to a draw: af­ 'ii'xe4 i..xe4 27 :xe4
ter 25 � g5+ 'it> g8 26 'iW h5 in the above An ideal outcome. Black will inevi­
line , Black can respond 26 . . . e3 ! and tably lose the e6-pawn within a couple
White has nothing more than perpet­ of moves (e.g. 27 . rt; f7 28 :%. ae l i.. b4
..

ual check. 29 :%.f4+) and he remains with an off­

20... 'iWf4 side knight.
Another rather unexpected move. 27...:%.ac8
21 �xe4 dxe4 22 g3 If White takes on e6 immediately,
Again a simple choice. In order to then this rook will penetrate to c2.
avoid losing the pawn on e4, Black's However, there is no rush.
queen must move onto the long diago­ 28 :%.ael
nal, but then White has various tactical Threatening to take twice on e6
ideas based on moving the d4-knight. with the rooks, thereby leaving the
22... 'iWeS d4-knight to cover c2.
After 22 ... °iff6 23 °ifd2 b4 24 .t xe4 28 ....tb4 29 :1e2 .tc3 ::SO .txc3
Black will be lucky to escape with :xc3 31 :%.xe6 1-0
only the loss of a pawn. This may appear a little premature,
23 'iWg4 but Black will lose the a6-pawn as
It is important to take into account well, and there is no point continuing
any finesse which might make the win with two pawns less.
easier. Here White has the choice be­
tween 23 'iWe2 and 23 'ii'g4 h5 24 'ife2. DAUT
The latter is clearly superior for two
reasons. Firstly the protruding pawn This acronym means: if in doubt,
on h5 is a weakness, as it will be unde­ 'Don't Analyse Unnecessary Tactics' .
fended; secondly, the weakening of Tactical analysis is an error-prone
the g5-square may allow White to oc­ activity. Overlooking one important
cupy it later with his knight, creating a finesse can completely change the re­
permanent danger to Black's king. sult of the analysis. If it is possible to
This is an example of the method of decide on your move on purely posi­
comparison (page 44) . tional considerations then you should
23 bS
••. do so; it is quicker and more reliable.
There is nothing better. 23 . . . i.. b 4 24 There are, of course, many positions
l:le2 is no different, while 23 . . . fxe6 24 in which concrete analysis is essential,
22 SECRETS OF PRACTICAL CHESS

but even in these cases you should not

analyse specific variations more than
necessary.
The following example is a mar­
ginal case.

J. Nunn M. Pribyl
-

Bundesliga 1 99516
Giuoco Piano

1 e4 es 2 ll:if3 ll:ic6 3 .tc4 lLif6 4 d3

.t c5 S 0-0 d6 6 c 3 0-0 7 ll:ibd2 a6 8
i. b3 i.a7 9 h3 ll:id7 10 i. c2 rs 1 1 indeed look promising on purely stra­
exfS l:xfS 1 2 d4 lt f8 tegic grounds, but the wide range of
Black has played the opening in a moves at White 's disposal indicates
rather provocative manner, aiming to that a complete analysis will be very
open the f-file and create counterplay time-consuming.
by attacking f2. However, the cost has In the end I decided on a straight­
been retarded development. forward forcing continuation which
13 ll:ie4 gives White a slight positional advan­
After 1 3 d5 ll:ie7 14 ll:i g5 lLif6 1 5 tage. In fact White could have secured
ll:i de4 lLi f5 White has secure control a larger advantage by another line, but
over e4, which would warm the heart that is the risk one takes in making a
of anyone who plays against the King's decision not to look too deeply into a
Indian. Unfortunately for White, the position. However, I would certainly
black bishop is on a7 rather than g7, have taken my time had there not been
and the pressure on f2 gives Black rea­ an advantageous alternative.
sonable counterplay. 14 lLiegS
13 ... exd4 (D) Here is an analysis of the other pos­
With quite a few of White's pieces sibilities:
pointing at Black's naked kingside, 1 ) 1 4 i. g5 'ii'e 8 1 5 lt e l lLide5 1 6
there is clearly potential for a forcing ll:i xd4 .t xh3 ! 1 7 gxh3 i. xd4 1 8 cxd4
continuation, but which move should ll:if3+ 19 'it> g2 ll:ixel + 20 'ii'xel ll:i xd4
he choose? 14 ll:ieg5, 14 lLifg5, 14 and Black is slightly better.
i. b3+ and 1 4 .t g5 are all reasonable 2) 14 'ii'd3 ll:ice5 1 5 ll:ixe5 ll:ixe5
continuations. I would regard this as a 16 ll:if6+ 'ii'xf6 17 'ii'x h7+ 'it> f7 de­
borderline case for deciding whether fends.
to spend the time to analyse every­ 3) 1 4 .t b3+ 'it> h8 1 5 ll:i fg5 'ii'e 7
thing thoroughly. The position does (not 15 . . . ll:ide5 16 lLixh7 !) 16 ll:ie6 ( 1 6
 A T THE BOARD 23

'6'h5 g6 1 7 1i'h6 l0 f6 and 1 6 l0 xh7 example 17 . . . i. f5 fails to 1 8 'ii b3+

'i'xe4 17 l0 xf8 l0 xf8 1 8 'iih 5+ l0 h7 and 1 9 i. xf5 . Now, however, 18 ... i. f5
1 9 'iif7 i. xh3 20 gxh3 dxc3 are fine is a threat.
for Black) 16 . . . l0f6 17 l0 xf8 'iixe4 1 8 18 l0xh7! (D)
:e t 1i'h4 1 9 l0e6 dxc3 and Black has Definitely best. The net effect of
two pawns for the exchange - the posi­ this is to exchange the h7- and d4-
tion is unclear. pawns, further eroding the defences of
4) 14 l0 fg5 ! and now: the black king.
4a) 14 . . . d5 15 l0e6 ( 1 5 'iih5 h6 16
l0e6 is also good) 1 5 ... 'ii'e 7 1 6 l04g5
l0f6 17 l0 xf8 wins the exchange.
4b) 14 ... l0 de5 15 l0 xh7 ! : f5 1 6
l0eg5 d3 1 7 i. b3+ ( 1 7 i. xd3 l0xd3 1 8
'ii x d3 is unclear a s the knights have
no way back from their advanced po­
sitions) 17 ... d5 18 g4 ! :n 19 l0 xf7
l0xf7 20 °ii'xd3 l0ce5 21 'ii'xd5 'iixd5
22 i. xd5 'iii> xh7 23 :e t and now, with
l:.+2.0. vs 2 l0 , White has the better
ending.
4c) 14 . . . dxc3 15 'iih5 ( 1 5 l0e6 'iie7
16 l0 xf8 cxb2 1 7 i. xb2 l0 xf8 is un­ 18 i.fS
•.•

clear) 1 5 . . . h6 1 6 °ii'g 6 with a decisive Not 1 8 . . . 'ito> xh7 1 9 'ii'xg6+ 'ito> h8 20

attack. i. g5 with a decisive attack, for exam­
4d) 1 4 . . . h6 1 5 l0e6 fie7 1 6 l0 xf8 ple 20 . . . 'iie7 2 1 l:. ae l fig7 22 i. xf6
l0 xf8 1 7 l:. e l i. e6 and Black has a l:. xf6 23 l:.e8+ l:.f8 24 l:. xf8+ °ii'xf8 25
pawn and a slight lead in development °ii'h7#.
for the exchange, but this does not pro­ 19 l0xf6+ 'ii'xf6
vide sufficient compensation. Forced, as 1 9 . . . : xf6 20 fi b3 i. xc2
14 l0f6 15 'iid3 g6
••. 21 'iixc2 i. xd4 22 i. g5 costs the ex­
After 1 5 . . . h6 1 6 l0 h7 lt e8 1 7 'ii'g6 change.
White has a winning attack, so this 20 'ii b 3 i.xc2 21 Vxc2 i.xd4 22
move is forced. However, the weaken­ i.h6 l:f7 23 l:adl
ing of the dark squares around Black's This position is the almost forced
king is serious when his dark-squared consequence of White's decision at
bishop is far away on the queenside. move 14. The immediate threat is 24
16 l0xd4 l0xd4 17 cxd4 d5 l:hd4, and the d4-bishop is pinned
This is almost forced, or else Black against the d5-pawn.
cannot complete his development, for 23 l:b7? (D)
••.
24 SECRETS OF PRACTICAL CHESS

A serious error, which costs Black switch his bishop to the long diagonal,
the game immediately. Other moves: or bring his rook into the attack, and
1) 23 ... <ith7 24 i.e3 i. xe3 25 fxe3 the game will be over. The lines are
'fle7 26 l:r. xf7+ 'ii'xf7 27 e4 dxe4 28 quite straightforward and forcing, so
'ii'xe4 is unpleasant for Black. The there is little possibility of error.
b7-pawn is attacked, and White threat­ 24 l:r.xd4! 'ii'xd4 25 'ii'x g6+ �h8
ens both 29 l:r.fl and 29 'ii'g4, followed 26 .l:r.el
by 30 :d7. Black is helpless against the threat
2) 23 . . . c5 24 i.e3 i. xe3 25 fxe3 of 27 l:r.e8+.
'ii'c 6 26 l:r. xf7 � xf7 27 e4 ! d4 (White 26 :d7
•••

wins a pawn after 27 . . . dxe4 28 :n + The only way to limp on. Now White
cl; g7 29 'ii'c 3+ '3; g8 30 l:r. f6 'ii'e 8 3 1 can win two pawns by 27 l:r. e8+ l:r. xe8
'ikc4+ cl;g7 3 2 l:r. e6 'ii'd7 3 3 'ii'xe4 28 'ii'x e8+ � h7 29 'ikxd7+ � xh6 30
'ii'd4+ 34 'ii'xd4+ cxd4 35 l:r. d6) 28 'ii'xc7, but Black has a passed d-pawn
'ii'c4+ cl;g7 29 b4 ! with strong pres­ and in a queen ending a strong passed
sure. pawn can counter-balance the loss of
3) 23 . . . i. xb2 24 l:r. xd5 l:r.e8 25 several pawns. At any rate, White
l:r. fd l i. e5 is Black's best line. How­ would have to take care, so it is worth
ever, I still prefer White because his looking for a killer move.
king is completely safe, while Black's 27 'ii' b5 ! 'ii'd 3
has only one pawn for protection. There are simply too many threats .
27 . . . '3;g8 28 l:r.e3 is hopeless, and after
27 ... 'ii'c4 28 i.f4+ '3;g8 29 :es+ White
wins the black rook for nothing.
28.l:r.e6
Once the rook enters the attack it is
all over.
28 :gs 29 i.g7+! �xg7 30 'ii'h6+
•••

1-0
It is mate next move.

Quite apart from the possibility of

miscalculation, there are psychologi­
cal traps lying in wait for the avid cal­
Here is a case in which it most defi­ culator. Suppose that, in a slightly
nitely is worthwhile looking for a favourable position, you see a sharp
forced win ! After the sacrifice on d4 and complicated line; it takes you half
Black will have no pawns at all de­ an hour to investigate all the variations
fending his king. White only needs to which may arise, but you discover that
 A T THE BOARD 25

the upshot is at best 'unclear' . It is prepared to sacrifice the rook on gl to

then incredibly hard to write your achieve it. The game continued with
analysis off as a waste of time and start the complex tactical variation 19 f4 !
thinkin g about alternative ideas. Even­ l0f3 20 'iig2 l0 xg 1 2 1 e5 0-0 22 .i. d3 !
tually you convince yourself to play .i. xe5 ! 23 fxe5 'ii'xh4 24 l:h g l 'ii'f4+
the tactical line with an argument like 25 �bl ! 'ii'xe5 . Even for a grandmas­
'well, he probably won' t find his way ter, working one's way through this
through all the complications, and line is far from easy, but that is not the
even then he doesn' t have any advan­ end of it. In the resulting position
tage' . Thus you end up playing a move Black has a rook and three pawns for a
which is objectively not the best, and it bishop and a knight, on paper a con­
is amazing how often one's opponent siderable material advantage. How­
does indeed find his way through the ever, Anand showed that after 26 l0c5 !
complications once they have been White's initiative prevents Black from
forced on him. co-ordinating his pieces. He went on
to win very nicely.
One can hardly criticize Anand's
choice, as it did give him the advan­
tage in a very forcing way, but for or­
dinary mortals (and normal GMs such
as myself) this continuation would be
a distinct leap in the dark. If faced with
the diagram position, it would be rea­
sonable to spot 19 f4 and think "Well,
that may be good, but it's very compli­
cated. I'll analyse it if I have to, but is
there another line which gives me the
advantage and involves less risk?" The
V. Anand G. Kamsky
- move 19 .i.e2 is natural ; it threatens
Linares 1 994 f4 and e5 without any sacrifice at all .
Black cannot blockade White's pawns
In this position a great deal depends with 1 9 . . g5 because 20 hxg5 hxg5 2 1
.

on White 's mobile e- and f-pawns. If 'ii'xb4 (or 2 0 'iWxb4 gxh4 2 1 l: xd4) is
he can set them moving by f4 and e5, very good for White. Anand did not
gaining tempi with both moves, then play this move because of 19 . d3, . .

Black will be dead lost. At the moment, which is indeed the only reasonable
however, Black threatens a fork on f3. reply. However, White can then con­
Anand decided that forcing through tinue 20 'ii'e3 ! and Black is in big trou­
f4 and e5 was so important that he was ble. The threat of f4 and e5 is renewed,
26 SECRETS OF PRACTICAL CHESS

and the pawn on d3 is hanging. The

continuation might be 20 . . . i.. xh4 2 1
i.. x d3 'ilc7 2 2 lll b 6 lll x d3+ 2 3 .:r. xd3
.:r.d8 24 .:r. xd8+ 'iFxd8 25 llld 5 and
Black is losing. He cannot castle ow­
ing to 26 'ii'xh6, the g7-pawn is hang­
ing, White threatens 'ii'c5 and the
bishop is dominated by the powerful
centralized knight. This variation is
fairly straightforward to calculate, and
probably not really necessary as after
22 lll b6 it is already clear that White
stands very well. Thus White could A. Shirov B. Gelfand
-

have gained at least as much advan­ Dos Hermanas 1 996

tage as in the game with the simple
move 1 9 i..e 2. It is easy to imagine This does provoke the following
that having calculated the difficult and combination, but Gelfand probably
attractive line with 19 f4, White was did not realize how dangerous the
not eager to find an equally good sim­ coming sacrifice would prove to be.
ple alternative, even though moves 17 . . . b4 would have been a safer alter­
like 19 i.. e 2 and 20 'ile3 are child's native.
play to someone such as Anand. 18 1:.xeS!
This involves the sacrifice of a
Safety-nets whole rook.
18 i.. d6 19 i..f4 gS?!
.•.

When you are thinking about a com­ After 1 9 ... i.. x eS 20 i.. x e5 'ii'd8 2 1
plex and lengthy tactical line, espe­ l:r.fl (2 1 i..c7 ll\hs ! ) Black would come
cially one involving sacrifices, it helps under very strong pressure, for which
to have a safety-net, i.e. an alternative he has only a minute material advan­
line which you can adopt if, half-way tage as consolation.
through your intended continuation, Shirov did not analyse 1 9 . . . ll:\ c4 ! ?,
you discover that it doesn't work. but perhaps this would have been a
The simplest type of safety-net is way for Black to reach a reasonable
when you have a perpetual check in position. The main line runs 20 l:r. c5
hand (D) . (20 i.. x c4 bxc4 21 llld4 g5 ! is much
This looks like a fairly normal Naj­ better for Black than the game, as with­
dorf Sicilian, but Shirov found an out the bishop on d3 White's threats
imaginative tactical idea. are far less dangerous) 20 . . . i.. xf4 2 1
17 l:r.fS! i..c8 .:r.xc7 i.. x g3 22 hxg3 lllxb2 and White
 A r THE BOARD 27

has an active rook, but his pawns are variations can indeed be carried much
shattered. The position should be at further, but in this position there are
least equal for Black. numerous options for both players,
20 l:tc5! and this increases the complexity by a
The stunning point of White's com­ large factor.
bination. Shirov based his sacrifice largely on
20 gxf4 21 °iWh4 i.xc5 22 'il'xf6+
••• intuition. In this position, with Black's
cio>g8 (D) king seriously exposed, he must have
felt that there would very likely be a
way to play for a win. Moreover, at the
back of his mind there was the com­
forting knowledge that if in fact there
was no winning attempt, he had the
safety-net of perpetual check to fall
back on.
23 lt:Jxc5 'lixc5 24 e5!
A risk-free method of playing for a
win. This introduces the d3-bishop
into the attack, and possibly the c3-
knight as well, while all the time re­
taining the option of perpetual check.
The first critical moment. White 24 i.b7?!
••.

has no trouble forcing a draw, for ex­ Gelfand slips up and now gets into
ample by 23 'i!Vxh6, when Black has no serious trouble. The critical line was
reasonable way to prevent perpetual 24 . . . lt:Jd7 ! 25 'i!Vf5 and now:
check (23 . . . 'lid6 24 'lig5+ 'i\i'g6 25 1 ) 25 . . . l:t d8? 26 'i!Vh7+ clt f8 27
'ii'xc5 is very good for White). 'lixh6+ c;i;>e8 (27 . . . c;i;> e? 28 'i!Vh4+ c;i;>e8
Shirov now thought carefully about 29 lL'ie4 'i\i'xe5 30 l:t e l wins) 28 lL'ie4 !
the various options for playing on. 'lie? 29 lt:Jd6+ c;i;> e7 30 lL'i f5+ cio>e8 3 1
This raises the question as to why he e6 winning.
embarked on the combination if he 2) 25 . . . l:te8 ! 26 'i!Vh7+ c;i;>f8 27 i.e4
had no idea how he was going to con­ (once again, White has a draw by 27
tinue at this stage. The answer is that 'lixh6+ c;i;> e7 28 'i\i'g5+ c;i;> f8, but can
even a leading grandmaster such as play on without risk) 27 . . . lt:Jb6 (it is
Shirov is human. We are already six essential to cover d5 ; 27 . . . 1'b8 loses to
moves away from the start of the com­ 28 lL'id5 ! ) 28 'ili'xh6+ c;i;> e7 29 l:td l (29
bination, and it is not possible to cal­ 'ili'xb6 'ili'xb6 30 lt:J d5+ c;i;> e6 3 1 lt:J xb6
culate complex tactics accurately an l:t b8 gives White no advantage) and
indefinite distance ahead. Very forcing now:
28 SECRETS OF PRACTICAL CHESS

2a) 29 . . . :%.b8 30 'iff6+ catf8 3 1 l:dS ! When the tactics have to work
'ifc7 (3 1 . . . �xd5 32 �xd5 wins) 32
e6 ! :%.xe6 33 l: d8+ 'ii'xd8 34 'ifxd8+ If you initiate tactics which involve a
cat g7 35 'ii'g5+ <iti f8 36 'ii'xf4 with a large commitment and have no safety­
clear advantage for White. net, then you have no margin for error
2b) 29 . . . i.e6 ! 30 'iff6+ cat f8 3 1 h4 at all. Thus you have to be absolutely
and it is still possible for Black to go sure that your idea works, and it is
wrong, for example 3 1 . . . :%.ac8 32 h5 probably worth double-checking eve­
'ii'f2 33 h6 � d7 34 'ii'g 7+ cat e7 35 rything before you commit yourself.
'ii'g 5+ f6 36 exf6+ � xf6 37 �d5+
.t xd5 38 .t xd5 and White wins. How­
ever, after 3 1 . . . 'iff2 32 h5 l: ad8 ! 33
:%.xd8 'ife l + 34 �h2 'ii'g3+ the game
finally ends in perpetual check, al­
though oddly it is Black who gives it!
Although he could have reached a
draw, the task confronting Gelfand
was enormous. He had to defend very
accurately for a long time against an
opponent who has taking absolutely
no risk, since he always had a perpet­
ual check in reserve.
25 'ifxh6 rs 26 exf6 :%.f7 A. Yusupov L. Portisch
-

The only defence, as 26 . . . 'ifc7 27 Tunis Interzonal 1 985

'ii'g 5+ ! �h8 28 :%.el would be hope­
less. In this position Black has already
27 'ii'g6+ �f8 28 'ii' b6+ �g8 29 moved his king and so cannot castle
l:fl ! l:e8 30 'ii'g6+ �f8 31 'ifh6+ kingside. White has just attacked the
� g8 c-pawn with �b5 and 1 8 . . . i. d6 would
Now Shirov could have played 32 be a normal reply. Yusupov certainly
.t h7+ :%. xh7 33 'ii'g6+ � f8 34 'ifxh7 has some compensation for the pawn
.t xg2+ 35 � xg2 'ifc6+ 36 l: f3 'ifxf6 as Portisch's king is stuck in the cen­
37 �e4. After 37 . . . 'ii'g7+ 38 'ifxg7+ tre, but Black's pieces are fairly well
� xg7 39 l: xf4 �c4 40 �c5 ! Black placed and he is, at any rate, not worse.
would have been two pawns down in However, Portisch now decided to
the ending. play a combination.
Shirov chose a different continua­ 18 ... a6?
tion and won after further complica­ The idea is that if White takes on c7
tions. then his knight will be trapped.
 A r THE BOARD 29

19 .!LJxc7+! �d7
There is no way back; after 1 9 . . . �fl
20 .:r. n <iti g8 2 1 .!LJd4 Wes 22 .!LJcxe6
White wins a pawn and Black' s h8-
rook is blocked in.
20 .!LJxa6 l:la8
The immediate acceptance of the
piece is bad for Black after 20 . . . bxa6
2 1 Wxa6 l:l b8 (2 1 . . . l:l c8 22 'ii' b7+ .:r.c7
23 .:r. ad l + .t d6 24 .:r. xd6+ � xd6 2S
l:. d l + 'ii'd 5 26 'ili'b3 .:r.bs 27 l:l xd5+
exd5 28 'ii'c3 is very good for White)
22 .:r.ed l + and now: was a pawn up in the diagram posi­
1 ) 22 ... �c7 23 l:l ac l .tcS (23 . . . l:lb6 tion) and, just as important, he has
24 'iFa7+ .:r.b7 2S .:r. xc6+ � xc6 26 been prepared to expose his own king
.!LJd4+ <t;c7 27 l:l c l+) 24 .!LJd4 l:lb6 25 seriously. In view of the total lack of
1i'a4 .t xd4 26 'ili'a7+ l:.b7 27 'ii'xd4 any safety-net, Portisch should have
and wins. made absolutely sure that there was no
2) 22 ... .t d6 23 .:r. ac l .:r.hc8 24 l:lc3 hole in his calculations.
and there is no way out of the d-file 22 b4 !
pin. White wins by some combination Yusupov pinpoints the flaw in the
of .:r.cd3 and 'ii'a3 . whole concept. Black does not now
Portisch was certainly aware that win a piece and his position has been
taking the knight with the b-pawn irreparably damaged.
would expose his king too much, and 22...l:lxa6
the text-move was the point of his After 22 . . . bxa6 White wins by 23
idea. He aims to take the knight with a :ac t <t;b7 24 .:r.xc6 � xc6 2S .!LJd4+.
piece, leaving the b7-pawn to defend 23 b5 .:r.a3 24 bxc6 b6
his knight and provide his king with Trying to keep some lines closed.
adequate defence. 24 . . . bxc6 loses to 2S l:l ac l i.. c S 26
21 l:ledl+ <t;cs (D) .!LJd4 'fi'eS 27 'ili'c4 ! and the attack
The critical moment. Portisch has breaks through.
staked the game on his ability to take 25 .:r.abl i.. c5
the knight with a piece. If he fails, then Black's position is hopeless owing
he will inevitably lose because the con­ to his fatally exposed king. The finish
cessions he has made in order to trap was 26 .:r.b3 l:la5 27 'fi'c4 l:lxa2 28
the knight are so large. He has been l:.xb6 .txe3+ 29 �bl .txb6 (or
prepared to go a pawn down (two 29 . . . 'fi'd5 30 l:. b8+ � xb8 3 1 'iF b4+
pawns have been sacrificed, but he �c7 32 1i'b7+ <t;d6 33 1i'd7+ �cs 34
30 SECRETS OF PRACTICAL CHESS

:t xd5+ exd5 35 'ike7+ � xc6 36 'ifxe3) Attack, but the same general principle
30 'ii'xa2 �c7 31 l:r.d7+ �xc6 32 applies in many openings. For exam­
'ifa4+ 'ii?c5 33 lL!d2 1-0 ple, it is not unusual for a player to
seize a long-term strategic advantage
I mplicit commitments in return for piece activity or a lead in
development. In this case the player
Almost every move in chess involves with the better development has taken
some sort of commitment. A pawn on an implicit commitment to under­
move cannot be reversed and with each take rapid action. Advantages such as
advance the pawn permanently loses a lead in development are inherently
the ability to control certain squares. temporary, because when the oppo­
Even a piece move is a commitment; if nent has brought out all his pieces the
the piece turns out to be badly placed advantage disappears. A common mis­
on its new square, it may have to go take is to take on such a position, but
into reverse, with consequent loss of not to appreciate that the long-term
time. However, the degree of commit­ chances lie with the opponent. The re­
ment is important. A piece sacrifice sult is a fatal lack of urgency. Here is
involves a greater degree of commit­ an example.
ment than a natural developing move.
We have already discussed the more J. Nunn J. Mellado
-

obvious types of commitment earlier. Leon 1 997

However, there is a more subtle type French Defence
of commitment, which we call im­
plicit commitment. Very often, a cer­ 1 e4 e6 2 d4 dS 3 eS c5 4 c3 'ii'b 6 5
tain type of commitment is bound up ll:if3 .td7 6 .te2 .tbs 7 c4 i.xc4 8
in a player's choice of opening. A i.xc4 dxc4
player as White adopting the Velimi­ I had noticed in my database a
rovic Attack in the Sicilian (one of the game won by my opponent with this
main lines of which runs 1 e4 c5 2 lL!f3 rather offbeat line (instead of the usual
lLic6 3 d4 cxd4 4 lLi xd4 lLif6 5 lLJc3 d6 8 . 'ifb4+). Players are often unduly
..

6 .t c4 e6 7 .t e3 i. e7 8 'ii'e2 a6 9 0-0-0 affected by the result of a game. If a

'ifc7 10 i. b3 0-0 1 1 g4) is committing player has won a game with a particu­
himself to an all-out kingside attack, lar line, he will very often repeat ex­
which may involve sacrifices, and if actly the same line, even if it is rather
he is hesitant about giving up material dubious. After looking at 8 . . dxc4 be­
.

then he has simply chosen the wrong fore the game, I quite liked the result­
opening. Of course, this is an extreme ing positions for White, so I decided to
example - there are few opening varia­ go down the same variation.
tions as one-sided as the Velimirovic 9 d5 lLie7
 A T THE BOARD 31

Black does not completely equalize trouble feeding across to the other side
after 9 . . exd5 IO 'ii'xd5 .!LJe7 1 1 'ii'xc4.
. of the board. In the endgame the pawn
10 d6 is more likely to promote, as there are
This is the critical continuation. In fewer pieces available to keep it under
Mellado's earlier game White had control. Black would probably have to
played IO .!LJc3, which allows Black to assign one piece to keep a watchful
swap everything on d5, with a more or eye on the pawn, leaving him a man
less equal position. short elsewhere. The only caveat here
10... .!LJec6 1 1 0-0 (D) is that White should not exchange too
many pieces in an endgame, since then
Black could use his king as the block­
ading piece.
In return for White' s long-term as­
set, Black's minor pieces can converge
to attack the e5-pawn and for some
time White will be tied down to its de­
fence. Moreover, Black's development
is quite good, especially as White will
have to invest some time in regaining
the pawn on c4. However, these com­
pensating factors are all temporary,
and given time White will complete
This is a typical case of implicit his development and reorganize his
commitment. The extra pawn is not forces to defend e5 efficiently. Black
relevant, as White will win back the has therefore taken on a heavy implicit
c4-pawn in a few moves. The key fac­ commitment either to keep White off­
tor is that Black has allowed White to balance or to convert his temporary
drive a wedge into the middle of his advantages into something more per­
position. The protected passed pawn is manent before White puts his house in
not going to promote in the near future order. If Black has not achieved some­
- there are simply too many pieces on thing concrete within the next half­
the board for that - but it is an asset dozen moves, then we can safely pre­
both in the middlegame and in the dict that he will be in trouble.
endgame. In the middlegame it stifles White, on the other hand, can con­
the activity of Black's pieces and cuts tent himself with more modest play.
communication between the queen­ All he need do is consolidate his po­
side and kingside. This could help sition and bring all his pieces into
White to mount a kingside attack, for play and Black will 'automatically' be
Black's queenside pieces would have worse. Of course, this might have been
32 SECRETS OF PRACTICAL CHESS

quite difficult if Black had continued the other must play to stir up trouble
accurately, but nevertheless White's quickly.
position is easier to play. He has a very 11 llld7 12 l:r.el g6 13 llla3
••.

clear-cut aim, while Black must try to The first task is to regain the pawn
generate counterplay ' somehow' , i.e. on c4.
he has to formulate a plan from scratch, 13. i. g7
..

whereas White's plan is handed to him Black cannot hang on to the pawn
on a plate. This is a typical situation in by 1 3 . . . 'ii'a6 14 •e2 b5 because 15 b3
such 'long-term advantage vs piece breaks up the queenside pawn struc­
play' situations. The piece-play side ture and regains the pawn in a more
has much more work to do, at least to advantageous way.
begin with, to find the best plan. If he 14 lllxc4 'ii'a6 15 'ii'e2 (D)
succeeds, then the other side may also
have to think carefully about how to
'
contain his oppont'.nt:s counterplay,
but if he does not succeed the game
can be quite easy.
To my surprise, Black spent very
little time over the next few moves and
only started to think when he was al­
ready in considerable difficulties. Evi­
dently he did not appreciate that this is
the critical phase of the game and ordi­
nary developing moves will not be
enough.
From all this we can learn an im­ 15 ... 0-0?
portant lesson. At the end of the open­ A natural move, but one which al­
ing, spend a few minutes deciding lows White to consolidate his central
whether one player has the better pawns. Black should have continued
long-term chances. This may be obvi­ 15 . . . b5 16 lll a3 (relatively best, as
ous simply from the opening variation White must not block in his bishop)
chosen (for example, in the Exchange 1 6 . . . l:r. b8 (after 16 . . . lll cxe5 17 lll xe5
Ruy Lopez it is obviously White who i. xe5 1 8 ll:\ xb5 0-0 1 9 a4 White has
has the better long-term chances), but some advantage; Black cannot take on
if it is at all unclear then it is worth d6 because of the d-file skewer, so
spending time on this question. The White manages to support his danger­
strategy for the two players may then ous passed pawn) 1 7 i. f4 0-0 and
be quite clear: one side must aim to White has more difficulties than in the
consolidate and contain his opponent; game because his knight is offside at
 A T THE BOARD 33

a 3. After 18 l:. ad l , for instance, Black covering the queening square, but we
c an continue 1 8 . . . c4 and try to plant a can already see that Black is running
knight on d3 (for which White will out of ideas.
probably have to give up the ex­ 19 ..e4
change). A typical move from the player
16 .i.f4 lLlb6 who has time on his side. It does not
If now 1 6 . . . b5, then White replies create any immediate threat, but tidies
1 7 lLlcd2; thanks to Black's inaccurate up White's position by unpinning the
15th move, the knight can retreat to c4-knight. Black also has the worry
the centre rather than to a3 . that the white queen might now switch
17 b3 to the kingside and form the basis of a
White must not allow the exchange direct attack.
of queens. His own queen is well 19 l:r.ac8 20 h4
•••

placed in the centre of the board, while Both relieving White's back rank
Black's is marooned on the queenside, (his king would .be very safe on h2)
so such an exchange would clearly fa­ and making another threatening ges­
vour Black. ture against Black's king.
17... lLldS 20 bS 21 lLle3 (D)
•••

The transfer of the knight to d5 has White aims to open the d-file by re­
blocked the d-file and prevented White moving Black's best-placed piece, the
backing up his pawn with a rook at d l , blockading knight on d5.
but i t has also relieved the pressure on
the e5-pawn.
18 J.g3
Now the defects of Black's earlier
play start to become apparent. Cer­
tainly he has developed his pieces, but
he has done nothing to prevent White
doing the same. All other things being
equal, the position will favour White
because of his superior structure. More
and more, Black struggles to find con­
structive moves. He can only under­
mine White' s pawns by . . . f6, but then
White would swap on f6, leaving the 21 lLlcb4?
.••

e6-pawn hopelessly weak. The correct plan is indeed to use the

18... h6 second knight to restore the d-file
Black wants to prevent White sup­ blockade, but this move is a tactical er­
porting his d-pawn by a later J. h4, ror which leads to the loss of material .
34 SECRETS OF PRACTICAL CHESS

2 1 . . . � xe3 22 'ii'xe3 �b4 was the right 29 ... :%.b2 30 �d4 :%.d2 31 �ab3
way to put the plan into action. After :%.b2 32 'i!Yxb4 'ii'a6 33 'ii'c3 :%.a2 34
23 :%. ac l �d5 24 "ii'e2 c4 25 bxc4 bxc4 �cs 'ii'c8 35 :%.cl 'ii'a8 36 'ii' b 3 :%.d2
26 �d2 'iWxa2 27 :%. xc4, for example, 37 'iWb4 :%.a2 38 'ilb7 :d2 39 'iWxa8
Black has avoided any immediate dis­ :%.xa8 40 �cb3 1-0
aster, but White's long-term advan­
tages remain intact. As the pieces are Positional thinking
exchanged, the d-pawn poses a more
and more serious danger, while Black's Although precise analysis plays some
bishop on g7 is practically dead. part in most chess situations, posi­
22 a3! �c3 23 axb4 'iWxal 24 'ii'c2 tional thinking is just as important.
(D) Even in sharp situations, positional
Winning two knights for a rook and factors can still play a major part; for
a pawn. This might not be decisive ex­ example, there is little point in playing
cept for the fact that the structure of a combination winning the exchange
the position very much favours minor if the result is a position riddled with
pieces over rooks. weaknesses (see Sax-Stean on page
105).
Unfortunately, if you have little
natural talent for it, positional think­
ing is one of the most difficult aspects
of chess to learn. One often hears of
how a certain player has a 'natural
feel' for the pieces and instinctively
puts them on the right squares. This is
either an innate ability or something
which is learnt by experience. If you
have played over and studied thou­
sands of games, and seen all the differ­
ent patterns and plans which can arise
24 'iWa3 25 'ibc3 cxb4 26 'iWd3
... from the main openings, then you al­
:%.c3 27 'iWxbS :%.xb3 28 �c4 'ifa2 29 ready have a good start when it comes
�aS to positional play.
The most convincing win. White Most players do not have the time
positions his knights both to prevent for this sort of study and must focus
the advance of the b-pawn and to move their efforts where they will do the
to c6 or b7, assisting the advance of most good. This really means con­
the d-pawn. 29 �fd2 :%.d3 30 'iWxb4 a5 centrating on the types of middlegame
would be less clear. position which can arise from your
 A r THE BOARD 35

opening repertoire. Most openings Accepted, Nimzo-Indian, etc. If such

give rise to certain distinctive central pawn structures arise in your reper­
pawn structures. Some pawn structures toire, it pays to look not only at games
are particular to just one opening. with 'your' opening, but at others
Look, for example, at the following which give rise to identical central
diagram: pawn structures. If you have a data­
base program with a 'position search'
facility, then you can use this to find
which openings result in a particular
pawn structure.
Once you have identified the main
pawn structures relevant to your open­
ings, then try to find games resulting
in these structures. Once again, a data­
base program makes this easy. If there
are too many games to play over, then
restrict yourself to games by grand­
masters - there is no point playing
over games in which the players adopt
You don't need to see the position quite inappropriate plans. It is much
of the pieces to know that it's almost better if the games are annotated, pref­
certainly a Winawer French. On the erably with words rather than sym­
other hand, look at the following dia­ bols. Languageless annotations are
gram : fine for displaying tactical analysis,
but when it comes to explaining the
plans for both sides, there is no substi­
tute for a clear explanation in words
by someone who really understands
the opening . Opening books which
emphasize general ideas and plans
may also be of help.
For each pawn structure, try to play
over about fifty games. This should be
enough to give you an idea both of the
typical plans adopted by both players,
and of which plans tend to succeed.
The idea is not really to learn what to
It could be virtually anything: Caro­ play in specific positions, but to see
Kann, 2 c3 Sicilian, Queen's Gambit what the two players are trying to do,
36 SECRETS OF PRACTICAL CHESS

and how each side tries to frustrate the point, which comes into play later in
other's plans . By looking at all these the game, is that White's al -rook may
games in quick succession, you will emerge via a3 .
notice connections and similarities be­ Of course, this slightly unusual idea
tween them which would not be ap­ is hardly enough to win the game by it­
parent if you looked at them over a self, but it puts Onishchuk off his
period of months. When these posi­ stride. Instead of the normal patterns
tions arise in your games, you will in this system, he has to work out a
then have a much better idea of what plan for himself. As we shall see, Ad­
you should be trying to do. ams has a much better grasp of the re­
Even grandmasters depend on this quirements of the position.
type of knowledge, and if they are 7... i..e6 8 :%.el i..xc4 9 dxc4!
thrown into unfamiliar territory, then Once again taking the chance to
they immediately start to play much steer the game along original paths.
less strongly. The following game is The mechanical reply would be 9 .!LJxc4,
an illustration of this, but it also dem­ when the knight could later drop back
onstrates a second valuable lesson: the to e3 . However, Adams chooses a dif­
importance of sticking consistently to ferent move, which greatly strength­
a plan. ens his grip on d5 . He has already
formulated his plan for the next phase
M. Adams A. Onishchuk
- of the game: to increase his control of
Tilburg 1997 the light squares as much as possible,
Two Knights Defence with particular reference to d5 and f5.
Rather unusually, this is the only pl�
1 e4 es 2 .!LJf3 .!LJc6 3 .tc4 .!LJf6 4 d3 he needs to win the game ! The first
.te7 5 0-0 0-0 6 a4 d6 7 .!LJbd2 step is to transfer his knight from d2 to
Adams is playing this quiet system e3 .
in a slightly unusual way. White nor­ 9 . l:e8 10 00
..

mally plays c3 at some stage, so as to Preparing .!LJe3, taking aim at both

preserve his c4-bishop from exchange critical squares.
(he can meet . .!LJa5 by i.. b5, and then
.. 10.. i.. fS?
.

drop it back to a4 and c2 if Black A definite error. One of White's

chases it with his pawns). Adams has problems is that the immediate .!LJe3 is
countered the threat of ... .!LJa5 in a dif­ impossible because the e4-pawn would
ferent way, by playing a4. This not be hanging. Thus White would have to
only enables him to maintain the spend some time preparing .!LJe3. By
bishop on the a2-g8 diagonal, but may allowing White to play i.. g 5 and
also form the basis of a space-gaining .t xf6, Black not only lets White solve
operation on the queenside. A third the problem of developing his bishop,
 A T THE BOARD 37

but also allows .!LJe3 without loss of up with a concrete line in which a5
time. Already one can see the outlines turned out to be a useful move, but he
of a good knight vs bad bishop position must look to the long term. Given that
shap ing up. he is aiming for light-squared control,
Black should have played 10 . . . h6. the possibility of a6, forcing . . . b6,
He need not worry about l 1 .!LJ g3 .i.f8 would be helpful. Black can, of course,
12 lLJrs, because the knight can be ex­ prevent this by playing . . . a6 himself,
pelled by 12 .. .� h7 followed by . . . g6. but this would not be without possible
1 1 .i.gS h6 defects. White will at some stage play
Having played ....i.f8, it would have lLJd5, and Black would like to be able
been psychologically very difficult to to expel the knight with . . . c6. If he has
return to e7, although that might have already played . . . a6, then the knight
been the best choice. The move played can settle on b6.
ultimately turns out to weaken Black's Thus Black is presented with a
kingside. choice of evils, which never makes for
12 .i.xf6 ..Wxf6 (D) an easy decision.
14... .!LJe7 15 :a3
Another useful move. Once again,
White is looking forward to the day
when Black will have to play . . . c6,
leaving his d6-pawn weak. Now the
rook can come to d3, stepping up the
pressure against d6 without loss of
time. The firepower could even be in­
creased by l:r. el -e2-d2. Note that when
one is lining up the heavy artillery
along an open or half-open file, it is
usually better to have the rooks in
front of the queen. Here White can
13 .!LJe3 'ii'e6 achieve this ideal formation very effi­
White has now completed the first ciently.
step of his plan. Both these ideas (gaining a queen­
14 a5 side space advantage and developing
In this static position, it is possible the rook via a3) were already inherent
to create plans stretching over a rela­ in White's 6th move, so Adams's play
tively long time, since the pawn struc­ has been very consistent.
ture is likely to remain unchanged. 15... g6? !
The text is a typical positional move. Black would like to improve the po­
White would be hard-pressed to come sition of his bishop, but he is being
38 SECRETS OF PRACTICAL CHESS

unrealistically optimistic. He will al­ 1 9 l:r.e2 b6 20 axb6 axb6 2 1 hxg6 fxg6

most certainly have to play . . . c6 at 22 l:r.ed2 ll:i c8 23 b4, followed by b5,
some stage, and then the bishop will gives White control of the central d5-
be needed on f8 to protect the pawn on square.
d6. If Black could force through . . . f5 18 a6
he would gain some counterplay and Forcing a further slight weakening
thereby justify . . . g6, but against Ad­ of the light squares.
ams's accurate play it proves impossi­ 18 b6 19 ll:idS l:[d7 20 hxg6 fxg6
.••

ble to realize this plan. 21 lLih4!

It would have been better to play Planning l:r. g3, when the weakness
1 5 . . . c6 at once; after 1 6 l:r. d3 l:r. ad8 1 7 of g6 will force . . . g5, giving White the
l:e2 ll:ig6 everything would be de­ f5-square. There is little Black can do
fended and White would have to find a to counteract this plan.
further 'mini-plan' to improve his po­ 21 ... c6 22 ll:ixe7+ l:r.exe7 23 l:r.g3 gS
sition. 24 ll:ifS (D)
16 h4!
An excellent and unexpected move.
Unlike the advance of the a-pawn, the
aim is not so much to gain space as to
weaken Black's kingside pawns. If he
replies 1 6 . . . hS, then 1 7 lLi gS 'ii'f6 1 8
ll:id5 ll:ixd5 1 9 cxd5 leaves White with
the advantage. Black cannot expel the
knight by . . .f6 because it would just
hop into e6; the best he could do would
be to exchange it off with . . . i. h6, but
White would retain the advantage be­
cause of his queenside space advan­
tage. An ideal position for White. Six of
16 ...i.g7 Black's seven pawns are positioned on
Black ignores the h-pawn, but after dark squares, and the knight occupies
h5 and hxg6 White has achieved two a beautiful outpost; in addition the
notable objectives: preventing Black's pawn on d6 is weak. Black is lost.
possible counterplay by . . . f5, and cre­ 24 .ltf7 25 l:r.d3 i.f8 26 b3 dS
••

ating a new weak pawn at g6. This desperate bid for counterplay
17 h5 l:ad8 costs material, but in any case White
Again a difficult choice. 17 . . gxh5. could step up the pressure easily
1 8 ll:ih4 is clearly bad, and now it is enough, for example by l:.e2-d2.
rather late for 17 . . . c6, as 1 8 l:d3 l: ad8 27 'ii'g4 'ito>h7
 A T THE BOARD 39

27 . . . dxe4 28 .!LJ xh6+ wins the ex­ wrong pieces; committing yourself to
change. weakening pawn advances.
28 cxd5 cxd5 29 .!LJxh6! 'ifxg4 2) Make sure your plan is realistic.
Or 29 . 'ifxh6 30 l:t h3 winning the
. . There is no point in embarking on a
qu een. five-move plan if your opponent can
30 .!LJxg4 .i.c5 31 l:te2 1-0 wait for the first four moves, and then
Black will be two pawns down for stop your plan by playing one move
nothing. himself.
3) Make sure your plan is not tacti­
There comes a point in the vast ma­ cally flawed. Even if what you are
jority of games when your acquired aiming for is worthwhile, this will not
knowledge will be exhausted and you help if your opponent can mate you
will have to rely on your own re­ while you are executing it.
sources. This point normally arises in
the early middlegame. The next step is Having decided on a particular
to formulate a plan. You may already plan, you have to strike a balance be­
have some idea of the options avail­ tween consistency and flexibility. On
able from your opening study; this the one hand, pointless changes of plan
will help you to narrow your search. In are very damaging. If you play two
any case, it is time for some serious moves of one plan, then three moves
thought; the plan you choose now will of another, then four moves of a third,
go a long way towards defining the you will probably be back more or less
whole shape of the struggle to come. where you started ! The ideal situation
In some positions, for example, those occurs when you formulate a plan and,
with a blocked centre, it may be ap­ while you are executing it, your oppo­
propriate to construct a long-term plan nent does little either to prevent it or
which may require ten or twenty moves to develop counterplay. Then you can
to execute. More likely, your plan will produce an elegant positional game
be much shorter-range, lasting per­ which is dominated by a single strate­
haps five moves. This applies particu­ gic thread. However, games like this
larly in relatively open positions. normally only arise between players
Here are some tips on making a of very different strengths (the Adams­
plan: Onishchuk game above is an excep­
l ) Make sure your plan is benefi­ tion in this respect) . More often your
cial. There is no point aiming for a tar­ plan will be blown off course because
get which does not actually enhance your opponent will interfere with it in
your position. Typical misguided plans some way. While consistency is a vir­
are: attacking on the wrong part of the tue, sometimes you have to be as prag­
board; aiming for the exchange of the matic as a politician in changing your
40 SECRETS OF PRACTICAL CHESS

plan. If your opponent has blocked 1 1 bS?!

..•

Plan A, but at the cost of creating a Weakening the queenside like this
weakness elsewhere on the board, it appears rather dubious. 1 l . . . l:.e8 is a
would be foolish to stick to your origi­ more solid alternative, which has been
nal intention, ignoring the new situa­ played in a number of games .
tion. Instead you should rethink your 12 lL!e3 'iid7 (D)
strategy. Most games are like this: the
players formulate a series of mini­
plans and strike a balance between
forwarding their own plans and inter­
fering with those of the opponent.
Eventually the balance is upset, either
by one player's plan succeeding deci­
sively, or by the game dissolving into
tactics.

J. Nunn - V. Hort
Wijk aan Zee 1 982
Giuoco Piano
The game has now deviated from
1 e4 eS 2 .tc4 lL!c6 3 lLif3 .tcS 4 c3 known theory, so it is time for White to
lLif6 5 d3 a6 6 0-0 d6 7 lL!bd2 0-0 8 think about his long-term plan and im­
.tb3 .ta7 9 :et .te6 10 lL!c4 h6 1 1 mediate 'mini-plan' . Since the posi­
a4 tion is still fairly fluid, it is difficult to
White has adopted a quiet opening create any long-term plans; trying to
system which avoids an early d4 in the anticipate what will happen more than
hope of playing it at a more effective a few moves ahead is fruitless, as too
moment later. While this system is many things might happen to throw a
very solid, it has the defect that if plan off course. However, the mini­
Black plays in equally solid style, the plan is much easier; as mentioned ear­
game may simply peter out to a draw. lier, one of White's ideas in this open­
White's last move aims for a5, gain­ ing is to play d4 in the middlegame,
ing space on the queenside and await­ when it may be more awkward for
ing a more favourable moment for the Black. White cannot play d4 immedi­
thematic move lLie3 (aiming at the ately, as Black could take on b3 and
squares d5 and f5). Readers will al­ then on d4, so 1 3 .t c2 is indicated.
ready be familiar with this type of Then d4 will be a real inconvenience,
strategy from the Adams-Onishchuk as Black must meet the threatened
game given above. fork by d4-d5.
 A T THE BOARD 41

13 J.c2 l:fe8 14 d4 exd4 17 h3!

A major concession by Black, giv­ The start of an alternative plan for
in g up his strong-point in the centre. supporting e4, based on �fl -g3 . White
1S �xd4 first of all rules out the possibility of
White could perhaps have played . . . J. g4, which might be annoying once
15 cxd4, although it is more compli­ the e3-knight has moved away.
cated after either 1S . . . d5 1 6 eS �e4 17 ..."ilc7 18 00 J.d7 (D)
(when 17 �xdS � xf2 18 �f6+ gxf6
19 catxf2 fails to 19 ... �xd4) or 15 ... �b4
16 J. b l c5 ( 1 6 . . . dS 1 7 e5 �e4 1 8
�xdS � xf2 1 9 �f6+ gxf6 2 0 cat xf2
fxe5 is unclear) 17 d5 J. g4 18 �xg4
�xg4 19 h3 �5 20 �xe5 l: xe5 .
15. . �xd4?!
.

After this White's advantage is safe.

Black would have preferred not to
make this exchange, but he was proba­
bly worried about the pressure on bS.
However, he could have made a bid for
counterplay by 1 S . . . b4 !?.
16 cxd4 c6 Now e4 is defended, and at first
The first mini-plan is over and has sight a natural continuation would be
achieved a definite success, but now it 19 �g3, further covering the central
is time for the next. White has a per­ pawn and making a move towards
manent advantage in his greater cen­ Black's king. However, one should al­
tral control, if only he can maintain his ways take into account the opponent's
'two abreast' centre intact. Thus his intentions and see if there is a possi­
next mini-plan should consist of con­ bility to frustrate them. Black's recent
solidation. The immediate problem is moves, removing the bishop from e6,
the e4-pawn; Black need only move suggest that he intends to play . . . cS .
his e6-bishop to attack it awkwardly. After 1 9 � g3 c5 20 dS c4, for exam­
White cannot defend it by 'ii'f3, be­ ple, Black activates his queenside ma­
cause the queen must defend the d­ jority while White, thanks to the bishop
pawn. One solution is to play f3 , but on a7, has trouble getting his own cen­
the weakening on the a7-gl diagonal tral majority moving.
is ugly, especially in view of the lurk­ 19 b4!
ing bishop on a7 - Black would have A very ambitious move, which is
good chances for counterplay by only justified because White's central
means of a timely . . . c5. structure is already secure. Now if
42 SECRETS OF PRACTICAL CHESS

Black ever plays . . . c5, White will re­ while all the white minor pieces are
ply bxc5 and d5, gaining a massive well placed to create threats on that
central majority. Moreover, the c l ­ side of the board. I decided simply to
bishop can now move to b2, setting up move my pieces towards Black's king
further latent threats against Black's by .!LJg3 and 'iif3 . At this point White
kingside. In a sense, this move falls could think about a breakthrough by
under the heading of 'consolidation' , d5 or e5, or simply play .!LJf5, gaining
because it i s mainly directed against the two bishops.
the possibility of Black breaking up 22 .!LJg3 :ae8 23 'ii'f3 .!LJh7
White's centre by . . . c5. Black could have tried the tricky
19 bxa4
••• 23 . . . a3 , hoping for 24 :xa3 .!LJd5 !, but
Black now threatens to attack the White would just play 24 .!LJf5 .
b4-pawn along the half-open file, and The text-move takes the sting out of
White's first priority is to defend it. d5 or e5, so White goes for .!LJf5.
The pawn on a4 can be recaptured 24 .!LJfS i.xfS 25 'ii'xfS (D)
later.
20 i.b2 'iib 6 21 i.c3 lte7
After 2 1 . ..c5 White would ignore
the hanging piece since 22 dxc5 dxc5
23 i. xf6 'iixf6 24 'iixd7 cxb4 is rather
messy. Instead he would continue with
the thematic reply 22 bxc5 dxc5 23 d5,
when the possible counterplay against
f2 is not a serious worry, e.g. 23 ... c4 24
'iid2 (24 :e2? .!LJxe4 ! draws) 24 ... :ab8
25 :e2, and White is ready for e5 .
After the text-move, White must
create his next mini-plan. The centre
is secured and Black cannot attack it Black's next move anticipates e5 by
with . . . c5 or . . . d5 , the latter because White.
the reply e5 would give White an auto­ 25 ... .!LJfS
matic kingside attack. White now has the two bishops to
Black evidently intends to restrain show for his efforts, but the knight on
White's centre by doubling rooks on f8 is an effective defender of the king­
the e-file, which leaves open the ques- side and direct attacking attempts do
. tion as to how White should make not appear promising. White must
progress. My eye was naturally at­ take care, or else Black's latent pres­
tracted to Black's kingside; there is sure along the lines a7-gl and e8-e l
only a single knight defending it, might be unleashed.
 A T THE BOARD 43

I therefore decided to switch to the

queenside, and build up against the
a6-pawn. Note that this plan would not
have been effective before as Black
could simply have played . . . i.. c 8.
26 l:lxa4 .!l:ie6 27 l:ldl .!l:ic7? !
Black decides to pre-empt White's
attack on the a-pawn and hopes to oc­
cupy a good square at b5, but now the
knight is dangerously far away from
the kingside. It would have been better
to leave the knight on e6, ready to move
to c7 or f8 as necessary. After 30 . . . 'ii'g 6 3 1 'ii'c3 i.. b 6 32
28 i.. b 2 l:.aal Black would not have a good an­
The immediate 28 e5 g6 is ineffec­ swer to the threats of 33 'ii'xc6 and 33
tive as 29 'ii'f4 is met by 29 . . . .!i::i d 5. e5.
Now, however, 29 e5 is a threat. Note 31 e5
how White has again changed his plan Now White has more advantages to
to utilize the circumstance that Black's point to: the e-pawn is no longer under
knight has gone to the queenside. pressure, and White can create threats
28 ... 'ii'bS along the bl -h7 diagonal. The imme­
Black does not want to be mated diate threat is 32 i.. c l 'ii'h4 33 g3
and, now that a6 is defended by the 'ii'xh3 34 i.. f5 .
knight, he switches his queen to the 31 .!l:ie6?
••.

king side. Alarmed by the growing danger to

29 'iFf3 'ii'g5 his king, Hort, who was in severe
While in positional terms Black's time-trouble, decides to jettison his a­
plan is quite reasonable, it has a con­ pawn in order to bring the knight over
crete defect: the queen is short of to the kingside. In many cases, if your
squares. pieces are tied down to defending a
30 .:r.as! (D) weak pawn, then letting the pawn go
Now Black is obliged to make a fur­ is a way to activate the pieces and de­
ther concession. Note that the greedy velop counterplay (see page 1 36). In
30 'ii'c 3 d5 3 1 'ii'xc6 dxe4 32 i.. xe4 this case, however, it makes little dif­
would be disastrous for White, since ference; the position remains more or
32 . . . l:l xe4 33 'ii'xc7 i.. b 8, followed by less the same except that Black is a
. . . .:le2, would give Black a crushing pawn down.
attack. 3 l . . .g6 would have been a more re­
30 dS
••• silient defence; after 32 g3 .:r.e6 33
44 SECRETS OF PRACTICAL CHESS

'ii'd 3 'iihS 34 <iti g2 l:.b8 35 i.c3, threat­ making one threat after another until
ening f4-f5, White would be much Black collapses.
better but not yet winning. 4S l:.f4 �h8 46 Wf3 i.f8 47 l:.f6
32 l:.xa6 l:.c8 33 g3 l:.b7 34 'iid3 g6 l:.bc7 4S :a2
3S i.c3 i.b6 After l:. af2 followed by e6 there
White has taken the pawn and met will be no defence.
Black's attack on the b-pawn. Now he 48 ... 'ii h7 49 e6 i.g7 SO l:.xf7 l:.xf7
must turn to a mini-plan for exploiting Sl exf7 'iig6 S2 :f2 i.f8 S3 i.d2
his advantage. The natural plan is the <ith7 S4 bS cxbS SS 'ii'xdS 'iid3 S6
one which has been in the position <itb3 i.g7 S7 i.b4 1-0
ever since White played e4-e5, viz. the
advance of the f-pawn. The method of compa rison
36 �h2 i.d8 (D)
Sometimes it is possible to avoid pre­
cisely evaluating the consequences of
the various options available. The rea­
son is that you are really interested in
relative assessments and not absolute
ones. If, for example, you know that
move A is '0.2 pawns' better than
move B, then you would prefer A to B
- it doesn't matter whether A is 0.3
pawns better for you or 0.1 pawns
worse for you, it is only the relative
evaluation which matters.
In practice, of course, this type of
31 .:.n logic is hard to apply if there are many
Ultra-cautious in time-trouble. 37 different continuations. If you have al­
f4 lll xf4 38 gxf4 'ii xf4+ is of course ternatives A, B, C, D and E then com­
unsound after 39 � h l , but I wanted to paring A with B, C with E, D with A
make really sure there were no com­ and so on is going to lead to a logical
plications ! conundrum rather than a decision. It is
37 'iie7 38 'iWe3 'ii'f8 39 f4 ll:\g7 40
••• much simpler to make an absolute as­
fS i.gS 41 'iie2 ll:\xfS 42 i.xfS gxfS sessment that A is a bit better for
43 h4 i.e7 44 :xrs 'iig7 White, B is equal, and so on, and then
Black is quite lost. In addition to the pick the move with highest value.
minus pawn, three of his four remain­ It follows that the method of com­
ing pawns are weak and his king is ex­ parison, as we call it, works best when
posed. The rest is easy and consists of there are very few options (three at
 A T THE BOARD 45

most) and the types of position result­ option of playing . . . 'ii'h4+ first. The
ing from these options are rather simi­ only difference between these two lines
lar, and so are more easily compared is that in one White's g-pawn is on g2
with each other than evaluated abso­ and in the other on g3 . There is no
lutely. question that the additional move g3
The most common case arises when helps Black. If White exchanges his
you have the possibility of an interme­ b5-bishop on c6, then he will have Jess
diary check which forces some kind of control of the light squares in any
concession. case; the weaknesses created by play­
ing g3 would then be quite serious. It
A. Meszaros Y. Zimmerman
- may well happen that White will
Balatonbereny 1 994 change his mind and play the bishop
Nimzowitsch-Larsen Attack back from b5, but in this case Black
has gained time.
1 b3 e5 2 .tb2 lL!c6 3 e3 d5 4 .tb5 5 ... 'ii h4+ 6 g3 �e7 7 fxe5 .txe5 8
.td6 5 f4 (D) lL!c3 lL!f6
8 d4 ! ? 9 lL!d5 'i\i'c5 is an interest­
. . .

ing alternative.
9 lLif3 .tg4 10 .te2 .td6 1 1 lL!b5
0-0 12 lL!xd6 'iixd6
White has the two bishops, but
Black's development is very comfort­
able. The position is equal.
13 lLid4 lL!xd4 14 .txd4 .th3
Black starts to take advantage of the
weakened light squares on the king­
side.
15 .tfl .td7 16 .tg2 c5 17 .txf6?
Too ambitious. White plays to win
This is a known theoretical posi­ a pawn, but his lack of development
tion. White's opening is designed to and weakened kingside make it a risky
exert pressure on the e5-square. Black endeavour. 17 .t b2 would have main­
cannot play 5 . . .exf4 owing to 6 .t xg7, tained the balance.
and defending the pawn with 5 . . .f6 is 17 ... 'iixf6 18 .txd5 .tg4 ! 19 'ii'c l
rather ugly - after 6 �h5+ g6 7 'ili' h4 l:tad8 20 .txb7 (D)
Black's kingside is weakened and his 20 ... l:txd2! 21 <it>xd2?
development is impeded. Now Black gains a clear advantage.
Therefore Black should defend the White could still have held on by 2 1
e5-pawn by . . . 'i\i'e7, but he has the �xd2 'ili'xal + 2 2 <it> f2 �f6+ 2 3 <li> g2
46 SECRETS OF PRACTICAL CHESS

2 7 . . . 'ik b2 2 8 l:. e l 'ii'x a2 29 'iii>h2 J. f3

30 :hgl i. xg2 3 1 :xg2 'ii'xb3, when
Black has the advantage, but White
might be able to draw since Black's
queenside pawns are disconnected.
24 J.h3+ 25 'it> gl 'ikxc2
•••

A vivid exploitation of the weak­

ness induced by 5 . . . 'ikh4+.
26 'ii'c l 0-1
White resigned without waiting for
26 . . . :g2+.

l:.d8 24 'ii'a5 ! (24 'ii'c l 'ii'c 3 !) 24 . . . 'ii'e7 Whilst occasions in which one can
(after 24 ...'ii'b6 25 'ii'xb6 l:.d2+ 26 'ito>fl consciously use the method of com­
axb6 Black will regain his pawn, but parison are not all that common, in
in the meantime White can extract his some sense it is being used all the
rook and the result should be a draw) time. When one is thinking 'does that
25 .i f3 .i xf3+ 26 'it> xf3 l:.e8 27 :e l knight belong on c4 or g3?' , this is
'ife4+ 2 8 'ito> t2 'ii'x c2+ 2 9 l:.e2 with a really a comparison. There is no at­
near-certain draw. tempt to evaluate the consequences
21 l:.dS+ 22 'it>el 'ii'c3+ 23 'ito>fl
•.. of these two alternatives in absolute
l::ld2 (D) terms; one is simply asking which is
better. In this wider sense comparison
is an important part of chess thinking.

Making your opponent think

Sometimes it is better to pass the re­

sponsibility for a difficult decision on
to your opponent. You may be con­
templating, for example, the move
. . . h6 which allows a piece sacrifice
J. xh6 ! ?. You look at the sacrifice a lit­
tle; it is a complete mess, extremely
hard to evaluate. However, the alterna­
24 •et? tives to ... h6 are about equal. You may
Losing by force. The last chance decide to play . . . h6 without further
was 24 'ii'x d2 'ikxd2 25 h3 J.e2+ 26 thought, and put the responsibility of
'it> g l 'ii'x c2 27 .i g2 (27 : h2 'ii'c 3 ! ) evaluating the sacrifice onto your
 A r THE BOARD 47

opponent - he will have to spend time

coming to a definite conclusion in­
stead of you. There is nothing more
frustrating than thinking for half an
hour, deciding that the sacrifice is not
sound, playing . . . . h6 and meeting with
the reply l:r. ad l within 30 seconds. In­
deed, there are many opponents who
would use the logic that if you thought
about the sacrifice for so long and still
allowed it, then it must be unsound.
Playing your move fairly quickly
short-circuits this logic . Indeed, your J. Nunn - S. Conquest
opponent may wonder if you have Hastings 1 9960
even seen the sacrifice. Obviously, this
type of logic can be overdone; it would increase the pressure on e6 to such as
be ridiculous to allow a sacrifice lead­ extent that . . .f6 will never be possible,
ing to a clear-cut win. Moreover, if you and it also removes any possibility of
were playing a dangerous attacking Black playing . . . l0 f6 and . . . .t h5 . He
player, then provoking him to make a might, of course, play . . . l0 f6-e4, but
sacrifice is probably not a good idea. the knight could be pushed away again
A 'making your opponent think' by f3 and he would be left with the
move usually works best if you are in a same problems.
slightly inferior position. Then your The only way Black can improve
opponent will be undecided about his position is to play . . . f6, intending
whether to continue positionally, in . . . i. f7, or, if White is careless enough
which case his advantage might prove to allow it, . . . e5.
insufficient for a win, or to enter the 17...f6
tactics, when, at the cost of some risk, Conquest played this very quickly,
he might end up with a larger advan­ which is a good practical 'make your
tage. If your opponent has a large ad­ opponent think' decision. White can
vantage to start with, then he will take the pawn by 18 i. xe6, but this in­
probably avoid tactics since he has volves pinning the bishop, which
good winning chances in any case. might have dire consequences. He can
Here is an example (D ). also take the pawn by 1 8 i. xd6 'ii'xd6
Black stands badly, mainly because 1 9 l0f4 .t f7 20 l0 xe6, which is simi­
of the miserable position of his bishop larly complicated. Finally, White might
on g6. White intends to swap bishops decide to ignore the offered pawn and
on d6 and then play l0 f4. This will continue positionally, although in that
48 SECRETS OF PRACTICAL CHESS

case Black can play . . . i. f7 and his 25 'ii'd 3 'ii'xd5 26 'ifxd5 cxd5 27 l:.d4
prospects will have improved as a re­ and the weak black pawns give White
sult of releasing the bishop from the a winning rook and pawn ending.
trap on g6. 20 l:.d3! (D)
I was faced with a difficult decision Not 20 i. xd6 'ifxd6 21 i. f7 l:. xe l
and it took me just over twenty min­ 22 .:r. xe l .t xf7 23 'ii'xf7 'ifxd4 and
utes to make up my mind to go for the Black even has the advantage.
immediate capture of the pawn. I also
considered continuing positionally, but
it seemed to me that although White
would hold an undoubted advantage,
making progress would be far from
straightforward. I had plenty of time
on the clock, and so I decided to ana­
lyse the pawn capture thoroughly. Fi­
nally I concluded that White would
emerge with a clear advantage, but I
still felt a little nervous as I took the
pawn !
18 .txe6!
The other capture, 1 8 i. xd6 'ii'xd6 20... lLJxe6
19 lLJf4 .t f7 20 lLJ xe6, is less accurate There is nothing better, e.g.:
in view of 20 . . . iLJb6 (20 . . . i. xe6 2 1 1 ) 20 ... b5 2 1 'ii' b 3 lLJxe6 22 l:. xe6
.t xe6 iLJf8 2 2 d 5 cxd5 2 3 l:.xd5 'ii'xd5 .:r.xe6 23 'ifxe6 .t xf4+ 24 lLJ xf4 l:. e8
24 i. xd5 l:. xe2 25 .:r. xe2 .:r. xd5 26 25 'ii'b3 l:. el+ 26 �d2 .:r.d l + 27 �c3
l:. e8+ � c7 27 l:. xf8 wins for White) i. f7 28 'ii' b4 and White consolidates
2 1 l:. d3 iLJd5 22 .t xd5 'ifxd5 23 l:. e3 his extra pawn, while retaining a posi­
l: xe6 24 l:. xe6 .t xe6 25 'ifxe6 'ii'xd4 tional advantage.
26 'ii'e3 with only a slight advantage 2) 20 . . . lLJxe6 2 1 l:. xe6 .t xf4+ 22
for White. lLJ xf4 .t f7 23 d5 .t xe6 24 lLJxe6 'ife5
18 iLJfS
••• 25 lLJxd8 .:r. xd8 26 d6 'iWe l + 27 .:r.d l
After 18 ... .t f7 19 'iWc4 lLJb6 20 'ii'h 3 'ifx f'2 2 8 'ife6 and the powerful passed
White keeps his material more easily. pawn is decisive.
19 'ifc4 .t hS 21 .:r.xe6 l:.xe6 22 'ifxe6 i.xf4+ 23
The main alternative is 1 9 lLJ xe6... iLJxf4 l:.e8 24 •xrs
20 l: xe6 .t xf4+ (20 'iff7 21 l:. xd6
. . . There is no reason why White should
and 20 . . . .t f7 2 1 l:. xd6 l:. xd6 22 'ii' b4 not take a second pawn. Black gets a
win for White) 2 1 lLJ xf4 .t f7 22 d5 couple of checks, but his counterplay
i. xe6 23 lLJxe6 'ii'd 6 24 lLJ xd8 l:. xd8 is soon quashed.
 AT THE BOARD 49

24 l:r.el + 25 �d2 l:r.dl + 26 'it>c3

.•• the consequences are not too serious,
.i. g4 27 'ii'cS then you have an oversight; if the up­
Applying DAUT. 27 'ii'x g4 'ifa5+ shot is catastrophic, then you have a
28 b4 'ii'a 3+ 29 'it> c4 'fi'xa2+ 30 � c5 blunder.
'fi'xc2+ 3 1 <ilid6 may also win, but there Why do chess players blunder? Hu­
is no point in analysing such a line man fallibility is obviously one rea­
when a safe alternative exists. son, but the majority of blunders are
27 l:r.al 28 l:r.e3 (D)
..• not purely the result of chance. There
are several factors that can greatly in­
crease the probability of a blunder and
if one is aware of these factors, then
one can take particular care during
these 'high-risk' periods.
The most common cause of blun­
dering is a previous oversight. Let us
suppose that your opponent suddenly
sacrifices a piece and that this sacri­
fice comes as a complete surprise to
you. A typical train of thought starts :
"Oh, * * * * ! I completely overlooked
that. It looks strong; perhaps I will
Black's counterplay has dried up, even have to resign. How could I have
and it is White's turn to go on the of­ overlooked such a simple idea? How
fensive. stupid. And I had such a winning posi­
28 .i. cS 29 l:r.e7 'ii'd8 30 tDe6
.•• tion." Not very constructive. Making a
Doubtless there were other ways to serious oversight is a huge blow to
win, but forcing exchanges is a risk­ one's self-confidence, it sends the
free approach. mind into a whirl and turns the stom­
30 i..xe6 31 :xe6 l:r.xa2 32 'ii'd6+
•.• ach upside down. Nobody's brain can
1-0 function properly in such a state. The
While Black is retrieving his rook only real solution is to take the time
from a2, White will make a meal of his necessary until one's mind and stom­
kingside pawns. ach have quietened down. The length
of time this takes depends on one 's
Oversights and b l u nders personality and will-power, and of
course how strong the opponent's
O versights and blunders are merely move actually is ! What is quite useless
two aspects of the same thing. If you is to have a major bout of self­
miss something, and by good fortune recrimination at the board - if your
50 SECRETS OF PRACTICAL CHESS

personality tends towards this, try to his minus pawn and crippling light­
save it until the end of the game (by squared weaknesses.
which time, hopefully, it may not be 1S... lllb4
necessary). This is another reason why Black not surprisingly tries to force
running short of time is a bad idea. If through . . . d5 before White tightens his
something unexpected happens, you grip.
do not have the possibility of a short 16 'ii'e4?
'time-out' to recover your composure. 16 'ii'b3 was much simpler. After
Here is an example of precisely the 1 6 . . . dS 1 7 a3 d4 1 8 axb4 dxe3 1 9 fxe3
wrong reaction, taken from one of my Black has very little to show for his
own games. two pawns.
16... dS!
J. Nunn D. Cox
- I hadn' t realized that this move was
Norwich Junior International 1 974 possible, although in retrospect it is
Sicilian, Pelikan hard to imagine any other follow-up to
Black's previous move. At once I saw
1 e4 cS 2 ll:\f3 e6 3 d4 cxd4 4 lllxd4 that the position had become rather
lllf6 S lllc3 ll:\c6 6 llldbS d6 7 i.f4 eS more murky than was really necessary
8 i.gS a6 9 llla3 i.e6 10 lll c4 i.e7? and the knowledge that I had made a
11 i.xf6 gxf6 12 llle 3 .:r.cS 13 lllcdS bit of a mess of it was the trigger for
fS? ! 14 exfS .txdS 15 'ii'xdS (D) what followed.
17 "ifxeS f6 18 "ifd4 (D)

Black's handling of the opening has

left a great deal to be desired and At this point I was getting slightly
White has a large advantage. One more worried, not about 1 8 . . . ll:\xc2+ 1 9
tempo and White would play c3, leav­ lll xc2 .:r. xc2 2 0 "ifa4+ l: c6 2 1 .t xa6 !
ing Black with no compensation for 0-0 22 .t bs, followed by 23 0-0, with
 A T THE BOARD 51

a clear two-pawn advantage, but about i. c5 22 'iFh4 or l 9 . . . i. c5 20 'ii'g 4 'ii'a5

18 ... i.cS. After 19 'ii'd2 'iFe7 Black has 2 1 0-0.
some nasty threats, while White has
fallen somewhat behind in develop­ Warni ng signals
ment. In fact White should be able to
retain the better position by 20 0-0-0 ! , One of the most common reasons for
but o f course White n o longer has the blundering is failure to take warning
same clear-cut advantage as at move signs into account. Tactics very rarely
15. strike out of a blue sky. There is nor­
18 .:r.xc2! ?
.•. mally some underlying weakness, visi­
This is not a particularly good move, ble beforehand, which the tactic seeks
but it had a devastating effect on me to exploit. When you see such a poten­
because I had completely missed it. tial vulnerability, it pays to look very
Having now overlooked two of my op­ carefully for a possible tactic by the
ponent's moves, I went into a panic. I opponent exploiting this weakness.
saw horrible threats such as 1 9 . . . 'ii'aS Once I played 1 00 games against
or 19 . . . ll:ic6 followed by . . . i. b4+ and Mike Cook at 10 minutes (for him) vs
quickly decided that I had to get my 5 minutes (for me). At that time Mike
kingside pieces out as rapidly as possi­ was about 2300 strength. About half­
ble. I saw that after 19 i. d3 ll:i xd3+ 20 way through the series (which I even­
'iFxd3 'ii'a5+ (or 20 . . . i. b4+) I would tually won 88- 12) he explained his
have to play 'it>fl, but the resulting po­ disappointment:
sition did not look too bad. My king "I thought that I would see lots of
could find safety after g3 and � g2, advanced strategic concepts in these
freeing the other rook; in addition games but actually all I have learnt is
Black's d-pawn would be very weak. LPDO."
Unfortunately my brain had stopped "LPDO?"
functioning. "Loose Pieces Drop Off."
19 i.d3?? i.c S 0-1 During the remaining games I saw
I am sure almost all players have what he meant. Most of the games
had this experience at one time or an­ were decided by relatively simple tac­
other - disorientation caused by one tics involving undefended pieces, when
mistake immediately leading to a sec­ the LP would duly DO.
ond, much more serious, error. A calm Now, whenever I see such a combi­
look at the position after 18 . . . lt xc2 nation (they occur quite often), I think
would have shown that White is still 'LPDO' .
clearly better after 19 i. e2 (threaten­
ing 20 i. h5+ followed by 0-0), for ex­ Looking at the diagram overleaf,
ample 1 9 . . . .:r. g8 20 i. h5+ 'iii>f8 2 1 0-0 the LP on h8 should have been a
52 SECRETS OF PRACTICAL CHESS

M. Chandler - W. Beckemeyer L. Christiansen - A. Karpov

Bundesliga 1987 Wijk aan Zee 1 993

warning sign to Black that he should White's queen had earlier moved
look for tactics especially carefully. from dl to c2, so this is also an exam­
16... lbb6? ple of overlooking a switchback (see
Walking right into it. page 57).
17 lbdxbS! axbS 18 'ii'd4 We will consider two other warning
and White won easily. signals, but virtually any tactical weak­
ness can give the opponent a chance
What is surprising about this exam­ for a combination. Note that by ' tact­
ple is that the initial move of the com­ ical weakness' I do not mean isolated
bination is one of the most familiar pawns, bad bishops and other such
tactical blows in the Sicilian. It seems long-term weaknesses; these are ' stra­
that the idea of it being followed up by tegic weaknesses' . A tactical weak­
anything other than i. xb5+ or lb xb5 ness is a short-term vulnerability such
just didn't occur to Black. as an undefended piece, a potential pin
or fork, or a trapped piece. Tactical
Even ex-world champions (or cur­ weaknesses may exist only for a single
rent ones, depending on your perspec­ move, so if your opponent creates one,
tive) can fall victim to LPDO (D): you should look for a way to exploit it
Black already has an LP on h5. He straight away.
added a second one by . . . One of the most common errors
u ...i.d6 amongst beginners is to allow an e-file
Christiansen looked amazed. pin of the queen, but even GMs can
12 'ii'd l l-0 underestimate the significance of a
LPDO! queen + king line-up.
 A T THE BOARD 53

Indeed, these defects are so serious

that this move falls into the 'has to
work' category (see page 28). The po­
tential e-file line-up and the fact that
Black has committed himself so heav­
ily are clear warning signals to which
Seirawan did not pay sufficient atten­
tion.
8 �f2!
It suddenly turns out that 8 . . . 'ii xe4
is impossible because of 9 i. bS+ �d8
10 .:r.e1 and the black queen is pinned
B. Spassky - Y. Seirawan against a mate on e8. Black is forced to
Montpellier Candidates 1 985 change his plan completely, and all the
defects of the move . . . 'fie7 are ex­
The diagram position arose after 1 posed.
e4 eS 2 f4 exf4 3 tllf3 tll e7 4 d4 dS S 8 ....tg4 9 hS tllh4 10 i.xf4 tll c6 1 1
tllc3 dxe4 6 tllxe4 tllg6 7 h4 (D), when i.bS 0-0-0 12 i.xc6 bxc6 13 'f#d3
Seirawan faced the problem of dealing tllxf3 14 gxf3 i.fS 15 'iWa6+ 'it>b8 16
with the threat 8 h5, which would re­ tll cS .tc8 17 'iWxc6 l:lxd4 18 .:r.ael
gain the sacrificed pawn on f4. Since lhf4 (18 ...'f#xc5? 19 i. xc7#) 19 'fibs+
Black had based his whole strategy on �a8 20 'f#c6+ 'it>b8 21 l:be7 i.xe7
maintaining this pawn, it would be a and White won easily.
bitter blow to be forced simply to give Of course 8 �f2 is a very unusual
it up. move and one can easily imagine how
7...'f#e7? Seirawan came to overlook it. In Open
This is an example of a major com­ Games, White does not normally vol­
mitment. Black's aim is to inconven­ untarily deprive himself of the right to
ience White 's knight; indeed, one's castle, and in most positions it would
instant reaction is that White must be a waste of time for Black to start
play 8 'ife2 to avoid losing the knight analysing king moves to the second
to . . . f5 . However, if it turns out that rank. How, then, can one expect to
White is not seriously inconvenienced spot moves such as �f2 in advance?
by the pin, then Black has done quite a This really is the point of the warning
lot of damage to his own position: the signs - they tell you when to take time
queen obstructs the development of out to look for unusual tactical motifs,
Bl ack's kingside, takes away the last which might only be justified because
flight square from the knight on g6 you have a potential vulnerability in
and sets Black up for an e-file pin. your position.
54 SECRETS OF PRACTICAL CHESS

Our final warning sign is the weak consider it a possibility for his oppo­
diagonal . Here is a vivid example: nent. However, an extra piece provides
pretty good compensation for the two
bishops !
Curiously, a similar 'diagonal dis­
aster' occurred in another game from
the Open Group of the same event.

Ye Rongguang - L. van Wely

Antwerp Crown Group 1997

The diagram position arose after 1

d4 lbi'6 2 lLlf3 g6 3 i.. gS i.. g7 4 li:lbd2
0-0 S e3 d6 6 i.. c4 cS 7 c3 (D). Now A. Stefanova - S. Giddins
Van Wely decided to develop his c8- Antwerp Open 1 997
bishop at b7, and played 7...b6??. The
weakness along the h l -a8 diagonal After 1 d4 dS 2 lLlf3 lLlf6 3 i.. f4 cS
is quite apparent, but Van Wely had 4 e3 i.. fS S c3 Black decided to con­
overlooked the simple continuation 8 tinue his development with the natu­
i.. xf6 i.. xf6 9 i.. d S, winning a piece. ral move S e6? (D). This ignored the
•••

The game ended 9 i.. a6 10 i.. xa8 d5

•.• danger of leaving the light squares on
1 1 c4 dxc4 12 0-0 cxd4 13 exd4 i..xd4 the queenside exposed by developing
14 lLlxc4 1-0. This horrible disaster the c8-bishop to f5, and then prevent­
was the direct result of not taking ac­ ing it from returning by . . . e6. Black
count of the danger sign 'trapped rook doubtless took into account lines such
on a8 ' . Had he looked a little longer, as 6 i.. b 5+ lLl c6 7 lLle5 l: c8 8 'ii'a4
Van Wely would surely have seen 'ii' b 6, when there are no more pieces
i.. x f6. The problem here is psycho­ White can bring to bear on c6. How­
logical: White would never normally ever, he failed to take into account that
play an anti-positional move such as the b8-knight is really the only piece
i.. xf6, giving up the two bishops for that can interpose on the b5-e8 diago­
no good reason, so Van Wely did not nal. Therefore 6 i.. xb8! put him in an
 A T THE BOARD 55

awkward situation. After the check on "ii'xd7 1 1 'iie2 IC.xg2+ 12 'iixg2 lhb8
b 5 Black will have to move his king, as 13 dxcS g6 14 IC.c3 i.. g7 Black had
7 . . . IC. d7 loses to 8 IC.e5. In the game some compensation for the piece and
Black chose 6 ... l:r.xb8 7 i.. b 5+ <l;e7 8 managed to draw in the end.
dxcS and soon lost. Perhaps 6 . . 'iixb8
. We deal with the question of how to
7 .t b5+ catd8 was slightly better, but in continue after you have made an over­
any case Black's position is hardly en­ sight under 'Defence' (see page 97).
viable. Here there were three factors
that led to Black's oversight: failure to 'Hard-to-see' moves
take into account the warning sign
'weakness along the b5-e8 diagonal' , Another type of oversight is the sim­
the anti-positional nature of i.. x b8 and ple 'blind spot' . In this case it is not a
the soporific effect of White's unam­ question of a misjudgement, but of a
bitious opening. move, either for oneself or for the op­
A similar motif occurred in the fol­ ponent, simply not entering one's mind
lowing diagram after the moves 1 d4 at all. This is usually because there is
lfil6 2 i.. g5 IC.e4 3 .tf4 d5 4 e3 i.. f5 5 something odd about the move, so that
f3 IC.f6 6 c4 c5? ! 7 cxd5 IC.xd5 (D). it doesn't fit into any of the standard
patterns. Here is a clear example.

M. Adams - L. Van Wely

Tilburg 1 996 U. Andersson - A. J. Mestel
London (Philips & Drew) 1 982
8 .txb8!
Black is obliged to surrender a piece, It is Black to play; Mestel contin­
as e3-e4 is coming. The consequences ued 29 l:r.d5 and White's material ad­
•.•

were less catastrophic this time as af­ vantage eventually proved sufficient
ter 8 ... IC.xe3 9 i.. b 5+ i.. d7 10 i.. xd7+ to win. However, Black could have
56 SECRETS OF PRACTICAL CHESS

forced instant resignation by 29 . . .: d2 ! ,

which either wins a rook or mates. Ad­
mittedly, Mestel was somewhat short
of time, but even so it is not hard to
calculate this simple combination,
provided that the idea crosses one 's
mind in the first place.
When two line-moving pieces are
facing each other, the options are usu­
ally limited to exchanging, allowing
the opponent to exchange, or moving
the attacked piece away. It is very un­
usual for the piece to move part of the M. Tai R. Fischer
-

way towards the enemy piece. The ex­ Cura�ao Candidates 1962
ception is when there is a very juicy
outpost somewhere on an open file, In this position White is aiming for
practically begging a piece to land on the exchange of light-squared bishops.
it. That was not the case here, so the Then he will have certain potential
winning move was a little hard to see. advantages: pressure against the iso­
Nevertheless, it is surprising that Mes­ lated a-pawn on an open file and the
tel did in fact play a move along the possibility of an eventual good knight
file, so he must have noticed that it vs bad bishop position with White' s
was impossible for White to take the knight occupying d 5 (for example, if
rook after such a move. Evidently the White plays l:r. a2, then �2-c4-e3-
fact that the rook was undefended on d5). Fischer found a simple and ele­
d2 formed just enough of a mental gant solution to Black's problems:
block to cause the oversight. 16...i.bS ! . This collinear move leaves
I call such moves along the line of White with three options : to leave the
attack (without capturing the attack­ bishop on b5, to play lllx b5, or to play
ing piece) collinear moves. They are i. xb5 . If White ignores the bishop,
even more difficult to spot when the then Black's next move will be . . . .t c6,
attacked piece moves away from the both attacking the weak e4-pawn and
attacking piece. Here is one example supporting the thrust . . . d5. To meet
from opening theory: 1 e4 cS 2 lllf3 d6 these threats White will be forced to
3 d4 cxd4 4 lllxd4 ll:\f6 5 lllc3 a6 6 play i. f3 , but this is hardly the most
i.e2 e5 7 lllb 3 .te6 8 0-0 lllbd7 9 f4 active square for the bishop; Black
'flc7 10 rs i.c4 1 1 a4 i.e7 12 i.e3 0-0 would obtain good counterplay by
13 a5 b5 14 axb6 lll xb6 15 .txb6 . . . a5-a4, when the b2-pawn is vulner­
•xb6+ 16 �bl (D) able. The second option, 17 lll x b5,
 A T THE BOARD 57

swaps off the knight which White was and now 24 . . . l:. xc3 would have left
hoping to put on d5 later, and moves Black a pawn up with a good position.
the weak a-pawn off the open file. The In the later game Unzicker-Fischer,
res ult would be a near-certain draw in Varna Olympiad 1 962, the German
view of the opposite-coloured bish­ grandmaster chose 21 h3 but lost in
ops, but this might be White's best op­ only five more moves: 21 ltcS 22
.••

tion. l:r.fel h6 23 'it>h2 i.. gS 24 g3? 1ka7 25

Fischer played 1 6 ... i.. b 5 in two 'it>g2 l:r.a2 26 'it>f1 l:r.xc3 0-1 .
games; both opponents chose the third Fischer's handling o f this game was
option and the games continued 17 very impressive.
i.. x bS axbS 18 ll:idS ll:ixdS 19 'iWxdS
:Ca4! (a second collinear move, put­ Switchbacks are another potential
ting pressure on e4; if White takes, blind spot. If you have just moved a
then his b-pawn will be in trouble) 20 piece from A to B, the idea of moving
c3 'ifa6 (D). it back from B to A can be hard to spot.
The third type of move which is
hard to spot is the 'hesitation move' .
This arises when a piece can move
from A to either B or C. You decide
that it is bad to move it to C and play it
to B instead. Next move it would be
good to play it from B to C, but you
don't consider it because this is pre­
cisely the move that you rejected a few
moments ago.
Here is an example (D):
Black has a fairly miserable posi­
tion. He is a pawn down, his king is
Not only does White have no ad­ somewhat exposed and White has a
vantage, but he is even slightly worse. passed d-pawn which bears constant
His queen looks impressive on d5, but observation. However, the opposite­
actually it is the knight which should be coloured bishops make it very hard for
there instead. The knight has no route White to win, and in addition Black
to reach d5, and indeed after . . . i.. g 5 it came up with an excellent defensive
will be virtually immobilized. Black idea:
can improve his position by . . . h6 and 1 h4!
...

.. . lk8-c4, for example. Now 2 i.. xh4 1if4+ 3 i.. g 3 'ii' h6+
The Tal game continued 21 l:r.adl leads to perpetual check, so I tried ...

:cS 22 lbcl b4 23 ll:id3 bxc3 24 bxc3, 2 d7 i..xd7

58 SECRETS OF PRACTICAL CHESS

J. Nunn R. Filguth
- Based on these lines, Black con­
World Student Team Championship, cluded that there was no defence and
Mexico City 1 977 so . . .
1-0
Not, of course, 2 . . . hxg3+ 3 'ii'xg3+ Filguth had overlooked the simple
followed by 4 d8'iF . 4 . . . 'iii>h5 ! , after which White has no
3 'ii'g8+ <itih6 winning chances at all. The reason: he
Black's best move. After 3 . . . �h5 4 had rejected playing the king to h5 the
.i. xh4 ! Black is more or less forced to previous move, but under the changed
play 4 . . . <iti h6, but then White has an circumstances it had become the only
extra tempo. In fact, even this would saving possibility.
only offer White marginal winning As an aside, I am one of those play­
chances because he has nothing forced ers who gets up from the board after
and the exchange of White 's danger­ almost every move (but even when
ous d-pawn for the enemy h-pawn has walking around, I am still thinking
clearly favoured Black. hard about the position). However, af­
After the text-move, 4 .i. xh4 i. e6 ter playing i. d6 I stayed at the board.
allows Black to defend comfortably. When players are completely win­
4 .i.d6 (D) ning, they often remain at the board in
My opponent appeared surprised by order to accept their opponent's resig­
this move, which has the nasty threat nation, and I suspect that Filguth thus
of 5 i. f8+ 'lfi> h5 6 'ii'h8+ � g6 7 'ii'g 7+ interpreted my unusual immobility. In
�h5 8 'ii' h6#. It is quite hard to stop fact, of course, I was frantically trying
this threat, for example 4 . . . 'ii'g 6 5 to find a good reply to . . . �h5 .
.i. f4+ 'iil' h5 6 'ii' h 8+, 4 . . . 'ii'g 5 5 i. f8+
� h5 6 'iF h7+ and 4 . . . 'ii'h7 5 .i. f4+ are I seem to be especially prone to
no help. overlooking 'hesitation' moves.
 A T THE BOARD 59

simple 1 9 'iW b3+ would have been

enormously strong. After 1 9 . . . 'ifxb3
(or else White swaps queens and takes
on d3) 20 axb3 l:ld8 21 lte3 White wins
the pawn, after which Black doesn ' t
have much to show fo r his two-pawn
deficit.
I didn't even consider 19 'ii'b 3+ for
the simple reason that I had rejected it
the previous move. It didn't occur to
me that the 'extra' tempo . . . d3 is actu­
ally very helpful to White, because it
J. Nunn - J. de la Villa severely weakens Black's d-pawn and,
Szirak lnterzonal 1 987 unlike the line after 1 8 'ii' b 3+, White
has not weakened the d4-square by
In this position White is a pawn up playing c3-c4.
but has yet to develop his queenside
pieces. Here I rejected 18 'if b3+ be­ It is worth trying to pinpoint repeti­
cause after 1 8 . . . 'it>h8 1 9 'ii'xb6 axb6 it tive errors in one's own play. These
is quite hard for White to continue his days, with the availability of strong
development. 20 d3 is impossible be­ computer programs, finding tactical
cause of the pressure on c3, and 20 c4 errors is quite easy. I always quickly
is met by 20 . . . d3 ! 2 1 .i. xd3 .i. d4 22 run my games past Fritz it is often
-

l:. fl .i. g4, which leaves White totally quite shocking what turns up !
paralysed. Black could just continue
with . . . l:. ae8 and . . . .i. e2, eventually Ti me-trouble
winning the pawn on f2 and setting up
possible mating threats along the h­ The advice here i s quite simple: don' t
file. get into time-trouble in the first place.
I therefore played 18 'iWc2! which, This immediately raises the question:
in addition to the attack on h7, threat­ what constitutes time-trouble? Some
ens to bring the queenside pieces out people panic when they have ten min­
by 19 d3 and 20 .i. f4. Black responded utes left for ten moves, while others
with 18 .. . d3! ?, based on the tactical appear quite calm even when down to
point 1 9 .i. xd3 ? 'ifxf2+ ! ! 20 <it xf2 their last thirty seconds .
.i.d4++ 21 �e2 .i. g4#. I actually played I will propose a definition: time­
19 'ti'xd3 .i. xc3 20 bxc3 'ii'xf2+ 2 1 trouble arises if you have less than one
�h2 'ifxel 2 2 .i. a3 ! and won after im­ minute left per move . In my opinion
mense complications. However, the this is the lower limit for guaranteeing
60 SECRETS OF PRACTICAL CHESS

an acceptable standard of play. It is of cost you a half-point. The piece you

course possible to play well even with hang later during time-trouble proba­
less time than this, but you are really bly is.
putting yourself in the hands of fate. If 3) Providing yourself with an ex­
no unexpected problems arise, no cuse. To those who do not suffer from
shocks, no really tricky decisions, then this particular ailment, it probably
you will probably survive more or less sounds ridiculous, but it is surpris­
unscathed, but how can you be sure ingly common. In these cases there is
this is going to happen? Those who are a (possibly subconscious) reluctance
addicted to time-trouble will show to admit that a loss is due to bad play;
you the remarkable games they won problems with the clock provide an al­
despite extreme time shortage; they ternative explanation. If you find your­
probably won't show you the much self saying to people after a game 'I
larger number which they lost because had a perfectly reasonable position
of the same time shortage. until I ran into time-trouble' then you
If you regularly run into time­ probably have this disease to some ex­
trouble, according to this definition, tent. It really is an absurd attitude. The
then you are doing something wrong clock is just as much a part of the
and it is almost certainly costing you game as the board and pieces, and los­
points. ing because of time-trouble is no dif­
The reasons for running short of ferent to losing because of weak play -
time are many. Here is a selection of it's still a zero on the score-sheet. The
the most common: only difference is that if you run very
1 ) Indecision. Probably the most short of time you will almost certainly
common reason. If you spend a long lose; if you play more quickly, you
time agonizing over decisions, then may lose because of weak play, but
you will inevitably run short of time. then again you may not. The best way
The pragmatic approach I recommend to improve is to gain over-the-board
in this book should help. If you find experience and if your games just be­
that, after lengthy thought, you usu­ come a time-trouble mess, then you
ally end up playing the move that you are probably not learning very much
wanted to play straight away, then from them.
learn to trust your intuition more. If you have time-trouble problems,
2) Worrying too much about very then in addition to the points men­
minor matters. There is no point think­ tioned above, bear in mind the follow­
ing for half an hour about a possible ing tips:
advantage or disadvantage of what a 1) Arrive for the game on time.
computer would call '0. 1 of a pawn' . The most obvious and simplest way of
This almost certainly is not going to saving time !
 A T THE BOARD 61

2) Don't spend a long time consid­ might overlook a powerful zwischen­

ering which opening to play. If you zug.
have the luxury of knowing more than 5) Chess is all about making deci­
one opening, then you should think sions. Postponing a decision doesn't
about this before the game. necessarily improve it. Try to get into
3) Don't think about theoretical the habit of asking yourself: is further
moves. Some players spend an inordi­ thought actually going to beneficial?
nate length of time reaching a position
with which they are already familiar. I won't give any advice about what
An extreme case was a game I played to do if you are in time-trouble, since
as White against Stohl in the Poisoned this situation shouldn't arise.
Pawn, an opening with which we were If your opponent is in time-trouble,
both very familiar. I played a line the main piece of advice is to stay
which was believed to lead to equality. calm. It is quite easy to become so ex­
He was so worried about the improve­ cited that you start bashing out moves
ment he felt sure was coming that he as quickly as your opponent and thus
thought carefully about each move, make no use of your time advantage.
trying to spot the 'improvement' in ad­ You should also avoid thinking "I
vance. By the time the improvement must try to win the game during his
arrived, well into the ending, he had time-trouble" and then embark on
used 90 of his 1 20 minutes reaching a some risky venture which you would
position he knew perfectly well. The never play under normal circum­
'improvement' , when it finally came, stances. This is only justified if you
was actually a rather modest affair, but have a very bad or losing position, and
thanks to Stohl's lack of time I won the the time-trouble really is your only
game. hope. Given a choice between a quiet
4) Don't use up time thinking about continuation and an equally good
obviously forced moves. Your oppo­ sharp continuation, then by all means
nent takes a piece; you suddenly no­ choose the sharp continuation, but an
tice that after your recapture he has a opponent in time-trouble is no justifi­
possibility that you hadn't noticed be­ cation for knowingly playing an infe­
fore. Don't think - just take back. rior move.
Thinking not only wastes time, it also Players in time-trouble tend to
alerts your opponent to the fact that spend most of their time analysing
there is something you are worried forcing, tactical continuations, so be
about. It he hasn't noticed it before, he especially careful not to allow a tacti­
probably will now ! Of course, you cal trick. Where strategy is concerned,
should be sure that the recapture really they will probably decide on a 'default
is the only possibility, otherwise you move' to play if your move carries no
62 SECRETS OF PRACTICAL CHESS

special threat. Very often this 'default every move until you become fed up
move' will be part of a plan for trans­ with it and spend a tempo ruling it out
ferring a badly-placed piece to a better completely. The most common mani­
square. If they are deprived of any festation of this form of laziness is
such obvious plan, then the result is preventing a possible check by the op­
usually confusion or a 'nothing' move. ponent. You have to calculate the check
With any luck, you can use the moves in every single line, and in the end you
before the time-control to gain a posi­ just pre-empt the check by moving
tional advantage which will set you up your king. There are, of course, many
nicely for next phase of the game. situations in which such a move is per­
fectly reasonable, but there are also
Laziness many in which the loss of time is im­
portant. Playing �b l after 0-0 (or
Laziness at the chessboard manifests 'it>b l after 0-0-0) may be just such a
itself in two ways. The first is when waste of time. When such a move is
you have a complicated possibility, justifiable, it is normally because of a
but decide not to analyse it on the concrete reason rather than because it
grounds that it would be too much is annoying to have to do a bit more
work. Note that is not the same as the calculation.
'DAUT' principle mentioned on page Here is an unusual example of lazi­
2 1 . That represents a conscious weigh­ ness:
ing of the balance between the time
spent and the possibility of finding a
better move. It comes into play most
often when there is a perfectly ade­
quate alternative which can be played
with little calculation. Laziness, on the
other hand, is a deliberate turning
away from a move which may tum out
to be very strong indeed. Losing a
game because you have overlooked a
move is one thing, but losing because
you have spotted the correct move and
could not be bothered to analyse it is
quite another ! J. Nunn Xie Jun
-

A second, more understandable, Hastings 1 9960

form of laziness is when your oppo­
nent has an irritating possibility. You In this position, it suddenly oc­
have to consider this same possibility curred to me that Black might play
 A T THE BOARD 63

. . . a3 as a pawn sacrifice at some stage. 21 .i.d7 22 .i.d3 'iFb6 23 'itii h l

•••

I started to calculate lines such as 2 1 exf4 24 .i.xf4 o!Dxd3 25 'iFxd3 'ii'f2 26

o!Df3 exf4 22 .i. xf4 a 3 2 3 .:r. xa3 l ha3 'iWd2 "ii'xd2 27 .i.xd2 l:Ue8
24 bxa3 and wondered if Black has Things have not gone very well for
enough compensation for the pawn. I White. His rook is still on the ridicu­
decided that she does not, because lous square a3, Black has the two bish­
White 's o!D d4 is annoying for Black, ops and White's e-pawn is weak. All
aiming at both c6 and f5. I started cal­ this adds up to a clear plus for Black
culating more and more lines and in and the main cause was the lazy move
each I had to take into account the pos­ 21 l:la3.
sibility of . . . a3 . In the end I just be­
came fed up with this, and decided to Determ i nation
rule it out.
21 .:r.a3? Of all the personal qualities which are
My reasoning was that since the important at the board, determination
rook is blocked in by both bishops is probably the most significant. Some
and unlikely to move along the first players, when confronted with an on­
rank, it might as well perform a useful erous defensive task, such as defend­
task (preventing . . . a3). Deep down I ing an inferior and long-drawn-out
was aware that this move looks highly ending, suffer a crisis of will-power.
artificial; my 'logic ' was really just a They see the gloom ahead of them,
rationalization for an outbreak of lazi­ with only the distant prospect of at
ness. most half a point, and become despon­
A more sensible piece of reasoning dent. The result is either a feeble resis­
would have been to note that Black's tance, or a decision to stake everything
queenside pawn majority will be very on a half-hearted swindle. When this
good in an ending, and so White must is brushed aside, they feel that they
play for a decision in the middlegame. can resign with good grace. Not sur­
For the moment White need not worry prisingly, this behaviour no recipe for
about . . . a3 , since the bl -bishop will success.
prevent the b-pawn promoting for sev­ Even very strong players can be­
eral moves. come so resigned to the prospect of in­
Thus 2 1 f5 is correct, aiming to evitable defeat that they do not notice
play o!D g4 or .!Dh5 to eliminate the de­ when a saving opportunity is presented
fensive knight. Since White's position to them.
in the diagram is not especially good, Here are a couple of notable recent
he should probably be happy with a examples.
sacrifice on h6 leading to perpetual The position overleaf is a simple
check. technical win for White, and it is hard
64 SECRETS OF PRACTICAL CHESS

47 . . . .:r.c3+ 48 cat g4 <l,; xe4 49 a6 :cs 50

a7 l:r. g8+ 5 1 <Ji;hS l:r. a8 and Black sim­
ply marches his king over and takes
the pawn.
47 <itie3 <itd6 48 'iii>d4 <itid7 49 �c4
<itic6 50 catb4 l:r.eS 51 l:r.ct+ �b6 52
.:r.c4 1-0

In order to understand this double

blunder, you have to take into account
the course of the game. Short had been
losing from an early stage, and Kaspa­
G. Kasparov N. Short
- rov had been exchanging down to a
London PCA World simplified endgame. Both sides were
Championship (9) 1 993 simply acting out the final stages of a
drama which had been decided in the
to imagine Kasparov going so far opening, and they knew full well what
wrong as to reach a drawn position, the result of the game would be. In this
yet that is what happened. frame of mind it is possible to let one's
45 .:r.at attention wander away from the posi­
Not a particularly accurate move, tion on the board. Nevertheless, it is an
although it does not throw away the astonishing error to occur at such a
win. White should simply put his rook high level.
on the fourth rank and march his king
to b4. Thus 45 :t d l + cat eS 46 l:r. d4 or
45 l:r. h6+ <l,;c7 46 .:r. h4 would have won
easily.
45 ... cates 46 e4??
An amazing blunder. One way to
win is by 46 cate2 <itie4 47 l:r. a3 (47 l:r. fl
and 48 l:r. f4 is also good) and Black
must either allow the king to reach d3,
or permit the advance of the pawn. In
either case White wins easily.
46 'iil'e 6??
••.

An equally serious error. Black can

draw by 46 . . . .l:r.cS ! 47 a5 (after 47 l:r. a3 A. Shirov J. Timman
-

.:r.c4 48 a5 l:r. xe4 49 a6 .:r.f4+ and Wijk aan Zee 1 996

50 . . . .:r.rs Black's rook makes it back)
 A T THE BOARD 65

Black had been under pressure for above it is quite easy to see that Black
some time and ended up in this rook only has a few checks and then White
ending a clear pawn down. can advance his pawn to the fifth rank
45 g4? ! (59 . . . 'ii'd6+ 60 � g7 'ii'd4+ 6 1 � t7
Both players seemed t o believe that 'ii'c4+ 62 W f8 'ii'b4+ 63 � g8), which
this position should be a clear-cut win would provide further evidence in fa­
for White, but in fact it is not so sim­ vour of assessing it as winning for
ple. The move played is probably al­ White.
ready an inaccuracy. After 45 'it> f3 4S...l:.e6
l:le6 46 l:td8+ (Shirov comments that Black must try to force White's
46 l:t c3 should win, but White's rook rook into a defensive position.
is tied to the defence of c6 more or le s s 46 l:.dS+
permanently and the win does not look White could still play 46 l:tc3.
at all simple to me) 46 . . . <itt7 47 l:td7+ 46...Wf7 47 l:td7+ l:te7 48 :Xe7+??
l:t e7 48 l:t xe7+ <it xe7 White can force This move throws away the win.
a winning queen and pawn ending by White could still have retreated his
49 � g4 � d6 50 � g5 � xc6 5 1 <it h6 rook and defended the c-pawn along
W b7 52 W xh7 c5 53 � xg6 c4 54 f5 c3 the c-file.
55 f6 c2 56 t7 c l 'ii' 57 f8 'ii' 'ii'c2+ 58 46 �e7 49 gS 1-0??
.••

'ii'f5 'ii'xh2 (after 58 ... 'ii'c6+ 59 W h5 Resigning is an even more aston­

'iWe8+ 60 W g4 'ikg8+ 6 1 W f4 White ishing blunder than Shirov's mistaken
remains two pawns up) 59 g4. The rook exchange. After 49 . . . W d6 Black
evaluation of endings with 'if+8 vs 'ii draws without undue difficulty:
is a simple matter at home, because 1 ) 50 f5 gxf5 5 1 h4 We5 52 'iii>f3 f4
one only has to consult a computer­ 53 h5 'iii>f5 54 g6 hxg6 5 5 h6 <it f6 56
generated database. The oracle in­ Wxf4 g5+ draws.
forms us that this one is indeed win­ 2) 50 h4 cat xc6 5 1 f5 W d6 52 f6
ning, but of course it is not so easy at Wd7 53 Wf3 cate6 54 W f4 W t7 55 cate5
the board. The general feeling amongst cate8 56 cat d5 W d7 57 <it c5 cat e8 and
GMs is that endings with 'ii'+g8 vs 'ii White cannot make progress, pro­
(and the defending king cut off from vided Black meets 'ii?c 6 with . . . <itd8. If
the pawn) are almost impossible to de­ the white king is on c5 or d5, then
fend in practice, even if some of them Black can play either . . . W e8 or . . . 'iii>d 7.
are drawn with perfect play (see page The fact that Black has this choice of
1 50 for more details). Thus even with­ moves means that these is no chance
out consulting the database, Shirov of White winning by somehow losing
would have been aware that this end­ a move, because only after <itc6 is
ing offers White excellent winning Black restricted to a unique square for
prospects. In the specific position his king.
66 SECRETS OF PRACTICAL CHESS

What makes Timman' s decision to The lesson to be learnt from this is

resign so surprising is that he could that while the game is in progress, you
have quite easily taken the pawn on c6 should be concentrating on the current
and waited to see Shirov's intention. position, regardless of what has hap­
Giving up so early is an indication of pened before. There are very few play­
defeatism - he felt that the rook end­ ers who have the technical skill to
ing was dead lost and, since the king convert advantageous positions into a
and pawn ending was a natural conse­ win without at some stage allowing
quence of the rook ending, then that the opponent a chance for counterplay.
too must be lost. The defender's task is to maintain his
concentration, so that when a possible
In these examples, both Short and saving opportunity arises, he does not
Timman were paying scant attention overlook it. There is no point in play­
to the position on the board and were ing on if all you are doing is goin g
heavily influenced by the preceding through the motions, waiting for the
play. right moment at which to resign.
2 The O pen i ng

B u i l d i ng a reperto i re 1 . . .c6: Solid positions, strategic,

safe king.
It would be nice to start the construc­ l . . .e6: Closed positions, strategic.
tion of an opening repertoire from 1 . . .e5 : Fairly solid, but can lead to
scratch, so that everything fits together almost any type of position, depend­
neatly, and all transpositions are taken ing on the follow-up. Petroff the most
care of. However, most players don't solid option.
have an elegant opening repertoire. 1 . . .d6: Sharp positions. Attack and
They play a motley collection of open­ counter-attack. However, White does
ings which they have accumulated, have solid options.
more or less by chance, over the years. l .. .d5 : Slightly off-beat, fairly solid
However, we will take the idealized but a little passive.
situation of someone starting from 1 . . . lbf6: Sharp positions . These
square one (al , presumably). days considered slightly off-beat.
The first step is to think about your 1 . .. lbc6: Slightly off-beat. Need to
personal style. Do you prefer open, consider what to play after 2 tDf3 .
tactical positions or closed, strategic
positions? Does an attack on your king Do the same with openings against
make you nervous, or are you happy 1 d4, flank openings and with White.
so long as you have a counter-attack? Then you have to consider how all
Do you prefer main lines, or some­ these fit together. If you choose the
thing slightly offbeat? Next, look at Pirc against 1 e4, it makes sense to
the various openings available, and consider the King's Indian against 1
see which ones fit in with your per­ d4. This is more flexible and will give
sonal style. For example, when choos­ you additional options later. For ex­
ing an opening with Black against 1 ample, you may decide later that you
e4, you might make some notes along would prefer not to allow the Siimisch
the following lines: against the King's Indian. To this end
you decide to meet 1 d4 with 1 . . .d6,
1 . . .c 5 : Open positions, tactical, at­ and against 2 c4 you will play 2 . . . e5 .
tack and counter-attack. Schevenin­ Many players who would quite like to
gen and Taimanov systems safer and play this system with Black are put off
less tactical than the Dragon, Najdorf by two things: firstly, White can play 2
and Pelikan. e4 and secondly White can play 2 M
68 SECRETS OF PRACTICAL CHESS

(it is true that Black can meet 2 l0f3 by line. Secondly, even if the worst hap­
2 ... .t g4, but many regard this as some­ pens and a blockbuster novelty fin­
what better for White). However, be­ ishes off a line, it is usually relatively
cause your openings have been chosen easy to switch to another line within
to fit together, neither of these prob­ the same opening. Mainstream open­
lems will bother you. The Pirc is al­ ings such as the Ruy Lopez or the Or­
ready in your repertoire, and after 2 thodox Queen's Gambit are not merely
l0f3 you can play 2 . . . l0 f6 and enter a single lines; they are massive com­
King's Indian, but with the Samisch plexes of different systems for both
having been ruled out. colours. Suppose, for example, you
Likewise, the Caro-Kann and the play the following line of the Chigorin
Slav fit together, and then you can an­ Defence in the Closed Ruy Lopez: 1
swer 1 c4 by 1 . . .c6, without having to e4 e5 2 l0f3 l0c6 3 i. b5 a6 4 .t a4 l0 f6
undertake any additional learning to 5 0-0 i.e7 6 l::te l b5 7 .tb3 d6 8 c3 0-0
cope with 2 e4. 9 h3 l0 a5 1 0 .t c2 c5 1 1 d4 'ilc7 1 2
There is a temptation to choose some l0bd2 cxd4 1 3 cxd4 i. d7 14 lOfl :acs
really unusual openings, because the 15 l0e3 l0 c6 16 d5 l0b4 17 .t bl a5 1 8
amount of study required is much less. a3 l0 a6. Suddenly you see a game in
However, I would advise against this. this line which seems to cast doubt on
Rarely played openings are usually Black's play. You need a replacement
rare precisely because they have some line, quickly. There is a wide range
defect. Sooner or later your opponents of possibilities to choose from; you
will start to exploit this defect and then could try 1 5 . . . l::tfe8 or, slightly earlier,
you will have to switch openings. If 1 3 . . . l0 c6. Perhaps delaying the ex­
you choose another very unusual open­ change on d4 and playing 12 ... i.d7 or
ing, the process will repeat itself. Af­ 12 ... l0 c6 might be an idea. All these
ter a few years, you will have spent are viable lines, so there is no need to
just as much effort as if you had cho­ panic. They all lead to the same gen­
sen a mainstream opening in the first eral type of position, so that the expe­
place, and you will have little to show rience you have gained with your
for your efforts. former line will not be wasted The
This problem doesn' t arise with a general principles governing play with
repertoire based on main lines. First of these Chigorin Defence pawn struc­
all, such lines, which have been tried tures will be still be valid in your new
and tested in thousands of grandmas­ line. Moreover, all the effort you have
ter games, are unlikely to be 'busted' put into learning how to combat ear­
in the first place. The worst that usu­ lier deviations by White (Exchange
ally happens is that a small finesse Spanish, Lines with 'ile2, etc.) will re­
leads to a minor reassessment of one main valid. Instead of facing a major
 THE OPENING 69

overhaul, only a minor modification that reviews of opening books tend to­
will be necessary. wards a certain blandness. Major pub­
Once you have chosen your open­ lishers are perhaps slightly safer than
ings, how best to study them? There is smaller operations, since they nor­
nothing better than a good book, mally have some sort of quality con­
which brings us on to our next topic. trol; having said that, they are also
under far more commercial pressure
U s i n g open i ng books to bring a book out on time, which can
lead to a rushed job. The saying 'Don't
These days there are books on virtu­ judge a book by its cover' is especially
ally every opening under the sun. valid for chess books . The covers are
Some are good, some are competent normally made by designers rather
and some are bad. Unfortunately, it is than chess-players; of course they
often not easy to decide which is should be checked, but even so howl­
which. In choosing a good book, re­ ers occasionally slip through. In the
views are an obvious guide, but these end, the reputation of the author is
are often not very helpful in the chess probably the best guide to the quality
world. This is not necessarily the fault of a book.
of the reviewers; I can testify from Having chosen a book, it is best
personal experience that reviewing first of all to gain an overview of the
openings books is a difficult and opening. Look first at just the main
thankless task. In order to assess an lines to create a mental picture of the
opening book properly, you really general structure of the opening. If the
have to be an expert in that particular book contains illustrative games, it is
opening. Of course, you can check to worth playing these over first; you will
see if it is up-to-date and well-written, probably already notice certain typical
but a really good openings book will themes repeating themselves. In most
contain all those finesses which only a major openings you will have a choice
practitioner of the opening will know. against each of the opponent's possi­
Cunning move-orders designed to bilities. As with the choice of the main
avoid particular lines, moves which openings, make a list of the various
caused a particular variation to be options and see which ones suit your
abandoned, but which were never ac­ style. Keep transpositional possibili­
tually played in a game - only a spe­ ties and move-orders in mind, to make
cialist will know whether such things sure that your opponent cannot cir­
are in the book or not. If one adds that cumvent your proposed repertoire by
chess book reviewers are normally un­ a simple variation in move-order. Sup­
p aid (apart from receiving a free copy pose, for example, that you are going
of the book), it is scarcely surprising to play a line of the Sicilian with 1 e4
70 SECRETS OF PRACTICAL CHESS

c5 2 ll:if3 e6, and are thinking about was published may throw doubt on
how to meet the 2 c3 Sicilian. You no­ some recommendations. In this case it
tice that the defence 1 e4 c5 2 c3 d5 3 is essential to have a source of alterna­
exd5 1Vxd5 4 d4 ll:i f6 5 ll:if3 .t g4 is tive possibilities.
currently popular, and decide to play More serious players who have a
that. It would be easy to overlook the chess database (see also page 1 66)
fact that White can play 1 e4 c5 2 ll:if3 may like to check this for recent devel­
e6 3 c3 instead. Of course you can still opments before playing a particular
play 3 . . . d5, but now the line with line. A database is also useful for see­
. . . .t g4 is impossible. There are ways ing which lines are currently popular,
round this particular problem, but it is and therefore where one should be fo­
important to have something ready in cusing one's efforts.
advance, and not to be caught out over At first you may find that your re­
the board. sults with a new opening are disap­
Notice that so far there has been no pointing. This is more likely with
detailed study. Everything has been strategic openings than ones based on
careful planning - good groundwork precise analysis. When I started play­
and thoroughness is the secret of suc­ ing the Sicilian Najdorf, my results
cessful opening preparation. Once you were very good. This is an opening in
have your repertoire mapped out, then which concrete knowledge of specific
you can study each line in detail. To lines is very important. I had just stud­
begin with, only study the main lines - ied the opening in great detail, and so
that will cope with 90% of your my knowledge was often better and
games, and you can easily fill in the more up-to-date than that of my oppo­
unusual lines later. nents. On the other hand, playing a
One important question is whether strategic opening requires a positional
it is better to study an opening book understanding which is better learnt
that offers a pre-selected repertoire by experience than from books. It may
(often called 'Winning With . . .' , or a be several games before you get up to
similar title), or one that offers a com­ speed with such an opening, but be
plete coverage of an opening. Cer­ persistent - your efforts will be re­
tainly both books have their place and warded in the end.
the ideal situation is to have one of
each type for a given opening. It is cer­ Books on offbeat openings
tainly useful to have someone do
much of the above-mentioned work A particular genre of books which de­
for you, but nevertheless some lines in serves special attention is that dealing
the proposed repertoire may not suit with dubious and rarely-played open­
you, or developments since the book ings. As in all categories of opening
 THE OPENING 71

books, there are good and bad exam­ Latvian (pages 72- 76 are based on this
ples. Typically, such a book will claim analysis). It is, of course, very rarely
that opening X is unjustly neglected, played and I would not expect a GM to
that recent games have shown X to be dare it, but this enhances its surprise
playable after all and that all sorts of value. It would certainly be embar­
hidden resources and novelties are re­ rassing to face it and plunge into deep
vealed for the first time in the book. thought on move three ! Moreover,
Unfortunately, 99% of the time the when a new book appears on an open­
reasons X is rarely played are entirely ing, it always encourages a few people
justified, the 'recent games' turn out to to take it up, so the chances of meeting
be encounters of little value between it were high enough to warrant devot­
unknown players, and the resources ing a little time to it.
and novelties will only stand a few When I was a teenager, there were a
seconds' close examination. The tricks couple of junior players who used the
which authors can use to make open­ Latvian which, if you don't know al­
ings such as X appear playable are ready, is characterized by the moves 1
many and varied, and are, of course, e4 e5 2 .!LJf3 f5. For many players this
revealed for the first time in this book! opening offers a huge advantage: it
It is hard to give details without provides a more or less complete de­
concrete examples, so that means I will fence to 1 e4. If you play a line of the
have to choose a couple of victims. In Spanish, then you have to worry about
this particular genre there is a depress­ various white systems in the Spanish,
ingly high proportion of poor books, such as the Exchange Variation and
and it would be easy to choose one of early d3 lines, plus the Scotch and 3
the really bad ones and have some fun. i. c4 - quite a body of theory to study.
In fact I have chosen one of the better Adopting the Latvian short-circuits all
examples, Tony Kosten's The Latvian that; apart from the King 's Gambit and
Gambit (B .T. Batsford, 1 995). This a few other openings, you don't have
book is unusual in that a strong grand­ to study anything else to meet 1 e4.
master has taken a look at a rarely The question is: can the Latvian be
played opening with a poor reputa­ considered playable?
tion. Tony was certainly taking on a The move 2 . f5 has a dubious ap­
..

real challenge with this title ! When it pearance; it doesn't develop a piece
arrived, I was quite baffled as to how and it weakens the diagonals h5-e8
the author had managed to fill up 144 and c4-g8, the latter being important if
pages; I had thought a detailed refuta­ Black intends to castle kingside later. I
tion would take I O pages at most. always viewed 3 .!LJxe5 as the most
I decided to use the book as a basis logical reply - Black should be pun­
for a couple of hours' analysis on the ished for not defending his e-pawn
72 SECRETS OF PRACTICAL CHESS

with 2 . . lllc 6. I had scored well with

. 4 lll x c6 is another obvious move,
this reply in my youth, so I decided to but after 4 . . . dxc6 I felt that this was
concentrate on it. rather playing into Black's hands. All
My first surprise came when I ex­ Black's minor pieces can immediately
amined the list of contents. Chapter 6 come into play, while White still has
dealt with (after 3 lll xe5) "3 . . .ll:\c6 and to move his d-pawn to bring his c 1 -
other third move alternatives for Black". bishop out. Moreover, White has no
I was stunned - I hadn't realized that minor pieces on the kingside, and if
3 . . . lllc6 was even vaguely possible. In White intends to castle on that side the
fact Tony justifiably dismisses the deficiency might allow Black to work
"other third move alternatives" but de­ up an attack. Neither of these lines is
votes over nine pages to 3 . ll:\c6 (D).
.. particularly clear, but the point of this
My first reaction on seeing an unfa­ initial evaluation is not to assess each
miliar move in a book is not to look at possibility rigorously, but simply to
the author's analysis, but to set the po­ decide which move looks best based
sition up on a board and decide what on general principles. If I had not
would be the most natural response. found a possibility that really struck
me as the most promising, then I
would have gone round again for a
second look at 4 'ii'h5+ and 4 lllx c6.
However, at this moment I noticed
a very tempting move: 4 d4 ! . This
seemed to me to fit in best with the
principle of rapid piece development
in the opening . White creates a 'two­
abreast' pawn centre and allows the
c l -bishop to be developed without
making any concession. The most ob­
vious reply is 4 . . . fxe4, but this just
loses after 5 lllx c6 dxc6 6 'ii'h 5+ and
I noticed that White could win the 6 ... g6 costs a rook, so Black must move
exchange by 4 'ii'h5+ g6 5 ll:\xg6 lll f6 his king. Nor is 4 . . . lllf6 possible, as 5
6 'ii'h4 hxg6 (6 . . . : g8 ! ?) 7 'ii'xh8 'ii'e7. lllxc6 followed by 6 e5 gives White an
In return Black has a substantial lead extra pawn and a good position. Black
in development. An assessment of this can, of course, play 4 ... lll xe5 but after
relatively complicated position would 5 dxe5 the pawn on e5 prevents . . . lllf6
take some time, so bearing in mind and Black's development is corre­
DAUT I decided to look for something spondingly awkward. If he tries to get
simpler. the pawn back by 5 . . .'ii'e 7, then 6 'ii'd4
 THE OPENING 73

looks very unpleasant, threatening 8 . . . 0-0 loses to 9 .t c4+. At the very

©c3-d5 . After five minutes I was still least Black will have to make several
unable to see any playable line for more moves with his queen and fall
Black against 4 d4, so I decided see hopelessly behind in development. In
what Kosten gave. His main lines are 4 addition one must also take into ac­
lilxc6 and 4 'ifh5+, with 4 d4 being count:
relegated to a small note, which con­ a) the dreadful positioning of the
tinues "4 .. .'ii' h4 ! 5 .t d3 fxe4 6 g3 'ifh3 pawn on f5, which has no positive fea­
7 .t xe4 lilf6" and ends up as ''fine" for tures at all and only weakens Black's
Black. Well, 4 ...'ii'h4 was certainly a position;
surprise ! However, I remained uncon­ b) the possibility of d5, forcing the
vinced. Black has violated virtually c6-knight to retreat; and
every principle of opening play, mak­ c) the plan of attacking c7 by lilc3-
ing weakening pawn moves, giving up d5 (or b5), possibly coupled with i.. f4.
a pawn and now developing his queen Summing up, Black's position is
before the other pieces - surely there probably just lost. There remained
had to be a way to gain a clear advan­ only one more point to be checked,
tage at the very least? Suddenly, I had viz. that Black could not exploit the
an idea. How about 5 lilf3 'ifxe4+ 6 temporarily active position of his
i..e2 (D)? queen by 6 . . . lilb4. However, it would
be a miracle if Black could get away
with breaking the one remaining un­
broken rule of opening play, that one
should not repeatedly move the same
pieces while leaving the rest of the
army at home.
The analysis, however, is fairly
simple : 6 . . . lilb4 7 0-0 'ifxc2 8 'ii'el
i..e7 (8 . . . 'ife4 9 lilc3 'ife7 10 'ii'd l lilf6
1 1 l:le 1 llle4 1 2 a3 lllc6 1 3 i.. d 3 wins)
9 lilc3 lil f6 10 lile5 (threatening 1 1
i.. d l ) 1 0 . . . f4 1 1 a3 lil c6 ( l l . . .d6 1 2
axb4 dxe5 1 3 dxe5 lilg4 1 4 lild5 .t d8
The more I looked at this the more it 15 e6 lilf6 16 lilxf4 0-0 17 i.. d 3 wins)
appealed to me. White threatens sim­ 12 i.. d 3 'ii'b 3 1 3 lilb5 lil xe5 14 dxe5
ply 0-0 and ltel . Black has insuffi­ and Black's position collapses. Even
cient time to develop his kingside this is presented in rather unneces­
minor pieces and get castled, for ex­ sary detail; Black's development is so
ample 6 . . . lil f6 7 0-0 .te7 8 :el and dreadful that it would not be amiss to
74 SECRETS OF PRACTICAL CHESS

condemn his position without con­ has an extra central pawn, and the po­
crete analysis. sition of his queen will turn out to be
Having done away with 3 . . . lbc6, I quite useful if he castles kingside, as
turned my attention to the main line, he will have instant pressure along the
3 .. .'ir f6, again focusing my efforts on f-file . Indeed, it quickly becomes ap­
the line I had played decades earlier, parent that Black will have a fine game
namely 4 lbc4 fxe4 5 lbc3. In those if he has a few tempi to complete his
far-off days the main continuation was development. After . . . .td6, . . lbe7 and
.

5 . . . .. g6, after which I racked up quite . . . 0-0 his king is safe, and the white
a few wins with 6 d3 i. b4 7 dxe4 knights are poorly placed. Black's
'iFxe4+ 8 lbe3 . Kosten considers this central pawns prevent the knights from
very good for White, and his main line moving forwards and the knight on e3
is 5 . . . °ii'f7 , a move which was practi­ blocks in the c l -bishop. A quick look
cally unknown in 1 970. I was aware at Kosten's analysis bears this out; in
of this move, and indeed had faced it most lines White feels obliged to sac­
once quite recently against Fritz in a rifice a piece by lbexd5 or lbc4 to dis­
man vs computer tournament. I won turb Black before he can castle. All
that game, but my handling of the these lines are very complicated so,
opening was rather cautious due to invoking DAUT again, I decided to re­
lack of knowledge. turn to the diagram position. I imme­
Kosten' s main line is 6 lbe3 c6 ! (D) diately wondered why White should
(his exclamation mark). not take the pawn on e4. After 7 lb xe4
d5 8 lb gS 'ii'f6 9 lbf3 I was again feel­
ing confident about White ' s posi­
tion. White has admittedly moved his
knights several times, but Black can­
not really boast about this, as so far the
only piece he has moved is his queen.
In terms of development, the two sides
are roughly equal and White's posi­
tion is without weaknesses - in fact,
Black's compensation for the pawn is
simply invisible. Kosten gives two
lines: 9 . i.e6, aiming to castle queen­
. .

side, and 9 . . . .t d6. Taking 9 . . . i.e6

His principal variation runs 7 d3 first, Kosten's continuation is " 1 0 d4
exd3 8 .t xd3 d5. Of course, White has lbd7 1 1 i. d3 0-0-0 (D) 1 2 c3 g5 1 3
a substantial lead in development, but 0-0 h5, not fearing 1 4 lb xdS ? ! .t xd5
Black has some strategic assets. He 1 5 i. xg5 "iig7 .
"
 THE OPENING 75

This line shows a typical ploy used 2 ) 1 3 . . . g4 1 4 cxd5 cxd5 1 5 c!Lle5.

by those attempting to make a very du­ 3) 13 ... dxc4 (relatively the best) 14
bious line playable: the totally irrele­ c!Lixc4 h6 ( 1 4 ... g4 15 c!Life5 i. xc4 16
vant move. Nobody could criticize the c!Lixd7 :%. xd7 1 7 .t xc4 is also very bad,
natural developing moves I O d4 and for example 1 7 . . . :%. xd4 1 8 'ii'b 3 c!Lle7
1 1 i. d3, but what on earth is 1 2 c3 for? 19 i.e3) 15 .t e3, to be followed by
The d4-pawn isn't attacked, it isn't some combination of °ii'a4, :%.c l and
likely to be attacked and White isn't c!Life5. Black is a pawn down for noth­
freeing his f3-knight to move any­ ing.
where. If White does nothing, then of If Black plays some other 1 2th
course Black will eventually work up move, for example 1 2 . . . i. d6, White
a serious attack on the kingside, but in again plays 1 3 c4 and in this case mat­
opposite-side castling positions speed ters are even worse as c!Lixc4 comes
is of the essence. The correct plan for with gain of tempo.
White is to play c2-c4 to develop coun­ Kosten's other line is 9 .t d6, and
...

terplay in the centre and on the queen­ here he gives IO d4 c!Lle7 1 1 c4 ! , which
side. is indeed White's strongest line. The
Suppose White starts with the most continuation is 1 1 . 0-0 12 'ii'b 3 'iii>h8 ! ?
. .

obvious move, 12 0-0, and Black plays (D) 1 3 i. d2 ! ? dxc4 1 4 .t xc4 c!Ll d7 1 5
as in Kosten's line with 12 . . . g5. White 0-0, which he assesses as favouring
plays 1 3 c4 (D) and Black is already in White (and therefore recommends
a very bad way: 9 . . . i.e6 as above).
1) 13 ... .td6 14 'ifa4 �b8 15 c5 .tc7 The point of 13 i. d2 is to prevent a
1 6 b4 with an extra pawn and a very check on b4, and therefore to threaten
strong attack. If Black plays 16 . . . a6, to win a pawn on d5 . It is also possi­
White can either sacrifice on a6 imme­ ble to play 13 i. e2 followed by 0-0,
diately or prepare it by 1 7 :%.b l . when White would also threaten the
76 SECRETS OF PRACTICAL CHESS

both honest and a strong player will

have some doubts about the lines he is
giving, and this often manifests itself
in phrases which are the chess equiva­
lent of a disclaimer in a contract.
After "6 . . . c6 !" in the above line
Kosten writes "Calmly leaving the
e4-pawn to its fate; what exactly does
Black obtain in return? Objectively,
very little: just a slight lead in develop­
ment and a lot of fun." Not exactly
confidence-inspiring for Black, but of
d5-pawn, and would have achieved course having taken on the book he
this purpose using only natural devel­ can hardly admit that the whole open­
oping moves. Therefore this line may ing is just unsound. The publisher
be slightly more accurate. Once again, would doubtless not be amused by a
Black's compensation cannot be de­ manuscript consisting of a ten-page
tected. The position resembles one refutation ! Similar key phrases are
from a normal opening such as the 'might appeal to a tactically-minded
Petroff, except that Black's f-pawn has player' , 'offers practical chances' and
unaccountably dropped off the board. ' Black's position is no worse than in
main-line openings such as the Ruy
When looking through 'dubious Lopez' . Depends which variation, I
opening' analysis, look out for the fol­ suppose !
lowing: Less honest authors are entirely
1 ) 'Nothing moves' by the oppo­ shameless about such matters. They
nent (i.e. the one facing the dubious recommend the most outrageously
opening), which only waste time. unsound lines without blushing even
2) Lines in which the opponent slightly. They would never play such
pretends he is in the nineteenth cen­ lines themselves, of course.
tury, co-operatively grabs all the ma­ My second example is also a rela­
terial on offer and allows a brilliant tively high-quality product from a reli­
finish. able GM author, Winning With the
3) Lines which are given without Giuoco Piano and the Max Lange At­
any assessment. tack by Andrew Soltis (Chess Digest,
4) Secret code words. 1 992).
Books with 'Winning With' in the
The fourth point perhaps requires title confer an extra responsibility on
some explanation. An author who is the author, in that the result of the
 THE OPENING 77

author's research is pre-determined. After 1 3 . . . 0-0 14 lLl xh7 (D) Black

If the author reaches page I 00 and can, of course, accept the sacrifice.
discovers that the opening he is rec­
ommending is unsound, then he is un­
likely to abandon the whole project.
Of course, this probably won't happen
if the opening chosen is a popular,
main-line system. 'Winning With the
Ruy Lopez' is an uncontroversial title;
many top GMs regularly play the Ruy
Lopez hoping to win. The problems
start to arise when 'Winning With. . .' is
combined with an offbeat and un­
popular opening, such as the Giuoco
Piano. You can be sure that if the es­
tablished theory of the Giuoco Piano Traditional theory has it that this is
favoured White, then lots of GMs a forced draw. Soltis offers some new
would play it, but of course it does not. ideas in the extremely complex posi­
The author is therefore undertaking to tions which result but, taking DAUT
discover something new which over­ into account, we will not look at these.
turns the prevailing opinion. Then the In any case, I think many players would
crucial question is how convincing prefer to avoid 14 . . .� xh7 if at all pos­
these new ideas are. sible; it has the defect that White can
The line of the Giuoco Piano rec­ force perpetual check virtually at will,
ommended in the above book runs I and of course a difficult defence in
e4 e5 2 lLlf3 lL!c6 3 i.c4 i. c5 4 c3 lLlf6 which a slight slip might be fatal is
5 d4 exd4 6 cxd4 i. b4+ 7 lL!c3 lL! xe4 8 not to the taste of many players. The
0-0 i. xc3 9 d5 (this is called the M9ll­ reasonable alternative to taking the
ler Attack) 9 . . . i. f6 10 : el lL!e7 1 1 knight is l 4 . . . i. f5, when the following
l:he4 d6 1 2 i. g5 i. xg5 1 3 lLl xg5 and moves are virtually forced: 1 5 lt h4
now there are basically two varia­ ( 1 5 lt xe7 "flxe7 1 6 lLlxf8 : xf8 is en­
tions. The old continuation is 13 . . . 0-0, tirely comfortable for Black) 1 5 . . . : e8
when White replies l 4 lLlxh7. The more 16 "f#h5 lL!g6 1 7 %:[d4 (D).
modern option, and the one which has Now Soltis gives l 7 . . . l:r.e5 1 8 f4
really put people off the M9lller At­ lL!xf4 ! 1 9 l:r. xf4 i. g6 20 "f#f3 � xh7 2 1
tack, is 13 . . . h6 ! , which leapt to promi­ i. d3 "which i s described as unclear by
nence following Barczay-Portisch, Harding and Botterill - a judgement
Hungarian Ch 1 968/9. Let's take these that apparently has yet to be tested.
two lines in turn. However, it is not easy for Black to
78 SECRETS OF PRACTICAL CHESS

situation. The first important point is

that Soltis notices, as Botterill does not,
that 2 1 . . . .i. xd3 loses Black's queen af­
ter 22 'iih 3+ � g8 23 lt h4. However,
Soltis does not add that the position
after 23 . . .'ii'xh4 24 'ii'xh4 lt xd5 offers
White essentially zero winning pros­
pects. Soltis in fact mentions Black's
best move, 2 1 . . .'ii'e 7, but then gives 22
'iih 3+ and stops (see point 3 above).
After 22 . . . 'it> g8 (D) it is quite im­
possible to see how White can even
defend against the threat of doubling equalize.
on the h-file, e.g. 2 1 . . . i. xd3 22 'iih 3+ !
� g8 23 lt h4 or 2 1 .. . 'iie 7 22 'ii h 3+
(not 22 lhf7? lt e l + 23 'it> t'2 'ii' h4+).
Black's best appears to be 2 1 . . . 'it> g8
22 i. xg6 fxg6 but 23 lt fl retains an
edge."
The Harding and Botterill book
quoted is The Italian Game (B .T. Bats­
ford, 1 977). However, in the later book
Open Gambits (B.T. Batsford, 1 986),
Botterill gives "20 'iif3 'it> xh7 2 1 i.d3
J. xd3 22 °ikxd3+ � g8 ... it is question­
able whether White has enough for the
pawn. However White can hold the 23 lt h4 is met by the simple tactical
balance with 20 'ii h3 'ii'c8 (Vukovic) point 23 . . . lt h5 ! 24 lt xh5 .i. xh5 25
21 lbf6+ gxf6 and now: 'ii'x h5 'ii'e3+, while after 23 .i. xg6
"a) 22 'iih4 f5 (intending ... 'iid8) fxg6 Black will force exchanges by ei­
retains tension, though I think Black is ther ... lte l + or . . . ltf8, leaving White a
better. pawn down. This brings us to another
"b) Simply 22 'iixc8+ lt xc8 23 warning:
lt xf6 is equal."
5) 'Winning With' authors display
What is the poor reader to make of great ingenuity in finding resources
all this? Soltis and Botterill give com­ for 'their' side, but often overlook
pletely different lines and offer differ­ even quite simple tactical defences for
ent evaluations. I will try to clarify the the 'other' side.
 THE OPENING 79

Botterill's other line for White, 20

'i!kh3, can be met by 20. . .'iWc8, although
I would feel a bit queasy about allow­
ing my kingside pawns to be broken
up by ll'i f6+ and then blocking in my
bishop by . . . f5. I would prefer to play
20 . . . l:r. h5 ! , which again depends on a
small tactical point: 2 1 ll'i f6+ 'ikxf6 !
(avoiding any damage to the kingside
pawns) 22 .:r. xf6 .:r. xh3 23 l: xg6 l:r.h4 !
and Black reaches a four-rook ending
with a clear extra pawn. White should
prefer 2 1 'ii' b3 'iii> xh7 22 'ii'xb7, when 2) 18 l:r. g4 ll'ie5 19 l:r. h4 f6 20 i. b5
22 . . . 'ii'b8 reaches a more or less equal i. g6 wins material.
ending, while 22 . . . .:r.b8 23 'ii'xa7 l:r.xb2 3) 1 8 .:r.c 1 l:r. e5 1 9 'iWh3 i. f5 20 g4
is unclear, Black's more active pieces 'ii'h4 ! and again White loses material.
balancing the long-term danger posed 4) 1 8 f4 'ii'e7 with a nasty check on
by the passed a-pawn. e3 to come.
The above analysis shows that after 5) 1 8 ll'i g5 (relatively the best
1 7 . . . l:r. e5 1 8 f4 White has a draw at move) 18 . . . 'ii'f6 19 'fl h7+ 'iii>f8 20 00
most, but I will be fair and point out an i.e4 with advantage to Black.
improvement for White: by 1 8 ll'i g5 !
'ii'f6 1 9 'ii'h 7+ <iti f8 20 'iW h5 he can Now we return to 1 3 . . . h6 ! , which
more or less force Black to accede to a most players cite as the reason for re­
repetition of moves by 20... cat g8. How­ jecting the M0ller Attack.
ever, going back one move, there is an Soltis recognizes the importance of
interesting possibility for Black to the line by devoting 1 6 pages of analy­
play for an advantage: 17 . . . i. c2! (D). sis to the position after 1 3 . . . h6. The
The point of this is quite simple. critical position arises after 14 'ii'e2
Black clears the fifth rank for his rook hxg5 15 .:r.e1 .t e6 1 6 dxe6 f6 17 l:r. e3
and threatens (amongst other things) (D).
18 . . . .:r.e5 19 'ii'h 3 'ii'e 7, followed by Soltis boldly states ''This move, at­
. .. l:r. e8, with a winning position since tributed to the Finnish correspondence
White is completely hamstrung by the player Juhani Sorri, rehabilitates the
trapped knight on h7 and can hardly new M0ller." We will take a look at
move a piece. two possibilities for Black, 1 7 . . . c6 and
The alternatives are: 1 7 . . . <itif8.
1 ) 1 8 l:r.d2 loses to 1 8 . . . ll'if4 19 'ii'g4 After 1 7 . . . c6 play proceeds 1 8 l:r. h3
.l xh7 20 'iWxf4 l:r.e4. l:r. xh3 1 9 gxh3 g6. Soltis surveys the
80 SECRETS OF PRACTICAL CHESS

�f7 wins for Black) 23 . . . <iti g7 Black

cannot be prevented from consolidat­
ing with ... 'ii'a5 followed by . . . .:r.h8, or
. . . d5 followed by . . . 'ii'd 6 and . . . lt h8
(note that 24 l:r.e3 loses to 24 . . . lLJf5). It
seems to me that Black is virtually
winning.
After l 7 . . . cat f8, Soltis's main line
runs 1 8 i. d3 cat g8 19 'ii'c2 l: h6 20
l: g3, and ends with the rather weak
comment that "White will now follow
with either l:r.ee3 and l:r.h3, or by
moves which have been tried in prac­ preparations for f2-f4". However, af­
tice and concludes that they favour ter 20 . . . d5 it is hard to see either of
Black. His suggestion is 20 'ii'd2 (D). these plans being effective, for exam­
ple 2 1 l:r. ee3 'ii'd6 22 .:r. h3 l:r. xh3 23
gxh3 (23 lt xh3 'ii'xe6 24 'ii'xc7 l:r. c8)
23 ... .:r.es or 21 l: fl 'ii'd6 preventing
White's f4. Again there are no practi­
cal examples.
This brings us to our next warning:

6) Do not trust lines which are not

based on practical examples. The more
examples there are, and the higher the
standard of the players, the more trust
you can place in the line.

White prevents . . . 'ii'a5, and threat­ The above analysis shows up the
ens 2 1 'ii'c3. Soltis analyses 20... d5, but weakness of the 'Winning With' ap­
he doesn't consider the reply 20. . . �f8. proach applied to offbeat openings.
intending . . . � g7 followed by gradual Theory in the Giuoco Piano gives a
consolidation with . . . d5 and . . . 'ii'd 6, unanimous thumbs up to Black; an
when White will be a pawn down in a author claims to have discovered
bad position. White's only active move something new which rehabilitates
is 2 1 h4, but after 2 1 . . .gxh4 22 °ii'h6+ one line. You build your entire reper­
<iti g8 23 'i!kxh4 (23 l:r.e4 fails to 23 . . .'ii'f8 toire on the basis of this claim. With
24 'i!Yxh4 d5, while 23 'ii'x g6+ lLJxg6 the Soltis book, this involves learning
24 e7+ d5 25 exd8'ii'+ lt xd8 26 Ji. fl something against the Two Knights
 THE OPENING 81

Defence and against the various other turns out that there is something wrong
systems Black might employ in the with 20 'ifd2. What do you do? It is no
Giuoco Piano. In fact your whole rep­ good switching to another line in the
ertoire is founded on one move, 20 Giuoco Piano, as these have long been
'ii'd 2, which has never been tested in dismissed as offering White nothing.
practice and which is analysed for just In fact, you may as well dust off your
over half a page. Then it suddenly books on the Ruy Lopez.
3 The M idd lega me

This is the part of the game in which it counterplay. It is no good starting

is hardest to set down rules and give some leisurely manoeuvres if your op­
good advice. For every rule one pro­ ponent has a passed pawn thundering
poses, there will be so many excep­ down the board. Only if both condi­
tions that the rule may be more tions are satisfied - a long-term advan­
misleading than helpful. Accordingly, tage and a lack of counterplay - can
we are going to deal more with the you afford to indulge in luxuries such
psychological side of the middlegame as improving your king position. How­
than with technical issues. ever, in this case you not only can take
your time, you should do so. You may
Good positions not be able to see a concrete reason
why a particular 'tidying-up' move
Congratulations, you have the advan­ might be useful, but you lose nothing
tage ! But what do you do next? The by playing it.
first step is to judge whether your ad­
vantage is of a short-term or long-term
nature. If it consists of better develop­
ment or attacking chances, then it is
probably short-term; if it is based on
better pawn structure or superior minor
pieces, then it is probably long-term.
Many advantages, such as control of
an open file, may be either short- or
long-term depending on whether the
opponent has a means of challenging
the asset. This step will provide you
with a clue as to whether you should
be thinking of quick action to exploit T. Petrosian W. Unzicker
-

your advantage before it disappears, USSR - West Germany match,

or whether you can afford to manoeu­ Hamburg 1960
vre slowly, tidying up your position
ready for further action. White's absolute domination of the
It is also important to take into ac­ c-file and pressure against the vulner­
count your opponent's possibilities for able a5-pawn give him a long-term
 THE MIDDLEGAME 83

advantage, and Black is so tied up as to 39 f4 'ito>h7 40 'ii'e 2 'ii'b 7 41 g4!

have no realistic prospects of counter­ hxg4 42 'ii'xg4 "ii'e7 (D)
play. Nevertheless, White still has to Now 42 . . . l:r. a6 may be met by 43
come up with a plan for improving his l:r. 6c2 ! and, thanks to f4, White can
p osition. He decides that a kingside swing both rooks to the kingside, giv­
pawn advance is the best way to pro­ ing a winning attack.
ceed, but if played immediately this
would expose his king. Petrosian there­
fore decides to play his king to the se­
cure refuge on a2 before proceeding
with his kingside plan. Only a com­
plete lack of counterplay could justify
such an extravagant manoeuvre, but in
this particular position White has ab­
solutely no need to hurry.
29 <itin <itigs
After 29 . . . l:r.b8, White would make
progress by 30 l:r.b6 l:r. d8 3 1 l:r.cc6 i. c7
32 .l:r. a6.
30 h4 hS 43 hS "ii'f6 44 'iii>a2 <itig7 4S hxg6
A very unpleasant decision. This 'ii'xg6 46 'ii'h4
gives White the chance of eventually The open files on the kingside, cou­
opening up the kingside with g4, but pled with all Black's other problems,
meeting h4-h5 with . . . g5 would have a prove too much.
similar effect, since White could later 46 i.e7 47 'ii'f2 <itif8 48 ll'id2 l:[b7
••.

play f4. 49 ll'ib3 l:r.a7 SO 'iFh2 i.f6

31 l:r.lc2 �h7 32 'iii>e l �g8 33 'ito>dl After 50 . . . i. d6 White wins by 5 1
�h7 34 �cl �g8 3S �bl 'iii> b7 36 l:xd6 ! l: xd6 52 f5 exf5 5 3 l: c8+ �e7
'ii'e2 'ii'b7 37 .:r.ct �g7 38 "ii'b S! 54 'ii'h 8.
Petrosian observes that 38 g4 hxg4 Sl l:r.c8 l:r.ad7 S2 lllc S! b3+ S3
39 'ihg4 l:r. a6 40 l:r. xa6 'iFxa6 41 h5 catxb3 l:[d6 S4 fS ! l:r.b6+ SS �a2 l-O
'ii'd3+ allows Black to develop some
activity, so he arranges to play f4 be­ One common error in prosecuting
fore starting his kingside play with g4. an advantage is to indulge in tactics
38... 'ii'a8 for the sake of it. There is a general
Or 38 . . . 'ii'xb5 39 axb5 a4 40 b6 feeling that the ideal game is one in
.:r.ad7 41 ll'ia5 l:r. a8 42 l:[ xd6 ! l:r. xd6 43 which you positionally outplay your
b7 l:[b8 44 l:r. c8 l:r. d8 45 l:[ xd8 l:[ xd8 46 opponent and then finish with a burst
lllc6 and White wins. of scintillating tactics. In fact, the
84 SECRETS OF PRACTICAL CHESS

ideal game is one in which you win

without allowing any unnecessary
counter-chances. If you can win with
a clear-cut combination, then that is
the best way to eliminate counterplay,
because the game is over ! However,
you should be really sure that your
combination does work; it is quite
easy to miscalculate complex tactics,
and if there is an element of doubt then
a purely strategic approach is proba­
bly better. Even if your combination
'works' (i.e. is not tactically unsound) J, Bellon - J. Nunn
then you should also be sure that your Zurich 1984
advantage in the resulting position is
greater than it is in the position before­ when a list of the threats shows just how
hand. A common error is to play an awful Black's position is: d6, fxg6,
elaborate tactical line gaining, for ex­ i. g5, .t c4, �g5. Consequently, Black
ample, the exchange for a pawn, only would have to try 20 . . . �f6 (20 . . .e4 2 1
to discover that there are fewer win­ fxg6 hxg6 2 2 d 6 and 2 3 i. c4 wins),
ning chances after the combination but 2 1 fxg6 hxg6 22 .t g5 e4 23 l:l ce l
than before. is crushing. One possible continuation
In the following example White, would be 23 . . . c4 24 .t xc4 � c5 25
with a strategically winning position, 'ii'g 3 �h5 26 'ii'e3 .t f6 27 .t xf6 � xf6
spotted a seductively beautiful combi­ 28 'ifd4 �cd7 29 � xe4 � xe4 30 d6 !
nation (D): and Black's position is tom to shreds.
This position is a Benoni gone hor­ Bellon's move gave me a momen­
ribly wrong for Black. White's pieces tary glimmer of hope, but then I real­
are all actively placed and he has the ized he intended 20 .. .'if xd6 2 1 �e4
two bishops, while Black's a8-rook is 'ife7 22 f5 with the same type of posi­
useless and his knights poorly placed. tion as after 20 f5. White, it is true, has
19 e5! given up a second pawn, but his knight
This is the thematic Benoni break­ has arri ved on e4 with gain of tempo.
through, so it came as no surprise. A quick look convinced me that there
19 dxeS 20 d6
••• was no real answer to the threats to f7
But this was unexpected. I had an­ arising after fxg6, .t c4 and, if neces­
ticipated the natural 20 f5 when Black sary, � g5. However, there was obvi­
is probably just lost. White threatens ously nothing better than to take the
to complete his blockade by �e4, pawn.
 THE MIDDLEGAME 85

20 'iixd6 21 'iixf7+? (D)

••• made the mistake of not asking the
question 'Is my advantage after the
combination greater or smaller than
my current advantage?'
21 ...ci;xf7 22 fxeS+ ci;g8
I would have preferred to approach
the passed d-pawn that is about to be
created, but after 22 . . . � e6 23 exd6
i. xb2 (23 . . . cat xd6 24 lbc4+ cate6 25
i. xd7+ � xd7 26 l:. f7+) 24 lt cd l the
threats against Black's exposed king
are too strong, e.g. 24 . . . i. d4 25 i. xd4
cxd4 26 lbf3 lbe5 27 l:. fe l l:. c5 28
i. xa6 bxa6 29 .:r. xd4 and Black cannot
I hadn' t seen this queen sacrifice at both save his pinned knight and stop
all, and when he touched his queen I the d-pawn.
couldn't imagine where it was going. 23 exd6 lbeS 24 lbe4
A second later I found out! It is of White's two active bishops, better
course desirable to avoid revealing to development and dangerous d-pawn
your opponent that you have over­ are undoubtedly enough to win, but al­
looked something, but in this case my though there are many very promising
body language must have broadcast continuations, there is no way to finish
the message far and wide. Black off instantly. After 24 i. xa6 bxa6
The question mark attached to this 25 .:r.xc5 (25 i. xc5 lbd3) 25 . . . .:r. xc5 26
move is perhaps a little harsh, because i. xc5 lb d3, for example, Black re­
White retains a winning position even gains the pawn on b2 (White should
after it. Yet from the practical point of avoid 27 b4? lbxc5 28 bxc5 i. d4+ ). In
view it is undoubtedly a mistake. 2 1 view of the doubled a-pawns, White
lbe4 would have led to a position in would be effectively almost a pawn
which White's win would require only up, but the win would not be guaran­
straightforward, obvious moves - he teed.
just has to take aim at f7 and Black 24 c4
..•

will soon collapse. The move played Keeping the pawn for the moment.
alters the character of the position 25 i.gS (D)
completely and presents both players 25 i. f4 is also very strong.
with new problems to solve. 25 . lbbS
..

White had noticed an attractive When you are desperate, it is some­

combination without tactical flaws times necessary to play ugly moves.
and leading to a won position. He then The threat of 26 d7 can only be met by
86 SECRETS OF PRACTICAL CHESS

26 . . . .!LJbd7 27 bxc4 a6 28 .t xd7 .!LJxd7

29 c5 ! i..d4+ 30 'iii>h l i.. x c5 3 1 .!LJxcS
l:1 xc5 (3 1 . . . .!LJ xcS 32 l:t fd l ) 32 l bc5
.!LJxc5 Black can regain the pawn, but
White wins by 33 lt c l .!LJd7 (33 . . . b6
34 l:1 xc5) 34 l:1c7.
It is clear from this variation that
White's queen sacrifice made the win
considerably more difficult; instead of
a clear-cut and straightforward posi­
tional win, White now needs to calcu­
late quite lengthy lines.
moving the a6-knight, but 25 . . . .!LJcS 26... .t:xf6 27 .txf6 a6!
loses to 26 .t xc4+ .!LJ xc4 27 l hc4 Again presenting White with the
i.. xb2 28 .!LJ xc5 l: bc5 29 l:1 xc5 i.. d4+ maximum difficulty. After 27 . . . .!LJbd7
30 <ithl .t xc5 3 1 d7 i.. b6 32 :c t and 28 l:tce l a6 29 .t xd7 .!LJ xd7 30 l:1e7
32 l:1c8, so the move played was the .!LJ xf6 3 1 l:1 xf6 the passed pawn and
result of a process of elimination. threat to double rooks on the seventh
If your position is objectively lost, rank would give Black no chance.
the most important rule is ' Keep the 28 i.. a4 .!LJbd7
game going' . This doesn' t mean that At last Black has managed to set up
you should play on for a long time in a some sort of blockade for the d-pawn,
resignable position; it means 'do not but it is not permanent because the
allow your opponent a simple forced knights lack any pawn support.
win' . The longer you can force your 29 :eel .!LJc6 (D)
opponent to work, the greater the Possible thanks to the interpolation
chances he will eventually go wrong. of . . . a6 and .t a4.
In the diagram position, White would
be looking for a knock-out blow and
would probably reject lines in which
he is 'only' a pawn up. Denying him a
quick win will eventually cause frus­
tration, a loss of objectivity and a pos­
sible error.
26 .!LJf6+? !
Once again this does not throw
away the win, but a pragmatic player
would have contented himself with
the gain of a pawn by 26 b3 . After
 THE MIDDLEGAME 87

30 i.c3
White decides to keep his bishops.
Once again there was a tempting alter­
native in 30 i. xc6 :xc6 3 1 :e7 �xf6
32 : xf6 :ds 33 l:tft7 lkxd6 34 l:t g7+
�f8 35 : xb7 (35 l:tef7+ �es 36 l:txb7
l: Sd7 offers Black fair drawing pros­
pects), but after 35 . . . h5 it is hard to say
whether White's advantage will nec­
essarily lead to a win.
30 bS
•••

The queenside pawn majority is

Black's only hope of counterplay. original notes (published, for exam­
31 i.dl! ple, in lnformator) I gave two possible
The correct plan; if the bishop can wins for White, only one of which
reach the b3-gS diagonal (via f3 or g4) works. I suppose I should be thankful
then Black will be mated. that I got 50% correct !
31 b4 32 i.g4
•.• The first 'win' was 34 i. xc6 l:t xc6
The simple move 32 i. d2 would 35 d7 and now 35 . . . l:t ccS 36 dxcS 'ii'+
also have been very awkward. Black's : xcS 37 :f2 c3 3 S l:t c2, followed by
queenside pawns aren't going any­ l:tb l . However, 35 . . . <it> g7 ! draws after
where because pushing them would 36 l:teS (36 l:te7+ <it>h6 37 :es doesn't
open the diagonals leading to Black's make any difference - Black can still
king. Thus White has time to play play 37 . . . l:tf6) 36 .. J:lf6 37 :bt c3 3S
i. f3-d5 or i. a4 and exploit his great l:tcS ! (White is slightly lucky even to
piece activity. draw ! ) 3S . . . l:td6 39 l:t xc3 l:tdS 40
32 ... bxcJ 33 i.xd7 cxb2 (D) l:txb2 l:. Sxd7 drawing.
Black's first real glimmer of coun­ My second line does indeed lead to
terplay, and the decisive moment of a win: 34 i. xcS ! l:t xcS 35 d7 l:t dS
the game. (35 . . .l:t c7 36 l:te8+ �g7 37 l:te7+ �h6
34 i.e6+? 3S dS'ii' �xdS 39 l:t xc7) 36 l:te6 ! (this
The earlier complications, coupled is the tricky move to see) and there is
with White's fruitless search for a no defence to the threat of l:t xc6.
knock-out, had left him short of time. 34...<it>g7 35 i.xc4
Just at the moment when he could It is now too late to take on cS: 35
have really used a few extra minutes, i. xcS l:t xc8 36 d7 l:tbS 37 l:teS bl 'ii' 3S
he was forced to move quickly and l: xbl l:t xbl + 39 � f2 l:t b2+ 40 <it> g3
threw the win away. The complexity l:t xa2 41 l:t cS l:t d2 42 l:t xc6 l:t xd7 43
of the position is such that in my l:t xc4 a5, followed by . . . l:t a7, gives
88 SECRETS OF PRACTICAL CHESS

Black a rook ending with excellent by 39 . . . lLJbS ! 40 a4 :t e l ! 41 axb5 l:tf8

winning chances. Besides, having re­ and the fl -rook is caught in a pincer­
fused to take on c8 the move before, it movement.
would be surprising if White changed 39 .:tcl (D)
.••

his mind now.

35 ... lLJd4
Suddenly everything starts to go
wrong for White. The bishop has to re­
treat along the a6-fl diagonal, since
36 i.. d 5 (36 i.. d 3 loses at once after
36 . . .l:C l ) 36 . . . l:t ab8, threatening to
play 37 . . . .:t c l , is very bad for White.
36 i.d3 l:tc3 37 i.. b l l:td8
Black can force a draw by 37 ... .:tc l
(intending 3 8 . . . l:td8) 38 i. d3 (cover­
ing e2 and so threatening 39 l:t xc l )
3 8 ... .:tc3, but now that things are turn­
ing his way he decides to play for a 40 i.. d3
win. If Black can take the d-pawn, White is almost without a reason­
then his own passed b-pawn will de­ able move, for example 40 h4 l:t f8
cide the game. (this is Black's main threat) 4 1 l:teel
38 l:te7+? (4 1 l:tfe l lLJ c6 42 l:te8 .:tf7 wins the d­
There are very few players who de­ pawn) 4 1 . . . lLJe2+ 42 <li>h2 l:td8 43 i.. d3
fend well when they have slipped from l:t xd7 44 i.. x e2 l:t xe l 45 l:t xe l l:te7,
a winning position into an inferior one. followed by . . . l:t xe2, with a winning
It requires exceptional self-control to rook and pawn ending.
forget about what has gone before and 40... lLJc2 41 i.xc2 0-1
j ust to concentrate on the current After 4 1 . . . l:t xc2 42 l:tb l :t e l + 43
situation. White could still have held :tel l:txe l + 44 l:t xe l l:t xd7, followed
the game by 38 l:t:f2 ! :t e l 39 l:tefl with by . . . .:tb7 , White's rook becomes im­
a dead draw after the exchange of b­ mobilized on bl and Black has an easy
and ct-pawns. The move played is a re­ win.
flex action to preserve the d-pawn, but
White should never have abandoned Bad positions
the first rank, except to attack the
pawn on b2. The first piece of advice is simple:
38 <it>h6 39 d7
••• don't give up hope. The history of
In my Informator notes I gave 39 chess is littered with won positions
l:tb7 as drawing, but in fact Black wins thrown away. Even world champions
 THE MIDDLEGAME 89

have been known to do this, and it is confusion' approach may be better. If

much more common at lower levels. you are aware of your opponent's
However, you should not just hope style, then this might also affect your
that your opponent is going to make a decision. One important point is that
mistake - you have to help him to do you should not change plans in mid­
so. Being determined to make the win stream. If you have decided on ' grim
as difficult as possible is the first step defence' then you should not lose pa­
in the right direction. A stiff resistance tience and switch plans later unless
is almost always unnerving for the your opponent carelessly allows you a
side with the advantage - he may well chance to break out. 'Grim defences'
have over-estimated the strength of his most often fail because the defender
own position and be anticipating a creates unnecessary additional weak­
quick finish. nesses, opening the door to the oppo­
There are two basic strategies when nent's pieces. The whole point of
confronted with a bad position. The 'grim defence' is to set up a solid posi­
first is to find some way to hang on, of­ tion which your opponent has to work
ten by liquidating to an endgame. The hard to break down; such positions do
attacker may not fancy winning a long not lend themselves to the creation of
endgame a pawn up and so may un­ active counterplay.
wisely continue to seek a middlegame If you decide to go for 'create con­
win. Even if he does go for the end­ fusion' then you should press the
game, a sudden switch from tactical panic button sufficiently early to give
middlegame play to technical end­ yourself a reasonable chance of suc­
game play can often prove disorientat­ cess. However, you should be sure that
ing . We call this the ' grim defence' your position is really bad enough to
response. warrant such drastic measures. In my
The second strategy is to seek to experience, it is far more common to
gain the initiative, even at material panic too early than too late.
cost, hoping to stir up complications Here is an example of each type of
and cause the opponent to go wrong. strategy, in both cases in response to a
We call this the 'create confusion' re­ stunning opening novelty.
sponse.
The choice between these alterna­ J. Nunn - W. Browne
tives depends mainly on the position Gjr/Jvik 1 983
on the board, but other factors can en­ Sicilian, Najdorf
ter the calculation. For example, if
your opponent is short of time he may 1 e4 cS 2 �f3 d6 3 d4 cxd4 4 �xd4
welcome simplification to an end­ �f6 5 �c3 a6 6 .tgS e6 7 f4 i..e7 8
game, so in this case the 'create 'iff3 'ilc7 9 0-0-0 �bd7 10 g4 bS 1 1
90 SECRETS OF PRACTICAL CHESS

.i.xf6 .!LJ:xf6 12 g5 .!LJd7 13 rs .!LJcs 14 though Browne failed to win the game,
h4 b4 15 .!LJce2 his novelty was voted the most impor­
In retrospect, one can say that White tant of the second half of 1 983. As the
should choose something else round similar vote in the first half of 1 983
about here, for example the sacrifice gave a lower score to the winner, it
1 5 fxe6 !?. would be fair to say that the most stun­
15 e5 16 .!LJb3 (D)
••• ning novelty played anywhere in the
world during 1 983 had just landed on
my board.
After I had recovered from the
shock, I settled down to decide on the
best reply. I saw that I could take the
knight, which leads ultimately to a
better ending for Black, or I could try
1 7 .i. g2, which sacrifices a pawn for
not very impressive compensation.
My choice was determined largely by
practical factors. I realized that Browne
would have analysed both these lines
carefully at home, so this was cer­
I was feeling quite happy with the tainly a crucial decision. I had spent a
opening. In previous games Black great deal of time sitting at the board
had invariably continued 16 . . . .i. b7, calculating these alternatives and so,
and I had analysed the resulting posi­ unusually when facing Browne, I was
tions and concluded that they favoured far behind on the clock. In view of the
White. time situation, I doubted my ability to
16 .!LJxe4!!
••• find my way through continuing com­
I was dumbfounded when this move plications which Browne would have
appeared on the board and for several analysed at home. I therefore decided
minutes I just couldn't see the point of to go for the ' grim defence' option,
it. After 17 Wxe4 .i. b7 1 8 :%.d5, fol­ when the many possibilities for both
lowed by .i. g2 if necessary, everything sides would inevitably mean leaving
seemed to be fine for White. Sooner or Browne's analysis within a few moves
later Black would have to take on d5 of entering the ending. Moreover, my
when White would have complete lack of time would be a less relevant
domination of the light squares. Then factor as the complications would be
I suddenly saw the idea (which will be less intense.
revealed in the game continuation) As an aside, there is very little you
and realized I was in trouble. Even can do to prevent the occasional
 THE MIDDLEGAME 91

shocking opening novelty - if you Tlris move is the key point which it
play sharp openings then it is an occu­ took me several minutes to see at move
pational hazard. 1 6 . The upshot is that Black gains a
17 'ifxe4 rook and two pawns for two knights.
In a later game Wedberg-de Fir­ In an ending, a rook and two pawns is
mian, Oslo 1 984, White did indeed try worth more than two minor pieces at
1 7 i. g2 i.b7 1 8 'iie3 d5 1 9 i. xe4 dxe4 least 90% of the time. Two bishops
20 tl'l g3 aS, but lost after great compli­ may sometimes hold the balance, but
cations. two knights have almost no chance.
17 i.b7 18 ltd5 ltc8 19 c3!
•.• 23 ltgl !
The only move, since 19 'it>bl i. xd5 White must try to keep Black's
20 °ikxd5 'ii'x c2+ 21 � al 0-0 ! leaves bishop passive. The material situation
White completely tangled up. He can would be the same after 23 lt xh4
try 22 f6 gxf6 23 gxf6 i. xf6 24 lt g 1 + il. xg5+ 24 �c2 il. xh4 25 cxb4, but
'it> h8 25 'iixd6, but after the simple Black would have a far easier time.
25 . . . il. g7 White has not solved any of His bishop can emerge via f2, his h­
his problems. pawn is already passed and his king
19 'iWc4 20 'ii'xc4
..• can obstruct White's queenside pawns
Not 20 i. g2 il. xd5 2 1 'iixd5 'iWxe2 by . . . 'it>d7-c7.
and Black wins. 23... bxc3
20 ltxc4 21 il.g2 il.xd5!
.•. After 23 . . . lt h2? 24 il. c4 ! White
Black must take straight away, be­ gains time by attacking the a-pawn.
cause 2 1 . . . lt xh4? 22 l1 xh4 i. xd5 (or 24 tl'lxc3 l1f4!
22 . . . i. xg5+ 23 l1d2 ! ) 23 lt g4 gives Black fastens onto the weak king­
White an extra piece. side pawns.
22 i.xd5 ltxh4! (D) 25 il.c6+!
An important finesse. After 25 il.e4
�d7 Black develops his h8-rook eas­
ily, while on 25 g6 fxg6 ! 26 fxg6 h5
Black has no need to develop his h8-
rook - it is already ideally placed be­
hind the passed h-pawn.
25 ...'it>fS
After 25 . . . 'it>d8? 26 tl'l a5 ! Black's
king becomes exposed, for example
26 . . . ltxf5 27 tl'lb7+ 'it> c8 28 tl'ld5
i. xg5+ 29 'it>bl and Black is obliged
to play 29 . . . .l:td8 in order to avoid a
worse fate. Therefore Black has to
92 SECRETS OF PRACTICAL CHESS

move his king the other way, but this White has made considerable prog­
blocks in the rook on h8 . ress since the last diagram in setting
26 i.e4 h5! up a light-squared blockade; of course,
After 26 . . . d5 27 i. d3 ! Black has he is quite happy to repeat moves. His
rather a lot of pawns attacked. The main problem is the insecure position
move played is best; White cannot af­ of the bishop on e4.
ford to let the pawn race down the 32 ll:lf3 .l:th3 33 ll:ld2 .l:th4!
board, so he has to exchange on h6, Black finds a way to play on. He
but then Black has activated the h8- both prevents 34 ll:lf3 and threatens
rook. 34 . . . .l:txd2.
27 gxh6 .l:txh6 28 ll:ld2! .l:th2 29 34 ll:lc3!
ll:ld5! This only apparently allows the
Utilizing a tactical point to occupy black bishop to emerge via d8 and b6.
the key d5-square. If now 29 . . . .l:t xd2, In fact, 34 . . . i. d8 could be met by 35
then 30 .l:t hl ! (30 ll:l xf4? .l:td4 3 1 .l:t hl ll:lc4.
i. g5 wins for Black) 30. . .g6 (30...�e8? 34... dS!
3 1 ll:l xf4 .l:td4 32 i. c6+ and White A combination liquidating to a rook
wins) 3 1 ll:lxf4 .l:td4 32 ll:ldS .l:txe4 33 ending favourable for Black. How­
f6 i. xf6 and Black has only a slight ever, there is saying that 'all rook end­
advantage. ings are drawn' . While this is clearly
'Grim defence' doesn't necessarily not intended literally, it contains a
mean that you should ignore tactics large element of truth. Rook endings
completely. Tactics are not the sole are often tricky to judge, because in
preserve of the attacker. one position an extra pawn may be in­
29 ... .l:tff2 30 ll:lf3! .l:th3 31 lLld2 sufficient to win, while in another, one
.l:thh2 (D) player may have a decisive advantage
despite material equality. The reason
for this is that piece activity is very im­
portant in rook endings and can often
prevent the exploitation of a material
advantage. Likewise, if material is
equal a difference in piece activity
may decide the game.
35 ll:lxd5 .l:txd2 36 �xd2 .l:txe4 37
ll:lxe7 �xe7 38 .l:txg7 l:tf4?
This wins a pawn, but allows White
to activate his king and rook. By now
Browne had become short of time
himself, and so missed the stronger
 THE MIDDLEGAME 93

continuation 38 . . . <it>f6 3 9 .l:th7 a5, aim­ A counterpart to the check on move

ing to push the pawn to a4 before go­ 25 . White seizes his chance to force
ing after the f-pawn with . . . .l:t f4. In this the black king into a more passive po­
case Black would have preserved good sition.
winning chances. 48 <it>f8 49 .l:tel .l:td2 so .l:txe3 m+
•••

39 .l:tg8! Sl <it>eS .l:txb2 S2 .l:ta3 .l:tb6 S3 <t/fS (D)

White's first chance to play actively
since the opening !
39 ....l:txfS 40 .l:ta8 .l:tf2+ 41 <it>c3 .l:tf6
The only way to keep the extra
pawn, but now Black's king is driven
back.
42 .l:ta7+ <it>f8 43 <it>d3 .l:te6 44 <it>e4
<t/g7 4S <it>fS (D)
Countering Black's threat of . . . <it>f6,
. . . d6, . . . <t/e6 and finally . . . f5+.
.l:t

The exchange of another pair of

pawns brings White closer to the
draw. In view of the active position of
White's king and rook, Black's win­
ning prospects are minimal, and now it
only requires moderate care to steer
the game safely home.
S3 <it>g7 S4 .l:tg3+ .l:tg6 SS .l:ta3
•.•

.l:tf6+ S6 <it>eS .l:tb6 S7 .l:tg3+ .l:tg6 S8

.l:ta3 .l:te6+ S9 <t/fS .l:tf6+ 60 <it>eS .l:th6
4S e4
•.• 61 <tifS .l:tc6 62 .l:tg3+ <it>f8 63 .l:tb3
The only chance is to push the e­ .l:tc2 64 ::ta3 .l:tcS+ 6S <it>f6 .l:tc6+ 66
pawn, but White can round it up once <it>fS <it>e7 67 .l:te3+ <it>d7 68 .l:td3+ <it>c7
it has become disconnected from the 69 .l:tf3 <it>d7 70 .l:td3+ <it>e7 71 .l:te3+
rest of Black's forces. .l:te6 72 .l:tb3 <tid6 73 .l:tb7 .l:teS+ 74
46 .l:tc7 e3 47 .l:tcl .l:td6 <it>f4 .l:taS 7S .l:txf7 lha2 76 <it>e4 <it>cS
The only try, since after 47 . . . e2 48 lf2.lf2
::te l , followed by <it>f4-f3, White draws
easily. This was a good example of 'grim
48 .l:tgl+! defence' . Faced with a choice of evils,
94 SECRETS OF PRACTICAL CHESS

White decided to go for the inferior had continued 1 3 exd5 exd5 1 4 i. f4

ending. After placing as many diffi­ 'ii'd8 1 5 .!LJ a4 and here a draw was
culties as possible in Black's way, he agreed. After 1 5 ... .!LJ xa4 16 l:r. xa4 :es
was eventually rewarded with a slip Black's pieces are quite active and the
allowing the half-point to be saved. rook on a4 is misplaced, factors that
The next game provides a real con­ balance the slight weakness of the iso­
trast. lated d-pawn.
13 eS! !
J. Plaskett J. Nunn
- A complete shock for me. Objec­
Lambeth Open 1979 tively speaking, this move simply re­
Sicilian, Najdorf futes Black's opening strategy. I saw at
once that after 1 3 .. .'iWxe5 1 4 i. f4 my
1 e4 cS 2 .!LJf3 d6 3 d4 .!LJf6 4 .!LJc3 queen would be in trouble, so I looked
cxd4 S .!LJxd4 a6 6 i.e2 .!LJbd7 7 i.e3 to see where I could move my knight.
.!LJcS Unfortunately, Black's natural retreat
After the present game I gave up 1 3 . . . .!LJfd7 loses a piece after 14 b4, so
this line. The idea of playing the knight Black would have to play 1 3 . . . .!LJe8.
to c5 is to exert pressure on e4, but However, then 1 4 f4 f5 15 exf6 (best;
once White has played f3 this pressure if White tries to play on the kingside
is irrelevant, and then it is hard to see with h3 and g4, then the knight on e8
what the knight is doing on c5. might actually become useful with
8 f3 e6 9 'iFd2 'f/c7 10 0-0 j.e7 1 1 . . . g6 and . . . .!LJ g7) 1 5 . . . .!LJxf6 1 6 i. f3
a4 0-0 12 a5 dS (D) gives White a clear positional advan­
tage - the e6-pawn is weak and it is
hard for Black to bring his c8-bishop
into play. However, Black is by no
means lost and this would be a reason­
able attempt at a 'grim defence' strat­
egy.
In the late 1 970s Plaskett was prone
to time-trouble, and he had already
thought a long time before playing
1 3 e5. Given the choice between a
messy, complicated position and one
in which he could increase the posi­
tional pressure by straightforward
At this stage I was feeling quite moves, I would certainly have pre­
confident, because in an earlier game ferred the former. I therefore decided
Jansa-Nunn, B udapest 1 978 White to look again at the capture on e5 and
 THE MIDDLEGAME 95

in the end I found the continuation

played in the game.
13 ... 'ii'xeS 14 .tf4 'ii' hS 15 g4 'ii'g6
I didn't spend long on the moves
1 5 ... 'iFM and 1 5 . . . 'ii'h3 . It is quite easy
to refute the former: 15 . . . 'ifh4 1 6 .i. g3
'iFh3 1 7 lt fe l ltle8 ( 1 7 . . . 'ii'h 6 1 8 g5)
1 8 .tn 'ii'h6 1 9 'iFxh6 gxh6 20 ltlxd5
and Black's position is disgusting. Al­
though it is quite hard to find a clear­
cut refutation of 1 5 . . . 'ii'h 3, there is lit­
tle point in thinking about such moves.
White can regain his pawn immedi­ on the 'create confusion' method. He
ately, and there is almost no chance has two pawns for the piece and, ow­
that Black can survive with his queen ing to the attacks on b4 and g4, he is
stuck on h3 and none of his other sure to gain a third pawn. This is not a
pieces co-operating with it. Home coincidence, but results from the fact
analysis shows that 1 5 . . . 'ii'h3 1 6 l:l f2 that White's tactical operation de­
lLle8 17 b4 lLld7 1 8 ltlxdS .i.d8 1 9 lLle3 ! pended on the weakening pawn stabs
is very strong; e.g. 1 9 . . . .i. f6 loses to 20 b4 and g4. Given that Black obtains a
.tn WM 21 .t g3 'ii'gs 22 f4. material equivalent for the piece, you
16 b4 may wonder why he is worse. The an­
This was the point of White's pawn swer is simply development. All the
sacrifice. The knight has to move, and white pieces are in active play, while
then .i. d3 traps the queen. Black's queenside is still sitting at
16 tbce4!
•.. home.
Except, of course, if it moves to e4 ! 18 �hl
Black can avoid immediate material White decides to jettison the b4-
loss by 1 6 . . . e5 1 7 .i. xe5 ltle6, but after pawn, preferring to keep the pawn on
1 8 .i. d3 'ii'g 5 1 9 'ii'xg5 ltlxg5 20 h4 g4 which shields his king . After 1 8
ltle6 2 1 ltlfS he has a very bad ending, .t g5 lld8 ( 1 8 . . . .t xb4 i s also possible)
for example 2 1 . . . .t xb4 22 .t xf6 gxf6 Black has sufficient play.
23 lLl xdS .tc5+ 24 �hi . This doesn't 18....txb4
even qualify as a ' grim defence' - it is Not 1 8 . . . ltlxg4 1 9 .t xg4 'iFxg4 20
j ust very bad without any redeeming lt g l 'ii'h4 21 ltlxe4 b5 (the only hope
features. is to get the bishop to the long diago­
17 fxe4 dxe4 (D) nal) 22 'ii'g 2 g6 23 ltlc6 and White
This continuation is Black's best wins.
practical chance, once he has decided 19 .teS!
96 SECRETS OF PRACTICAL CHESS

An excellent move. The bishop position of Black's queen. However,

serves several useful functions on e5 ; the resulting complications are not in
it supports the knights on d4 and c3, White's favour.
threatens i. xf6 in some lines and rules This was the moment to switch
out any possibility of Black activating from tactics to calmer play. After 2 1
the c8-bishop by . . . e5. At first sight i. xf6 gxf6 2 2 axb6 .:r.ac8 2 3 .!LJa4 i. d2
Black can reply 19 . . . e3 20 "ii'xe3 i. xc3 24 °ii'g 3 l:r. fd8 25 c3 White's pieces
21 'ii'x c3 'ife4+ regaining the piece, may appear scattered and poorly co­
but White continues 22 .!LJf3 'ii'xe2 23 ordinated, but Black does not seem
l:r.fe l °ii'b 5 (23 . . . 'ii'f2 24 i. d4 loses the able to exploit this. One possible line is
queen) 24 i. xf6 gxf6 25 'ii'xf6 'ii'c6 26 25 . . . i. xc3 (25 . . . e3+ 26 i. f3) 26 .!LJ xc3
l:r.e5 h6 27 g5 with a decisive attack. l:r.xd4 27 i. xa6 and now 27 . . . i. xa6 28
Up to here, White has handled the l:r. xa6 .:r.d3 fails to 29 b7.
game well. He has not simply been A common error in wild games is
content to keep his material, which to overlook a positional continuation.
might have allowed Black to complete With the adrenaline pumping, and
his development and create strong mates all around, it is easy to develop
counterplay. Instead, he has kept his 'tunnel vision' that only sees tactical
pieces active and not been afraid to possibilities; then a calm liquidating
calculate tactical lines. line can easily be missed.
19 b6! ?
..• 21 ...exfS 22 gxf5 i. cS! (D)
Following the confusion policy. Black makes use of the protection
White's last move demolished Black's of c5. After 22 . . . .!LJg4 23 i. xg4 'ii'x g4
hopes of activating his light-squared 24 l:r. g l 'ii'f3+ 25 'ii'xf3 exf3 White
bishop on the c8-g4 diagonal, so now forces mate by 26 .:r. xg7+ <itih8 27
the idea is to develop it on the long di­ l:r. g8+ � xg8 28 l:r. g l#.
agonal. The choice of . . . b6 rather than
. . . b5 was based on two factors: firstly,
in some lines Black plays . . . i. c5 and
then it is useful to have the bishop de­
fended (see the game); secondly, the
possibilities of axb6 and . . . bxa5 give
White more to think about.
20 °ii'e3
Both unpinning the c3-knight and
preventing the e4-pawn's advance.
20 i.b7 21 .!LJfS?
•••

An incredibly ingenious idea, which

aims once more to exploit the poor
 THE MIDDLEGAME 97

23 'iFxc5 continuation was 27 l:. g l ! � xe5 (not

Forced, as after 23 fxg6 i. xe3 Black 27 . . . � xe3 ? allowing mate in five) 28
is material up. l:[fg3 (28 l:. f4 f6 29 axb6 is unclear; the
23 e3+! 24 i.f3
••. passed b-pawn is an asset, but White's
Black wins after 24 l:.f3 .i.. x f3+ 25 king is exposed and the e5-knight is
.i.. x f3 "ii'xf5 26 'iFxe3 l:. ae8. impossible to dislodge) 28 . . . °ii'xf5 ! 29
24 .txf3+ 2S l:[xf3 °ii'hS 26 Wxe3
...
.:r.xg7+ �h8 30 'ii'h6 (30 axb6 �g6 3 1
After 26 'ii'c6 l:. ac8 27 W b7 � g4 28 'ii'h6 'ii'f3+ 3 2 l:. g2 i s also a draw)
.i.. c 7 Black has the pleasant choice be­ 30 . . . °ii'f3+ with perpetual check.
tween 28 . . . � xh2 29 i. xh2 l: xc3 30 27 �xe3 28 �f6+ gxf6 29 l:.g3+
...

l:. g3 l:. xc2 31 l:. g2 .:r.xg2 32 'iFxg2 �g4 0-1

bxa5 and 28 . . . .:r.xc7 29 Wxc7 � xh2 30
'ikxh2 'iFxf3+ 3 1 Wg2 °ii'xf5 32 axb6 Defending well after having made
Wf6. In both cases Black ends up with an oversight requires especially cool
four pawns for the piece. nerves. We have previously discussed
26 �g4 (D)
••• the possible causes of oversights and
The unexpected point of Black's the warning signs which can indicate
play. After the queen moves, Black when danger is near. Suppose, despite
takes on e5, regaining his piece. this advice, you neverthel�ss overlook
a surprising and strong move by your
opponent. The first piece of advice is
to stay calm. It is all too easy to bash
out an instinctive response, either
through uncontrollable nervous agita­
tion or in an attempt to persuade your
opponent that you had foreseen his
move and had a reply ready. This is a
mistake. The correct approach is to
spend a few minutes just calming your
nerves. Don't get caught up a mental
loop of self-recrimination - you don' t
have time for this while you are at the
27 �d5?? board. Try to forget about the history
A horrible blunder in extreme time­ of the position, and just consider the
trouble - White plays for a mate which current state of affairs on the board. A
does not exist. Despite the downturn calm look will very often show that
in his fortunes, if White had kept a your opponent's move is not nearly as
clear head then he would probably strong as you feared at first and that
have saved the game. The correct there are still fighting chances. Then
98 SECRETS OF PRACTICAL CHESS

you can choose one of the defensive on your attack, there is no point hold­
techniques outlined above and con­ ing pieces in reserve, since the game
tinue the struggle. will be decided before reserve pieces
will be of any use.
Attack The following position is a classic
example:
1 '
l nviti ng everyone to the party

Most books dealing with attacks on

the king discuss the various typical
methods of breakthrough; sacrifice on
h7, sacrifice on g7, double bishop sac­
rifice, and so on. I will not deal with
these, partly because they are dis­
cussed at length in many other places,
but also because they represent only
the final stages of an attack. Most
players can manage a double bishop
sacrifice, provided that they first arrive
at a position in which such a sacrifice A. Nimzowitsch - S. Tarrasch
is possible. St Petersburg 1914
The main factor governing the suc­
cess of an attack on the enemy king is Black already has a local superior­
whether you can bring more attacking ity on the kingside, because White has
pieces to bear on his king position than no defensive pieces there at all. How­
he can muster for the defence. If you ever, 18 . . 'ii' M is ineffective owing to
.

have a large local superiority of force, 1 9 lllf3 and the kingside is shored up.
then a sacrificial breakthrough will of­ Tarrasch's next move brings the c6-
ten arise as a matter of course. You bishop into the attack as well.
still have to calculate that the sacrifice 18... d4!
works, but the odds will be heavily in Not only unveiling the bishop, but
your favour if you have enough wood also ruling out both the defensive ll:\f3
in the vicinity. and counterplay by 'ii'c 3.
This title of this section is Yasser 19 exd4
Seirawan' s catchy phrase for an im­ There is nothing better as White is
portant attacking principle. Incorpo­ unable to feed any pieces across for
rating every possible piece into an the defence. Now, however, 1 9 . . . 'ii'M
attack greatly increases its chance of 20 g3 'ii'h3 gets nowhere after 2 1 llle4.
success. If you have staked everything 19 ... .txh2+!
 THE MIDDLEGAME 99

The moment is ripe for the double counter these manoeuvres either de­
bishop sacrifice. fensively or by generating counterplay
20 'it>xh2 'ii h4+ 21 'it>gl �xg2 22 in another part of the board.
f3 In addition to bringing your own
Forced. After 22 'it> xg2 'iig4+ 23 pieces to bear on the enemy king, it is
'it>h2 %:td5 24 'iix c5 l:t h5+ 25 'ii xh5 also important to block the passage of
"ikxh5+ 26 'it> g2 'i\i'g5+ Black picks up enemy pieces to the threatened area.
the knight on d2, while 22 f4 'iig3 is Sometimes a sacrifice is necessary.
deadly.
22... l:tfeS
With the pawn on f3, 22 ... 'i\i'g3 may
be met by 23 ltJe4.
23 tt:Je4
A desperate bid for counterplay
based on the long diagonal and weak­
ness of f6. After 23 l:t fe l l:t xel + 24
l:t xel 'iixel + 25 'it> xg2 'iie2+ 26 'it>g3
%:td5 Black wins easily, for example 27
f4 l:t h5 (threatening 28 ... %:th2) 28 'ili'c l
�h2+ 29 'it>f3 l:t h3+ 30 'it>e4 'ili'g2+ 3 1
'it> e5 %:t e3+ 32 'it>d6 %:te6+ 33 � xc5
%:tc6+ picking up the queen. S. Dvoiris A. Khalifman
-

23 ... 'ii h l + 24 �f2 �xfl 25 d5 Russian Championship, Elista 1 997

25 l:t xfl 'ii h2+ wins the queen.
25...f5 26 'ili'c3 'iig2+ 27 �e3 %:txe4+ 15... h6 16 h4!?
28 fxe4 f4+ A brave decision, but it was proba­
Overlooking a quicker mate after bly made easier by the fact that 1 6
28 . . . 'i\i'g3+ 29 'it> d2 'ili'f2+ 30 'it> d l �h4 'iih5 i s rather bad for White, who
�e2#, but Black wins easily anyway. cannot play 1 7 � g3 on account of
29 'iPxf4 :rs+ 30 �es 'ii h 2+ 31 17 . . . tt:Jxe4 and the d l -rook hangs.
'iPe6 %:te8+ 32 'it>d7 �b5# (0-1) 16 'iicS
..•

Black can accept the piece at virtu­

Many familiar attacking manoeu­ ally any stage over the next few moves;
vres are designed to gain the local su­ in each case the verdict would be un­
periority which is usually necessary clear.
for a successful attack. When you play 17 'iig3 tt:Jh5
%:tel -e3-g3 or 'iid l -el-h4, you are cre­ Svidler suggests 1 7 . . . 'it>f8 ! ?, intend­
ating the preconditions for the attack ing to take on g5 and then retreat the
to break through. The opponent must knight to g8.
1 00 SECRETS OF PRACTICAL CHESS

18 'fi h2 hxgS?! is defending with very limited mate­

Khalifman finally decides to grab rial.
the bishop, but it turns out to be the 21 exd5
...

wrong moment to do so. Having said Other moves are hopeless, for ex­
this, it would be a far-sighted player ample 2 1 . . . lLJ xdS 22 .:r.h I lLJc3+ 23
who anticipated White's 2 1 st move. bxc3 'ti'xc3 24 'iF b7+ �f8 25 l:r.dfl
The safest continuation was 1 8 . . . 'fif2 'ii'g 7 26 l:r. xf7+ 'ilxf7 27 'fih8+ 'ii'g8
1 9 i. xe7 l:r. xe7 20 .:r.fl 'fig3, heading 28 l:r. fl + .t f6 29 'ifxf6+ and mate, or
for the exchange of queens and a 2 1 . . . .t xg5 22 l:r. h l lLJh5 23 gxh5 <lt g7
roughly equal position. 24 hxg6 l:r. h8 25 'ii'g 3 exd5 26 'iFxg5
19 hxg5 g6 20 g4 lLJxf4 (D) dxe4 27 'fig4 .t d7 28 'i'xe4 .t f5 29
'ii'xb7 .t xg6 30 l:r. hgl , followed by
l:r. dfl , with equal material and a crush­
ing attack for White.
22 l:r. hl lLJh5
Again a forced move; 22 . . . f6 allows
23 'ifh7+ 'it>f8 24 exd5 i. xd5 25 l: xd5
lLJ xd5 26 'ifxg6, while after 22 . . . f5
23 'ii'h7+ 'it>f8 24 exf5 the kingside
pawn-mass is immediately decisive.
23 gxh5 rt;g7 24 exd5
The key feature of the rest of the
game is the way White's blockade on
d5 prevents Black's queen from taking
Now Black appears to have every­ part in the defence.
thing under control. After 21 .:r.hl Black 24 i.d7
•••

can play 2 1 . . . 'ii'eS, using his queen 24 . . . l:r.h8 loses to 25 dxc6 l:r. xh5 26
like a Dragon bishop. 2 1 'fixf4 is also 'ii'f4, but perhaps Black could have
inferior after 2 1 . . . i. xgS, followed by put up slightly more resistance with
. . . 'ii'e5, when Black has an extra pawn 24 . . . .t bS. However, even in this case I
and a large positional advantage. think White should win by 25 hxg6
21 lLJdS ! ! .:r.h8 26 'iFf4 fxg6 (26 . . .f6 27 'ii'e4 'ii'c7
A n amazing move. Black's defence 28 'ife6 wins) 27 l:r. he l 'ii'c7 28 l:r.e6
is based almost entirely on switching (intending 'ii'e4) 28 . . . l:r. h7 (28 . . . .l:r. hS
his queen to the kingside. White is 29 'ife4 <ltf8 30 c4 gains access to fl ,
prepared to offer another piece to pre­ while after 28 ... .:r.dg8 29 .:r.de l .t d8 30
vent this. The result is that White is 'ifd4+ <ilt f8 31 i. a4 ! a rook will land
able to operate with his queen and two on either e8 or fl ) 29 l:r.de l and now
rooks on the kingside, whereas Black Black can try:
 THE MiDDLEGAME 1 01

1 ) 29 . . . 'iii> g 8 30 l:r. xg6+ l:r. g7 3 1 l:r.e7 33 l:r.el + � f4 34 l:r.fl + �g3 35

J:xg7+ 'ii? x g7 3 2 'ifd4+ 'iii>f7 (32. . .'iii>g8 'ii'xrs 'ili'e3 36 a3 l:r.eS 37 •rs 1-0
33 J: h l ) 3 3 g6+ <iti xg6 34 l:r. g l + i. g5
35 •e4+ winning for White. A common method of excluding
2) 29 . . . l:r.e8 30 c4 .t a6 3 1 •d4+ defensive pieces from the critical sec­
wins. tor is the pawn-wedge. A chain of
3) 29 . . . .t f8 30 c4 .t a6 (30 . . . .t d7 pawns extending deep into enemy ter­
3 1 l:r. xg6+ 'iii>h8 32 i. c2 i. g7 33 l:r. h6 ! ritory has the effect of cutting his posi­
.t xh6 34 gxh6 is lost for Black) 3 1 tion in half, and may in itself prove
.t c2 � g8 ( 3 I . . .t xc4 3 2 i. xg6 l:r. h3
. sufficient to prevent any pieces com­
33 .t f5 l:r. h5 34 J: h6 .t xa2+ 35 'iii>a l ! ing to the rescue of the beleaguered
wins, or 32 ...<itig8 transposing) 32 .t xg6 king. Then it is a matter of whittling
l:r. h3 (32 . . . .t xc4 33 .t xh7+ °ii'xh7+ 34 away the defensive pieces on the side
g6 .t d3+ 35 <iti al 'ii'd7 36 J:e7) 33 b3 of the wedge near the king, by sacrifi­
b5 (Black has nothing else) 34 l:r. e7 ! cial means if necessary. It is important
.t xe7 35 'ili'f7+ <ith8 36 l:r. xe7 l:r. hl + to watch out for a counter-sacrifice
37 � b2 :t h2+ 38 i. c2 and wins. breaking up the wedge, but otherwise
25 hxg6 l:r.h8 26 'ii'f4 rs such attacks are usually plain sailing.
26 . . .fxg6 27 l:r.hel is dead lost.
27 l:r.h6!
A nice collinear move (see page
56). The threat is simply 'ii'h2 and
l:r. h l , and thanks to Black's inability to
feed pieces to the threatened sector,
there is not much he can do to stop it.
27...l:r.de8
Alternatively, 27 . . . l:r. xh6 28 gxh6+
� h8 (28 . . . 'ii? x g6 29 'ii'g3+ .t g5 30
l: g l ) 29 c3 i. f6 (29 . . . l:r. g8 30 g7+ 'iii>h7
3 1 i. c2 'ii'c 8 32 l:r. el i. f6 33 J:e6 !
wins neatly) 30 i. c2 'ii'b 6 3 1 g7+ � g8
32 'ili'g3 ! (threatening 33 h7+ � xh7 34 J. Nunn A. Vydeslaver
-

g8 'ili'+) 32 . . . �f7 33 l:r. g l .t xc3 34 Leeuwarden Open 1 995

g8'ii'+ J:xg8 35 °ii'x g8+ <ite7 36 J:g7+
.t xg7 37 'iFxg7+ <itid8 38 h7. I had earlier sacrificed my h-pawn
28 .h2 in order to gain time for my attack.
Decisive. This had the positive effect of opening
28 .txgS 29 l:r.h7+ l:r.xh7 30
••• the h-file and allowing the f-pawn to
•xh7+ �f6 31 'ili'f7+ 'ii?e5 32 'ii'xd7 advance, but on the other hand Black's
1 02 SECRETS OF PRACTICAL CHESS

bishop has occupied a post on g5 that 23 ... i.xd4 24 i.h7+ <ifi'h8 25 °ii'xd4
is invulnerable to pawn attack. eS
20 f6! Or 25 . . . bxc3 (25 . . . f6 26 :%. h3) 26
White establishes his pawn-wedge. i. f5+ <iti g8 27 :%.db l gxh6 28 :%.xh6 f6
20...b4 29 'Wg4+ �f7 30 l:[ h7+ mating.
The tactical justification for White's 26 'WxeS f6 27 'iWh2 gxh6 28 'iWxb6
last move lies in the line 20 . . . gxf6 2 1 Black is helpless against the many
exf6 e 5 (2 1 . . . ll:i xf6 2 2 :%. xh6 i. xh6 23 threats.
.t xf6 'Wc7 24 l:[ h l wins) 22 :%.xh6 ! ! 28 ... 'We7 29 .trs+ 'iil'g8 30 .te6+!
.t xh6 23 'ii'g3+ <iti h8 24 'Wh4 'Wxf6 25 :r1 31 .:r.h1 1-0
i. xe5 and wins.
Since Black cannot remove the in­ In the following example, all the
truding pawn, he decides to continue above three elements (inviting every­
on the queenside. Now White's task is one to the party, excluding defensive
to remove the only piece left defend­ pieces and the pawn-wedge) come to­
ing Black's kingside, the g5-bishop. gether.
21 l:[bS!
Getting rid of the bishop by the
most direct method possible. The ma­
terial loss involved is irrelevant. Not
21 lLie2? i. b5 ! , when the exchange of
the d3-bishop greatly weakens White's
attack.
21 ... ll:ixf6
After 2 1 . . . bxc3 22 :%. xg5 hxg5 23
'ii'g 3 mate is inevitable within a few
moves, and is only delayed by one
move after 23 . . . l:[ b8 24 b3. 2 1 . . .gxf6
22 exf6 bxc3 23 :%.xg5+ hxg5 24 'Wg3
and 2 l . . . i. xf6 22 exf6 ll:ixf6 23 .:r.xh6 N. Short A. Chemin
-

bxc3 24 i. xf6 are no better, so Black European Team Championship,

decides to offer a piece to break up the Pula 1997
deadly pawn-wedge.
22 exf6 .txf6 23 :%.xh6! In this position White clearly has a
White should win in the end after very strong attack, but Black has cer­
23 i. xf6 'iWxf6 24 'iWxf6 gxf6 25 ll:ie2 tain counterchances based on White's
ril g7 26 .t xa6, but I decided to calcu­ weak back rank and the possibility of
late a tactical kill out to the end. The . . . ll:if2+. Chemin thought that he had
next few moves are all forced. prevented the sacrifice on h5, but . . .
 THE MIDDLEGAME 1 03

29 �xhS ! Over-sacrificing
. . . Short played it in any case. The
point is that after 29 . . . gxhS White does Most chess players love to attack.
not continue 30 'ii'xh5, which allows Pressing home an assault against the
30 . . . �f2+, but 30 e5 ! . This includes enemy king, sacrificing a couple of
the d4-rook in the attack, excludes the pieces and finally delivering mate is a
c5-rook from the defence and cements great thrill . . . provided it works.
the e5-f6 pawn-wedge. White would One of the great dangers of even a
then threaten the devastating 3 1 l: xg4, correct sacrificial attack is over-sac­
and Black wouldn't be able to do rificing. It very often happens that the
much: 30 . . . l:dS (30 . . . l: xe5 3 1 l:h g4 first sacrifice is sound, but then the
l:r. e l + 32 'ii'xel hxg4 33 :xh8+ 'it> xh8 player gets overwhelmed with the de­
34 'iFh4+ � g8 35 i. h6 mates) 3 1 sire to finish 'brilliantly ' and instead
i.e3 ! (the simplest, covering f2 and so of just bringing up all his pieces and
threatening to take on h5; after 3 1 mating his opponent, he goes on a
i. xd5? exd5 Black would defend g4, quite unnecessary sacrificial spree,
when the attack would be stopped) endangering the win. As the material
3 1 . . . � xe5 (3 1 . . . � xe3 32 'ii'g5+ 'it>f8 deficit increases, the opponent gains
3 3 : xh5 mates) 32 'ifxh5 i. xf6 33 more and more chances to return some
l:r. xd5 exd5 34 'ii'h 7+ � f8 35 i. cS+ or all of the material in order to fend
winning Black's queen. off the attack. Sometimes some quite
29 l:r.xgS 30 'ifxgS
.•. unlikely-looking moves become pos­
White could have won instantly by sible if there is enough spare wood to
30 � g7 ! l:r. h5 (30 . . . �f2+ 3 1 'ifxf2 jettison.
i. xg7 32 fxg7 'it> xg7 33 'ifd2 wins a
piece) 3 1 � xh5 gxh5 32 'ifgS+ <it f8
3 3 e5 �f2+ 34 � g l � xh3+ 35 gxh3
and Black has no defence to the threats
of 36 'if xh4 and 36 : h4 followed by
l:r. xh5.
However, even after the move actu­
ally played White retains a clear ad­
vantage.
30 �f2+ 31 �gl �xh3+ 32 gxh3
••.

i.c6?
Black collapses. This move fatally
weakens e6.
33 �f4 'it>h7 34 es i.f3 35 <it f2 M. Botvinnik V. Chekhover
-

'ii'c6 36 l:lc4 1-0 Moscow 1935

1 04 SECRETS OF PRACTICAL CHESS

White's minor pieces are ideally winning line is mundane, but it is effi­
placed to attack Black's king, except cient.
for the knight on h3 . The ugly cluster 24 ... <i;xf7 25 g6+ <i;g8? ? (D)
of black pieces on the queenside is not After this Black gets mated. He
doing much and is certainly of little should have played 25 . . <i,;f8 26 'i!Vxe6
.

help in defending the kingside. White llle 5 ! (Black's two extra pieces jus­
therefore decides, quite correctly, to tify this odd-looking move) 27 l:t xf6+
sacrifice his one poorly placed piece gxf6 28 °i6'h3 (thus far given by Bot­
in the interests of opening up Black's vinnik) and now either 28 . i. c5 ! or
..

king position. 28 ... i.b4 29 l:te l i.xg2 ! 30 <i;xg2 i. xel

22 lllg5 ! hxg5 3 1 dxe5 fxe5. The position is still very
Black must accept, since otherwise complicated, but Black, a rook and a
he cannot defend f7. piece up, has plenty of possibilities to
23 fxg5 lt:l8d7 (D) return some of the material. My own
After 23 . . . lt:l 6d7 24 lll xf7, followed view is that White would be struggling
by 'i!V h5, White has a crushing attack, to draw in these lines.
so Black offers to return the piece.

After the move played, White fin­

24 lll xf7? ished nicely:
This second sacrifice is not only un­ 26 'i!Vxe6+ <i;h8 27 'ii'h3+ <i;g8 28
necessary, it even endangers the win. i.f5 lllf8 29 i.e6+ lt:lxe6 30 'i!Vxe6+
After 24 lt:l xd7 l:t xd7 (or 24 . . . lll x d7 <i;h8 31 'ii'h 3+ <i;g8 32 l:txf6 i.xf6 33
25 l:t xf7 <i; xf7 26 'ili' h5+ and White 'i!Vh7+ �f8 34 l:tel i.e5 35 'ii' h8+
forces mate) 25 gxf6 i. xf6 26 l:t xf6 <i;e7 36 'ii'xg7+ <i;d6 37 'ili'xe5+ 'it>d7
gxf6 27 'ii'g4+ <i; f8 28 i. a3+ l:t d6 29 38 'ii'f5+ 'it>c6 39 d5+ 'it>c5 40 i.a3+
°ikg3 <i; e7 30 c5 ! White's attack breaks <i;xc4 41 'i!Ve4+ <i;c3 42 i.b4+ �b2 43
through without any difficulty. This 'ii' b l# ( 1 -0)
 THE MiDDLEGAME 105

Defence the game, but if there is no mate then

queen and rook may have to grovel
The principles for defending against back to the centre, with great loss of
an attack on the king are to some ex­ time.
tent the converse of those given above The defender can often exploit the
in the section on attack. The defender negative side of an attack by a suitable
should try to move his own pieces sacrifice to take the sting out of the on­
across to help the endangered king slaught.
while, if possible, obstructing the free
passage of the opposing pieces to the G. Sax - M. Stean
critical sector. European Team Championship
One particular motif which often Final, Moscow 1 977
arises in practice is that of the defen­ Sicilian, Scheveningen
sive sacrifice. The idea of the attacker
sacrificing material is a familiar one, 1 e4 cS 2 .!LJf3 e6 3 d4 cxd4 4 lLJxd4
but it happens almost as often that the .!LJf6 S .!LJc3 d6 6 i..e2 a6 7 0-0 i..e7 8
defender gives up material. Here I am i..e3 'ilc7 9 f4 0-0 10 g4 .!LJc6 1 1 gS
not talking about the situation in .!LJd7 12 rs .!LJdeS 13 f6 i.. d8 (D)
which the attacker has sacrificed, and
the defender is returning material, but
about cases in which the defender is
prepared to accept a genuine material
deficit.
The basis for such sacrifices very
often lies in the positional concessions
made by the attacker. In a Sicilian, it
may be very useful for attacking pur­
poses if White pushes his kingside
pawns to g5 and f6, but if the attack
collapses the white king may not ap­
preciate having had his defensive
pawn wall sent into the other half of A typical Sicilian position. The white
the board. Launching an attack usu­ kingside pawns have launched them­
ally involves a concession of some selves forward, but in return Black has
sort; it may be the creation of weak­ undisputed control of the square e5.
nesses, as in the case of a pawn ad­ 14 fxg7?!
vance, or it may be sending pieces off­ This looks like a mistake as there is
side. If White plays his rook to h3 and no need for White to commit himself
queen to h4, then mate on h7 will end to this capture so early. Recently this
1 06 SECRETS OF PRACTICAL CHESS

intuitive feeling was backed up by a White decides to grab the exchange,

practical test; Sutovsky-J.Polgar, Til­ but Black obtains tremendous com­
burg 1 996 continued 1 4 .i. d3 (ena­ pensation. It would have been better
bling the white queen to reach h5) to try 16 a3 followed by, for example,
14 . . . .!LJ xd4 15 .i. xd4 'ii'a5? ! 16 fxg7 doubling rooks on the f-file. However,
lj; xg7 17 <j;hl 1'. b6? ! 1 8 .t xe5+ 'ifxe5 I do not believe that Black has any
1 9 °ifh5 .te3 20 l:r. f3 .i. xg5 21 :gt f6 problems as a major threat from White
22 h4 1 -0. Subsequent analysis of this is still several moves away.
game suggests that although Black 16 hxg6 17 .i.h6+ 'lt>g8 18 .i. xf8
..•

need not have lost so quickly, White lj;xf8 19 .!LJxc6 'ifxc6 (D)
retains some advantage after any de­
fence.
14 'iPxg7!
••.

Bravely eliminating one of the dan­

gerous white pawns. After 14 . . . :e8? !
1 5 °ifd2 b5 1 6 .t h5 lLJe7 17 b3 ! .!LJ7g6
18 a4 ! bxa4 19 l:r. xa4 1'. b7 20 .!LJf3
White had a slight advantage in Khol­
mov-Spassky, Moscow Zonal 1 964.
Stean's move may appear suicidal,
but it turns out that White has prob­
lems continuing his attack because his
development is rather poor (all those
pawn moves on the kingside). He The transformation since the previ­
really needs his queen on the kingside ous diagram is remarkable. There is no
in order to create genuine threats, but trace left of White's advanced kingside
both 1 5 .t d3 (intending 'ii'h S) and 1 5 pawns. Now Black only needs to play
'ii'e l drop a pawn after . . . .!LJxd4 and . . . <t; g7 to secure his kingside com­
. . . .t xg5. Note that in the latter case pletely. He has a monster knight on e5,
Black justifies his play with a tactical and tremendous pressure on the dark
point: 1 5 'ikel .!LJ xd4 16 .i. xd4 .t xg5 squares. Finally, once he has played
17 'ii'g 3 � h8 ! and the bishop is invul­ . . . � g7, . . . .t b6 and . . . .i. b7, his rook
nerable. can swing across to h8 and White's
Thus White has to proceed with his king will be the one subject to attack.
attack much more slowly, but Black In return for these many positional ad­
only needs to play . . . b5 and . . . .i. b7 to vantages all White has to show is the
set up a further awkward threat against very small material advantage of rook
the e4-pawn. for bishop and pawn.
15 °ifd2 bS ! 16 g6? ! 20 .td3
 THE MIDDLEGAME 107

After 20 'ii'h6+ � g8 2 1 l bt7 �xt7 Material loss is inevitable.

22 .:r.n + 'it> e8 23 'ii g7 i.e7 Black de­ 33 .:r.bl .:r.xh3 34 'it>xh3 l0f4+ 35
fends easily. �g4 l0xd3 36 :e2 rs+ 0-1
20 �g7 21 l0e2 i.b7 22 l0f4
•••

.tb6+ 23 �hl .:r.b8 A defensive sacrifice can also prove

All Black's pieces are in attacking effective for psychological reasons.
positions; White is obviously in trou­ The attacker is mentally geared up for
ble. a possible sacrificial assault on the
24 °ii'g2 °ii'd7 25 l:lael 'ike7 26 'ii'g3 enemy king, and then suddenly he is
'ii' h4 27 �g2 (D) defending and trying to nurse his ma­
The exchange of queens would be terial advantage to an ending. This re­
no help to White. One line is 27 'ikxh4 quires a complete shift of mental gears
l:t xh4 28 .:r.e2 l0g4 29 l:[ f3 i. d4 30 b3 which can prove difficult to achieve.
g5 3 1 l0h3 f6 32 .:r.g3 b4 33 .:r.d2 �h6,
and now that Black has everything L. Ljubojevic A. Miles
-

prepared, he is ready for the deadly European Team Championship

breakthrough . . . f5 . Final, Skara 1 980
Sicilian, Dragon

1 e4 c5 2 l0f3 d6 3 d4 cxd4 4 l0xd4

lOf6 S l0c3 g6 6 f4 l0bd7 7 l0f3 'ikc7
8 i.d3 i.g7 9 0-0 0-0 10 �hl a6 1 1 a4
b6 12 'ikel i.b7 13 'ikh4 eS 14 .td2
l0c5 15 l:tael l:lae8 16 fxe5 dxe5 (D)

27 gS 28 l0h3 f6
.••

White has no active play at all, so

Black has plenty of time to tidy up his
position before making further prog­
ress.
29 l:te2 l0g6 30 l:teel i.d4
The changing of the guard. Now the
bishop comes to occupy e5 . A fairly typical position for the 6 f4
31 c3 i.e5 32 'ikxh4 .:r.m4 Najdorf (the game started as a Dragon,
1 08 SECRETS OF PRACTICAL CHESS

but later transposed into a Najdort). 20 g4? !

White has attacking chances on the White w as wrong t o take the pawn .
kingside, while Black can organize 20 ltefl would have been better, when
counterplay against the isolated e4- 20 . . . f5 2 1 exf5 ! i. xf3 22 lt xf3 is a
pawn. The long-term chances are with very dangerous exchange sacrifice.
Black, because of his superior pawn Therefore Black would have to con­
structure, so the burden of proof lies tinue more slowly, for example by
with White. 20 . . . 'iic S, but White would have an
The normal plan of attack is based edge.
on .i.h6 and c!Ll g5 but this is not very
, 20 c!Lif4 21 c!Lixf4 exf4 22- .i.xf4
...

effective here because White would be c!LieS 23 ltb3? !

losing a tempo with his bishop and so White will be forced to take on e5
Black would be able to meet 17 i.h6 by in any case, and it seems odd to put the
1 7 . . . c!Lih5. rook offside. 23 ltffl was better.
17 c!LlgS 23 gS 24 .i.xe5
...

This appears very dangerous, since The rook would only be useful on
1 8 lt xf6 is threatened, and 1 7 . . . h6 1 8 h3 if White could play 24 i. xg5 here,
lt xf6 ! i. xf6 1 9 'ii'xh6 .i. xg5 2 0 .i. xg5 but then 24 . . . c!Ll g6 25 'ii'hS lte5 wins a
lt e6 2 1 .i. c4 lt d6 22 c!Lid5 .i. xd5 23 piece.
i. xd5 gives White a crushing attack. 24 :xeS 25 'iig3 (D)
...

17 c!Licd7
.••

Therefore this move is forced. If

White's bishop were on h6 instead of
e3, then he would have a standard at­
tacking plan of g4 (preventing . . . c!Lih5 ),
followed by doubling rooks on the f­
file, but as it is White always has to
take . . . h6 into account.
18 :o h6 19 c!Llb3 c!Llb5!
Not 1 9 ... g5 20 c!Lixg5 hxg5 2 1 i. xg5
with a tremendous attack for White.
One line is 2 1 . . . :e6 22 c!LidS 'ii'd6 23
lt efl c!Li xd5 24 exd5 : g6 25 :h3 f5 26
'iih7+ rj;/f7 27 i. xf5 l:hg5 28 i. g6+ Again, there has been a remarkable
<l; e7 29 'ifxg7+ <itii d8 30 l:hf8+ c!Lixf8 transformation in the position. The
3 1 : h8, and wins. only trace remaining of White's attack
The text-move shields the pawn on is the misplaced rook on h3 . In return
h6 and prepares the following pawn for the pawn, Black completely domi­
sacrifice. nates the dark squares, while White's
 THE MIDDLEGAME 109

pieces are doing little apart from de­ put White's position under intolerable
fending the pawn on e4. Moreover, the stress.
advance g4 has seriously weakened 29 'ii'e l 'ii'cS 30 l:r.e3
White's kingside, and if Black man­ The tricky 30 .t xa6 leads only to
ages to play ... f5 under favourable cir­ self-destruction: 30 . . . .t xa6 3 1 exf5
cumstances then White will be in real .t b7+ 32 .!LJe4 Wxf5 33 l:r.e3 .i. d4.
trouble. 30 .l:r.fl 31 l:[g3
•.•

An additional point is that Ljubo­ 3 1 lle2 .t d4 32 .:t xf2 llxf2 33 tlJd l

jevi� is an attacking player who does :xc2 3 4 .i. xc2 'ifxc2 wins for Black.
not adapt well to positions requiring 31 ... .i.d4 32 .!LJds 'ii'd6
careful defence; this game is a case in Threatening 33 ... l:r. xh2+.
point. 33 .!LJe3 'ii'g6
25 Wcs 26 'ii'g l
... Avoiding the trap 33 ... .te5? 34 lLJc4
White should be thinking about how l:r. xh2+ 35 <iti xh2 .t xg3+ 36 � g2 ! and
he can draw, and his pieces are so pas­ White survives.
sive that the only realistic chance is to 34 .!LJg2 (D)
play lLJd5 at some stage, returning the
pawn to reach the haven of opposite­
coloured bishops. However, the im­
mediate 26 .!LJd5 is impossible because
of 26 . . . .t xd5 27 exd5 'ii'xd5+ 28 <iti gl
l: hel + 29 'ii'xel .t d4+ 30 <itifl °ii'h l +
and Black wins.
26... 'ii'b4 27 l:r.bl?
White has failed to adjust to the
changed situation. He tries to hang on
to his pawn, but putting another piece
on a bad square allows Black's initia­
tive to increase decisively. After 27
l:r. f3 ! 'ii'xb2 28 .!LJd5 he would still 34...'ii'h S?
have had good drawing chances. A pity. Black could have finished
27 fS!
••• the game immediately with the pretty
An excellent move . Black gets rid stroke 34 ... .i. xe4 ! 35 .i. xe4 'ikxe4, for
of his backward f-pawn and activates example 36 h3 •xel + 37 l:r. xel llfl+
the f8-rook without actually moving 38 'ii?h2 l:r. xe l 39 .!LJxel .te5.
it! 3S 'iFdl
2s gxrs :exes After 35 'ii'g l .te5 36 l:r. e3 .:td2, fol­
The additional threats which result lowed by . . . l:r.ff2, White would be to­
from the penetration of the black rook tally paralysed.
110 SECRETS OF PRACTICAL CHESS

35 .. .'ili'xdl + 36 l:r.xdl i.xe4 37 h4? h-file in order to deliver mate on h7.

Loses at once. The only chance was However, White must be careful be­
37 i. c4+ �g7 38 ll g4 i.f3 39 .:r.gxd4 cause Black threatens to defend h7 by
l:r. xg2 40 .t d5 i. xd5 4 1 l:r. xd5 l:r. xc2 . . . bxc3 followed by . . . 1kxe4. The solu­
(4 1 . . . .:r.ff2 42 .:r. d7+ <iti> g6 43 .:r. 1 d6+ tion appears simple: White must play
�h5 44 l:l h7 l:r. xh2+ 45 <iti g l may not his gl -rook to h4, so as to both defend
be a win), but even here Black has ex­ h7 and cover e4.
cellent winning prospects. That is indeed what happened in
37 .te5 38 i.xe4 i.xg3 39 �e3
•.• one of the first games to reach this po­
l:r.h2+ 40 'iPgl :Xh4 41 i.d5+ <iti>g7 42 sition, Chandler-Yudasin, Minsk 1 982,
c;#o>g2 i.f2 0-1 but after the further moves 18 .:r.g4
bxc3 19 l:r.h4 cxb2+ 20 .txb2 White
It goes without saying that the at­ was shocked by 20 ... 'iFxe4! 21 llxe4
tacker should strive to prevent such a i.xe4 (D).
defensive sacrifice, although this often
involves being quite far-sighted.

1 e4 c5 2 �f3 liX6 3 d4 cxd4 4 �d4

�f6 S �c3 d6 6 .tc4 e6 7 .te3 .te7 8
'ii'e2 a6 9 0-0-0 1kc7 10 .tb3 0-0 1 1 g4
�d7 12 l:r.hgl � 13 g5 b5 14 �xc6
�xb3+ 15 axb3 'ii'xc6 16 'ii'hS b4 17
.td4 .tb7 (D)

Once again, if we compare the dia­

grams then we can see the change that
has taken place. Black's light-squared
bishop is absolutely secure on the c2-
g6 diagonal and while it is there White
has no attacking chances on the king­
side. Indeed, it is now White who has
to think about defence, since c2 is very
weak and Black can open further lines
This is a theoretical position from by advancing his a-pawn. The posi­
the Velimirovic Attack in the Sicilian. tion is in fact favourable for Black and
White's aim is to transfer a rook to the White was soon in trouble : 22 .ta3
 THE MIDDLEGAME Ill

ltfc8 23 :d2 l:ab8 24 <itb2 l:b5 25 h4

i.f6+ 26 'iii>c l i.eS 27 'ii'g4 l:xb3 28
Wxe4 l:xa3 29 �dl . Now, if Black
had played 29 . . . l: ac3, it is doubtful
that White could have survived. In the
game Black played the less forceful
29 g6 and White eventually escaped
•••

with a draw.
However, that is not the end of the
story. James Howell realized that the
sacrifice on e4 is Black's only method
of preventing mate on h7, and so dras­
tic measures to prevent it are justified. Howell 's stunning innovation was
Howell-Wahls, World Junior Champi­ widely publicized and sent Black hunt­
onship, Gausdal 1 986 continued 18 ing for earlier alternatives. These days
lbdS ! (to block the queen's path to e4) you can only expect to get one point
18 exd5 19 l:d3 l:fc8 20 c3 (D).
..• from a piece of homework, but several
Now Black has no reasonable an­ years later Howell was the fortunate
swer to the threat of l: h3 . The finish recipient of a second point - Howell­
was 20 dxe4 21 l:h3 �f8 22 g6!
•.• E.Ragozin, Cannes Open 1 993 contin­
fxg6 23 'ii'xh7 <itie8 24 l:bg6 bxc3 25 ued as above up to move 24, and ended
1Vg8+ �d7 26 'ii'e6+ �d8 27 bxc3 24 <itd7 25 l:xg7 l:r.e8 26 i.f6 'ii' b5
...

.tf8 28 'ii'f7 .te7 29 'ii'xe7+ <itixe7 30 27 c4 'ii'c6 28 l:xe7+ l:xe7 29 'ii'xe7+

l:xg7+ 1-0. <itic8 30 l:r.h7 1-0.
4 The Endga m e

Many games are decided in the end­ King and Pawn endi ngs
game, especially between players of
comparable strength. Mastery of the King and pawn endings very often
endgame is just as important as profi­ represent the final phase of a game. Of
ciency in the opening and middle­ course, both sides may promote, in
game. Even though this truth has been which case the players can look for­
repeated over and over again, the end­ ward to a lot more fun, but the major­
game still remains a neglected area of ity of 'lt>+8 endings are decided in the
chess study, especially amongst club pawn ending itself. Unlike most other
players. types of position, the concept of an
In the past they had some excuse, as 'inaccuracy' is almost unknown in
many club games were decided by ad­ � +8 endings. Given sufficient energy
judication before the endgame was and skill, most positions are capable
reached, but quickplay finishes are now of being analysed to a definite conclu­
the rule rather than the exception. In sion. This means that errors can only
tournaments, too, the quickplay finish occur in half-point jumps. Also, un­
is the most common method of decid­ like other types of position, in which
ing long games. The practical effect is you may recover from a mistake and
that players can no longer rely on an gradually fight your way back into the
'if I get an endgame, I ' ll work it out game, a slip in a �+8 ending usually
over the board' attitude. The fast means the certain loss of half a point.
time-limit implies that you have to This means that accuracy is at a
know the correct method beforehand. special premium in this type of end­
Moreover, familiarity is very impor­ ing, so it is essential to be familiar
tant. If you have to ransack your mem­ with the main principles.
ory for some half-forgotten but vital There are three fundamental con­
snippet of information, the chances cepts in 'lt> +8 endings. Surprisingly,
are that you will have lost on time be­ however, not even all GMs are famil­
fore your memory cells release the iar with their correct application.
necessary information.
This chapter is therefore designed Opposition
as a quick guide to what it is abso­
lutely essential to know about the end­ This is the most basic principle and has
game. the widest application. The following
 THE ENDGAME 113

position provides us with a clear-cut �dS catb7, then White can win in two
example. ways:
1) If White now heads for the g­
pawn, he has gained a vital tempo: 4
�es <itib6 s �rs �cs 6 'iii>xgS <itb4 7
�fS �xa4 8 gS �b4 9 g6 a4 10 g7 a3
11 gS'ii'. It is worthwhile thinking for a
moment about why moving one file to
the queenside has actually gained a
tempo in this line. The point is that af­
ter 3 cat d5 the black king is not only
unable to approach the a-pawn, but
must actually move one square further
away. Thus White loses one tempo but
Black loses two.
The two kings face ('oppose') each 2) 4 �cs (White mops up the a­
other with the minimum possible gap pawn before heading for the kingside)
of one empty square between them. 4 cata6 S �c6 <ita7 6 catbS catb7 7
•••

White has the advantage because his <itixa5 �c6 8 <itb4 �b6 9 <itic4 with an
king is one square further advanced. easy win.
This advantage is sufficient to win if If the black king heads the other
Black is to move, because the black way with 2...<itid6, then White heads
king has to move to one side or the for the a-pawn, having gained an extra
other. This allows White's king to ad­ tempo because his king is one square
vance to the fifth and run to one or nearer the queenside: 3 <itibS <itie5 4
other of the enemy pawns. In this situa­ <itxa5 �f4 S �b4 <itixg4 6 aS �h3 7
tion we say that 'White has the opposi­ a6 g4 8 a7 g3 9 a8'ii' . This position is a
tion' . technical win; if you are unfamiliar
However, even in this relatively with the process, here it is: 9 g2 10•..

simple case there is still one finesse. 'iWf3+ <itb2 11 'ii'f2 'iii> h l 12 .h4+
After 1. ..�c6 White must not run di­ <itigl 13 catc3 �n 14 'ii'f4+ �e2 lS
rectly to the g-pawn with 2 � e5, be­ 'ii'g3 catn 16 'ii'f3+ <itigl 17 'ii'f4 cath1
cause after 2 . . � c5 3 cat f5 � b4 4
. 1s 'ii'h4+ <itigl 19 <itd2 �n 20 'ifet#.
rlt xg5 � xa4 5 <it f5 <it b4 6 g5 a4 both The situation after 1 ...<i&i>e6 is virtu­
sides promote at the same time and the ally symmetrical. Once again White
position is a draw. Instead, 2 <itc4! is must avoid the immediate dash for the
correct, again taking the opposition, pawn with 2 catc5?. The correct method
but under even more favourable cir­ is 2 �e4 <itif6 (2 . . . rltd6 3 cat f5 and
cumstances. If Black plays 2 <itb6 3
••• White promotes first; Black's a-pawn
114 SECRETS OF PRACTICAL CHESS

does not reach the seventh rank) 3 With White to play, one's first im­
<itii dS 'it;f7 4 'it; cS (4 'it;e5 also wins) pulse is to rush forward with the king
4 'it; e6 S 'iii> b S 'it;eS 6 c;k;xaS �f4 7
... by 1 'it> c3?, but this is a mistake: Black
'ii? bS �xg4 8 aS 'iii> h 3, transposing to replies 1 . . . �c7 ! and after 2 � c4 �c6
line 2 above. or 2 'ito>d4 � d6 Black gains the opposi­
While the details of this position in­ tion and White cannot make progress.
clude one or two subtle points, the The basic principle governing such
basic principle is clear enough: when cases of what is called the 'distant op­
White has the opposition, Black must position' is that when the kings face
give way with his king and permit each other with an odd number of
White's own king to advance. It is squares in between, then the player to
worth noting that if White is to move move loses the opposition. The situa­
in the diagram, then Black has the op­ tion when there is just one square in
position, but because of White's ini­ between the kings (discussed above) is
tially more favourable king position, then just a special case of this rule.
he can hold the draw: 1 'iii> c4 (1 'it>e4 In the above position, it follows that
also draws, but any king retreat to the 1 � c2! is the correct move ( 1 �b3? is
third rank loses, as Black can trans­ also bad, since after 1 . . . 'ito> d7 ! Black
pose into the above analysis with re­ can meet 2 'ito> c4 by 2 . . . <itic6 and 2 'ito>c3
versed colours) 1 � c6 (after 1 . . .�e5
... by 2 . �c7). Black replies 1 �d8.
.. ...

both sides promote simultaneously) 2 Now 2 � b3? can be met by 2 . �d7, . .

'it>d4 and Black has nothing better than and of course 2 �c3? 'iii>c7 and 2 <itid3?
to return to d6. <itid7 give Black the opposition. The
The situation becomes only a little correct move is 2 <iti d2!, again in ac­
more complicated when the kings are cordance with our 'odd-square' rule.
further back: At first sight White is not getting any­
where, since if Black just keeps mov­
ing his king up and down the first rank
by 2 �e8, White can apparently only
•••

follow suit on his second rank.

However, White can make progress
by carrying out what is called a 'by­
pass ' . This involves moving in the op­
posite direction to Black's king and at
the same time advancing. If it works,
then the result will be to regain the op­
position, but with the kings two
squares closer together. White can re­
peat the manoeuvre until there is just
 THE ENDGAME 115

one square between the kings and then lose nothing by allowing White's king
we have the situation discussed in the to penetrate in that direction. Simi­
previous diagram. Here White exe­ larly there must be a queenside target
cutes a by-pass with 3 'iii>c 3! . Now or Black could safely move his king to
there is a direct threat to play 4 'it> c4 the other side of the board. The by­
and 5 'it> bS, so Black has to head for pass manoeuvre is almost always per­
c6 by 3 'it>d7. White continues 4 'it>d3
••• formed on the file which is equidistant
and he has achieved his objective. between the two targets . In the above
Now the situation is simpler. 4 . . . c;t> c7 case this is the d-file.
and 4 . . . 'it> e7 lose because White just
heads for the g-pawn or the a-pawn
respectively, so Black's king must ad­
vance to third rank. White just op­
poses the enemy king and gains the
'close' opposition, winning as in the
previous diagram. Had Black played
2... 'itc8, White would have by-passed
on the other side by 3 c;t> e3 .

When I first saw this idea as a very

young player, one point really con­
fused me. On most files White was
content simply to maintain the opposi­
tion but then, suddenly, on one par­ In this position, which is superfi­
ticular file White would abandon the cially similar to the above, White to
opposition and perform a by-pass. play cannot win because there is no
How do you know on which file to by-pass. White does win if he takes
execute the by-pass? Eventually I was the opposition with his king on the
able to answer my own question and in fourth rank, but he cannot force this
doing so gained a deeper understand­ from the diagram. The critical position
ing of king and pawn endings. In posi­ arises after 1 'iii>d2 'iii>d7 2 'itd3 and
tions dominated by the opposition, the now Black must take care. 2 c;t>c6? 3
. . .

attacker has two targets. In the above <itc4 and 2 . . 'it>d6? 3 'it> d4 allow White
.

diagrams these are the pawns on as to gain the 'close' opposition, and in
and g5. This is perfectly natural: the addition Black's king must also be
nature of the opposition is that if ready to stop White lunging for the a­
Black's king goes to the queenside, pawn. It follows that 2 'it>c7! is the
•••

then White's king slips through to the only drawing move. This would be the
kingside and vice versa. If there were ideal time for a by-pass, except that
no kingside target, then Black would the e4-square, which White needs for
116 SECRETS O F PRACTICAL CHESS

the operation, is controlled by a black White to play can draw, but only if
pawn. he starts with the paradoxical move 1
In the above examples, the 'two-tar­ <it h l ! . Other moves fail, for example 1
get' situation was quite obvious, be­ � fl ? (this loses because White cannot
cause the targets were far apart. Cases maintain the close opposition) 1 . . . <it>d2
in which the targets are closer together 2 <it> f2 <itd3 (White would like play 3
are governed by the same principles, <it f3, but his pawn is in the way) 3 � g3
even if their nature is less transparent. �e3 4 � g2 �e2 5 � g3 �fl and Black
reaches a target. Other first moves fail
similarly: 1 <it>h2 �d2 ! (preparing the
by-pass) 2 � gl (2 <it g2 <ite2) 2 . . . �e3
(by-pass executed) 3 � g2 �e2 and
wins, or 1 <it g3 �el 2 � g2 �e2.
After 1 �h i ! Black cannot make
progress, as White can always main­
tain the opposition:
1 ) 1 . . . �e2 2 � g2 � d3 3 <it>h3 !
<it e3 4 � g3, etc.
2) 1 . . .<it> c l 2 �g l ! (the only move)
2 . . . �c2 (2 . . . g4 is met by 3 <it> g2 ! , but
not 3 fxg4? e4 4 �f2 � d2 and Black
H. Neustadtl, 1 890 wins) 3 � g2 <it> c3 4 � g3 � d3 5 �h3
and so on.
In this position the target squares
for Black are (obviously) f3 and (less We end this section on the opposi­
obviously) fl . The latter is a target be­ tion with the usual warning that while
cause if Black's king reaches fl then general principles can provide excel­
White is losing no matter where his lent guidance, in the end it is the spe­
king is (within reason) or who is to cific position on the board that matters
play. For example, if White's king is and a forced win overrides any other
on g3, then, with White to play, 1 <ith3 considerations.
� f2 2 � g4 � g2 wins easily. If Black The following diagram is very simi­
is to play then 1 . .. <it> g l forces the same lar to the Neustadtl position given
line. It follows that the opposition ma­ above. On this basis one might assume
noeuvres will take place along the the position to be a draw, because after
ranks (because the ranks are at right 1 �f6 �b6 or 1 � g6 � a6 Black gains
angles to the line j oining the two tar­ the distant opposition. However, there
gets) and that any by-passing will take is a slight difference because the kings
place on the second rank. are further apart than in the Neustadtl
 THE ENDGAME 117

point, because a new factor entered the

equation.

The Reti manoeuvre

There is no better way to explain this

idea than to give Reti's original exam­
ple, even though it is one of the most
famous endgame positions in chess
history.

A. Mandler
Prager Presse 1 929

position and this operates in White's

favour. After 1 �f6? 'it> b6 play pro­
ceeds exactly as before, but with the
kings on the g- and a-files a second
factor enters the position: the possibil­
ity of a breakthrough by c5. This only
works if Black's king is on the back
rank, so that White will promote with
check. Moreover, Black's king must R. Reti
be on a8, because otherwise c5 can be Kagan 's Neueste
met by . . . <l,;c7 (we saw this in line 2 of Schachnachrichten 1 921
the previous diagram). Thus White can
win by 1 'iPg6! �a6 (Black must main­ White appears to be two tempi short
tain the opposition; if he deviates, then of catching the h-pawn, but he never­
White gains the opposition himself theless manages to overhaul it by si­
and wins by reaching d6 or d8 with multaneously threatening to promote
his king) 2 'it>g7! �a7 3 <itg8! (3 'it>f8? the c-pawn. The analysis runs 1 �g7!
'it> b8 4 c5 'it> c7 !) 3 'it>a8 (Black is
•.. h4 ( 1 . . . � b6 2 <it f6 h4 3 <it e5 trans­
dragged to his doom; 3 . . . 'it> b8 4 � f8, poses to the main line) 2 'it>f6! 'iPb6
3 . . . 'it> b7 4 <itf7 and 3 . . . 'it> b6 4 'it> f8 all (2 . . h3 3 � e7 <itib6 4 'iP d6 and both
.

win for White, the last line being a sides promote at the same time) 3
by-pass) 4 c5 dxcS 5 e5 and promotes <itie5 ! 'it>xc6 (3 . . . h3 4 <Ji; d6 and again
with check. In this case the logic of the both sides promote) 4 'it>f4 and the im­
opposition broke down past a certain possible has been achieved.
118 SECRETS OF PRACTICAL CHESS

If you have not seen this position The use of a second threat as a kind
before, it is worth playing over the so­ of warp-drive for White's king occurs
lution several times to see exactly why in quite a wide range of situations. The
it works. Once again, the 'two-threat' above diagram is a second example in
concept is important, but this time it is which White's king goes on a lengthy
not a question of tempo play, but the detour before catching the enemy
exploitation of the fact that a diagonal pawn.
king move, if measured with a ruler, is In this case the white king is one
longer than an equivalent horizontal or tempo short of catching the h-pawn,
vertical one. In terms of catching the for example 1 a6? <itic6 2 catc8 � b6 3
h-pawn, the routes 'ii? h7-h6-h5 and 'ii?b8 <iti xa6 4 r3; c7 h5 and White is too
'iP g7-f6-e5 are identical. The advan­ late. However, by correctly exploiting
tage of the latter is that it activates a the threat of promoting the a-pawn
second threat, that of promoting the White can make up the time: 1 cat c8 !
c-pawn. It takes Black two tempi to <iti c6 2 <itib8 <itibS 3 r3;b7 (this is the key
deal with this threat, by taking the c­ move; the white king just makes it into
pawn with his king, and this is exactly the square of the pawn) 3 <iti xa5 4
. . .

the time White needs to catch up with catc6 and White is in time.
the h-pawn.
Despite this coldly logical explana­ Triangulation
tion, it is a remarkable idea and there
still seems to be a whiff of magic about
the position.

If it were Black to play in the dia­

gram then he would lose immediately,
as White's king penetrates to b6 and
L. Proke� takes the a-pawn. If it is White to
Sachove Umeni 1 947 move then his task is considerably
harder. The key is a king manoeuvre
 THE ENDGAME 119

which, by losing a tempo, transfers the with mystic numbers). The opposi­
move from White to Black. However, tion, triangulation and a little bit of
before starting this manoeuvre White brain-power are all that are needed
has to play what is, in any case, a forced for the types of position that arise in
move: 1 'iPdS. Now l . . �d8 2 <it>d6 is a
. over-the-board play. Of course there a
win even without the a-pawns, so few esoteric positions which cannot
1 'iPc8 is forced. White cannot then
... be solved by such straightforward
win with the direct 2 <it>d6 <it> d8 3 c7+ techniques, but in all my career I have
<it>c8 because 4 <it>c6 is stalemate - he never seen such a position arise in
must be more subtle. While the white practical play. The following position
king remains adjacent to c5 Black's represents about the limit of complex­
king cannot occupy c7, and so is re­ ity that one can anticipate over the
stricted to shuffling back and forth board.
along the first rank. If the white king
moves around the triangle c4-d4-d5
the effect will be to transfer the move
to Black: 2 'iPc4 �d8 3 �d4 �c8 4
'iPdS and now we have the position af­
ter the first move, but with Black to
play. White wins after 4 ... � d8 5 � d6
or 4 . . . <it> c7 5 �c5.
White's triangulation worked be­
cause of Black's lack of space; he
could not emulate White's manoeuvre
without stepping off the edge of the
board.
If you have looked at an advanced N. Grigoriev
book on king and pawn endings, you K novoi armii, 1 920
will probably have noticed something
called 'the theory of corresponding A position such as this would be an
squares ' . This is normally accompa­ easy win for White if his king were on,
nied by diagrams with lots of little say, d4, but here the win is difficult be­
numbers (or letters, or sometimes both) cause of Black's active king position.
on the squares. While this is an inter­ The first point to note is that after 1
esting subject from the theoretical d4? <it>e4 2 <it>c3 Black does not play
point of view, to be quite honest it is of 2 . . . <it> d5 ?, when 3 <it>d3 forces the en­
almost no practical value (not to men­ emy king back, but 2 . <it> f5 ! . If White
. .

tion that your opponent might object moves anywhere except to d3, then
when you start adorning the board Black just returns to e4, but after 3
120 SECRETS OF PRACTICAL CHESS

� d3 � f4 White cannot make prog­ <itdl ! (5 �d2 � d4 forces White to re-

ress, e.g. 4 <it d2 �e4 5 'iii>c 3 'iii>f5 !, etc. trace his steps with 6 � c2) S ...'it> fS 6
It becomes apparent that White 'it>e2 'it>f4 7 �f2 �eS (after 7 . . . 'it> f5 8
would be much better off in the dia­ �e3 �e5 9 d4+ 'iii>d 5 1 0 'iii> d3 White
gram position if it were Black to play. wins by driving Black's king back) 8
If Black's king goes anywhere except �e3 �dS (the critical moment; if
to f4, then White wins by 2 'it> e3 or 2 White is to make progress he must
d4, while after 1 . . . 'iii> f4 White plays 2 give up the b4-pawn) 9 d4! �c4 10
�e2 and the black king is gradually 'it>e4 'it>xb4 11 dS 'it>cS (forced, or else
forced back. In fact the win is still not White promotes before Black) 12 'iii>eS
too easy, but we will return to this (White must spend a tempo support­
point later as the first step is to work ing his pawn, so now both sides pro­
out how to lose a tempo. If White mote at the same time; however, . ) . .

plays 1 � c2, then Black must reply 12 ... b4 13 d6 b3 (interposing 1 3 ... <itc6
1 . . . � f4 (since l . . . 'it>e3 2 <itc3 loses at 1 4 <it e6 doesn't change the situation)
once, and after other moves White 14 d7 b2 15 d8'ii' bl'ii' 16 'ii'c8+ and
wins by either 'it>c3-d4 or 2 d4). Read­ 17 'ii'b8+, winning Black's queen.
ers may now recognize a pattern emerg­
ing: the situation is basically the same Expect the unexpected
as in the previous diagram rotated
through 90 degrees, with c3-e3 taking Readers will have noticed several odd
the place of c5-c7 and c2-f4 the place king manoeuvres in the preceding po­
of d5-c8 . Now the solution should be sitions. This is no coincidence; para­
apparent. While the white king remains doxical king moves are quite common
adj acent to c3, Black's king must re­ in king and pawn endings. It is very
main adj acent to e 3 ; in other words important to take into account the pos­
Black can only oscillate between f3 sibility of 'unnatural' king moves,
and f4 (e2 and f2 are out because the both for you and for your opponent. It
d-pawn advances). White only has to is impossible to give rules for finding
triangulate c2-b2-b3-c2, all the time such moves, precisely because they
staying adjacent to c3, and he will lose are often ' one-off' events which only
a tempo. The solution runs 1 �c2 �f4 work in that precise position. How­
2 �b2 <itf3 (2 . . . � e5 3 <it c3 <itd5 4 ever, this won't prevent me from giv­
'it>c2 'it> e5 5 'iii> d l transposes) 3 <itb3 ing a few entertaining examples.
'it> f4 4 �c2 (unlike the previous dia­
gram, there is still quite a lot of play Shirov had headed for the position
left in the position) 4 �eS (the most
•.. in the following diagram because it
resilient defence; 4 . . . 'it>e3 5 <it c3 and seems an obvious draw after, for ex­
4 . . . 'it> f3 5 'it>d2 lose more quickly) 5 ample 1 <it g5 � g2 2 f4 � f3 3 f5 'it>e4.
 THE ENDGAME 121

I. Rogers - A. Shirov M. Hebden - G. Flear

Groningen 1990 British Championship, Brighton 1980

However, after the surprising 1 'it> g3! agreed the players suddenly noticed
Shirov resigned. The threat is simply that White could play 3 'it> f5 ! (D) in­
to push the pawn to f6 and then run stead of 3 <it xf4.
with the king to take the f7-pawn, si­
multaneously defending White 's own
pawn. Black can only try to extract his
king from the box by 1 . . . 'it> g l 2 f4 'it> fl
3 f5 <ite2, but White can keep Black's
king at bay: 4 � f4 f6 (or 4 . . . 'it> d3 5
'it>e5 and wins after 5 . . . 'it>e3 6 f6 or
5 . . . <it c4 6 <it f6) 5 <ite4 ! � f2 6 <it d5 .
The backward move of White's king is
sufficiently counter-intuitive that even
such a strong player as Shirov over­
looked it.

In Hebden-Flear, Black had been a Suddenly Black is in trouble. After

pawn up for most of the game, but here 3 . . . 'it> d4 4 'iii> xf4 'it> d3 5 'it> f5 'iii> e 3 6
he abandoned his winning attempts <it g6 'it>f4 7 'iii> x g7 'it> g5 8 f4+ 'it> xh5 9
and played 1 .. . l:r.d5? ! , at which point f5 the f-pawn promotes, while 3 . . . �d6
the players agreed to a draw. Indeed, 4 <it g6 <ite6 (4 . . . 'it> e5 5 <it xg7 'it>d4 6
after 2 : xd5+ � xd5 3 'iii> xf4 �e6 'iii> xh6 'it>e3 7 <it g5 ! 'it> xf3 8 h6 'it> g3 9
there is clearly no point in continuing. h7 f3 1 0 h8 'ii f2 1 1 'ii'h l ; in this line
However, as soon as the draw was Black lost because he was not able to
122 SECRETS OF PRACTICAL CHESS

play his king to g2 at move 8) 5 <l,; xg7 trying to fulfil two tasks at the same
<l,; e7 6 <Ji; xh6 <l,;f6 is hopeless because time.
White can extract his king by using
up his one remaining tempo with the Chess is more than counting
h-pawn: 7 �h7 � fl 8 h6 <itii f8 9 <t> g6
� g8 10 'it> f5 'it> h7 1 1 'iii> g 5 followed by One technique applicable in king and
1 2 � xf4. pawn endings, which is often recom­
Having analysed both 3 . . . 'it> d4 and mended in textbooks, is that of 'coun­
3 . . .<Ji;d6 to a loss, one might be tempted ting ' . This name refers to the method
to assess the position as a win for of working out how many moves both
White, but remember: 'Expect the Un­ sides will take to promote a pawn. Of
expected' . What does Black need to course, this is only useful in 'race'
achieve in order to draw? First of all, situations, where each king marches
he must be able to defend his g7-pawn into the opposing pawn-mass with the
if White plays � xf4 followed by aim of creating a passed pawn as
� f5-g6. That means that his king quickly as possible.
must stay within three squares of f8 . If I strongly advise against using the
White adopts his alternative plan of technique of 'counting' .
heading straight for g7 without both­ There are several flaws with this
ering about the f4-pawn, then Black method, and its unrestricted applica­
has to be able to answer 'it>xh6 with tion can result in dreadful blunders .
. . . 'iii> xf3 ; one move slower and White The main problem is that this method
will win as in the bracket given above. cannot detect any of the finesses which
Since it takes three moves for White to make king and pawn endings so inter­
reach h6, it follows that Black's king esting and which arise quite often in
must stay within three squares of f3 . practice.
The first criterion implies that Black The first cause of confusion is that
must move to c5, c6 or d6 and the sec­ you have to take account of who moves
ond implies that he must move to c4, first. If White takes seven moves to
d4 or c5. Luckily for Black, there is promote and Black eight, when White
one square common to both these sets, promotes does Black reply by promot­
namely c5. Thus 3 . . . 'it>c5 ! ! draws, e.g. ing or by pushing his pawn to the sev­
4 'iii> g6 'it> d4, 4 'iii> xf4 <Ji; d6 or 4 <J;; e s enth? The answer is that it depends on
'it>c6. who moves first. In the heat of battle it
It is quite hard to describe exactly is quite easy to get this wrong and be
what this motif depends on - it cer­ 'out by one' (computer programmers
tainly has nothing to do with the oppo­ know all about being 'out by one' ).
sition. Perhaps it is closest to the Reti The second cause is that the count­
manoeuvre, in that Black's king is ing method gives you no clue as to the
 THE ENDGAME 123

relative whereabouts of the pieces at � c4 it is all over. White captures the

the end of the race. White may pro­ black pawns on f4, h4, and g3 while
mote and give check, or promote on Black takes the a4-pawn and walks
h8 and stop Black's promotion on al . back to the kingside."
These things are obvious if you run the Lasker does not consider what hap­
variation through in your head, but are pens if Black plays 2 . . . 'iii>d4. By count­
not obvious if you are just counting. ing (carefully ! ) one quickly finds that
The simple fact is that you actually Black's g-pawn is still on the sixth
have to calculate the variation to make when White queens and, moreover, it
sure that one of these special situa­ is White's turn to move - enough rea­
tions doesn't arise at the end of the son to abandon the position as lost, if
race and if you are going to do that, one does not take the trouble to visual­
then you may as well not waste time ize the resulting position.
on counting. If we actually play over the moves
2 . . . 'ifi>d4 3 a5 �e3 4 a6 'it> f2 5 a7 'iii> x g2
I will quote a position taken from 6 a8'if �h2 (D), then we arrive at the
Chess: The Complete Self-Tutor by following diagram:
Edward Lasker (it appears in all edi­
tions before 1 997), a popular chess
primer which advocates the use of the
counting method.

It only takes a few moments to real­

ize that this is a special situation. The
congested mass of kingside pawns
prevents White from pinning the g­
Lasker writes "In the position pawn diagonally and from giving
reached, the sequence 1 b4+ axb4+ 2 check on the h-file. In fact the g-pawn
'it> b3 forces Black to leave his pawn is unstoppable and the best White can
unprotected, giving White a passed do is to grab as many kingside pawns
pawn. After 2 . . . 'it>c6 3 � xb4 � b6 4 as he can, hoping for a favourable
124 SECRETS OF PRACTICAL CHESS

queen ending. However, even this hope 2 . . . <iti b4 3 � c6 'iii>a5 4 <iti b7 he cannot
is in vain: after 7 'ii'b8 (7 'ii'h8 g2 8 move to b6.
'iFxh4 g l 'ii' 9 'ii'xf4+ 'iig 3 1 0 'iixf5 There is far more to king and pawn
'ii'x h3 1 1 'ii'g4 ! ? 'ii'g 3 and 12 ... � g2 is endings than counting, and if you use
also a draw) 7 . . . g2 8 'ii'xf4+ �b l 9 this method you are asking for trouble.
'ii'xf5 g l 'ii' the ending is a sure draw.
Finally, if you don't calculate the Rook endi ngs
whole variation, how can you be sure
that the moves you are intending to Rook endings are the most common
play are actually possible? Here is a type of ending to occur in practical
simple example. play. The reason is that if there are no
open files, a relatively frequent occur­
rence, then exchanges of rooks are
rather unlikely. Minor pieces and
queens can be exchanged far more
readily. The ability to play rook and
pawn endings well is a great distin­
guishing feature between a master
player and a club player. In simultane­
ous displays, it is noticeable how the
club players routinely lose completely
equal rook endings, and how easily the
master manages to escape with a draw
from lost rook endings.
W. Schlage C. Ahues
- Part of the reason for this is that
Berlin 1 921 rook and pawn endings are genuinely
complicated, with a great deal of the­
It takes White five moves to take on ory. Experience is perhaps even more
a7 and it takes Black five moves to important; the master will probably
reach c7, so Black can meet � xa7 by have encountered many rook endings
. . . �c7 and the position is a draw. This in the course of his career, the club
was borne out by the game, which player far fewer. However, the situa­
duly finished 1 <itie6 �c3 2 <iti>d6 �d4 tion for club players is certainly far
3 �c6 <itieS 4 <itb7 <itd6 S <itixa7 �c7 from hopeless; indeed, this is one of
1/z-1/z . Right? Wrong ! If White had those areas in which a few hours'
played 2 �d5 ! then he could still take study can provide great dividends. We
on a7 in five moves, but Black would start with the ending of rook + pawn
take longer as the moves he wants to against rook, which is fundamental to
play are illegal, for example after all rook endings.
 THE ENDGAME 125

Rook and Pawn vs Rook fifth rank. Nevertheless, the position is

completely drawn because Black has
This ending is sufficiently complex positioned his rook on his third rank,
that your author was able write a 320- thereby preventing the immediate ad­
page book solely about this material vance of the white king. The only way
balance (and even then several inter­ in which White can get his king to the
esting positions had to be omitted ! ) . sixth is to play I e6, but then Black
However, many average tournament switches plans and plays 1 . . . l:b l , pre­
players have not studied even the fun­ paring to bombard the white king with
damentals of this ending. In fact, a few checks from behind. Now that the
basic principles can be quite effective pawn is on e6, there is no hiding place
in improving one's rook endgame for the king, and after 2 � f6 :n + 3
technique. �e5 l:.el+ 4 �d6 l:.d l +, etc., the posi­
If the defender's king can occupy tion is clearly drawn. If, in the diagram
the pawn's queening square, then the position, White waits with I l:l a7 then
ending is almost always a draw. The Black does likewise by either 1 . . . l:. c6
only exception is when the pawn is al­ or l . . . l:. h6.
ready far advanced and the defender's
pieces are badly placed. The general
drawing technique is quite simple and
is called the 'third-rank defence' .

The second important principle is

that of 'checking distance' . If the white
king has no way to hide from a check­
ing black rook, then the factor deter­
Here is a typical position. White ap­ mining whether the checks are a minor
pears to have made quite a lot of prog­ nuisance or a serious problem is the
ress. His rook is cutting the enemy distance between the king and the
king off on the back rank and his king rook. The further apart they are, the
and pawn have advanced as far as the better it is for the side with the rook.
126 SECRETS OF PRACTICAL CHESS

The above example makes this clear. try 4 :rs ( 4 'iii>e6 � g7 ! and 4 : as � g8
If it is White to play then he should do not help White) intending 5 <iti e7,
continue 1 'iii> g 6, which creates the followed by 6 f6. If Black keeps his
most awkward problems for Black. If rook in its current position then he will
White waits, then Black sets up the eventually lose, but Black can change
third-rank defence by . . . :b6. 1 �f6 direction by 4 l::l a l ! . If White now
.••

l::l b6+ leads to the same thing, while 1 plays a king move to free his pawn,
f6 : g t + is an immediate draw (note then Black starts checking from the
the large checking distance). side. There is no shelter, so in view of
After 1 � g6 it is easy for Black to the large checking distance White
go wrong, for example the passive line cannot effectively meet the barrage of
1 . . . l::lb 8 2 f6 � g8 fails to 3 l::l g7+ <itf8 checks. White can, of course, play
(or 3 . . . <ith8 4 l::l h7+ <iti g8 5 f7+) 4 :h7 other moves but they do not help, for
and White's plan of 'switching to the example after 5 l::le 8 (to meet 5 . . . l::la6+
other side' leaves Black defenceless. by 6 l::le 6) the simplest draw is to re­
Note that this line would be a draw if turn to fl with the rook.
White's pawn were on g6 and his king On the other hand, 2 . . . �e8? loses.
on h6, because there would be no After 3 l::la8+ <itid7 4 l::lf8, there is noth­
room to switch to the other side. This ing Black can do to prevent White
leads us to the important conclusion playing 5 <it g7 (or � f7, if the rook
that in the type of position in which leaves the f-file), followed by f6, and
Black has been forced to defend pas­ there is no way to stop the pawn.
sively with his rook on the first rank, Black's problem is that he lacks the
the result is a draw with an a-, b-, g- or lateral checks which saved him after
h-pawn, but lost with a pawn on one of 2 . . . <it g8. There is no space for his rook
the other four files. on the kingside, and he cannot give
Here we have an f-pawn so this de­ checks from the queenside because his
fensive plan is not good enough for own king gets in the way.
Black. Instead, he should play 1 ..l::lfl.. This principle is usually formulated
This ensures that after 2 l::l a 8+ <it e7 as 'move the king to the short side' in
White cannot play 3 f6+. White can order to leave the other side free for
only make progress by playing 2 'iii>f6, long-range rook checks, but in the end
which threatens mate. Black has to it all comes down to ensuring adequate
choose between 2 . . . �e8 and 2 . . . � g8. checking distance.
One move loses and the other draws. Here is another example (D):
The correct choice depends on en­ Black is to play, and he faces the
suring that Black has sufficient check­ threat of <it c4 followed by b5 . If he
ing distance in the subsequent play. plays I . . . l::lh5, then White advances his
After 2 � g8! 3 l::la8+ � h7 White can
... pawn by 2 'iii> a4. The only possibility
 THE ENDGAME 127

3 ... ltb7+ 4 �c5 l:r. c7+ 5 � b6 is hope­

less) 4 �a6 :as+ 5 �b7, followed by
b5, and White wins.
We already saw in the analysis on
page 126 how the defender must be
prepared to use his rook flexibly and
operate from different directions ac­
cording to changing circumstances.
Here is another, very important, exam­
ple of this:

to defend is to retreat the rook along

the b-file, so then when the white king
moves to a4 or c4, Black can start
checking. In order for this to work,
Black needs to have the maximum
possible checking distance and in fact
1 .:r.bS! is the only move to draw. Af­
...

ter 2 l:d4 (2 <iti c4 lt cS+ 3 � b5 l:r. bS+

4 <itc5 ltcS+ 5 �b6 :tbS+ is pointless
- the white king has to retreat; after
the text-move, however, this line is a
threat since White could play <itic7 at Black cannot move his king to the
the end) 2 <itieS (dislodging the rook
... queenside because after . . . <itie7 White
from the defence of the b-pawn) 3 l:r.d7 replies a7 and the king is caught be­
<it e6 (Black cannot wait; 3 . . . �e4? loses tween two stools. After either . . . <it d7
to 4 <it c4 :cs+ 5 <iti b5 .:r.bs+ 6 <it c5 or . . . <itif7 White replies .:r. hs, winning
l:r.cS+ 7 <it b6 l:r. bS+ S .:r.b7, followed by the enemy rook. It follows that Black's
b5; in this line White's rook effec­ king is trapped in an invisible box on
tively reduced Black's checking dis­ the kingside, consisting of the marked
tance by being able to interpose on the squares in the above diagram (plus g7,
seventh rank) 4 l:r. a7 �d6 5 �a4 �c6 the square the king is currently stand­
and Black is in time to prevent the ing on). Black can only use his rook to
pawn's advance. If the checking dis­ defend, and he must counter White's
tance had been even one square less threat of playing his king over to de­
Black would have lost, for example fend the pawn, releasing his rook and
1 . . .:r.b7? 2 �c4 ltc7+ 3 � b5 lt cS (try­
. clearing the way for the pawn. If Black
ing to regain his checking distance; checks from behind, the white king
128 SECRETS OF PRACTICAL CHESS

hides on a7 (this is why White must diagram, then Black has no time to set
not push his pawn to a7 prematurely) up the Vancura draw. The precise de­
and again his rook is freed. tails of how close White has to be to
Black to play draws by being flexi­ win are rather complicated and may be
ble with his rook: 1 .. . l:r. fl + 2 <ite4 l:r.f6 ! . found in specialized endgame books.
This change o f direction i s the key;
a7 is no hiding place when Black is The extra Pawn
checking from the side. Play might
continue 3 'iii> d5 : b6 4 'iii>c 5 (4 l:r. a7+ One of the most common endgame
<it g6 5 catc5 l:r. f6 is the same) 4 ... :f6 5 situations in practical play is a rook
<it> b5 (now White threatens to move his and pawn ending with an extra pawn
rook, so Black must start checking) on one side of the board, and the posi­
5 . . . l:r. f5+ 6 � c4 l:r. f4+ 7 �d5 l:r. f6 and tion otherwise symmetrical.
White cannot make progress. If White Here is a typical situation:
plays a7 at any stage, Black replies
. . . l:r. a6 followed by . . . : al , and if the
white king then approaches the pawn,
Black draws by checking the king from
behind (the nook on a7 is no longer
available). This double change of di­
rection (al to f6 and then back to al ) is
quite hard to see, and in fact it was not
discovered until 1 924 by Vancura.
Several years ago I was asked to
check some rook ending analysis, but I
couldn't understand any of it - White,
who had the advantage, seemed to be
forcing Black to reach a drawn posi­ An ability to win positions such as
tion, which Black then avoided ! Fi­ this is fundamental to successful chess.
nally, I looked at the date of the White may have won a pawn with a
analysis - 1 9 1 2 - and realized that be­ combination, or perhaps Black had an
cause the Vancura draw wasn't known isolated e-pawn which White picked
at that time, all positions of the type up with subtle strategic play - what­
shown in the previous diagram were ever your chess style, situations in
thought to be winning for White. Not which you win a pawn for essentially
surprisingly, this led to some odd no compensation are very common.
moves. The question then arises: how often do
I should add that if White's king is you convert the resulting positions
nearer to the a-pawn in the above into a full point?
 THE ENDGAME 129

I showed this position to some of like it, nor do standard works on the
the world's leading grandmasters, in­ endgame such as Levenfish and Smys­
cluding Karpov, Kramnik and Anders­ lov's famous Rook Endings.
son, and asked them to estimate their Fine's old Basic Chess Endings
winning chances as White, assuming contains perhaps the most helpful ad­
an opponent of equal strength. The an­ vice. He recommended that if the de­
swers were all in the range 80-90% . fender's king is on the side where the
The general feeling was that such a po­ pawns are balanced then you should
sition should obj ectively be a win, and adopt the following plan:
the 'missing' 1 0-20% represented the 1 ) place the king and rook on the
probability of making a mistake one­ best possible squares;
self rather than the probability of run­ 2) advance the pawns on the other
ning into exceptionally good defence. wing as far as convenient, without ac­
A search of 500,000 games re­ tually setting up a passed pawn;
vealed 94 clear-cut examples of this 3) transpose to a won ending with
type of structure (3 vs 3 on one side, 3 an outside passed pawn.
vs 2 on the other, no doubled pawns What he means is that you should
and pawns not blocked). The result not necessarily create the passed pawn
was 72 wins, 19 draws and 3 losses( !). as quickly as possible, as some posi­
The 77% success rate was slightly tions with an extra outside passed
lower than the GMs thought, but at pawn are drawn. Instead, you should
their level technique is better and so, continue to manoeuvre until the passed
assuming that the position is objec­ pawn can be created under favourable
tively won, they would be more likely circumstances, for example with your
to round up the point. They probably rook behind it.
also wouldn't lose a rook ending with If the defender's king is on the side
a clear extra pawn ! with the potential passed pawn, then
Unfortunately, such situations are the decisive manoeuvre is the penetra­
poorly covered in endgame theory tion of the attacking king on the oppo­
books, which tend to concentrate on site wing.
more simplified positions. While they Having played over all the 94 ex­
devote a lot of attention to the situation amples mentioned above, I failed to
with 3 vs 3 on one side and an outside find any clear-cut examples of this
passed pawn on the other wing, the winning plan. Either the ending was
situation in which there is a majority misplayed by one side or the other, or
rather than a passed pawn is hardly the defender seemed to assume that he
covered at all. The massive Encyclo­ would lose in the long run and em­
paedia of Chess Endings, with 1 727 barked on a desperate bid for counter­
positions, doesn't contain anything play. Thus readers will have to make
130 SECRETS OF PRACTICAL CHESS

do with Nunn vs Fritz5 from the dia­ the black rook may switch behind the
gram on page 128: pawns, and this could prove very awk­
ward if Black has pawns on a4 and b4.
l h4 Now that Black has advanced a pawn
We are in the second of Fine's cases to a4, White can conveniently halt the
mentioned above, so White 's long­ pawns by a3 and c3, which also makes
term aim will be to penetrate amongst his quecnside safe against an attack by
the queenside pawns with his king. Black's rook.
However, at the moment this pawn 4 ... l::td l 5 h5
structure offers no avenues of approach In rook endings it is quite often a
for the king, so White must find a way good idea to have such a chain of
to induce the pawns to advance. One pawns . If the white rook defends the
plan would be to play 'ito>el and then at­ base of the chain, then all the pawns
tack the pawns by .:r.e3-a3, etc., forc­ are secure against enemy rook attack.
ing them to move forwards. Then 5 ... l::t b l 6 c3 g5
White could play lld3 and 'it> d2, cross An attempt to obstruct White on the
the d-file with his king and then try to kingside.
get amongst the pawns. 7 hxg6+ hxg6 8 'ittf4 l::td l (D)
Perhaps this plan is the most sys­
tematic, but I decided on another idea,
that of advancing the kingside pawns.
If Black remains passive, White will
gain a great deal of space on the king­
side and should eventually be able to
create a passed pawn there under fa­
vourable circumstances. If Black tries
to generate counterplay by advancing
his queenside pawns, then White will
again have the possibility of switching
his king to the other side of the board.
1 ... b5
I doubt that a human would defend 9 'it>e3
in such forthright style ! Fritz decides This is where Fine's advice came in
to gain space on the queenside, but it is helpful. The immediate attempt to ad­
creating precisely the sort of openings vance with the king leads nowhere,
that will help White later on. since after 9 'it>e5 c6 Black is ready to
2 g4 as 3 'itg3 a4 4 a3 drive the king back with . . . l::t d 5+. In­
It is a good idea not to let the queen­ stead, White intends l::td2, followed by
side pawns advance too far. Later on 'it>d4-c5. In fact Black can prevent this
 THE ENDGAME 131

plan, but White loses nothing by try­ 16 l:xa3 17 :c7+ cate6 18 <it>xg6
•••

ing it before considering other ideas. <itd6 19 :cs <it>d7 20 :cs

9 ... c;i?f6? Black cannot continue his attack on
An error, allowing White to execute the rook as the b5-pawn is hanging.
his plan. 9 .. :n was also bad since 10
. 20 l:b3 21 rs a3 22 f6 :bl 23 f7
•..

l:f2 J:d 1 1 1 J:d2 :n 1 2 l: d7+ c;i;>f6 13 :n 24 :XbS a2 25 l:aS c;i;>e7 26 l:xa2

l: xc7 l:bl 1 4 l:c6+ transposes to the l:f6+ 27 <iti>gS 1-0
game.
The correct move was 9 . . . <ite6 ! , This, incidentally, provides an ex­
when 1 0 l:d2 l: e l + 1 1 �d4 <it d6 pre­ ample of using computer programs for
vents the king reaching c5. Instead 1 0 training purposes. It is possible to set
<ite4 i s probably best, followed b y f4 up simplified, idealized positions and
and the possible creation of a kingside play them out against the computer.
passed pawn. The best method is to play the same
1o :d2 :n position several times for both col­
1 0... l:e l + 1 1 �d4 l:fl 12 �c5 l: xf3 ours, trying different ideas to see what
1 3 <it> xb5 is hopeless - Black's queen­ works and what doesn' t. This kind of
side pawns are too weak. experience cannot be gained in over­
l l l:dS the-board play. We will return to this
Now White can exploit the queen­ subject in Chapter 5 .
side pawns with the rook instead of the
king. Many endings with an extra pawn
11 c6 12 l:cS :bt 13 l:xc6+ <it>f7
••• eventually reduce to the situation in
14 <itf4 l:xb2 ts �gs which the pawns are equal on one side
This more or less forced sequence and one player has an outside passed
has left White very well placed. He will pawn on the opposite wing . The re­
inevitably win the g-pawn, thereby mainder of the section deals with this
gaining two connected passed pawns common occurrence. For the sake of
on the kingside, whereas Black cannot simplicity, let's assume that White
create two passed pawns on the queen­ possesses an extra pawn on the queen­
side. side. White's rook will need to defend
1S l:b3 16 f4!
••• the pawn, and there are three possible
A small finesse avoiding Fritz's positions for the rook: behind the
trap. After 1 6 l: c7+? �e6 1 7 <it> xg6 pawn, to the side of the pawn, and in
<it> d6 ! Black either perpetually attacks front of the pawn. This list is arranged
the rook or drives it away from the de­ in descending order of desirability
fence of c3, when Black can also gain from White's point of view.
two connected passed pawns (e.g. 1 8 The situation in which White's rook
l:c8 catd7 1 9 :bs <it>c7, etc.). is behind the pawn is a familiar one.
132 SECRETS OF PRACTICAL CHESS

The pawn is continually threatening to situation is just as if he had lost the op­
advance, so Black has to block the position in a king and pawn ending.
pawn with his rook. Not only does the After . . . c;i? d5 , White plays � b4 and
rook become tied down, but Black is supports his pawn with the king; after
left open to the possibility of zugzwang. . . . �b5, White can head for the king­
The classic example of this situation is side pawns with his king. Black can
also one of the best: only avoid these lines by playing pawn
moves on the kingside, but White can
always move his rook up and down, so
that Black is sure to fall into zugzwang
sooner or later.
S9 ... liPb5 60 liPb3
Black is helpless, so White can af­
ford to gain time on the clock by re­
peating moves.
60...c;tcS
After 60 . . . l:l xa5 6 1 l:l xa5+ c;i;> xa5,
White wins by 62 �c4 liPb6 63 liP d5
�c7 64 � e5 c;1;> d7 65 c;i;> f6 c;1;>e8 66 f4
!iPf8 67 f5 gxf5 68 � xf5 and 69 <iti g5.
A. Alekhine J. Capablanca
- 61 c;1;>c3 �bS 62 c;1;>d4
Buenos Aires World Now the king heads for the vulner­
Championship (34) 1 927 able enemy pawns. Black's king is
too far away to prevent this, so Capa­
54 l:la4 blanca decides to use his king to
In accordance with the above prin­ blockade the a-pawn and his rook to
ciple, the rook is better behind the fight against the penetration of the
pawn than to the side of it. white king. This is indeed the best de­
54...c;i?f6 fence, but however Black plays he
White's immediate threat is to run cannot overcome the fatal handicap of
his king to b5 and lift the blockade of having one or other of his pieces fully
the pawn. Black can prevent this using occupied blockading the a-pawn.
his own king, but he quickly runs into 62....:r.d6+
zugzwang. If Black tries to arrange a perpetual
55 liPf3 �es 56 �e3 hS S7 c;i;>d3 attack on White's rook by 62 ... �b4 63
!iPdS 58 !iPc3 �cs S9 l:la2 l:l al c;i;> b3, then 64 c;1;> c5 <itib2 65 'iil' b 5
This is the reason why having the wins.
rook behind the pawn is so favourable. 63 !iPeS l:le6+ 64 liPf4 c;i;>a6 6S <itigS
Since Black cannot move his rook, the l:leS+ 66 <itih6 :rs (DJ
 THE ENDGAME 133

The alternative is to defend the f7- 7 1 . ..gxf5 72 � xf5 l:. c5+ 73 � f6 l:. c7
pawn along the second rank by 66 ...l:.e7 74 l:.f3 � xa5 75 l:.f5+ wins easily.
67 ril g7 l:l d7 68 � f6 l:k 7, but White 70 gxrS 71 �h6 r4 72 gxr4 l:.dS
•••

wins after 69 l:.e2 � xa5 70 l:. e7 l:.c2 73 'it>g7 :rs 74 l:.a4 'it>bS 7S l:. e4 �a6
7 1 l:. xf7 l:. xf2+ 72 � xg6 l:. g2 73 l:. f3 . 76 �h6 l:.xaS
Black could have put up more resis­
tance by 76 . . . �b7, but White can still
win with 77 l:.e5 l:. xf4 78 � g5 l:. fl 79
l:.f5 ! l:.gl + 80 <it xh5 'it>a6 8 1 �h6, and
so on.
77 l:.eS l:.al 78 'it>xhS l:.gl 79 l:.gS
l:.hl 80 :rs �b6 81 :xr7 �c6 82
l:.e7 1-0

If you have an extra outside passed

pawn and your rook is behind the
pawn, then you should have very good
winning chances. As with any pro­
67 r4?! posed rule, there are exceptions (for
By exposing the pawn on g3, Ale­ example, a position with rook on al
khine makes the win more difficult. and pawn on a2 against rook on a3
After 67 <j;; g 7 l:.f3 68 'iii> g 8 f5 (the only would be far less favourable ! ) , but in
chance, as 68 . . . l:. f6 69 f4 l:.f5 70 � g7 normal situations such positions are
wins immediately) 69 � g7 f4 70 � xg6 extremely favourable for White. The
fxg3 7 1 fxg3 l:. xg3+ 72 'it> xh5 l:.c3 73 further away from the kingside the
l:. a4 ritb5 74 l:. f4 � xa5 75 'it>g5 the h­ pawn is, the better for White, but even
pawn decides the game without diffi­ with a c-pawn his winning prospects
culty. would be good.
As Speelman points out, White can When the rook defends the pawn
even avoid the minimal counterplay of from the side (we will assume the rook
. . . f5-f4 by continuing 67 'it> h7 l:. f3 is to the right of the pawn), Black will
(67 . . . l:. f6 68 'it> g8 l:. f5 69 f4) 68 'it> g7 place his rook behind the pawn. Then
l:.f5 69"f4, winning straight away. much depends on the kingside pawn
67 :cs 68 l:.a3 l:.c7 69 'it>g7 l:.d7
•.• structure. The white rook is vulner­
70 rS?! able to attack by Black's king, and the
Once again introducing unneces­ best situation arises when the rook can
sary complications. 70 'iii> f6 l:.c7 7 1 f5 occupy an invulnerable square on the
(the l:. e3-e7 plan is less effective now, kingside where it is defended by a
since Black can target the g3-pawn) pawn.
134 SECRETS OF PRACTICAL CHESS

<it d3 with an easy win as the king ad­

vances to support the b-pawn.
The course of the game is a warning
against trying to be too clever in tech­
nical positions. If you see a systematic
and safe winning procedure, don' t
worry about the fact that there may be
a quicker win; just play it.
49 .l:r.h2??
The idea is that Black has to check,
or else White switches his rook behind
the pawn, but then White's rook can
B. Spassky - Zhu Chen return to h5 and he has released his
Veterans vs Women, king from the back rank immediately.
Copenhagen 1 997 49 .:r.ct + 50 <it g2 f5!
•••

The point Spassky had overlooked.

Here White has an ideal position on Black breaks up White's favourable
the kingside. His rook can settle at f5, pawn structure on the kingside and
creating a cosy little cluster of pieces draws.
defending each other. Black can only 51 .:r. h1
break this up by playing . . . h5 at a mo­ After 5 1 gxf5 l:r.bl 52 l:r. h5 'it> f7 53
ment when both her rook is attacking .:r. xh6 .:r. xb5 54 f6 White has no win­
the pawn on f3 and the white king is ning chances as his rook is completely
not defending it. However, it turns out immobilized. 5 1 l:r. h5 fxg4 52 fxg4
that this possibility is of little value to l:r. bl 53 'it> f3 l:r. b4 is no better because
Black. White's king cannot move to the
White can win by simply playing queenside without immediately aban­
his king to c l . If the black rook stays doning the g-pawn.
on the second rank, White pushes the 51 l:r.c2+ 52 'iii>g3 l:b2 11z.112
..•

b-pawn. Otherwise the king has to be

released and it can then advance. The The final case, when the white rook
analysis runs 49 'it> fl l:r. c3 (49 . . . 'iii> g 6 is in front of the pawn, offers the few­
50 'iii>e l l:r.b2 5 1 � d l l:r. f2 52 .:r.f5 h5 53 est winning chances. Indeed, most po­
'it>el wins a pawn) 50 �e2 .:r.b3 5 1 .:r.f5 sitions are drawn. Even in the most
'it> g6 52 � d2 h5 (the only moment favourable case (that of a b-pawn) it is
when this is playable, but it is too unclear whether White can win unless
slow) 53 �c2 l:r.e3 (53 . . .hxg4 54 <it xb3 he has an additional advantage, such
<it xf5 55 fxg4+ wins) 54 l:r. xh5 .:r. xf3 as a weak enemy pawn structure on
55 b6 l:e3 56 b7 l:r.e8 57 .:r.b5 l:r. b8 58 the kingside.
 THE ENDGAME 135

The situation with an extra pawn Endings with 3 vs 2 (and no passed

and all the pawns on the same side also pawns) on one side are also drawn;
occurs quite often in practice. Here is again the same principles apply. If you
a typical position: remove the pawns on e3 and f7 from
the above diagram, Black's safest first
move is again 1 . . .h5 ! . White to play
could try 1 g4, but here his winning
chances would only be very small.

Positional advantage

We have already mentioned that piece

activity is extremely important in rook
and pawn endings (see page 92). A
rook on the seventh usually ties the
opponent's pieces down to defensive
positions, while a king penetrating into
Black's ideal defensive pawn struc­ the opposing pawn-mass can make
ture is f7-g6-h5 . White's only chance mincemeat of them. Another point is
is ultimately to create a passed pawn, that a passive rook is very often per­
and this pawn formation means that manently passive. Here is an example,
White can only make one at the cost of taken from Levenfish and Smyslov's
several pawn exchanges. First he must classic book Rook Endings.
play g4, when Black can swap his h­
pawn. Then White will aim for f5, but
another pair of pawns disappears. It
follows that once Black has achieved
this optimal set-up, White's winning
chances are minimal. Thus if Black is
to play in the diagram, 1 . . .h5 ! is the
right move.
If White moves first, then of course
he plays 1 g4 ! . While the resulting po­
sition is still theoretically drawn, there
is no doubt that in practice Black's
task is rather difficult. Black cannot
prevent White from gaining a lot of St Petersburg London
-

space on the kingside (<it g2-g3, h4-h5, Telegraph match, 1886- 7

etc.) and his defence is not easy.
136 SECRETS OF PRACTICAL CHESS

At this point London conceded the g-pawns) 4 l:r. a8 (4 :a7+ �e6 5 a5 l:r. a4
game. Chigorin gave analysis indicat­ 6 a6 �e5 7 f4+ �d5 8 l:r. a8 : a3+ 9
ing that White could win and the posi­ �fl �e4 is also a draw) 4 ... catg7 5 a5
tion has been subject to considerable l:la4 6 a6 'iii>h6 7 f4 (trying to prevent
debate ever since. Levenfish and Smys­ the king's entry; 7 a7 <it> g5 is a clear­
lov gave the verdict that Black could cut draw) 7 ... g5 8 f5 (8 fxg5+ fxg5 9
draw, but only with very precise play. �d3 <j;; g 7 is also drawn) 8 . 'iii>g7 and
..

The white rook can range freely White cannot make progress. If White
across the fifth rank, while Black's is plays a7 at any stage, then there is no
restricted to just three squares. First of shelter for White's king on the queen­
all, defending passively is no good. side. However, if the pawn stays on a6
White can move his king towards the then Black can safely take the g4-
vulnerable a-pawn and the defence pawn and return to a4.
eventually breaks down: I . . :a? 2 �e4.

l:r. a6 3 � d4 l:r. a7 4 f4 :a6 (or 4 . . . �d6 5

l:r.b6+ �e7 6 �c5 :c7+ 7 <itb5 :c l 8
g5 f5 9 l:r. xg6 l:r. fl I O � xa5 l:r. xf4 1 1
l:lf6 .l:r.fl 1 2 'iil' b6 f4 1 3 a5 .l:r. gl 1 4 .l:r.xf4
l:r. xg5 1 5 a6 : g6+ 1 6 'iil'b7 l:r.g l 1 7 l:r. b4
and wins) 5 : b7+ cat d6 6 f5 gxf5 7
gxf5 l:r. a8 8 l:r.b6+ 'iil'e7 9 'iil'c5 l:r.d8 I O
:bs and White wins the a-pawn while
retaining both his pawns. With his
pieces so actively placed, this means a
sure win.
Very often in such situations, it is
better to jettison the weak pawn and U. Andersson R. Hilbner
-

try to activate the rook, and here this Johannesburg 1 981

plan just about draws for Black. Note
that Black must not delay; if he allows In this position White has a perma­
White's pieces to become any more nent positional advantage in the shape
active before giving up the pawn, then of Black's backward c-pawn. Never­
the game will be past saving. The best theless, it requires a high standard of
line is 1 . l:r.c6! 2 l:r.xa5 :c4+ 3 'iil' e3
. . endgame technique to convert this sin­
�f7! (another difficult move to find; gle positive factor into a win.
the king has the option to moving to h6 23 g4
and g5 for a counter-attack against It is moves like this that are the key
White 's pawns, while being protected to successful endgame play. It is not
from rook checks by his own f- and just a random pawn advance but part
 THE ENDGAME 137

of a well-thought-out plan for putting

Wbite's central pawn majority to use.
The obvious plan is to play e4 at some
stage, but for this White needs the sup­
port of the king (note that the rook
should stay on the c-file to prevent
. . . c5). White could try e3, 'ito>e2-d3 and
then e4, but Black could counter this
plan by playing . . . f5 at some point.
The move played not only makes it
much harder for Black to play . . . f5, but
also opens up a second possibility for
activating the white king, based on White finally achieves e3-e4, and
l:r.c2 and 'ito> g3-f4. gives Black a difficult choice. If he ex­
23...aS 24 h4 g6 25 l:k2 h6 changes on e4, then White's rook gains
After 25 . . . f5 26 gxfS gxfS 27 <iti g3, access to fS to attack the backward f­
followed by 'ito> f4, White ties Black pawn, while otherwise Black has to
down to the defence of fS, and then defend the dS-pawn.
plays e3 and l:. g2, when his rook can 33 l:.d7
••.

penetrate along the g-file. After 33 ...dxe4+ 34 fxe4 l:r.e6 3S l:r.fS

26 'ito>g3 gS 'ito> c7 36 es fxeS 37 dxeS 'ito> d7 (37 . . . cs
This is Black's counter-plan. He 38 'ito>e4 'ito>c6 39 l:.f6 'ito> d7 40 <itidS wins
prevents White's king occupying f4. easily) White can pursue a policy of
27 h5 constriction, e.g. 38 'it> d4 <iti e7 39 b3
It would be wrong to exchange on 'ito>e8 40 a3 'ii? e7 41 b4 axb4 42 axb4
gS as Black's rook would be in a posi­ 'ito>e8 43 l:. f3 rite? 44 l:l a3 and White's
tion to occupy the h-file. rook penetrates.
27 f6 28 <itif2
••• 34 a3 l:.d8 35 l:lcl
White now returns to the original White's intention is to play b4 so as
plan o'f playing his king to d3, because to provide further support for the rook,
the change in the kingside pawn struc­ but for the moment he simply waits
ture has made . . . fS virtually impossi­ until the game is adjourned at move
ble (if Black tries to prepare it with 40.
. . . l:r.f8, then White plays e4 immedi­ 35 ... l:r.d7 36 l:lc2 l:.d8 37 l:.c3 l:r.d7
ately). 38 l:.c2 l:r.d8 39 l:r.cl l:r.d7 40 l:.cS l:r.d8
28 'ito>c7 29 e3 'iii>b 6 30 <itie2 l:.c8
..• 41 b4
31 .l:r.cS Now White undertakes positive ac­
Of course White prevents . . . cS. tion.
32 l:r.e8 32 'ito>d3 l:r.e7 33 e4 (D)
••• 41 ... axb4 42 axb4
138 SECRETS OF PRACTICAL CHESS

Black is in zugzwang. 46 'ifi>fS cxdS 47 lhdS l:r.e3

42 l:.d7
••. Black's only chance is to counterat­
After 42 . . . dxe4+ 43 fxe4, activating tack, but Andersson has everything
the rook by 43 .. J:l a8 is too slow, as worked out.
Black's kingside pawns are vulnerable 48 �xf6 l:r.xf3+ 49 'iti> g7!
and White would just play 44 l:r. f5 . This is the crucial finesse. Black
Otherwise a rook move along the file draws after 49 'iti> g6 .l:tf4 50 � xh6
is the only possibility. l:r. xg4 5 1 <it> g7 l:r. xb4 52 h6 l:r. h4 53 h7
43 exd5 <it>a5.
Exploiting the absence of the rook 49... l:r.f4 so l:r.d6+
from the first rank. Now 43 . . . cxd5 The point behind White's last move.
loses to 44 l:.c8, for example 44 . . . f5 45 He takes the h6-pawn with his rook,
gxf5 l:r. f7 46 l:r. h8 l:r. f6 47 l:r. h7 ! <it> c6 48 and at the same time cuts the black
'ifi>e3 <it>d6 49 l:r. b7 � c6 50 .l:ta7 �b6 5 1 king off along the third rank.
l:tg7, followed by l:t g6. SO.. 'ifi>c7 51 l:r.xh6 l-O
.

43...l:r.xd5 44 'ifi>e4 White wins after 5 1 . . . l:r. xb4 52 l:r. g6

Making use of the extra support for l:r. xg4 53 h6 l:r. e4 (however Black
the rook provided by b4. plays, he has to give his rook up for the
44 ...l:r.dS 45 d5 (D) pawn within a few moves, for example
after 53 .. J lh4 54 h7 White threatens
l:r. h6, so Black has to take the pawn im­
mediately) 54 h7 l:r.e7+ 55 'ifi>h6 l:r.e l
5 6 h8'ik l:r. hl + 5 7 'i; g7 l:r. xh8 58 'ifi> xh8
g4 (58 . . . b4 59 'iti> g7 b3 60 l:r. xg5 <i; c6
61 l:r. g3 wins) 59 'iti>g7 g3 60 'ifi> f7 g2
(60 . . . b4 61 l:r. xg3 <it> c6 62 l:r. g5) 6 1
l:r. xg2 <it>b6 6 2 l:r. g5 'iti>a5 6 3 <i; e6 and
the white king is easily in time.

M i no r- piece endi ngs

This is really four endings rolled up in
White cannot penetrate to f5 as his one; knight endings, bishop vs knight
d-pawn is hanging, but once it has been endings and bishop endings, with ei­
exchanged Black will be unable to de­ ther same- or opposite-coloured bish­
fend his kingside pawns. ops. A detailed coverage of each of
45 .. JleS+ these endings is not possible here, so I
45 . . . cxd5+ 46 l:r. xd5 l:te8+ 47 'ifi>f5 will concentrate on the most useful in­
l:r. e3 transposes to the game. formation.
 THE ENDGAME 139

Knight endings a draw since 1 'iil'b 6 again gives a per­

petual attack on the knight. If the
pawn is on a4 and knight on a5 then
White loses even if he can attack the
knight with 1 rltb5 on the first move;
Black just plays l . .. � g5 and when he
promotes on hl , White is prevented
from promoting on a8 .
Now suppose that the pawn is on b7
and knight on b8, with Black to play.
The position is drawn by repetition:
neither side can deviate from the se­
quence 1 . . . ll'i a6+ 2 <it b6 ll'ib8 3 'iii> c 7
(3 <ita7? ll'id7 would win for Black)
Here is a common situation - the 3 . . . ll'ia6+.
knight is single-handedly holding back Now move the pawn to b6 and the
a passed pawn. There are three possible knight to b7. Black to plays wins by
outcomes to such a struggle: either the 1 .. . ll'i a5, because the white king takes
knight loses the battle against the white far too long to make it round to attack
king and pawn, and the pawn pro­ the knight (note that even 1 �d6 is im­
motes. Then White wins. Alternatively, possible because of l . . . ll'ic4+). How­
the white king cannot achieve more ever, White's loss here was due to the
than a perpetual attack on the knight; poor initial position of his king on c7.
then the result is a draw. Finally, it may If the king starts on c6, then the posi­
be that the knight can not only restrain tion is a draw since 1 . . . ll'i a5+ 2 'iii> b 5
the pawn, but can do so while provid­ and 1 ... ll'id8+ 2 �d7 both lead to repe­
ing a continuous supply of spare tempi. titions.
Then Black has time to advance his h­ With the pawn on b5 and knight on
pawn and win. b6, it doesn't matter where the white
The diagram position is a draw. The king is; Black always wins since he
knight can prevent the pawn's promo­ can generate an unlimited supply of
tion but without providing any spare tempi. If White attacks the knight by 1
tempi: l 'iii>b7 ll'ib5 2 �b6 ll'id6 ! 3 �c6 � c6, then 1 . . . ll'ia4 wins as before,
ll'ic8 4 <i;;c 7 ll'ia7 and so on. while 1 'iil'c 5 is met by 1 . . . ll'ic8, and 2
If the pawn is on a7 and the knight �c6 fails to 2 . . . ll'ia7+.
on a8, then of course White wins by 1 With a c-pawn, Black wins even
'iil'b 7. with the pawn on the seventh rank. If
Moving the other way, with the pawn Black's knight is on c8, he can meet
on a5 and knight on a6, the position is �d7 by . . . ll'i a7 and � b7 by . . . ll'i e7,
140 SECRETS OF PRACTICAL CHESS

and the white king has to talce the long queenside. Thus after 57 ... 'it>h5 58 'it>c3
way round to attack the knight again. �a6 59 � xf5 'it> g4 60 �e3+ 'it> xh4
White wins by 6 1 �xd5 'iii> g 5 62 �c4
Moving on to knight vs knight end­ � f5 63 �b5 � b8 64 �b4 'iii>e6 65
ings, the most important point to note �c6 �d7 66 a6. The line 57 . . .f4 58
is the devastating effect of an outside 'it>e2 �h5 59 'it> f3 � xh4 60 � xf4 is
passed pawn. The knight is such a similar, for example 60 . . �h5 61 �e5
.

short-range piece that if it is prevent­ � g6 62 � d6 � f7 63 'iii>c 5 � a6+ 64

ing the advance of a passed pawn on � b6 �b8 (or 64 . . . �b4 65 'it> b5) 65
the queenside, then it cannot exert any �b7 �d7 66 �b3 .
influence on the kingside. Thus the 58 �e3 �cs 59 'iii>f4 �3+ 60 �e3
side with the outside passed pawn will �b4
be effectively a piece up on the other If Black repeats by 60 . . . �c5, then
side of the board. Knights are particu­ White continues 61 �f3 �f6 (6 l ...�h5
larly ineffective against rook's pawns : 62 �f4 � g6 63 � e5) 62 h5 � g7 63
� d4 � h6 (63 . . . �f6 64 h6 'iii> g 6 65
�xf5) 64 �f4 'it> xh5 65 � xf5 winning
as in the previous note.
61 �f4 �d3+ 62 <itiif3 �b4 63 �e2
�f6 64 �f4 d4 65 'it>e2 'it>f7
Not 65 . . . 'iii>e 5 66 �3+.
66 �dl �f6 67 �d2 (D)

S. Dolmatov A. Beliavsky
-

USSR Championship, Odessa 1 989

Here Black is fighting against two

rook's pawns. Despite material equal­
ity and the small number of pawns,
White has a winning position.
S7 �d4 �a6 Black is in zugzwang and must go
White is normally happy to swap back with his king.
his h-pawn for Black's f-pawn, since 67 �f7 68 �2
•••

his king will then be faster to reach the Now the d-pawn is lost.
 THE ENDGAME 141

68 ... �g6 69 lLJxd4 f4 70 'it>e2 'it>hS

71 �f3 �xh4 72 �4
This lLJ+A vs lLJ ending is a simple
win.
72... �hS 73 <itiieS �g6 74 'it>d6 'it>f7
75 'iii>cS lLJa6+ 76 �b6 lLJb4 77 lLJc6
lLJdS+ 78 'iii>b7 �e6 79 a6 'it>d7 80 a7
lLJc7 81 lLJeS+ �d8 82 lLJc4 lLJa8 83
lLJb6
White avoids the trap 83 � xa8??
'it>c8 drawing.
83... lLJc7 84 �c6 1-0
S. FlolJr J. Capablanca
-

Bishop vs Knight endi ngs Moscow 1935

The relative value of the bishop and endgame players of all time, just about
the knight is a familiar topic in every managed to save the game, but I would
phase of the game, and the basic prin­ doubt my ability to do likewise ! He
ciples are the same in an ending. adopted what is undoubtedly the cor­
On average, a bishop is worth a lit­ rect plan, namely to avoid allowing
tle more than a knight. A bishop is at any more pawns to be fixed on light
its best in an open position with a fluid squares. He therefore played . . . b6,
pawn structure. Pawns fixed on the ... a5, ... .i.d7 and . . . f6, and waited to see
same colour as the squares on which if White could make progress.
the bishop moves are an obstruction; It is worth noting that if Black had a
the more such pawns exist, the more dark-squared bishop, then the position
restricted the bishop is and the weaker would be a comfortable draw. If, on
the squares of the opposite colour. the other hand, Black had a further
Knights favour blocked pawn struc­ pawn weakness, for example the pawn
tures and like stable, invulnerable on b7 were on b5 (so that White could
squares on which they can settle with­ fix the queenside pawns by playing
out being disturbed by enemy pawns. b4), then he would be lost.
A typical such situation arises when a It is worth adding one point which
knight occupies the square in front of applies particularly to the endgame.
an isolated pawn, as in the following Bishops prefer positions with unbal­
diagram. anced pawn structures. If both sides
We will not discuss this very com­ create a passed pawn, a bishop can sup­
plicated endgame in detail here. Capa­ port the friendly passed pawn while at
blanca, who was one of the greatest the same time holding up the enemy
142 SECRETS OF PRACTICAL CHESS

passed pawn. Knights, being short­ knight in this open position, and his
range pieces, have to be committed to king is slightly more active.
one or the other task. 30... hS
Players often underestimate the ad­ There is nothing wrong with this
vantage which is conferred by having move. It allows a possible penetration
a bishop against a knight, even in posi­ by the white king via f4 to g5, but in it­
tions with symmetrical pawn struc­ self this is not dangerous as the simple
tures. reply . . . rilf7 secures the kingside pawn
structure.
31 .i.el llles 32 .i. f2 (D)
White aims to provoke a pawn move
on the queenside, so as to provide a
possible entry route for White's king
on that side also.

z. Ilincic - G. C abrilo
Cetinje 1992

I suspect that many players would

abandon the position as a draw, but in
the game White won. Ilincic anno­
tated his game in both Informator and 32...ll:\c4?
the Encyclopaedia of Chess Endings, This leads to an exchange of pawns,
claiming that the diagram position is which helps the defender, but allows
actually winning for White. As we the white king to occupy the fifth rank,
shall see, this claim is rather far­ which is a far more important factor.
fetched, but it is true that White's ad­ Ilincic gives 32 . . . a6 33 cat d4 llld7 34
vantage is sufficient to justify playing .i. g3 but stops here with the claim that
on. White is winning. However, Black can
26...<itg8 27 'iil'd2 �f7 28 'iil'e3 <ite6 just move his knight, meeting 35 �cs
29 .i.c3 g6 30 c;i;> e4 by 35 . . . ll:\d7+, and further progress is
White has two very slight advan­ not easy.
tages : his bishop is better than the 33 .i.xa7 lllxb2 34 ritd4 ritd6
 THE ENDGAME 143

Certainly not 34 . . . � f5? 35 � c3, 45 liP xg6 'ii?e6 (45 . . . b5 46 'iii> xh5 wins
winning after 35 . . . �d l + 36 'iii> c2 or as in the note to Black's 35th move) 46
35 . . . lb a4+ 36 'it> b3 b5 37 'it> b4. <iti xh5 liP f7 47 � h6 ! 'ii? g8 (or 47 . . . b5
3S i. b8+ liPe6 48 !iP h7 b4 49 g4 b3 50 i. c l lb c4 5 1
One of the problems facing the side g5 lbe5 5 2 h4 lb g6 5 3 h5 �f8+ 54
with the knight is that a 'race' situation 'iii>h6) 48 liPg6 b5 49 h4 �c4 50 i. f6 b4
normally strongly favours the bishop. 5 1 h5 b3 52 h6.
Here 35 . . . liPc6 36 'it>e5 lbc4+ 37 'it> f6 42 a4 b6?
� xa3 38 � xg6 �c4 39 liP xh5 b5 40 This loses straight away, although
i. f4 b4 4 1 'it> g6 b3 42 i. c l b2 43 the position was in any case very diffi­
i. xb2 �xb2 44 h4 wins for White. cult for Black. After 42 . . . lb b2 43 a5
36 <itcS 'iPd7 �d3 44 i.e3 �e l 45 g3 lbd3, White
Black also cannot afford to let the continues 46 'iii>d4 ! �b4 (46 . . . �e l 47
white king get at the b-pawn, e.g. 'iii>e 5 'iii>e7 4 8 i. g5+ <iti f7 49 'iii> d6) 47
36 ... catf5 37 <iti b5 ! 'ii?e4 38 a4 �e3 39 cate5 and Black cannot play 47 . . . 'ite7.
a5 � f2 40 i. e5 �d3 4 1 i. d4+ liP xg2 43 i.e3 �b2
42 !iP b6 and wins. 43 . . . 'it> c7 44 'iPe6 is also an easy
37 <itdS win.
By now Black is in serious trouble. 44 i.xb6 �a4 4S i.d4
White's main plan is to cover e7 with The knight is trapped and can only
his bishop and then play �e5-f6. be rescued at the cost of both Black's
37 lLld3 38 i.g3 �b2 39 i.el �d3
... kingside pawns.
40 i.d2 lbb2 41 h3 (D) 4S h4 46 �es 'iii> e7 47 'iPf4 <ite6
•••

48 <itgS 'it>dS 49 i.f2 lbb2 SO i.xh4

�dl Sl <itxg6 �e3 S2 g4 'it>e4 S3
i.gS �c4 S4 i.f6 lbe3 SS <itigS �f3
S6 i.d4 1-0

Bishop endi ngs

Considering the case of same-coloured

bishops first, the outside passed pawn
is again a very powerful force, even if
not quite so effective as in knight end­
ings. I would therefore normally ex­
pect endings with 3 vs 3 on one side
41 �d3
••• and an extra outside passed pawn on
After 41 ... liPe7 42 i. g5+ �d7 White the other side to be winning, although
wins by 43 � e5 lbc4+ 44 liP f6 � xa3 if the passed pawn is not too far away
144 SECRETS OF PRACTICAL CHESS

from the remaining pawns there is a passed pawns are better than passed
small chance of a blockade. pawns with only one file between
With all the pawns on the same side, them, but two widely-separated passed
a 3 vs 2 or 4 vs 3 ending with fluid pawns are best of all.
pawns should normally be a draw. 3) If you are trying to draw an
Since bishops are heavily influenced opposite-coloured bishop ending, then
by fixed pawns, one of the most im­ it is better to have your pawns on the
portant questions is how many pawns same-coloured squares as your bishop.
are fixed on the same-coloured squares This is practically the only type ofpo­
as one's bishop. The more there are, sition in which this reversal of the nor­
the worse it is, all the more so in that mal principle applies. The reason is
the opponent's bishop will automati­ that your drawing chances depend on
cally be a ' good' bishop. Two pawns setting up an unbreakable fortress.
fixed on the same colour as one's Everything has to be defended, and a
bishop may be a fatal weakness, espe­ pawn which is capable of being at­
cially if they can both be attacked si­ tacked by the enemy bishop will just
multaneously by the opposing bishop. drop off.
This completely immobilizes one's
own bishop, and zugzwang may not be Here are a couple of examples illus­
far away. trating these principles:

Opposite-coloured bishop endings

are often quite tricky because many of
the principles that apply in a wide
range of endings break down in this
case. Here are some of the main differ­
ences:
1) Material advantage is less im­
portant than usual. Endings with one
extra pawn are usually drawn. Even
two extra pawns may not be sufficient
to win. For this reason reduction to an
opposite-coloured bishop ending is
often a useful drawing resource in a Y. Piskov J. Nunn
-

desperate situation. Bundesliga 1992

2) What is important is the ability
to create passed pawns. If you can cre­ Black is in a difficult position be­
ate two passed pawns, then you have cause his king is more exposed. While
good winning chances. Connected there are still heavy pieces on the
 THE ENDGAME 145

board, the opposite-coloured bishops to avoid the blockade of the pawns,

pose a particular danger for Black. and so they should occupy squares of
37 .1f6 the opposite colour to his bishop. This
Threatening mate in two. 3 7 . . . .l:txb8 usually means that they must be de­
is no defence because of 38 �h8+ fended by the king.
cj;f7 39 �g7+ �e8 40 i/e7#. 43... .1d3 44 c5 .1c4 45 d6 <tie6 46
37... ilhS ! <tin <tid1
This is the safest move. Even though The blockade is set up. The a- and h­
Black loses two pawns, the resulting pawns can be defended by the bishop,
opposite-coloured bishop ending is a and the passed pawns are totally im­
clear draw. It is possible that Black mobile.
might have got away with 37 . . <tif7 38
. 47 <tJg3 .1e6 48 h4 <tic6 112.112
.l:tb7 ilxf6 39 .l:t xd7+ .l:te7, but the ad­
vantage of opposite-coloured bishop
positions is that if they are drawn, then
they are usually completely drawn.
38 'ii'xh5 gxh5 39 .l:txe8+ .1xe8 40
.1e7 .1g6 41 .1xc5 cj;f7!
Not 4 1 . . . .1 d3? 42 d6 .1 f5 43 .1 xa7
and White obtains three passed pawns.
42 .1xd4
After 42 .1 xa7 .1 bl 43 a4 d3 44
<tif2 .1 a2 Black wins one of the white
pawns, again with an easy draw.
42 a6
.•.

Threatening to win a pawn by means This is a slight modification of a

of . . . .1 d3 followed by . . . .1 c4. game I played in a simultaneous dis­
43 a3 play during 1 977.
White has two extra connected Black defended the g-pawn in the
passed pawns, and if he could support only way possible, by heading for h4.
them with his king then he would be 1. .tel
••

winning. However, his king is too far At first sight this is a dead draw. If
away and by attacking the pawns with White's king attacks the g5-pawn,
his bishop, Black can force them to ad­ then Black plays . . . .1 h4 and waits
vance to dark squares, when it is easy with his king, while if White's king
to blockade them. The defender must approaches the c-pawn, then Black's
keep his pawns on the same-coloured bishop is freed and so he can just keep
squares as his bishop, but the opposite his king on c7. This is in accordance
holds true for the attacker. He must try with the principle that you need two
146 SECRETS OF PRACTICAL CHESS

passed pawns to have a chance of win­

ning an opposite-coloured bishop end­
ing.
2 �f6!
White 's plan is to force the bishop
to h4 and then play g3. If Black replies
. . . ..t xg3 , White continues WxgS and
has his second passed pawn. If Black
replies . . . fxg3, White plays ..t g2 and
the black bishop is imprisoned. When
the white king approaches the c-pawn,
Black will be forced to release his
bishop by . . . g4, again giving White with his bishop and running to h8 with
another passed pawn. his king. Then it will not matter if
However, this plan must be imple­ Black has to surrender his bishop for
mented accurately. After 2 �fS ..t h4 3 the c-pawn, as he has a 'rook's pawn
g3 fxg3 4 ..ll g 2 <it>d6 White experi­ and wrong bishop' draw. However,
ences certain difficulties. He cannot White can win by 5 . . . ..tel 6 h4 �e7 7
move his king to f6 or e4 and after 5 hS and now:
<it' g4 <it> c7 he is not making progress. 1) 7 ... ..t c3 (using the bishop to hold
In fact, this is very much like a king back the h-pawn, but the passed pawns
and pawn ending, since both sides can are too far apart for this to work) 8 h6
only move their kings (if White moves <it>d6 9 h7 .Iles 10 <it> fS ..t h8 1 1 <it> xf4
his bishop, then . . . g2 draws at once). ..t b2 1 2 <it>fS ..t c3 1 3 � g6 <it>e7 1 4
White can try to exploit the fact that d7 i. dl ..tb2 1 5 ..t a4 (threatening 1 6 c7)
is not available for Black's king by 5 15 . . . <it>d6 16 <li>f7 winning the bishop
<it> g6 <it>e6 6 �h7, but his subtlety is to for the h-pawn.
no avail : after 6 . . . <it>e7 7 <it> g7 �e8 2) 7 . . . ..t as 8 <it> g6 <it> f8 9 ..t ds (of
White cannot 'by-pass' by 8 � f6. course White must prevent . . . 'it> g8 at
The key is to lose a tempo before all costs) 9 . . . f3 10 h6 f2 1 1 h7 fl 'ii 12
the bishop arrives on h4. h8'iW+ <it>e7 1 3 'iie s+ <it> f8 1 4 'iid6+
2...i.h4 3 <Jif5 <Jid6 <it>e8 15 'iid 7+ <it> f8 16 'iig7+ <it>e8 1 7
White has lost a tempo and now the ..t f7+ <it>d8 1 8 c7+ ! ..ll x c7 1 9 'iif8+
g3 plan works. <it'd? 20 ..lle 8+ winning the queen.
4 g3 (D) 5 ..t g2
4 fxg3
..• Black is in zugzwang and must go
After 4 . . . i. xg3 5 <it> xgS, the win is back with his king.
still not straightforward as Black has a 5 � c7
...

possible drawing plan of covering c7 Or 5 . . . <it>e7 6 'iti>eS.

 THE ENDGAME 147

6 'iii>eS g4 most common in practice (the most

Otherwise the king reaches d6 and common being rook and pawn vs
that would really be Black's last chance rook). It is also by far the most compli­
to play . . . g4 before White promotes cated, so much so that it has proved too
his pawn. daunting for the maj ority of authors
7 hxg4 and very little has been written about
White wins easily by supporting it. Before the advent of the computer
the passed g-pawn with his king. database, Averbakh provided the best
It is worth noting that Black only coverage. However, the complexity of
loses in the diagram as he cannot de­ the ending was only emphasized by
fend the g5-pawn from d8 or e7 . If his the fact that his seventy pages of
king were on the apparently inferior analysis dealt almost exclusively with
square b8, then 1 . . . .i. a5 would draw. certain simple cases of the pawn on
the seventh rank.
I must emphasize that the special Now that the oracle of a perfect da­
principles mentioned above apply tabase has been constructed, one would
only to pure opposite-coloured bishop have expected someone to provide a
situations. As soon as extra pieces are far more detailed explanation of the
added to the equation, normal service principles behind the ending. How­
is resumed. Many players assume that ever, the database has revealed that
endings with rooks and opposite­ Averbakh's efforts only scratched the
coloured bishops are almost as drawn veneer of this ending and most of the
as pure opposite-bishop positions, but real work remains to be done.
this is not so. An extra pawn in an end­ I suppose I could stop here with the
ing with rooks and opposite-coloured comment 'nobody understands 'ii'+8
bishops normally confers good win­ vs 'ii'' , but this does seem a little cow­
ning chances; adding a pair of knights ardly.
is similar. If even more pieces are The first point to make is that if the
added then the attacking potential of defending king is in front of the pawn
opposite-coloured bishops starts to then the position is almost always
come into play, and then one cannot drawn. There are exceptions in which
talk about a drawish influence at all. a winning exchange of queens can be
forced, but they are very few indeed.
Queen endi ngs Thus the only interesting case is
when the defending king cannot move
Queen and Pawn vs Queen in front of the pawn. Dealing firstly
with the a-pawn, this not surprisingly
Of all the fundamental endings with offers the fewest winning chances of
piece + pawn vs piece, this is the second any pawn. The diagram below shows a
148 SECRETS OF PRACTICAL CHESS

typical situation with the pawn on the optimally over the board? Our objec­
sixth. tive is only to establish some general
principles helpful in practical play and
for this the 50-move rule is not espe­
cially relevant.
This diagram already reveals many
of the important principles governing
'ii'+8 vs 'ii' . We can ignore the 'special
case' squares a7 and a8, which are
only marked because White to play
can mate in one. There are basically
two drawing zones. One lies near to
the pawn and these squares are drawn
because with the enemy king so close
to the pawn, Black can very often ex­
We will display several diagrams of change queens and catch the pawn with
this type, the aim being to convey a his king. Virtually all 'ii'+8 vs 'ii' posi­
great deal of information in a visual tions (except those with the pawn on
form. The various stars indicate the re­ the seventh) have such a zone, which
sult when the black king stands on that we call the proximity zone. When the
square. A white star indicates that the black king lies in the proximity zone,
position is a win with White to move, White is restricted in his attempts to
but a draw with Black to move. A black avoid perpetual check, as he can only
star indicates that it is a win whoever interpose his queen in favourable cir­
moves first. Unmarked squares are cumstances. There is a popular myth
drawn whoever moves first. that when defending 'ii+ 8 vs 'ii, it is
In the case of squares that are under always best to have your king as far
attack by White, then the position is away from the pawn as possible; as we
only legal with Black to move and a shall see, this is true only in certain
black star indicates that it is a win; no circumstances.
star indicates that it is drawn. The second drawing zone is in the
A few positions are illegal because lower right-hand comer, i.e. diago­
the kings stand next to each other; nally opposite the pawn. White's main
these are also unmarked. weapon in preventing perpetual check
Readers should note that we take no is to interpose his queen in such a way
account of the 50-move rule in our dis­ as to force the exchange of queens, ei­
cussion. With optimal play, some of ther by checking or by pinning the en­
the positions are drawn only because emy queen. When the black king is
of this rule, but who is going to play poorly placed, for example on f7, then
 THE ENDGAME 149

there are many chances to achieve this L. . rt;g3 ! (although it would be amaz­
and Black's options are consequently ing if anyone realized over the board
restricted. White can set up a position that 1 . . . �h3? is losing). If the king
with his king on b7 and queen on c6. starts on e4, there is a choice between
Then l . . . 'iWe7+ is answered by 2 'iWc7, l . . . rt; dS and l . . . � f3 . but I would al­
1 . ..'il b4+ 2 � c8 'ii'f8+ 3 rt;c? ends the ways go for the proximity zone if pos­
checks, 1 . . . 'ii'b 3+ 2 �c8 and 1 . . . 'ifb l + sible, as the draw is far simpler.
2 �a8 likewise. The toughest check to Summing up, the proximity zone is
meet is l . . . 'ii'b2+, but even in this case the best location for the black king,
White can manoeuvre so as eventually then the bottom-right comer, then the
to force Black to give one of the infe­ bottom-left comer. The top right cor­
rior checks mentioned above. In the ner is by far the worst.
end Black's checks dry up, and White If we move the pawn back to a5, the
can push his pawn. This problem af­ drawing zones expand enormously:
flicts Black whenever his king is in the
top right-hand comer. The lower left­
hand comer is also not ideal, but this
largely depends on Black's queen po­
sition. If Black can occupy a good
square with his queen, then this is suf­
ficient to draw. The bottom right-hand
comer is the best of the distant areas;
White has very few chances to inter­
pose with check and so Black has few
problems drawing.
While much of the play is governed
by general principles, there are a few
anomalies. The isolated loss with the The lower-left squares are marked
black king on h3 (when 1 'iWd4 ! is the only because White has the possibility
only winning move) is hard to explain, of "ifa3+, forcing an immediate ex­
as is the loss with Black to move when change of queens, so we can fairly say
his king is on c2. that Black is only in danger if his king
The point about the two drawing is in the upper-right comer.
zones is emphasized if we imagine We can summarize by saying that
the black king starting somewhere on with an a-pawn, the pawn needs to be
Black's fifth rank. With the king on on at least the sixth rank in order to
d4, the only drawing move is 1 . .. <itid5 ! , have decent winning chances.
entering the proximity zone. When Not surprisingly, the b-pawn offers
the king is on h4, the only move is far more winning chances . There are
150 SECRETS OF PRACTICAL CHESS

two main reasons. First of all, the

white king has the squares on the a-file
to aid him in his efforts to avoid per­
petual check. Secondly, an exchange
of queens is much less likely to lead to
a draw, even when Black's king is quite
close to the pawn, so the proximity
zone is much reduced in size. Here is a
typical case with the pawn on b6 (D):

king on certain d-file squares, but

only if it is his turn to move, and even
then the draw requires great accuracy.
With the king on d8, Black's unique
drawing move is 1 . .. 'ii'g l ! , with the
king on d7 again there is only one
move, 1 . . . 'ii'h l ! , and with the king on
d6, 1 . . . 'ili'd5 ! is the only saving possi­
bility.
The drawing zone in the bottom­
We can see how vastly different this right comer is relatively large, and it is
situation is. If White is to move then clear that against a b-pawn, if you can­
he wins no matter where Black's king not bring your king in front of the
is located. With Black to play there is a pawn, then this is the area to head for.
small drawing zone in front of the We do not give the details in the case
pawn, and a slightly larger one in the of the pawn on b4; the situation is
bottom right-hand corner. somewhat similar to that of the pawn
With the pawn on b5 White's win­ on a5, i.e. White can only win if the
ning chances are of course less than black king is near to the top-right cor­
with the pawn on b6, but still greater ner.
than with the pawn on a6 (D): To summarize, with a b-pawn the
Here we see the familiar pattern pawn should be on at least the fifth rank
emerging, but with some differences. to provide good winning chances.
The proximity zone is much smaller The c-pawn provides the greatest
than with the a-pawn; it includes all winning chances of any pawn. One
the squares in front of the pawn, but reason is that the drawing zone in the
little else. Black can draw with his bottom-right corner disappears when
 THE ENDGAME 151

the pawn is on at least the fourth rank, It follows that if you have various
so in this case Black is left only with methods of liquidating to 'i!V+8 vs 'ii' ,
the proximity zone. the c-pawn is the one to go for. The de­
The situation with the pawn on c6 fender's only chance is to bring his
needs no diagram, because Black is king near to the pawn, but even with
losing except if his king is in front of the pawn as far back as the fourth rank
the pawn. When the pawn is on c5, the there are still good winning chances.
drawing chances are still very slim, Central pawns offer significantly
except if the king is in front of the fewer chances than with the c-pawn.
pawn: Once again, there is no distant draw­
ing zone and the defender should have
his king somewhere near the enemy
pawn. Here are two typical diagrams:

Even with a pawn on c4, White's

winning prospects are surprisingly
good:
152 SECRETS OF PRACTICAL CHESS

In the second diagram the drawing 'ii'b 3-c3-c5 and drive the enemy queen
zone has expanded considerably and away. The pawn can then be advanced
encompasses quite a large part of the and, if necessary, the manoeuvre can
board. The square a4 is rather anoma­ be repeated all the way down the
lous (White to play wins only with 1 board.
�c5 ! ) . This would be the winning tech­
Just as with the c-pawn, the defend­ nique in the case that the kings are on
ing king should be as near to the pawn the kingside and White 's king is safe
as possible. It is worth noting that it is from perpetual check. Note that we
much better for the defender to have are assuming that Black cannot ex­
his king on the short side of the pawn change queens. This is normally a safe
than the long side. assumption, but there might be cases
in which Black could swap on c5, take
The extra Pawn the c-pawn with his king and return to
the kingside in time to save the game.
Just as in the case of rook endings, In that case White might do better to
queen endings with an extra outside play 'ii' b3-a3-a5 in order to advance
passed pawn are relatively common. It the pawn. The problem with this is
is hard to say whether the winning that it badly decentralizes the white
chances are greater or less than in the queen, and while the queen is away,
case of rooks. Objectively speaking, I Black's own queen can take up an ac­
suspect that they should be greater, but tive post, increasing his chances of
there are far more chances for things perpetual check.
to go wrong with queens on the board ! If the white king is not safe from
The main danger is that of perpetual perpetual check, then the win is much
check, so king safety is an important more complicated, if indeed it is pos­
factor. sible at all. The plan is again to try to
Unlike the corresponding situation push the pawn using the above ma­
with a rook, a queen is capable of noeuvre, and when the black queen
shepherding a pawn to the promotion starts checking, the white king will
square all on its own. Suppose, for ex­ run to the queenside, where his pawn
ample, White has a passed pawn on b4 and queen stand ready to provide shel­
supported by a queen on bl , and that ter. It is important to bear in mind that
Black has blockaded the pawn with his a barrage of random checks from the
queen on b5 . If the queens were re­ enemy queen very rarely amounts to a
placed by rooks, then the pawn could perpetual, provided that one's own
not advance without either zugzwang queen occupies a central position and
or the approach of White's king. With there is at least one pawn for shelter.
queens, however, White can continue Therefore White should have no
 THE ENDGAME 153

hesitation about running with his king 55 'ii'g8+ 'ito>f3 56 'ii'g3+

to the queenside. Here is an example: Black wins after 56 'ii'f8+ 'it>e4 57
"ifxh6 'ii'e 3+ 58 "ifxe3+ 'it xe3 .
5 6 'ite4 57 "iig6+ �d4 5 8 'ii'd6+
.••

Once again the h6-pawn is invul­

nerable.
58... 'it>c3
With a slightly different version of
the same idea: 59 'ii'xh6 'ii'd l + 60 � g2
"ifd2+.
59 'ii'a3+ 'ito>d2 60 'ii'a2+
Or 60 "if b2+ and the checks run out
after 60. . .�d3 6 1 'ii'bl+ �d4 62 "iib6+
'ito>c3 63 'ii'a5+ (or 63 "iff6+ 'it> c2 64
"iff5+ 'it>d2 65 'ii'd5+ transposing)
G. Kieninger - E. Eliskases 63 . . . 'it>d3 64 'ii'f5+ (64 'ii'd5+ �c2 65
German Championship, "iff5+ �d2) 64 . . . � d2 65 "ifd5+ � c l ,
Bad Oeynhausen 1 938 for example 66 "ifc5 'ii'e l + 6 7 'it> g2 c3
and the pawn edges forwards.
It is Black to move and he first of all 60 �d3 61 'ii'a6
•.•

brings his queen nearer the centre with Realizing that there is no perpetual
gain of time. check, White decides to halt the pawn's
49 'ii'c2+ 50 �gl "ifdl+ 51 'ito>g2
••• advance by pinning the pawn: this is
"ife2+ 52 �gl 'ii'e3+ 53 'it>g2 "ife2+ 54 also a typical motif in queen and pawn
�gl c4 endings.
Black has achieved all he can with 61 hS
.••

checks and now takes the opportunity As a matter of fact, this position
to push his passed pawn. This allows would be a win even without the h­
White to start his own barrage of pawns, but their presence certainly
checks, but Black's king can eventu­ makes Black's task simpler. White has
ally evade the checks by moving to the no threat, so Black simply creeps for­
queenside . One unusual feature of the ward with his h-pawn. When it arrives
position is that Black is prepared to on h3, White will have mating threats
leave his h6-pawn undefended. Nor­ to contend with in addition to the
mally, every pawn has to be protected passed c-pawn.
to avoid being gobbled up by the en­ 62 "ifb5 'it>c2! (D)
emy queen, but in this special case the The immediate 62 . . . h4 would be a
possibility of . . . 'ii'e 3+ provides an in­ mistake because of 63 'ii' b l + 'ito> d4 64
direct defence. "if b6+ � c3 65 "iff6+ 'it> d2 66 'ii'd4+
154 SECRETS OF PRACTICAL CHESS

and the h4-pawn is vulnerable. It is 66 'ii'e3+

...

true that Black's position is so strong Black decides to improve his queen
that he might win even after losing his position before advancing the c-pawn
h-pawn, but understandably he does further. If Black plays . . . h3, he must
not wish to test this theory. start to take care about possible stale­
mates.
67 'iii>fl 'ii'f3+ 68 'iii>g l 'ii'g4+ 69
�f2
Or 69 'iii> f l � e3 70 'iWc5+ 1i'd4 7 1
'ii'e7+ 'ii'e4 7 2 'ii'c 5+ 'iii>f4 7 3 'ii'd6+ (73
'ii'c7+ � g4 74 'ii'g7+ <itih3) 73 . . . � g4
and the checks run out.
69 'ii'd4+ 70 <itifi <itid2 (D)
...

70 . . . �e3 7 1 'ii'e 8+ 'ii'e4 would have

been a little quicker.

63 'ii'a4+
White must start checking, as Black
threatened to check on g4 and then
play ... c3.
63 �d2 64 'ii'a2+ �d3 65 'ii'a6?
.•.

This makes life easy for Black as

his h-pawn is able to advance another
square. 65 'ii'bl + would have offered
more resistance, but Black would es­
cape from the checks in the end:
65 . . . �c3 66 'ti'al+ � b4 67 'ii'b l+ �c5 71 'ii' b 4+ 'iWc3 72 'ii'b5 'ii'd3+ 73
68 'ii'f5+ �d4 69 'ii'd7+ (69 'ii'f6+ <itif2 'ife3+ 74 <itin 'ii'f3+ 75 'iii>gl c3
'ii'e5 70 'ii' b6+ � c3 and 69 'ii'f4+ �d3 Although the process is rather slow,
10 •rs+ 'ii'e4 7 1 'fin + 'iii>d2 72 'ii't2 + the pawn gradually advances. White
�d l also win for Black) 69 . . . 'iii>c 3 70 can give several checks between each
'if g7+ �c2 7 1 'ifg6+ 'iii>d2 72 'ii'd6+ pawn move, but there is no perpetual.
'ii'd3. 76 'iFa5
65 h4 66 'ii'b 5
••. The alternative pin 76 'ii' b4 loses
After 66 'ii'd6+ Black wins as in the after 76 . . . 'ii'e 3+ 77 �fl 'ii'e2+ 78 � gl
previous note: 66 . . . 'iii>c2 67 'ii'g6+ 'ii?d2 h3 79 'ii'd4+ (79 'ii'd6+ ..d3) 79 . . . <iti c2
68 'ii'd6+ 'ii'd3. 80 'ii'a4+ � bl 81 'ii'b4+ (8 1 'ii' b 3+
 THE ENDGAME 155

'ii'h2) S l ...'i'b2 S2 'iWe4+ �al S3 'ii'a4+

•a2.
76 'iWe3+ 77 �n 'ii'e2+ 78 �gl
...

h3 79 •dS + 'it>cl 80 'ii'g5+ 'it>b2 0-1

Where the pawns are on one side of

the board, the winning chances with
an extra pawn are probably greater
than with rooks. An ending with 4 vs 3
offers fair winning chances, and even
3 vs 2 is sometimes won in practice,
although I have no doubt that theoreti­
cally it should be a draw. worth noting that after 1 ll aS+ i. gS 2
ll a7, for example, 2 . . . i. e6? would
Common end ings without allow 3 ll h7+ <it gS 4 lie? and White
wins (checking distance again).
Pawns If we move the bishop from b3 to
Rook vs minor piece b4, then White wins even if Black
moves first, for example 1 . . . 'iii> g S 2
The ending rook vs bishop (without l:r.aS+ .i.. f8 3 l:r. bS and mate next move.
pawns) arises occasionally. There are The third important point is the fol­
three important things to know about lowing position.
this ending. First of all, in a general
initial position this ending is totally
drawn. Secondly, if the defender's
king is trapped in the corner then he
can draw if his bishop moves on
squares of opposite colour to the cor­
ner square; otherwise it is lost. The
reason is quite simple (D).
Here White can try 1 llaS+, but af­
ter 1 . . .i.. g S he must lift the stalemate.
•

This involves either letting the black

king out of the corner, or unpinning
the bishop. The only other reasonable
winning attempt is l ll h7+ � gS 2 In this case Black's king is near the
.:r. b?, attacking the bishop and threat­ 'wrong' comer and in fact the position
ening mate on bS, but Black has the is lost whoever moves first. The im­
saving move 2 . . . .i.. c2+. However, it is portant point is that the result is the
156 SECRETS OF PRACTICAL CHESS

same no matter where Black's bishop

is positioned, so long as it is on a dark
square.
The winning method runs 1 ltfi
i.b2 ( l . .. i. h4 2 lt f3 i. d8 3 ltf4 forces
the bishop to a bad square, winning af­
ter 3 . . . i.e7 4 .:r. a4 or 3 . . . i. b6 4 l::tb4
i. c7 5 :lc4) 2 l::tf2 i. g3 (Black tries to
keep the bishop sheltered behind the
white king for as long as possible;
2 . . . i. g l 3 lt g2 also drives the bishop
out) 3 lt g2 i.d6 (3 . . . i. h4 and 3 . . . i. f4
are impossible because White wins 'iii> g 8 ! is also a draw) 4 ...� g8! 5 .:r.g7+
the bishop with a discovered check) 4 'iPhS 6 lt gl (this is the key moment;
ltd2 i.e7 5 lta2 and again White wins Black must play carefully) and now:
the bishop. 1 ) 6 ... .!Llh7? 7 <iti g6 ! � g8 (7 ... .!Llf8+
'iii>
8 f7) 8 l::t g 2 (purely a waiting move)
Rook vs knight is another ending 8 ... .!Llf8+ 9 <iti f6+ ! �h8 10 'iii>f7 wins.
which occasionally arises in practice. 2) 6 . . . .!Lle6? 7 'iii> g 6 ! .!Llf4+ (after
It can even result from some endings 7 . . . .!Llf8+, 8 <iti f7 wins) 8 <itif7 wins.
of rook vs pawn in which the pawn has 3) 6 . . . .!Lld7 ! (the drawing move, be­
to underpromote to avoid mate. While cause it allows Black to meet � g6 by
the details of the ending are signifi­ . . . � g8) 7 � g6 (7 .:r.d l .!Llf8 ! ) 7 . . .<iti g8 !
cantly more complicated than with s :n (8 .:r. g2 <iti f8) s . . . .!Llf8+ ! 9 <iti f6
rook vs bishop, the one basic principle .!LJh7+ ! IO <itie7 <iti g7 ! and Black draws.
is clear enough: if the king and knight Whilst this marginal position re­
are not separated, then this ending is quires accurate play, the general situa­
almost always a draw. Even if the de­ tion (king and knight next to each
fending king is on the edge of the other and away from the edge of the
board (this situation arises automati­ board) is a comfortable draw.
cally in the underpromotion cases),
then the position is usually drawn. Lost positions of the type that arise
Black's king is in a relatively unfa­ in over-the-board play generally fall
vourable position near the corner of into two categories . The first arises
the board, but there is just enough when the king and knight are next to
space to hold the draw. However, ac­ each other, but are poorly placed. One
curate defence is required. Play might obvious case is when Black's king is
continue 1 'iPf6 .!Llh7+ 2 'iii> g6 .!Llf8+ 3 trapped in a corner. A less obvious
�h6 �hS 4 :r1 (4 l::t g 7 .!Lle6 ! s l::tf7 possibility is when Black's king is on
 THE ENDGAME 157

the edge of the board and the knight 83 lllc4

•••

occupies g7 (b7, b2 and g2 are equiva­ Or 83 ... catg3 84 l:lc7 ! �f3 (84 . . . ll:\a4
lent, of course). The position with the 85 catd5 lllb 2 86 l:r. c l ! ll:\ a4 87 'iPd4
black king on g8 and knight on g7 is wins) 85 l:r. c3+ ! (a surprising winning
one of the worst defensive formations move, but a typical motif in such end­
and is quite often lost if the white king ings; moving to either the second or
is nearby (for example, with the king fourth rank allows White to chase the
on f6 and rook on al Black loses even knight) 85 . . . 'iii>e 2 (85 . . . 'ite4 86 :tc6 !
if he moves first). lll a8 87 'iP d6 followed by l:r. a6) 86
The second category consists of po­ l:r. c6 ! ll:\a4 87 l:r. c4 ! lll b 6 88 :b4 lllc 8
sitions in which the king and knight 89 <itid7 and the knight is trapped.
are separated and cannot join up. 84 l:r.f3+! 'iii>g4
Sometimes the win is quite simple, but White wins easily after 84 ... 'iP g2 85
in others subtle play is required to l:r.c3 ! ll:\a5 86 'iPd5 lllb7 87 l:r. a3 llld 8
keep the king and knight apart while at 88 l:r. a7, followed by l:r.d7.
the same time making progress. If the 85 l:r.d3 'iii>gS
knight cannot be trapped directly, then The knight cannot move, for exam­
it is important to keep the defender's ple 85 . . . ll:\a5 86 �d5, 85 ... lllb6 86 l:r.b3
king restricted. The attacker some­ ll:\c8 87 :b7 or 85 . . . lllb2 86 l:r.d2 ! .
times has to use his own king for this 8 6 'iPdS! lllb6+
purpose, even though it is counter­ The alternative is 86 . . . lllb2 87 :d4
intuitive to move it away from the en­ <itif5 88 <itic6! (not 88 � c5? � e5 ! with
emy knight. zugzwang; White must use triangula­
tion to ensure that this position arises
with Black to move) 88 . . . 'iii> e5 89 �c5 !
cate6 90 'iPb4 � f5 9 1 �b3 ! (not 9 1
cat c3? �e5 and again White i s in zug­
zwang) 9 1 . . . �e5 92 l:r. d8 and wins.
87 �es ll:\c4+
Black has little choice, for example
87 . . . � g4 (87 . . . ll:\ c8 88 'iPe6 lllb 6 89
l:[d4 wins similarly) 88 <itie6 <iti g5 89
l:[d4 'iii> g6 90 :b4 or 87 ... ll:\ a4 88 l:r.b3
ll\ c5 89 l:[bS ll:\ a6 90 'iPd6+ � f6 9 1
l:[b6.
88 'iii>e4 (D)
A. Karpov - L. Ftacnik Karpov could have won more rap­
Thessaloniki Olympiad 1988 idly by 88 � e6, e.g. 88 . . . <iti g6 89
l:[ g3+ <ith5 90 <it d5 lllb6+ 9 1 �e5
158 SECRETS OF PRACTICAL CHESS

lbc4+ (alternatively, 9 1 . . . �h6 92 l:td3 Or 90 . . . lbb2 (90 ... lba3 9 1 �d3) 9 1

<lt g7 93 l:td4 <lt f1 94 �d6 � f6 95 'iii>e3 <itif5 9 2 <it d2 <lte5 9 3 l:. b4 and
l:. f4+ <itii g5 96 l:t b4) 92 'itf4 (keeping wins.
the black king confined is more im­ 91 'iii>eS lbc8 92 <lte6 lba7 93 <ltd7
portant than running after the knight) 1-0
92 ... lbd6 93 : g7 'iii>h6 94 l:.e7 lbc4 95
:c7 lbd2 96 l:.c2 lbb3 97 <ite5 <lth5 98 Rook and minor piece vs Rook
<ltd5 .
This occurs quite often in practice.
The general position is a draw, whether
the minor piece is a bishop or a knight,
but the practical winning chances de­
pend very much on which minor piece
is involved.
The ending of l:.+lb vs : should be
a simple draw as there are only a few
winning positions. Almost all of these
have the defending king badly placed
near a comer, and nothing like this can
be forced from a normal starting posi­
tion. The general feeling amongst
88 lbb6?
••• GMs has been that there is no point in
Ftacnik collapses. He should have carrying on this ending because there
tried 88 . . . <ltf6, when White can only are no real winning chances.
win with extremely precise play : 89 However, the following ending may
l:l.d4 ! lb a5 90 l:l. a4 ! lb b7 9 1 l:. a6+ ! cause this judgement to be modified
<lt g5 92 <ili>e5 ! lbd8 93 l:. f6 ! (this posi­ (D):
tion would be a draw with White to 70 <ithS
play) 93 ... lbb7 (93 . . . <iti g4 94 <itd5 lbb7 A perfectly reasonable move, but
95 l:. a6 wins) 94 l:. f8 lbc5 (94 . . . lb a5 70 :rs+ lbf4 7 1 l:. g8 would have been
95 <ltd5 lbb3 96 :n <iti g4 97 :d1 <it f4 slightly more comfortable for Black.
98 <lt c4 lb a5+ 99 <iti b5 picks up the After 7 1 . . . l:. h l + 72 <ltg5 l:. g l + 73 'it>h4
knight) 95 :ds � g4 96 l:. d5 lbb3 97 the result is either stalemate or the
'iii>e4 lbc 1 98 'iii>e3 and wins. white king escapes from the edge of
89 l:.d8 lbc4 the board.
White threatened 90 <ltd4, and if 70...lbg3+ 71 <lth6
89 . . . lba4 then White wins by 90 <ltd4 Thanks to the unfortunate position
lbb6 9 1 <lt c5 lba4+ 92 <ltb4. of White's rook, her king cannot es­
90 l:.d4 lbb6 cape from the edge: 71 'it> g6 lbe4+ and
 THE ENDGAME 159

7S :n + was simpler, for example

7S . . . lllf5 79 'iti> gS ll g7+ SO <J:; f8 :a7
S l <J:;es.

J. Polgar G. Kasparov
-

Dos Hermanas 1 996

now 72 'it>f5 and 72 <J:; f7 drop the rook.

However, 7 1 <J:; g6 llle4+ 72 ct;h6 ! was 78... :dS ! 79 :at ??
probably safer than the text, as on e4 This is actually the losing move.
the knight is less well placed to shield The only move to draw is 79 : n + !
Black's king. and after 7 9 . . . lll f5 one line runs SO
7t ... lllf5+ 72 <J:;b7 'iti>f4 73 :b8 J::l'.f2 : d4 Sl <J:; gS ! .
:g7+ 74 ct;bS 79... llle6!
White's king is now uncomfortably Kasparov, once given his chance,
placed in the corner of the board. The finds the only winning move.
position is still drawn, but care is SO :a6 'iti>f7 St :a7+ 'iti>g6 S2 :as
needed. :d7 S3 l:l'.b8 llc7 84 <J:;g8 J::l'.cS 85 l:l'.a8
74 ... :d7 75 :es libs 86 ct;b8 :b7 S7 J::l'.cS lllc7
As an example of how close White 87 . . . :b6 is the winning idea. The
is to losing, 75 : fS? would lose after plan is to play . . . lll g 5 without allowing
75 . . . 'iti> g5 76 : as 'iti> g6. One line runs White to check on the sixth rank. Then
77 : gs+ 'iti>h6 1s :g 1 :ds+ 79 : gs ... lllh7-f6 (possibly with . . ..:t e6 to keep
:d3 (this square is chosen to avoid the king confined) will fatally restrict
tricks such as 79 . . . :d2 SO : g2) SO the white king.
: g l :f3 S l : g4 llle7 S2 ll h4+ <J:; g6 Kasparov's move retains the win,
S3 ll h6+ <J:; f7 S4 :h7+ <J:; fs S5 ll h l but loses time.
tll gS S 6 <J:; h7 <J:; f7 S 7 'iti>hS tllf6 with a 8S l:l'.g8+ ct;b6 89 :gt
quick mate to come. It would probably have been a bet­
75 ... <J:;gS 76 ll e6 llld4 77 :et <J:;f6 ter practical chance to try S9 :fS and
7S :dt?! (D) see if Kaspa{ov could find the winning
160 SECRETS OF PRACTICAL CHESS

idea (89 . J:Cb6 ! ) mentioned in the pre­

.

vious note.
89 l:tb8+ 90 l:tg8 �e8 0-1
•••

As 9 1 l:[ f8 cat g6 92 l:t g8+ catfl wins.

It is up to the individual whether to

continue this ending, but I would only
regard it as worthwhile in the case of a
quick-play finish.
If one is the defender, then the main
advice is perfectly obvious: try to keep
your king away from the edge of the
board or, failing that, try to keep it I. Rogers M. Illescas
-

away from the comer. However, even Spanish Team Championship 1 996
if you do find yourself in a less favour­
able position, don't despair - even with along the edge of the board away from
the king in the comer many positions the white king. Thus Black avoids the
are drawn. most dangerous situation, that in which
White creates mating threats by hav­
The ending of l:t+i. vs : is one of ing his king directly opposite Black's.
the most common pawnless endings 78 <itid5
to arise in practice. In general it should Of course 78 <itid3 is answered by
be a draw, but the defence requires 78 . . . �aS .
considerable accuracy and grandmas­ 78 �a3
•••

ters have been known to lose it even in It is worth pointing out that the Co­
the days of adjournments. There are chrane Defence is most effective when
two basic drawing techniques, and it is the black king is near the mid-point of
worth knowing both of them because one edge. If, for example, the same
the position you find yourself in may formation were set up along the third
be more suited to one rather than the rank, then the corresponding move
other. would take the black king to a2, dan­
gerously near the corner.
The following diagram is the basic 79 i.d3 l:b4
position of the 'Cochrane Defence' . Black takes the chance to release
The black rook is pinning the enemy his king from the edge.
bishop and this prevents the white 80 l:tb8 l:g4 81 i.c4 'iii> b4 82 .ie2
king approaching its counterpart. The l:g7
only way to unpin is to play cat d5 or Now Black can meet 83 � d4 by
�d3, but then the black king moves 83 ...:ct7+. Black should try to maintain
 THE ENDGAME 161

a good checking distance for his rook, 109 ltd3+ �cl 1 10 ltdS <itb2 1 1 1
and it is helpful to position the rook so �e3 <itc3 1 12 .te4 :h8 1 13 :cs+
that it can check along both files and <itb4 1 14 <itd4 :h4
ranks. Cochrane again.
83 :bs+ �c3 84 ltc8+ 'iii>d2 8S l lS ltcl <itbS 116 :bl+ tf2.1/z
.tf3 �e3 86 ltc3+ �d2 87 :a3 :d7+
88 �c4 ltc7+ 89 �d4 ltd7+ 90 i.dS
ltd8 91 :a2+ �dl (D)

J. Norri S. Atalik
-

European Team Championship,

All White's efforts have only re­ Pula 1 997
sulted in the Cochrane Defence reap­
pearing on the lower edge of the In this position White is in no posi­
board. tion to use the Cochrane Defence. Any
92 �e4 <itcl 93 .tc4 ltd2 94 lta8 attempt to switch the rook behind the
ltd7 9S .i.d3 <itb2 96 ltb8+ �c3 97 black king would lose, for example
:lc8+ <itb4 98 .tc4 lth7 87 ltc8? ltb2+ 88 <iti fl ltf2+ 89 cat e l
Ensuring the maximum checking <itid3 would lead to a lost 'Philidor po­
distance. sition' (see the next diagram). There­
99 i.e6 lth4+ 100 �d5 :h5+ 101 fore White has to use the second of the
�d6 lth6 two drawing techniques, the ' second­
Even when there is no Cochrane rank defence' .
Defence, pinning the bishop is an ef­ 87 ltd2
fective defensive tactic. The basic idea is to defend pas­
102 ltc4+ 'iii>a3 103 :c3+ �b2 104 sively with both king and rook on the
ltg3 'iii> c2 lOS �es lth8 106 i.fS+ second rank. White will just oscillate
'iii>d2 101 lta3 :es+ 108 �f4 :rs with his rook between c2 and d2 until
Again a useful pin. Black undertakes positive action.
1 62 SECRETS OF PRACTICAL CHESS

87 ...ltbS 88 l:C2 lth2+ 89 <itidl position' . This winning position arises

This is the first of the two main relatively often as a result of inferior
points behind the second-rank defence. defence, but it is quite tricky to win. If
Black can drive the enemy king to the you don't know the correct method, it
edge of the board with a check, but is easy to go round and round in cir­
then his own rook is attacked so he has cles, until you give up in frustration.
no time to approach with his king.
89 lthl + 90 �e2 lth2+ 91 <itidl
•..

lth3 92 �e2 i.c3 (D)

- - - -
wm - - -
- - - -
- - - -
- -·- -
m m m ma
m : m� m m Philidor, 1 749
- - - -
This is the characteristic Philidor
This is Black's other winning at­ position. The kings face each other
tempt. At first it seems that White is in with Black's king trapped on the edge
zugzwang as any rook move loses in­ of the board, and his rook restricted to
stantly (93 ltc l lth2+ 94 <iti dl <itid3). one file by mating threats.
93 'iii>d l! 1 ltf8+ ! lte8 2 :f7!
Now 93 . . . <iti d3 is answered by the First White improves the position
stalemate defence 94 .:r.d2+ and the of his rook with gain of tempo, and
king has to retreat. confines Black's king to the first rank.
93 ... lthl + 94 �e2 'iii>d4 95 'iii>f3 The immediate threat is 3 lta7.
Given the chance, White's king 2 .:r.e2
•••

slips away. White can force Black's rook to c2

95 JID + 96 ltf2 ltal 97 �g4 :as
•• by playing .:r. a7 and this side-to-side
98 ltg2 1/2-1/2 switch, with possible checks on d7
thrown in, gives rise to so many possi­
If you have the rook and bishop, bilities that confusion is easy. What is
there is one winning position which hard to grasp is that White can only
you must be aware of - the 'Philidor win by means of appropriate bishop
 THE ENDGAME 1 63

manoeuvres; the bishop appears ide­ between el and c l , but this gives
ally placed on d5, so the idea of mov­ White the chance to transfer his rook
ing it away is counter-intuitive. from h7 to b7 with gain of tempo.
Black cannot run with his king, as 4 . l:tcl (D)
..

2 . . . <itic8 loses to 3 lt a7 ltd8+ 4 <it c6 4 . . . <ltc8 loses to 5 l: b2 :td l 6 lt h2

<it b8 5 ll b7+ � a8 6 ltb l <lta7 7 �c7, catb8 7 :Ca2.
forcing mate or win of the rook. Thus
he must play his rook down the e-file,
in order to meet :Ca7 by a switch to the
c-file. It turns out that e2 is the best
square for Black's rook and e3 is the
worst; e l is somewhere in between.
The win after 2 . . . lle3 runs 3 ll d7+
<itie8 (3 . . .<iti c8 4 lt a7 wins at once, be-
cause 4 ... ltb3 is impossible) 4 ll a7 catf'8
5 :t f7+ <ite8 (White has forced the
king from d8 to e8 with gain of tempo)
6 ltf4 (threatening 7 i. c6+) 6 . . . <iti d8
(6 . . . ltd3 fails to 7 lt g4 because Black
lacks the reply . . . ltf3) 7 .i.e4 (the final s .i.b3
blow for Black; d3 is covered) 7 . . . �e8 This is the key idea, without which
8 .i.c6+ and mates in two more moves. White cannot make progress, and it
If Black plays 2 .. J:te l , then 3 i. f3 explains why el is an inferior square
is a reflection of the position after to e2. This position is a genuine zug­
White's fifth move in the main line. zwang and if White were now to play,
After 2 . . .lt e2, White's ultimate aim his quickest win would be 1 .i.e6 ltd l +
is to force Black's rook to the third 2 i. d5 ltel 3 .i. b3, passing the move
rank. to Black.
3 :th7 5 ltc3
•.•

A waiting move to force Black's Or 5 . . . cat c8 6 :t b4 (threatening 7

rook to the slightly inferior square e l . i.e6+) 6 . . �d8 7 :t h4 lte l (7 . . . 'ii?c 8 8
.

After 3 . . . :te3 White wins a s i n the pre­ i. d5) 8 i. a4 � c8 9 i. c6 ltd l + 10

vious note. i. d5 ! catb8 1 1 ll a4 winning.
3... llel 4 ltb7 6 .i.e6
White 's winning line only works Now that Black's rook has been
when his rook is on f7 or b7. The forced onto the inferior third rank,
side-to-side manoeuvres are typical of White transfers his bishop back to d5
this ending ; Black is forced to follow with gain of tempo.
suit and oscillate with his own rook 6 ... ltd3+ 7 .i.d5! ltc3
164 SECRETS OF PRACTICAL CHESS

Here the win is analogous to the It should be noted that in most cases
note to Black's second move. the speeding up of play will favour the
8 l::ld7+ Wc8 side trying to win, because he can go
8 . . �e8 9 : g7 and f3 is out of
. round and round several times, each
bounds. time hoping for a mistake, whereas the
9 l::l h7 'it>b8 10 l::lb7+ 'lt>c8 1 1 l::lb4 defender only has to go wrong once.
�d8 12 .tc4 <iltc8 13 .i.e6+ and White Here is a quick round-up of the most
mates. common endings where the result
might be affected:
Quick-play fin ishes 1 ) Rook and bishop vs rook. The
drawing techniques are fairly well­
These days the normal method of fin­ known, but to apply them in practice
ishing a long game is the 'quick-play requires considerable thought. I would
finish' in which, from a certain point not care to defend this in a quick-play
on, the players have a fixed amount of finish.
time to complete the game, no matter 2) Rook and knight vs rook. Should
how many moves this might take. In be comfortably drawn at a normal
international competitions, the quick­ time-limit, but in a quick-play finish it
play finish normally comes into effect might be worth continuing (as Kaspa­
at move 60, and players are generally rov showed).
given half an hour (sometimes an 3) Queen and pawn vs queen. Of
hour), in addition to any time they still course, some positions are objectively
have, to complete the game. lost, but even many of the drawn posi­
Quick-play finishes only affect a tions require very accurate defence. I
relatively small percentage of tourna­ imagine that the 'marginal' drawn po­
ment games, but it is important to be sitions would be almost impossible to
aware of the impact such a finish may defend in a quick-play finish, and even
have on certain endgames. There are some of the fairly comfortable draws
many endings in which the odds would would probably be difficult.
favour a draw at a traditional time­ 4) Rook and pawn vs rook. Posi­
limit, but in a quick-play finish the tions which are drawn by the standard
balance of probabilities is changed. 'third-rank' defence would still be
Quick-play finishes also put a pre­ drawn in quick-play. In more complex
mium on memorized knowledge; after positions, there would of course be
the exhaustion of six hours' play, and more errors, but in these endings the
with limited time on the clock, work­ attacker also has to play accurately, so
ing out a complex ending is very diffi­ the accelerated tempo probably makes
cult - you just have to know how to the task equally difficult for both play­
play certain positions. ers.
 THE ENDGAME 165

If one can liquidate to a variety of positional advantage but not necessar­

different endings, then the fact that ily a win.
there is a quick-play finish might af­ It is perhaps unfortunate that modi­
fect one's decision as to which to aim fications to the time-limit can cause
for. If one evaluates l:l +.t vs l:I as an what amounts to an alteration in the
80% chance of a win (which seems evaluation of certain endings, but that
reasonable) then one might prefer that is the price one has to pay for the
to a line in which one has a clear elimination of adjournments.
5 U si ng a Com puter

Computers are wonderful tools. In Game data bases

half a century they have moved from a
few specialized applications to occupy When non-chess players think of a
a central position in our society. Mod­ grandmaster making use of a com­
ern civilisation probably could not puter, they probably think of playing
function without them, so dependent programs. However, at this level the
have we become on their services. main use of a computer is to access a
With the growth of the Internet, a fur­ large database of games.
ther communications revolution as There are two main uses for a data­
great as the introduction of radio and base. For the professional, examining
television is upon us. the games of prospective opponents is
Computers have also had an impact part of the job. If you are participating
on chess. Early chess computers were in a round-robin tournament, some of
laughably weak but, after Deep Blue's this may be done before the event, but
defeat of Kasparov in 1 997, nobody is in a Swiss event it can only be done
laughing now. For a relatively small when the pairings are known, which
price, you can buy a chess program in normally does not allow much time
a shop which, running on a standard for preparation.
PC, can defeat virtually anybody ex­ The second use is to look at the
cept for !Ms and GMs. You can also games played in specific opening
purchase a database of up to a million lines. If you are intending to play a
games. particular variation in the afternoon, a
The question arises as to how these quick scan of the database to see if
tools can best be used for personal im­ there are any recent games is very
provement. The discussion that fol­ helpful. You might get a new idea, or
lows refers to the database program you might see something potentially
ChessBase and the playing program dangerous for your opponent. For
Fritz. This is not because ChessBase home preparation, a database is also
have paid me a wad of money to men­ extremely useful. Instead of searching
tion their products, but because these through dozens of Informators, New
are the products I am most familiar in Chess Yearbooks and other standard
with. The discussion is as general as references, you can call up all the
possible and doubtless applies to other, games in a particular line with a few
similar, products. keystrokes. The games can then be
 USING A COMPUTER 1 67

merged into a single game with varia­ current position. A method of pasting
tions, so that you can easily see the this analysis into the game is highly
general structure of the variation. If desirable.
the database program has an interface That deals with the program, but
to a playing module, then at a key­ what about the data? Nobody is going
stroke you can set the computer to to enter a million games by hand, so
work analysing any given position. one is utterly dependent on commer­
The range of features available is cial offerings. However, here the situa­
truly astounding, and these days any tion is much less satisfactory than with
serious player is virtually obliged to the programs. Even the best databases
use a database. contain a significant number of errors.
The main features which I regard as Poor-quality data is commonplace;
essential for a database program are: indeed, sometimes it is so poor as to
1) It should handle databases of up make the data practically useless.
to a million games without struggling. Some 'commercial' databases are lit­
Of course, you will need a powerful tle more than games collected from
computer to handle such large data­ every available source, and just lumped
bases, but even so operations on very together - the 'kitchen sink' approach.
large databases can be rather slow. Typical problems are:
2) It should have facilities for en­ 1 ) Inconsistent spelling of names.
tering both variations and text annota­ This is perhaps the most irritating. If
tions, and manipulating these. you are considering buying a data­
3) It should run under the current base, have a look to see if Korchnoi is
version of Windows. DOS is dead; spelt more than one way in the data­
forget anything running under DOS . base. If it is, then just forget it. There
Mac users will have to make do with is nothing more irritating than missing
what is available. a critical game because you have
4) It should support openings keys failed to guess whether the player is
of unlimited depth, and should have ' Korchnoi' , ' Korchnoj ' , ' Kortschnoi'
facilities for users to modify and ex­ or any of his other close relatives. The
pand these. same applies to tournaments; you may
5) The ability to merge several find games from the same tournament
games into one game with variations described as 'Wijk aan Zee' , 'Wijk' ,
is critical. Once you have used this 'Hoogovens' , 'WaZ' , etc.
feature, you will not want to do with­ 2) Duplicate games. This is often a
out it. result of problem 1 . Sometimes whole
6) There should be an interface to a tournaments are duplicated because of
playing program, so that you just have some minor difference in the spelling
to hit a key to see the analysis of the of the event name.
168 SECRETS OF PRACTICAL CHESS

3) Incorrect results. This is one of you may say. Well, a Trabant and a
the most common errors. When enter­ Rolls-Royce are both cars with four
ing data, most programs have a default wheels. Unfortunately, the defects of a
option. If the operator forgets to enter database are not visible by looking at
the result, you will just get the default the shiny surface of the CD-ROM; you
result - with a two-thirds chance of actually have to use it before you spot
being wrong. You may even find com­ the problems.
binations of these errors, for example At the end of the day much of the
Korchnoi beat Bareev, but in an amaz­ blame for this situation must be laid at
ingly similar game Kortschnoj lost to the door of the chess players them­
Barejew. selves. Copying of data is common­
4) Incorrect moves. The above er­ place. Why should a company invest
rors are often fairly noticeable, but in­ thousands of man-hours producing
correct moves are trickier to spot. Of high-quality data if it is just going to
course, if the error is such as to leave be stolen (because that is what copy­
one player's queen en prise then you ing is in this case)? Software piracy is
might be a bit suspicious, but if it is the illegal - don' t do it.
wrong rook to d 1 , then you might never
know unless you compare the game Once you have a database you will
with another source. Once, when deal­ probably want to keep it up to date.
ing with a collection of games that There are various commercial services
contained many errors, I came across a which offer regular 'top-ups' for your
particularly grotesque example. I later database. These days much of the in­
mentioned it to the grandmaster con­ formation can be downloaded from
cerned: the Internet, although then you have
"Do you know how your game was all the problems of inconsistent names,
mangled in this game collection? Ac­ forgetting which tournaments you
cording to their score, your opponent have already included, etc. For really
could have mated in two, but instead up-to-the-minute material, the Inter­
left his queen en prise with check", I net is unbeatable, but it requires con­
laughed. siderable effort to keep on top of the
"But that really happened", he re­ flow of data. Unless you really need
plied. new games on a week-by-week basis,
the pre-packaged commercial offer­
Employing a poor-quality database ings are probably better. The Internet
causes immense frustration and waste is wonderful for e-mail, news and spe­
of time; the small amount of money cific enquiries for which you can use a
saved is not worth it. 'But they are search engine. Otherwise it strikes me
both databases with the same games .. . ,' as being a great time-waster.
 USING A COMPUTER 169

Playi ng progra ms of the ' game' and see where the stu­
dent could have played better. Most
The availability of cheap but very players don't have a suitable opponent
strong playing programs for the home on hand for such exercises, but the
PC provides many new opportunities computer can perform a similar job
for training and self-improvement. If and can also help in the post-mortem
you already have a computer, I would analysis. Even though computers play
strongly recommend buying such a endings less strongly than the middle­
program. The most obvious use, that game, they can still be useful. As dis­
of checking over one's own games, is cussed on page 1 3 1 , if you have trouble
also one of the most useful. It is amaz­ winning rook and pawn endings with
ing how often a computer check re­ an extra pawn, play out such a position
veals missed tactical ideas. Even if against the computer. It is a good idea
nothing shattering is found, the com­ to play out the same position several
puter often suggests interesting alter­ times, trying different plans. You will
native ideas which were overlooked soon develop a feeling for which plans
during the game. It is important not work and which are ineffective. If you
only to play over the game itself, but experience trouble winning at all, you
the tactical lines on which your deci­ may get some ideas by reversing the
sions were based. For this reason, you colours !
should only consider buying a pro­ The computer does have several
gram that supports the entry of varia­ limitations; in the areas of positional
tions, and can afterwards store the judgement and plan-forming it is of
games and variations in a database. practically no help, and long-term sac­
You may well find a pattern in the rifices are almost always rejected by
ideas you have missed, and this will the machine. Analysing the Najdorf
provide you with useful information Poisoned Pawn with the aid of a com­
about which areas of your game need puter is a futile exercise; White's com­
special attention. pensation is so long-term that the
Another use is to play out training computer does not see it at all, and
positions against a computer. The fa­ resolutely assesses every position as
mous Russian trainer Mark Dvoretsky winning for Black.
recommends the method of 'playing Curiously, I find playing normal
out' interesting positions . The idea is games against computers much less
that if, for example, one of his students helpful. Computers have a particular
is weak in tactics, Dvoretsky would style of play and one soon learns how
set the student to play tactical posi­ to avoid their strengths and exploit
tions against a strong opponent. After­ their weaknesses. This knowledge is
wards they would go over the course of no value against human players,
1 70 SECRETS OF PRACTICAL CHESS

who (normally ! ) have a very differ­

ent set of strengths and weaknesses.
Moreover, it is quite easy to become
depressed playing against the com­
puter. All one really learns is that it is
common to overlook tactical points.
Playing programs are just starting
to have a major impact on grandmas­
ter chess. Here is one example:

A. Shirov L. van Wely

-

Monaco Amber (rapid) 1 997

Sicilian, Najdorf 15 ltlxe6!!
An absolutely stunning innovation,
1 e4 cS 2 ltlf3 d6 3 d4 cxd4 4 ltlxd4 which gives White a clear advantage
ltlf6 5 ltlc3 a6 6 .te3 e6 7 g4 h6 8 f4 in every variation. After the game Shi­
bS 9 .t g2 .i.b7 10 gS rov announced that this move had not
This move was introduced in the been discovered by him, but by Fritz.
game lvanchuk-Topalov, Las Palmas Sure enough, when I set up the posi­
1 996 as an improvement over the pre­ tion on my Fritz, the program found
viously played 1 0 a3. ltlxe6 in less than a minute. It is per­
10...bxgS 11 fxgS ltlbS haps slightly surprising that Shirov
The lvanchuk-Topalov game con­ used up his innovation in a rapid event,
tinued 1 1 . . .b4 12 ltla4 lLlh5 1 3 0-0 but he probably felt that since anyone
lLld7 14 g6 lLlhf6 15 c3 lLle5 16 gxf7+ with a Fritz could find this move, it
'iP xf7 1 7 cxb4 lt h4, and now 1 8 .t f4 would only be a matter of weeks be­
would have been clearly better for fore someone else played it.
White ; in the game Ivanchuk played 15 'iPxe6 16 eS .txg2
•••

1 8 'ii' b 3 and lost quickly. The text is a Black is lost after 1 6 . . . ltl xe5 1 7
counter-improvement by Gavrikov. .t xb7 ltb8 o r 1 6 . . . 'ii'c7 1 7 exf6 ltl xf6
12 g6 ltlf6 13 gxf7+ 'iPxf7 14 0-0 1 8 ltld5 .i. xd5 1 9 lt xf6+ 'iP xf6 20
ltlbd7 (D) 1i'xd5.
Now Kulaots-Gavrikov, Hallsberg 17 exf6 ltlxf6 18 'iPxg2 ltc8
1 99617 continued 1 5 e5 .t xg2 16 exf6 18 . . . We8 is no improvement, owing
.t xfl 1 7 ltlxe6 'if xf6 1 8 ltlg5+ 'iP g6 to 1 9 :t xf6+ 'iPxf6 ( 1 9 ... gxf6 20 "ifd5+
1 9 'ii'd5 and here Black could have �e7 2 1 'ii'b 7+ � e6 22 : e l wins) 20
gained a large advantage by 1 9 . . . :M "ifd5 � g6 (20 . . . 'ii'xe3 2 1 ltle4+ rile7
20 ltlce4 l:r.g4+ 21 ltl g3 lt xg5 22 22 1i'xd6+ �e8 23 1i'e6+ ritd8 24
.t xg5 Wxg5 23 Wxa8 .th3. ltd l + mates) 2 1 'ii'd3+ � f7 22 :n+
 USING A COMPUTER 1 71

'iii> g 8 23 •ds+ � h7 24 l::t f3 j,,e 7 25 line was refuted by a computer, but it

l::t h 3+ � g6 26 'ife6+ j,, f6 27 l::t g 3+ may be the first in which the human
'iii>h7 28 'iff5+ and wins. 'innovator' was honest enough to ad­
19 'iff3 j,,e7 20 :&el 'iii>f7 21 'ifdS+ mit it! The question naturally arises as
'iii>f8 22 'iffS 'ifc7 23 j,, d4 'ifb7+ 24 to how many opening innovations are
�gl (D) capable of being found by a computer.
Of course 24 ltld5 .J:. xc2+ 25 'ifxc2 Every four months a panel of leading
'ifxd5+ 26 'ife4 l::t h5 27 'ifxd5 l::t x d5 grandmasters votes on the most im­
28 l::t f4 is also good, but Black could portant opening innovations. The re­
put up considerable resistance in the sults are published in lnformator. At
ending. the time of writing, the most recent is­
sue for which the results of this vote
are available is Informator 68. I set
Fritz4 to work on the positions in
which the top 15 novelties of Informa­
tor 68 were played, to see how many
the machine would find. In each case I
allowed sufficient time for a complete
analysis down to 1 1 ply (although
many lines were taken far deeper, of
course). The result: Fritz found 3 of the
15 innovations.
Of course, there are many problems
with such a test, for example some of
24 .J:.c4 25 .J:.f'4 'ifc8 26 l::te6
•.. the 'innovations ' may in fact not be
All White's pieces are in perfect good, and will perhaps be refuted in
central positions, while Black's are the near future. Another point is that
scattered randomly around the edge of sharp, clear-cut innovations tend to
the board. head the Informator list, because a
26... l::txd4 subtle finesse which can only be ap­
Losing at once, but after 26 . . . � f7 preciated by specialists in a particular
(or 26 ....J:.h6 27 lLlrl5) 27 ltle4 l:.xc2 28 opening line will not score well in
ltl xd6+ j,, xd6 29 l::t xf6+ gxf6 30 'democratic' voting. In a couple of
'iixf6+ 'iii>e 8 3 1 'iff7+ the result would cases the innovation was Fritz's sec­
be the same. ond choice, which would certainly
27 l::txf6+ 1-0 alert a human operator to the fact that
there was something of interest pres­
It isn't clear if this is the first in­ ent. However, to summarize I would
stance in which a top-level opening say that Fritz was only of value for
1 72 SECRETS OF PRACTICAL CHESS

finding opening innovations in certain

types of position. Having said that, to
find 3 of the world's top 1 5 innova­
tions in less than a day is quite an
achievement !
Here are two of Fritz's successes
and one of its failures.

0. Korneev - Y. Piskov
Linares Open 1 996

at the top of Fritz's list within 2 sec­

onds, and in 75 seconds it was being
assessed as winning for White. The
game concluded 20 l:r.xfS 21 tDxe6+
•..

'ito>e7 22 'ii'xfS lLJxd4?! 23 lLJxd4 'ii'xd4

V. Miluydas - S. Muraviov 24 l:r.fel+ 1-0
corr 1 994
In the following example, Fritz
Here Fritz took less than five seconds failed to find Timoshenko's innova­
to find the innovation 24 tDh3+ ! ! •.• tion although, as we shall see, it is not
(exclamation marks as given i n Infor­ clear that the new idea was any im­
mator; 24 . . . tD xe6 would have been provement over the move previously
good for White) 25 'it>g2 lDgS ! . The played.
game concluded in spectacular fash­
ion with 26 'iWf2! .i.f3+ 27 'ii?n f4 28 G. Timoshenko - B. ltkis
g4 'iWh3+ 29 'it>el tDxe6 30 lDeS lDgS Baile Herculane 1 996
31 lba6 ! ! .t.xg4 32 .i.c4 ! ! tDe4 33 French Defence
'iWf3!! 'iWh4+ (33 . . . .t. xf3 34 tDt7+ 'ito> g8
35 lDh6+ is also a draw) 34 'ii'g3 and 1 e4 e6 2 d4 dS 3 tDc3 tDf6 4 eS lDfd7
the players agreed to a draw in view 5 f4 cS 6 tDf3 tDc6 7 .t.e3 cxd4 8
of 34 ... fxg3 35 tDn+ 'it> g8 36 tDh6+. lLJxd4 'iWb6 9 'iWd2 'iWxb2 10 l:tbl 'iWa3
1 1 .i.bS tDxd4 12 .i.xd4 .t.b4 13 0-0
In the following position from the a6 14 l:r.b3 'ii'aS 15 l:r.fbl .t.a3 (D)
French Defence, 20 tDfS ! ! was already 16 f5! !
 USING A COMPUTER 1 73

The innovation (with lnformator's survive after the obvious 22 . . . lLJ xeS,
exclamation marks), deviating from but unfortunately this move isn't men­
16 i. xd7+. Fritz probably failed to tioned at all in Timoshenko's notes. I
find it for two reasons: firstly, it is not cannot see anything better than 23
clear if it is really better than talcing on 'ii'c3 lLJg4+ 24 cate2 l:l a8 25 lLJxc8 h5
d7 ; secondly, the combination would 26 'iib4+ cat g8 27 'ii'xb7 l:l a2 with a
be simply too deep for Fritz, even were likely draw.
it sound. 22 :as? 23 'ii'g S!
.•.

16 axbS 17 :xa3! 'ii'xa3 18 �bS

••• Now Fritz agrees that White is win-
'ii'xa2 19 lLJd6+ �f8 20 l:lal 'iix al+ ning.
21 i.xal l:lxal+ 22 <itif2 (D) 23 f6 24 'ii'hS g6
•••

This is really the critical moment of Or 24 ... �e7 25 'ii'f7+ �d8 26 fxe6 ! .
the whole combination, and we are al­ 25 'ii b6+ catg8 2 6 lLJe8 �f7 27
ready 1 2 ply away from the starting lLJd6+ �g8 28 lLJe8 �f7 29 'ii'g7+
point. Fritz still feels that Black can catxe8 30 'ifxh8+ lLJf8 31 exf6 1-0
I nd ex of N a mes

Numbers refer to pages

Adams 36, 55 Filguth 58

Ahues 1 24 Fine 1 29
Alekhine 1 32 Fischer 56, 57
Anand 1 2, 25 Flear 1 2 1
Andersson 5 5 , 1 3 6 Flohr 1 4 1
Atalik 1 6 1 Ftacnik 1 57
Averbakh 147
Gelfand 26
Beckemeyer 52 Giddins 54
Beliavsky 1 40 Grigoriev 1 19
Bellon 84 Gunst 1 0
Botvinnik 103
Browne 89 Hebden 1 2 1
Hort 40
Cabrilo 1 42 Howell 1 1 1
Capablanca 1 32, 1 4 1 Hilbner 1 3 6
Chandler 5 2 , 1 1 0
Chekhover 103 llincic 142
Chemin 1 02 Illescas 1 60
Chigorin 1 36 Itkis 1 72
Christiansen 52
Cochrane 1 60 Kamsky 25
Conquest 47 Karpov 52, 1 57
Cook 5 1 Kasparov 64, 159
Cox 50 Khalifman 99
Crouch 7 Kieninger 1 53
King 8, 14
de la Villa 59 Komeev 172
Dolmatov 140 Kosten 7 1
Dvoiris 99 Kotov 7
Kuligowski 1 1
Eliskases 1 53 Kuzmin, G. 1 2
 INDEX OF NAMES 1 75

Lasker, Ed. 1 23 Sax 105

Lautier 1 2 Schlage 1 24
Levenfish 1 29, 1 35, 1 3 6 Seirawan 1 5, 53
Ljubojevic 107 Shirov 26, 64, 1 2 1 , 1 70
London (city) 1 35 Short 64, 102
Smyslov 129, 1 35, 1 36
Mandler 1 1 7 Soltis 76
Mellado 30 Sorri 79
Mestel 55 Spassky 53, 1 34
Meszaros 45 Speelman 1 3 3
Miles 107
Miluydas 172 Stean 105
Muraviov 1 72 Stefanova 54
Sveshnikov 1 2
Neustadtl 1 1 6
Nimzowitsch 98 Tai 5 6
Norri 1 6 1 Tarrasch 98
Nunn 1 1 , 1 5 , 1 9, 22, 30, 40, 47, 50, Timman 64
58, 59, 62, 84, 89, 94, 1 0 1 , 144 Timoshenko 1 72
Tisdal! 5, 9
Onishchuk 36
Unzicker 57, 82
Petrosian, T. 82
Philidor 1 62 Van der Sterren 1 9
Piskov 144, 172 Van Wely 54, 5 5 , 1 70
Plaskett 94 Vydeslaver 1 0 1
Polgar, J. 1 59
Portisch 28 Wahls 1 1 1
Pfibyl, M . 22
Prokd 1 1 8 Xie Jun 62
Psakhis 8, 1 4
Ye Rongguang 54
Ragozin, E . 1 1 1 Yudasin 1 1 0
Reti 1 1 7 Yusupov 28
Rogers 1 2 1 , 1 60
Zhu Chen 134
St Petersburg (city) 1 35 Zimmerman 45
I nd ex of O peni ngs

French Defence 30, 172

Giuoco Piano 22, 40, 76
King's Gambit 53
King's Indian Defence 1 5
Latvian Gambit 7 1
London System 54
Nimzowitsch-Larsen Attack 45
Ruy Lopez 1 9
Sicilian, Dragon 1 07
Sicilian, Najdorf 56, 89, 94, 170
Sicilian, Pelikan 50
Sicilian, Scheveningen 1 05
Sicilian, Velimirovic Attack 1 10
Torre Attack 54
Trompowsky 55
Two Knights Defence 36
OOAl�I BllTI
What is the best way to improve your chess results? Memorising an opening
encyclopedia, learning endgame theory, solving puzzle positions . . . there
must be an easier way. How about making the most of your existing talent?

In this book, John Nunn helps you to do precisely that. Drawing upon more
than three decades of experience, he provides advice that will help players
of all standards, playing styles and temperaments to achieve improved
results. His methods take into account psychological factors and are firmly
based on good common sense and the objectivity that has made John Nunn
one of the world's favourite writers on chess.

Topics include:

• Defending difficult positions

• How to study the openings

• How to make decisions

• Avoiding common m istakes

• When to calculate

• Coping with pressure

• Avoiding time-trouble

Dr John N u n n has been one of the world's leading grandmasters for nearly
twenty years and regularly competes in both grandmaster tournaments and
domestic events. He has won four gold medals in chess Olympiads and
finished sixth overall in the World Cup in 1 989. He has twice won the British
Chess Federation Book of the Year Award.

Other titles from Gambit Publications:

1 01 Chess Opening Surprises S.T.A.R. Chess

Graham Burgess Paul Motwani

Vishy Anand: My Best Chess Games The GAMBIT Gulde to the Bogo-lndian
Vishy Anand Michael Rohde

How to Beat Your Dad at Chess

Murray Chandler

Gambit Publ ications Ltd is:

Managing Director: Mu rray Chandler GM £1 4.99 $ 1 9.95

Chess Director: Dr John Nunn GM
Editorial Di rector: Graham Burgess FM ISBN 1 -901983-0 1 - 3

For further information about

Gambit Publications, write to us at:
Gambit Publications Ltd
69 Masbro Road, Kensington
London W 1 4 OLS, England.
O r send a n e-mail to: 1 0061 7.2702 @ compuserve.com 9 7 8 1 90 1 9830 1 2 >