Bishop v. Knight - The Verdict PDF

Bis hop versus Knig ht:
T he Verdict
Steve Mayer
B. T. Batsford Ltd, London

First published 1 997
Steve Mayer 1997
ISBN 0 7 1 34 82 1 5 X
British Library Cataloguing-in-Publication Data.

A catalogue record for this book is
available from the British Library.
All rights reserved. No part of this book may be

reproduced, by any means, without prior permission
of the publisher.
Edited by Graham Burgess and typeset by Petra Nunn for

Gambit Publications Ltd, London.
Printed in Great Britain by
Redwood Books, Trowbridge, Wilts
for the publishers,
B. T. Batsford Ltd,
583, Fulham Road,
London SW6 5BY
To my mother; Gloria Mayer
A BATSFORD CHESS BOOK

Edito rial Panel: Mark Dvoretsky, Jon Speelman
Commissioning Editor: Paul Lamford
Gene ral Manager: David Cummings
Contents
Symbols 4
Preface 5
Acknowledgements 8
1 Some Characteristics of Knights and Bishops 9

2 Two Bishops and the Steinitzian Restriction Method 12
3 Chigorin and the Knight Pair - The Traditional
Case for Success 29
4 The Rest of the Story - Chigorin usually
lost with the Knight Pair 47
5 The Problem Knight 64
6 The Problem Bishop 74
7 The Over-rated Knight 89
8 The Over-rated Bishop 99
9 Changing the Colour of a Bishop 108
10 Increasing the Speed of Your Knights 118
11 The Bad Bishop 134
12 The Sacrifice for Active Bishops 148
13 The Unexpected Exchange 165
14 Shattered Pawn Positions 177
15 The Ruy Lopez Ending 193
16 The Grindable Ending - Rook and Bishop vs
Rook and Knight 201
17 Capablanca's Theorem - 'iY+1tJ is better
than 'iV+.t in the Ending 209
Index of Players 219
Index of Openings 222
Index of Endings 224
Symbols
+ Check
++ Double check
# Checkmate
x Capture
!! Brilliant move
Good move
!? Interesting move
?! Dubious move
? Bad move
?? Blunder
+- White is winning
White is much better
White is slightly better
= Equal position
=+= Black is slightly better
+= Black is much better
-+ Black is winning
Ch Championship
(n) nth match game
(D) Diagram follows
1-0 White wins
112-112 Drawn game
0-1 Black wins
Preface
I'll let you in on a little secret: the a value of 3 . The program played
most common material imbalance quite well, but even the program
in chess is that of bishop versus mer realized that there isn't 'really'
knight. Intellectually, most chess a difference of 0.4 points between
players realize this, but it is easy to the minor pieces. Telling the pro
forget in the face of the conven gram to accord them different val
tional 'value scale ' of the pieces. ues was simply the most expedient
After all, for most of us, the litany method of factoring in the effects
of ' one, three, three, five, nine, the of differences in the respective
game' was the first bit of strategy pawn structures.
we were taught, generally within As a chess teacher who has
fifteen minutes of learning the dis worked with perhaps a thousand
tinction between checkmate and chessplayers over the years, I am
stalemate. accustomed to the plaintive 'Which
In fact, bishops and knights are is really better? ' that starts as soon
different pieces. They move differ as a novice chess player has
ently from each other and it is learned enough to ask tough ques
highly unlikely that they share the tions. My answer, which always
same value. This was recognized strikes the questioner as evasive, is
as least as far back as the mid-nine ' It depends on the position.' Curi
teenth century and a great deal of ously, despite the importance of
energy was expended trying to this subject, it appears to have re
fine-tune the value that should be ceived very little detailed coverage
accorded the minor pieces. in chess literature . Having written
However, there's a big problem: this book, I understand this pau
the performance difference be city: it's a huge subject domain and
tween the bishop and knight can be any single volume on the subject
very marked depending on various must be considered a preliminary
aspects of the position, particularly to a lifetime of individual work. In
those connected with pawn struc deed, a different writer might have
ture. I once worked on a computer selected wholly different examples
program which accorded the bishop and chosen different themes to em
a value of 3 .4 points and the knight phasize.
6 Preface
It will probably be helpful to the two-fold. First, chess really does

reader if he/she understands the in come down to concrete variations,
tent of this book. It's meant primar so it seems to me that it would be a
ily as a middlegame primer on the bit dishonest to present games as
topic of bishop versus knight. The 'model' play and then not bother to
ending is touched upon specifically point out where things went wrong
(and lightly) in three chapters, but (or could have been improved) .
not in the context of single-piece Secondly, the analysis frequently
minor-piece endings. This topic has sheds light on the thematic ideas
been covered quite well for some that are illustrated in particular
time by a number of great writers chapters, thus serving to strengthen
and players, including Averbakh, the reader's conceptual under
Fine, et al. I strongly urge you to standing of the material. I've also
study bishop vs knight endings to tried to avoid the inflationary 'page
gain a better understanding of the filling ' of such obvious comments
minor pieces. as 'The knight plays to e5 because
I also tried to avoid having this it is then well-placed in the centre.'
turn into an openings primer. The The writer must assume a certain
exchange of bishop for knight oc amount of knowledge on the part of
curs in a variety of openings, e.g., the reader. It's possible that some
the Nimzo-Indian, the Trompow readers will feel a bit lost at times,
sky Attack, the French, several Open perhaps because I've assumed
(and Closed) Sicilian variations, et something they don' t know, but I
aI. It's common sense that any open trust that anyone reading this work
ing which has been played at the understands that centralized knights
highest levels as long as the Nirnzo are normally well-placed, doubled
Indian must be fundamentally pawns are frequently weak, it's
sound, so the resulting middlegame usually best to have pawns in front
positions must also be fundamen of your castled king position, etc.
tally sound. Consequently, I have The reader should remember
made some effort to avoid over that a single game may illustrate
loading what is intended to be a several different ideas. In at least a
book on the middlegame with spe few cases, I feel that the choice of
cific openings that can be studied which chapter to place a particular
with the aid of specialized opening game was arbitrary, i . e . , the same
works. game could have been featured in
I have adopted a more analysis three or four different chapters.
intensive approach than most books I didn't think that it was necessary
of this type. The reason for this is to include chapters on 'obvious '
Preface 7
cases of minor-piece superiority (or made a remark along the lines of

inferiority), so there are no chap ' the bishops are an advantage you
ters devoted to bishops in wide can win with ' . While everything
open positions or that sort of thing. depends on the position, I think it's
It is my hope that the space gained true that the bishops are better than
has allowed for discussion of ideas the knights in a wider variety of po
that might be a bit less shop-worn, sitions than the knights are better
such as over-rated pieces, problem than the bishops. Of course, I ' m
pieces, knight speed, 'changing the not sure this does us much good, as
colour' of a bishop, etc. we only get to play one position at
Oh, I nearly forgot to answer a time.
which minor piece is really better.
One of my favourite quips comes Steve Mayer
from GM Tony Miles, who once Virginia, 1 997
Ac knowledgements
No book is written in a vacuum, so this book can hardly be an exception.

A number of people offered advice, encouragement and aid along the way,
so I'd like to thank them. Of course, any errors or misconceptions are
strictly the author's.
A number of friends, both chess players and non-chess players, offered
advice and encouragement. These include Brian Deatly, George French,
Roger Mahach, Bill Mason, Bill Robinson, Sal Rosario, John Stannard,
Anna Tecson, Richard Terry and Jessica Wilder. My apologies to anyone
I've forgotten.
John Fedorowicz was kind enough to offer some comments and in
sights into the exciting game Miles-de Firmian, Manila Interzonal 1 990.
My boss and friend, David Mehler, was always encouraging and my job
with the U . S . Chess Center makes it much easier to justify writing chess
books.
l owe a lot to Graham Burgess. His analytical diligence and editorial
skills have made this book better than it was in its original form.
Finally, I ' d like to thank and dedicate this book to my mother, Gloria
Mayer.
1 Some C h aracteristics of
Knights a nd Bishops
It's useful to consider some charac The knight is a short-range

teristics of the minor pieces. Let's piece, i.e., unlike the rook, bishop
start with the knight on an open or queen, it can ' t swoop from one
board. end of the board to the other in a
single move.
The knight at al has only two

possible moves, while the knight at The knight' s method of moving
f3 has eight moves. This gives us is not related to that of any piece
the range of two squares to eight other than knights. This means that
squares for a knight's choices. It' s it can attack as many as five impor
also possible for the choices to tant pieces without being subjected
change quickly, e.g., the knight at to counterattack.
a2 will triple its choices to six The knight can reach squares of
squares the moment it moves, both colours and has the potential
while the knight at f3 could move to reach all 64 squares.
to h2, in which case its options The knight changes the colour
would drop to three squares. These of the square it sits upon every time
dramatic increases and decreases in it moves. Further, it attacks squares
mobility suggest that knight moves of an opposite colour to that upon
should be considered carefully. which it sits. David Bronstein makes
10 Some Characteristics of Knights and Bishops
the point that a knight occupying a toward the edge of the board should
hole in a position with a weakened always be considered carefully, as
colour complex actually attacks al the knight's decrease in mobility
legedly healthy pawns situated on may make it a target for trapping
the opposing colour complex. operations by the opponent.
It's harder to speak of general

characteristics of the bishop, as so
much depends on pawn structure.
Of course, its biggest limitation i s
that it can only reach half the
squares on the board. However, it
has the advantage of being a long
range piece.
As every beginner knows, the

knight is able to jump over other
pieces. This allows it to perform
defensive duties without suffering
a drop in offensive power. The
knight makes an excellent block
ader.
If there is some reason that the

bishop wishes to control g8, it does
it as well from the distant a2 as it
would from the adjacent f7 .
We also note that bishops have
their greatest mobility in the centre
and their least mobility on the side
of the board. Therefore, the bishop
at e5 has thirteen choices , while
the bishop at a2 has only seven
The knight does well in closed choices.
or semi-closed positions. A knight B ishops are less dynamic than
move to the edge of the board or knights in that they can move and
Some Characteristics of Knights and Bishops II
still continue to protect or attack even see it attack along two diago
some of the same squares they hit nals.
before they moved. By compari
son, the knight loses all touch with
its previous set of squares the mo
ment it moves.
The bishop pair make up for

many of the deficiencies of the sin
gle bishop. Here the centralized
bishops hit 26 squares, while even
Bishops have the further advan a centralized queen hits only 27
tage that their mobility allows them squares. When there is a contest
to provide a variety of functions. between two bishops and a bishop
Here we see a bishop preventing and knight, one of the most typical
the advance of passed pawns along methods of countering the bishop
two diagonals. If we were to add pair is to exchange the like-col
more pieces to the board, we might oured bishops against each other.
2 Two Bishops and the
Steinitzian Restriction
Method
Steinitz is commonly credited with Board, Reti points to the following

the discovery that two bishops ver game as probably the earliest dem
sus a bishop and knight or two onstration of Steinitz's discoveries.
knights is often an advantage.
Whether the first world champion Rosenthal - Steinitz
truly ' discovered' this can be de Vienna 1873
bated, as earlier masters already
recognized that two bishops can 1 e4 e5 2 t2Jf3 t2Jc6 3 t2Jc3 g6! ? 4
form a potent attacking force. d4 exd4 5 t2Jxd4
However, it cannot be disputed that The sharper 5 t2J d5 ! ? was intro
Steinitz' s play did much to codify duced in Rosenthal-Steinitz, Lon
the technical exploitation of the don 1 883 (see page 2 1 below) .
two bishops in open and semi-open 5 g7
positions, nor can it be denied that 6 e3 t2Jge7
his voluminous writings did much 7 c4 d6
to spread the glad tidings to such Neishtadt points out that 7 . 0-0
. .
younger masters as Charousek, Las 8 WHd2 t2J e5 9 e2 d5 ! would lead

ker, Pillsbury and Tarrasch. The to positions similar to the game but
very term 'the advantage of the two with Black having saved a tempo
bishops' (or the simpler 'the two by advancing the d-pawn in one
bishop s ' ) quickly passed into the move (cited by Euwe in his book
lingua franca of chessplayers. From Steinitz to Fischer).
8 0-0 0-0
Richard Reti is among the 9 f4? ! (D)
authors who credit Steinitz with A modern master would shy
working out the method of how to away from this move, as it accom
use the two bishops to advantage in plishes nothing aside from the per
an open or semi-open position. In manent weakening of the e3- and
the classic Masters of the Chess e4-squares . It does give the illusion
Two Bishops and the Steinitzian Restriction Method 13
B W
of increasing White's space advan operate in close proximity to the
tage but Steinitz is able to demon opposing fo rces. In order to be
strate through concrete play that come lastingly effective, it must
the white centre can be destroyed. find protected squares near the en
9 lLlaS! emy's camp, mostly squares pro
10 i.d3 dS! tected by pawns, inasmuch as other
1 1 exdS lLlxdS pieces a re in the long run not suit
12 lLlxdS 'iUxdS able for the protection of the
There can be no doubt that knight. Itfollows the refo re that in
Black is better, as he has exposed completely open positions without
the weaknesses on the e-file. Note, pawns, the bishop is superior to the
too, the fact that the pawn at f4 knight, a fact that is confirmed by
helps ensure that the white queen's the results of the endgame theory.
bishop is bad. Conversely, the knight is superior
13 c3 lidS to the bishop in closed positions,
14 'ifc2 lLlc4 on the one hand because the pawns
1S i.xc4 'iUxc4 (D) are in the bishop's way, and on the
Steinitz has solved the problem other hand because the pawnsform
of the offside queen 's knight and points of supportfor the knight, as
acquired the two bishops. Reti has remarked above.
a great deal to say about this posi The method created by Steinitz
tion: to turn the advantage of the two
In contrast to the far-reaching bishops to the fullest possible ac
bishop, which can become e ffec count, is applicable only to posi
tive from a distance, the knight, in tions such as [Rosenthal-Steinitz,
order to become effective, has to Vienna 1873] which are nt;ither
14 Two Bishops and the Steinitzian Restriction Method
closed nor completely open, but in the opposing bishop, because the
which the re are still points of sup remaining bishop may then en
port for the Knight, protected by counter difficulties due to its in
the pawns, as here for example d4 ability to cover half the squares on
and e5. The method then cons ists the board.
in advancing the black pawns in 17 lbfJ b6
such a way that these points of sup 18 lbeS 'ii'e6
port become unsafe for the knight 19 'iHfJ ..ta6
which the reby is condemned to a 20 Ilfe1 f6! (D)
passive role and becomes quite in Black keeps a slight edge after
effectual. 20 . . . ..txeS ? ! 2 1 ..tf2 ! , but then he
The continuation of the present has created opposite-coloured bish
game, along with this chapter's ops and sold the bishop pair too
other examples, will bear out the cheaply.
accuracy of Reti's general descrip
tion of what might be termed the
Steinitzian Restriction Method. It
is worth noting that the keeper of
the bishops must be prepared to
move pawns to restrict the oppos
ing knight or knights. This requires
delicate timing so as to avoid creat
ing exploitable weaknesses in the
camp of the bishops. Further, many
of these pawn moves will have at
least the temporary effect of re
stricting the bishops, so care must w
be taken that it will be possible to
unveil the bishops at a later time. I noted above that the player
16 'iHf2 cS! with the two bishops must be pre
Steinitz steals the d4-square pared to make 'weakening' pawn
from White ' s knight. An additional moves and to block his bishops
benefit of the text move is that it when applying the Steinitzian Re
prevents the white bishop from striction Method. The text, which
opposing the black king's bishop steals the eS-square from the white
from the central d4-square. Indeed, knight, is an example of this.
the most effective remedy against 21 lbg4 hS!
the two bishops in such positions is This additional pawn move is an
the exchange of one of the them for even better illustration of the need
to move pawns when applying the 24 i.b7

Steinitzian Restriction Method. It 25 'tWg3 lid5!
seems odd to move the h-pawn, 26 litxd5! 'tWxd5 (D)
because Black does not have any
near-term opportunity for a king
side attack, nor is the knight at g4
generating any concrete threats.
However, Steinitz recognized that
on g4, the knight would still pre
sent Rosenthal with tactical oppor
tunities against the weakened f6
and h6-squares.
22 lbf2 'ift7
23 f5
In From Steinitz to Fischer, Euwe
wrongly gives this a question mark. W
23 g5
It may appear that Black's strat 27 lidl??
egy has ended in disaster, because This goes down without a fight.
his once proud king's bishop is Rosenthal had an excellent oppor
now buried behind its own pawns. tunity to obscure his positional dis
However, the fS-pawn cannot be advantage by playing 27 i.xgS !
defended easily with the g-pawn, fxgS 2S 'ifxgS. White then threat
so Steinitz has good chances of ens 29 lie7, when Black would be
winning it. Further, even if White forced to return the piece with
somehow manages to keep the fS 29 . . . 'ifxg2+. Black has a couple of
pawn in place, it should still be pos options, the second of which is
sible later to redeploy the bishop quite complex:
onto the open bS-h2 diagonal, e.g., a) 2S . . . l:.fS allows Black to de
. . . i.g7-fS-d6. fend the second rank, but White can
24 l:.adl force an immediate draw with 29
White's queen will now be lite7 lif7 30 lieS+ I HS 3 1 lie7 =.
forced from its active post on the b) 2S . . . 'iff7 29 lbd3 ! (29 l:.e7?
long diagonal, but the attempt to lieS ! defends) sets Black a difficult
create counterplay with 24 WHc6 ? ! defensive task. White has two
i.b7 2 S WHe6 WHxe6 2 6 fxe610ses the pawns for the piece and is threaten
e-pawn after the further 26 . . . lid6 ! ing to win with 30 lie7 . Play might
27 e7 lieS , when the pawn will be continue 29 . . :iYfs (29 . . . i.dS ? 30
Surrounded and captured - Mayer. f6 ! ) 30 lie6 ! ? (intending 3 1 l:.g6
and 32 f6; 30 tZJf4! ? is also interest These variations pass without

ing) 30 . . . l:ld8 3 1 f6 (D), with the mention by both Euwe and Reti. In
further possibilities: Reti 's case, this may have been a
function of his fundamentally
narrative style of annotation. The
narrative style is a powerful in
structional tool, as it allows the an
notator to illustrate general rules,
but it is less apt when analysing a
particular position. Chess is first
and foremost a game of analysis
and variations. Even very bad posi
tions usually offer at least one op
portunity to complicate the game; a
narrative style of annotation too
B often obscures this.
27 'fUxfS (D)
b l ) 31. . .'iVf7 ? 3 2 tZJe5 ! is a fit Steinitz 's strategy finally pays
ting victory for the once humiliated dividends in the form of a pawn.
knight. More important yet is the fact that
b2) 3 1 ...':d7? ! 32 fxg7 (32 tZJe5 Black's pieces will soon hold full
is also strong) 32 . . . l:txg7 33 ':g6, sway over the board.
and something has gone wrong
with Black's position.
b3) 3 1 . .Jhd3 32 ':e8 ! .l:r.d l +
33 'itf2 ':d2+ ! (an amusing echo of
White 's 3 2nd move; neither rook
can be taken) 34 'itfl .ta6+ 35
'it g l .l:r.d l + 36 'itf2 ':fl+ 37 'itg3
':xf6 - Mayer.
Black will emerge with a rook
and two bishops for a queen and
two pawns. White ' s king position
could prove a problem, but Black's
own king will also face threats of W
perpetual check. I would prefer to
play B lack, but White still has 28 'Wic7 .idS
chances , whereas he is simply lost 29 b3 l:te8
if he doesn' t sacrifice the bishop . 30 c4 .trT
31 i.el :e2
32 :n 'ilVc2
33 'ifg3 'ii'xa2
34 'iWb8+ '.1i>h7
35 'ifg3 .tg6
36 h4 g4
37 lbd3?? 'ifxb3
38 'ifc7 'ilVxd3
0-1
The irony of the knight's fate is
amusing.
W
R6ti cites another example that
has come to be regarded as a clas Black intends to play 9 . dS, ..
sic example of the Steinitzian Re eliminating White's spatial advan

striction Method. In my view, it is a tage . We will return to this posi
more straightforward example of tion shortly, as it was the subject of
the power of the two bishops, but commentary by both Steinitz and
as a chess game, it is considerably R6ti.
les s interesting than Rosenthal 9 'ii'd 2?! d5
Steinitz. 10 exd5 lLlexd5
11 lLlxd5 'ii'xd5!
Englisch - Steinitz White meets 1 1 . . .lLlxdS with 1 2
London 1883 i.gS .
12 i.e2 lbg4!
1 e4 e5 2 lbf3 lLlc6 3 .tb5 g6 4 d4 Steinitz prefers to obtain a con
exd4 5 lLlxd4 crete advantage, i.e., the pair of
R6ti thought that S .tgS was so bishop s , rather than to play for
powerful that it would practically ' activity ' with 1 2 . . . lb e4 ! ?, which
refute Black's opening play. Smys leaves him with no obviously fa
lov, among others, has demonstrated vourable continuation after 1 3
that Black has an acceptable game 'iVd3 . The text i s a good example of
after S . . . f6, followed by kingside Steinitz's principle of 'the accumu
development and playing for . . . dS lation of small advantage$' .
to dissolve the white centre. 1 3 i.xg4 .txg4
5 i. g7 14 lLlb3 'ii'xd2
6 i.e3 lLlf6 This move passes without com
7 lLlc3 0-0 ment by R6ti, but in the tournament
8 0-0 lLle7!? (D) book for London 1 88 3 , Steinitz
prefers 14 . . :iWc4 ! , with the idea of litd6 wins the exchange) 20 . . . i.xd4
1 5 . . . l:.ad8 and 1 6 . . . c5. 2 1 cxd4 a5 .
IS liJxd2 l:.ad8! (D) Still, Steinitz gives 1 9 liJd4 as
There is no reason to play superior, with the idea of bringing
1 5 . . . i.xb2 1 6 .:tab 1 , when White's the knight to the more active f3-
rook lands on the seventh rank. square. Englisch probably rejected
it out of a desire to keep the a-pawn
shielded on the a2-g8 diagonal.
19 cS
20 i.gS f6!
Black consistently steals diago
nals from the sole white bishop.
Another advantage of the text is
that it allows the black king to
come toward the centre.
21 i.f4 <i;t7
22 f3 gS! (D)
w
16 c3
This shields the b-pawn on the
long diagonal, but the weakness of
d3 will have consequences.
16 l:.fe8
17 liJb3 b6
18 h3 i.e6
19 Jiifdl
Black has obtained a very large
lead in development which allows w
him to use tactical means to main
tain the advantage of the bishops. Once again limiting the white
For instance, on strictly strategic bi shop. Englisch decides to cede
grounds, White would have liked the d-file, since 23 i.e 3 ? ? drops
to play 1 9 i.d4 here, as it would a piece after 23 ... l:.xd l + 24 l:.xd l
neutralize the black king's bishop. i.xb3 . Reti agrees with this deci
However, then Steinitz would ce sion, because he didn ' t think that
ment his advantage with 19 ... i.xb3 the h2-b8 diagonal is particularly
20 axb3 (20 i..x g7 ? i.c4 2 1 i.f6 attractive. Steinitz thought that
White should have played 23 i.c7, best squares and is ready to begin
wh ich indicates a preference for the decisive breakthrough.
the h2- b8 diagonal. 30 f2 gxf4!
23 ':'xd8 l:xd8 31 .txf4 i.g5
24 i. e3 h6 32 .txg5 hxg5
25 l:tel f5 Black has eliminated the white
26 f4 i.f6! bishop, which was useful for de
Steinitz maintains kingside ten fending the entry square at d2. Fur
sion. An exchange at g5 will not be thermore, Black has the additional
particularly attractive for White, as possibility of creating a passed
Black then has the option of secur pawn with .. .f4.
ing a space advantage with . . . hxg5 33 e3 f6 (D)
or of challenging the white bishop,
which has become a key defender
of d2, by . . . i.xg5 .
27 g3 a5!
This gains space and threatens
to destroy the white queenside by
advancing the pawn down to a3 .
28 lbc1 a4
29 a3 .tc4! (D)
w
34 h4 gxh4
Steinitz is in the pleasant situ
ation of having multiple good con
tinuations. For instance, he could
also have secured the outside passed
pawn for himself with 34 . . . f4+ 35
gxf4 gx h4. I suspect that Steinitz
w chose the text-move as he had al
ready worked out the game con
Steinitz ensures that the knight tinuation to a dead win.
will be subject to exchange if it at 35 gxh4 lite8+
tempts to re-enter the game via e2. 36 f2 l:txel
Black now has his pieces on their 37 xe1 e5
38 ltJe2 2) 9 h3 c6 10 J.d3 d5 1 1 exd5

This results in a lost king and ltJexd5 1 2 ltJxd5 ltJxd5 13 c3 ltJxe3
pawn ending, but otherwise White 14 fxe3 WIgS , and Black's advan
is effectively playing without his tage was indisputable in Macken
knight. zie-Steinitz, London 1 883 (0- 1 in
38 J.xe2 48 after mutual blunders).
39 xe2 f4 3 ) Steinitz advocates the odd
40 c4 g4 and apparently pointless 9 f3 in the
41 e3 f4+ tournament book for London 1 883.
42 e4 f3 4) Reti offers 9 e5 ltJe8 1 0 J.f4.
43 e3 g3 Black can then play to undermine
0-1 the white centre with 1O . . . f6 or
1 O . . d6, but 1 1 l:te l should still
.
The position after 8 ... ltJe7 ! ? (D) keep some advantage for White. I.n
is very interesting, as Steinitz' s re any event, White should perhaps
sults at London 1 883 indicate that play 9 e5 even if it leads to no more
it is surprisingly difficult for White than an equal game, as the other
to prevent the freeing . . . d7-d5 ad continuations (with the possible
vance: exception of ' 5 ' ) allow Black to
free his game and obtain good
squares for his pieces.
5) 9 J.g5 h6 10 J.h4 c6 1 1 J.e2
g5 1 2 J.g3 d5 13 e5 ! ? ltJe4 14 ltJxe4
dxe4 1 5 c3 ltJg6 1 6 'iUc2 ltJxe5 1 7
i.xe5 J.xe5 1 8 'iUxe4 l:te8 1 9 'iVd3
c5 20 ltJf5 J.xf5 2 1 'iVxf5 J.xh2+
22 xh2 ':xe2 23 l:tad l 'iUb6 24
l:td7 :f8 25 liUd 1 Wlxb2 26 llxf7
l:txf7 27 "i/ig6+ f8 28 lld8+ 1 -0
Dvoirys-Malaniuk, USSR Cham
pionship 1 989. This interesting
W game seems to be the only modern
grandmaster game featuring the
1 ) 9 J.c4 d5 10 exd5 ltJexd5 1 1 position after Black's eighth move.
ltJxd5 ltJxd5 1 2 i.xd5 'iUxd5 , and Strangely, the position after
Black had obtained a typical ad 8 . . . ltJe7 is not to be found in ECO
vantage (the two bishops) in the C, so we must be satisfied with
game Sellman-Steinitz, London Dvoirys-Malaniuk in the above
1 883 (0- 1 in 32 moves). note. It is also noteworthy that the
same position arose in Blackburne of the games at London 1 883. For

Steinitz, London 1 8 8 3 , with the instance, in the following game,
insignificant difference that the Rosenthal gives Steinitz a lesson in
white bishop was at e2 rather than the power of two bishops versus
at bS (that game arose out of a bishop and knight. The charac
Three Knights). teristics of the pawn structure are
Steinitz's games with and com not identical to Rosenthal-Steinitz,
mentary on the position arising af Vienna 1 87 3 , but the fact remains
ter 8 . . . lDe7 ! ? are intriguing, as they that Rosenthal clearly knew how to
raise questions about nineteenth use the bishops.
century chess. It is clear that Stei
nitz considered the position after Rosenthal - Steinitz
8 . . lDe7 ! ? acceptable for Black; it
. London 1883
is also clear that he considered the
typical positions where Black ac 1 e4 e5 2 lDf3 lDc6 3 lDc3 g6 4 d4
quires the bishop pair extremely fa exd4 5 lDd5 .tc5?! (D)
vourable. The thoughts of his
opponents are far less clear, since
they kept allowing the same type of
position to arise. There are two
possibilities : that his opponents
considered White's position sound
or that the level of opening prepa
ration in 1 883 was extremely poor.
The second scenario seems more
likely.
The question of how quickly
Steinitz 's contemporaries picked
up his ideas is one that arises as a W
matter of historical curiosity. Many
a simplistic formulation of chess Steinitz had a materialistic ele
history has it that such great con ment to his style; it doesn't show to
temporaries and successors as favour in this example. The d-pawn
Charousek, Lasker, Pillsbury, Tar will prove difficult to hold in any
rasch, and Zukertort understood event, but by developing the bishop
much or all of Steinitz's discover to cS , the holes at h6 and f6 become
ies, but that 'lesser masters ' did important weaknesses.
not. This strikes me as simply Steinitz demonstrated Black's
wrong, despite the evidence of some correct path three weeks later in
another game Rosenthal-Steinitz, 14 .ib3?

London 1 88 3 : S . . . .ig7 ! 6 .igS The bishop should have re
lLlce7 , when ECO C considers the treated to e2, as now Black could
position equal. force the exchange of White's
6 .tc4 d6 king's bishop by Steinitz's 14 . . . cS ! ,
7 .ig5 f6 with the threat IS . . .c4.
8 .ih4 'it>f8 14 lLlh6?
Steinitz gives 8 . . . hS, threatening 15 0-0 g5?!
to win a piece with 9 . . . gS and Steinitz attempts t o secure his
1O . . . h4, as 'the correct play ' . dark squares as a line of defence
9 'i!t'd2 'it>g7? against a white attack with f3-f4.
White threatened 1 0 Jt.. xf6 lLlxf6 Such a plan is also known from the
II 'i!t' h6+, but Steinitz gives 9 . . gS
. Modem Benoni, but it has obvious
10 .ig3 lLleS as fine for Black. drawbacks , e.g., the weakening of
10 b4 .i b6 fS and the general weakening of
11 lLlxb6 axb6 the kingside.
12 lLlxd4 (D) Steinitz gives the immediate
IS . . lLlef7 as correct. Note that
.
I S . . . cS? 1 6 f4 ! favours White.

16 .i g3 lLlerT
17 :ae1 c5?
This weakens the queenside to
no purpose, as Black no longer
threatens to play . . . cS-c4. The di
rect 1 7 . . . .id7 is superior.
18 lLlb5 I:.d8
19 f4!?
Rosenthal elects to play for a di
rect kingside attack, but simply 19
B I:. d l (Steinitz) wins the backward
d-pawn.
And so Rosenthal has regained 19 g4
his pawn and obtained the bishop 20 lLlc3 .te6
pair. Steinitz 's position, on the other 21 lLld5 Jt.. xd5
hand, has little to recommend it, as 22 Jt.. xd5
his development lags and his king Now White has obtained two
side is weak. bi shops for two knights and has
12 Wlie7 chances of breaking through on
13 f3 lLle5 practically any part of the board.
22 c7 28 exf6+ xf6
23 bxc5 29 c3+ g6
Steinitz queries this move and 30 l:t el Ilxel+
su gge sts 23 'ic3 , intending 24 e5 , 31 'ixel 'id7 (D)
as superior. In fact, there is nothing Steinitz queries this move and
wrong with the text. offers 3 1 ... 'ie7 32 J.e4 WHe6 33 "ii'e2
23 . bxc5 (D) f6 as best play, when it is his
view that White's isolated queen
side pawns give Black the better
game. Given that the position is still
a middlegame, I think it is prema
ture to award Black the advantage.
W
24 liIbl?!
Steinitz doesn't comment on
this move, but it is surprising that
both he and Rosenthal missed the W
direct and logical 24 e5 ! , with the
point being that White mates after 32 lIb6 lLJc6
24 . . . dxe5 25 fxe5 fxe5 26 l:txe5 ! 33 J.h4!
lDxe5 27 WHg5+. Black would have 'A very fine move, which pa
had to sit tight after 24 e5 ! , but then ralyses Black's game' - Steinitz.
White has the pleasant choice of 33 lDxh4
exposing the king with 25 exf6+ 34 'ixh4 'if5??
(this is probably best) or playing The text allows White to mop
for a passed pawn with 25 e6 - up. For better or worse, Black had
Mayer. to try 34 . . . h6! 35 .te4+ g7 36 f5,
24 l;Ia7 when the g4-pawn will drop -
25 l:tb5 l;Ie8 Mayer.
26 l:tfbl lLJd8 35 J.c4! xf4
27 e5 lDf5 36 J.d3+ f7
37 'ixh7+ e6 and Englisch-Steinitz, was not lost

38 l:xc6 (D) on the younger members of what
has come to be known as the Clas
sical School. Tarrasch especially
showed a thorough command of
the Steinitzian Restriction Method.
B. Richter - Tarrasch
Nuremberg 1 888
1 e4 e5 2 liJf3 liJc6 3 .tb5lLJf6 4 0-0

liJxe4 5 e2 liJd6 6 i.xc6 dxc6 7
'tWxe5+ 'iVe7 (D)
White uses a pin to win a piece.

The rest is given solely for the re
cord: 38 :ii'e 3+ 39 n 'tWf4+ 40
.
'it'gl 'ie3+ 41 'it'n 'iVf4+ 42 'it'e2

'tWe5+ 43 'it'dl bxc6 44 'tWxa7 'iVxh2
45 .tn! 'igl46 'ia6 'id4+ 47 'it'c1
'if4+ 48 'it'dl 'id4+ 49 'tWd3 'iVal+
50 'it'd2 'iVxa2 51 'iVc4+ 'ii'xc4 52
i.xc4+ d5 53 .te2 g3 54 'it'e3 'it'e5
55 c4! 'it'f5 56 'it'f3 'it'e5 57 'it'xg3 W
'it'd4 58 'it'f4 dxc4 59 g4 1-0
The preconditions for the Stei
Not to belabour the obvious, but nitzian Restriction Method have
Rosenthal's use of the two bishops, been established. Black has ob
along with his exploitation of the tained the bishop pair and is ready
weakness of d5 , and the to and fro to swap queens . White should un
' tacking' from kingside to queen doubtedly have kept queens on by
side to kingside, is quite reminis 8 f4 or 8 a5 , probing the black
cent of Steinitz's play at its best. queenside, when he might have
had some opportunity of benefiting
The strength of the bishop pair from the state of Black's kingside
in such positions, as in the games development. Unfortunately, Rich
Rosenthal-Steinitz, Vienna 1 873 ter accommodated Tarrasch with . . .
8 lite!? ! xe5
9 litxeS+ i.e7
10 d3 f6!
White probably thought that
Black would have difficu lty cas
tling. In fact, Tarrasch demon
strates that his king is better placed
than if it had already castled, as the
position is tending toward an end
game and his king is already in the
centre.
Black's central pawn position W
or lack thereof - is typical of the
Steinitzian precedents, but here would still leave him with a some
White still has a d-pawn. This gives what crippled queenside pawn
him potentially better chances than structure. Instead, he prefers to
the absence of both centre pawns, play the knight to the outpost at d4.
as the d-pawn may help support the The knight will be so strongly
knight at a variety of central posts, placed that Richter will eventually
e.g., e4 or c4, although neither of be forced to exchange it and un
these squares is absolutely secure. double B lack's pawns, in which
11 :e1 cJ;;f7 case Tarrasch will have added a
12 i.f4 g5! substantial space advantage to the
Englisch-Steinitz has already bishop pair.
familiarized us with this method 17 t2Jd2 t2Jd4
of limiting the opposing minor 18 i.xd4
pieces. Richter surrenders his remain
13 i.d2 i.g4 ing bishop, as he hopes that his
14 t2Jd4 eS! knight will have better opportuni
The white knight is expelled ties of finding play in the resulting
from the centre and Black prepares semi-closed position.
to use d4 as an outpost for his re 18 ... exd4
maining knight. 19 a3 e5!
15 t2Je2 t2Jb5! B lack has the possibility of set
16 i.e3 l':tad8 (D) ting up what Hans Kmoch called
Tarrasch shows keen positional 'the quart grip ' . Generally, this will
judgement. He could obtain two involve advancing his queenside
bis hops versus two knights by the pawns to d4, c5, b5, and a4, at which
immediate capture at c3 , but that point various sacrificial attacks
will allow Black to queen a pawn. 22 ':e2 b5

We will soon see an example of the 23 ':ael J.rs
quart grip in action. 24 lL'lge4 ':g8
The quart grip is a special case 25 lL'lb3 ':c8
of what Nimzowitsch termed 'the 26 lL'led2 J.d6
qualitative pawn majority ' . The N ow might have been a good
typical pawn majority is quantita time to slip in 26 . . . h4, with the idea
tive, e . g . , three pawns to two in a of taking en passant if White later
particular sector of the board. The plays for control of e4 with g2-g4
qualitative pawn majority, as the (c.f. the note to White's move 30).
name implies, is a case where one's 27 lL'le4 i.rs
pawns are substantially superior in 28 lL'led2 f5
quality to the opposing pawns in a Tarrasch takes the e4-square
particular sector of the board. The from the knights and cramps the
quart grip, which derives its power white position even further.
from a spatial advantage, is one of 29 Ile5! J.d6?? (D)
the typical examples of the qualita
tive pawn majority.
20 lL'lg3 h5! (D)
w
Thi s should hav thrown away
w the fruit of all Black's labours, in
spite of the fact that Tarrasch at
Tarrasch increases his kingside taches no special marks to the
space advantage . Note that there move in his annotations in Drei
are no exploitable holes in Black's hundert Schachpartien. Simply
position, despite the fact that he has 29 . . . h4 would have maintained the
made a number of pawn moves. advantage, as the white rook does
21 f3 J.d7 not have an exploitable target.
30 ':5e2?? Momentarily preventing the ad

Now the game reverts to its stra vance of the a-pawn to a4, which
te gic course and White once again would complete the formation of
has a very difficult position. Rich the quart grip on the white queen
ter should have jumped at the op side.
portunity to play 30 ':d5 ! l:Ig6 31 ':ab8
(30 . .:c6 3 1 ttJxc5 wins a pawn)
.. 32 ttJab3 h4
3 1 g4! hxg4 32 fxg4. The black po 33 cJth1 ':g6
sitio n is then surprisingly difficult, 34 cJtg1 i.e6
as the automatic 32 .fxg4 33 ttJe4
.. 35 ':f2 ':a8
reveals that the black pieces have 36 ':fe2 as
lost their cohesivene s s . Further, 37 ttJb1 a4 (D)
his p osition is riddled with weak The quart grip is finally in place.
nesses, for example, c5, d6, f6, and Black's winning plan is to create a
g5 . Black might try 32 .:f6, but
... passed pawn on the queenside.
simply 33 lIn is very good for
White, as it's impossible to prevent
the white knight from reaching e4
in one way or another - Mayer.
Curiously, Tarrasch never real
ized how flawed this game is. He
gives simply "After 30 lI d5 ':c6,
the white rook would be very much
endangered. "
As with Rosenthal-Steinitz, Vi
enna 1 87 3 , it must be stressed that
even very good games nearly al
ways have a 'blip ' where a tactical w
interlude could have thrown a
spanner in the strategic works. The 38 ttJ3d2 c4
difficult task facing the defender in 39 ttJf1 ':c8
a strategically inferior position is 40 cJth1 c3
remaining alert for an opportunity 41 bxc3 dxc3
that may never arise and - even 42 ttJe3 b4
more difficult - judging when the 43 ttJc4 i.xc4
time is right and taking decisive ac 44 dxc4 litxc4
tion if the tactical chance arises. 45 lle3 bxa3
30 ':a8 46 ttJxc3 i.b4
31 ttJa5 0-1
A sterling example of the Ste degree of simplification has oc

initzian Restriction Method. I re curred. This view is borne out an
gard this game as even better than ecdotally by the discovery that of
such prototypes as Rosenthal the three 'classic' games examined
Steinitz, Vienna 1 87 3 , and Eng in this chapter (Rosenthal-Steinitz,
lisch-Steinitz, since Black had to Vienna 1 87 3 , Englisch-Steinitz,
address the fact that White had a and Richter-Tarrasch) , the only
centre pawn. Further, Tarrasch's one that couldn' t have been played
method of gaining space across the much better by the loser was Eng
entire board is quite impressive, lisch-Steinitz, which is also the
even though the players exchanged most materially reduced of the
blunders at moves 29 and 30. trio.
In my view, whenever a player
Verdict: The Steinitzian Restric substantially overestimates his or
tion Method had a major and last her prospects with the bishop pair,
ing impact on the perception of the it' s likely that an exaggerated be
relative strength of bishops and lief in the efficacy of the Steinitz
knights. However, as Reti pointed ian Restriction Method must take
out, it is most effective when a fair at least some of the blame.
3 Chigorin and the Knight
Pair - The Traditional Case
for Success
You have probably heard that Mikh games in this and the following
ail Chi gorin was a 'champion of chapter.
the knights ' who showed, say, a Yet examples of Chigorin 's De
'marked preference' for knights fence in Chigorin's oeuvre are rela
over bishops. This received wis tively rare. Much more common
dom is part of every chess player's are examples of Chigorin playing
cultural education; we all 'know' the white side of the Evans Gambit,
that Chigorin favoured knights an opening in which the bishops
over bishops, in contrast to Stei can generate tremendous attacking
nitz, his great rival. force for the gambiteer. Similarly,
Is this view correct and where while Chi gorin often played closed
did it come from? There are two openings, e.g., the Closed Sicilian
way s of demonstrating if Chi go as White or a form of proto-King's
rin actually favoured knights over Indian Attack against the French,
bishops . The first approach uses he did not make any special effort
his games as ' texts ' that reveal his to exchange his bishops for knights
thoughts on chess strategy. The in these games, despite the closed
believers in Chigorin's alleged nature of the resulting positions. If
knight fetish invariably point to we go solely on the basis of Chi
his queen's pawn defence, which gorin's games, then it's clear that
runs I d4 d5 2 c4 lLlc6, with a sub he 'preferred' bishops to knights.
sequent . . . .tg4. (Chi gorin was also A second approach is to turn to
known to play I d4 d5 2 lLlf3 .tg4.) Chigorin's writings and see what
Chigorin's Defence normally re he says on the subject of bishops
sults in Black giving up one or both and knights. Chi gorin was a pro
bishops for knights, in return for lific writer and probably addressed
a lead in development and a mar the relative value of the minor
ginally superior pawn structure. pieces in his long writing career.
We shall examine many of these Unfortunately, he wrote primarily
30 Chigorin and the Knight Pair - The Traditional Case for Success
in Russian and most of his writings on his play believed that he fa

appeared in newspapers and peri voured knights over bishops. In
odicals, e.g., the magazine publish stead, certain lesser and less
ed by the St Petersburg Chess Club. attentive writers jumped from the
These writings have probably not special case of Chigorin's Defence
survived the travails of twentieth to the much broader claim that
century Russia; I am certainly un Chi gorin favoured knights over
aware of any comprehensive 'com bishops. The claim sounded rea
plete works' of Chi gorin's writings. sonable, perhaps, to those not ac
Therefore neither his games quainted with Chigorin's games.
nor his writings can be blamed for Today it has become the chess
the myth that Chigorin favoured equivalent of an urban legend.
knights over bishops. The third and Chi gorin has been an underap
final possibility is that influential preciated figure in English-speak
commentators may have written ing countries. Fortunately, a great
that he favoured the knights over deal was done to rectify this by the
the bishops. Yet an examination of publication of Mikhail Chigorin
the 'usual suspects ' , for example, Selected Games in 1987. The
Botvinnik, Grekov, Kotov, Lasker, authorship of this excellent book is
Reti, Steinitz, Tarrasch, et aI, fails granted to Bogoljubow, who pub
to turn up any claims of this na lished a book on Chi gorin in Rus
ture. sian in 1 926. In fact, Mikhail
In 1 996, I asked the world-wide Chigo rin Selected Games is much
readership of the U senet news more than Bogoljubow 's work, as
group ' rec.games.chess.misc' if Colin Leach (the book's editor)
they knew where the 'Chigorin fa also collated a lot of annotations by
voured knights' party line may other sources into something that is
have started. No one could supply much more than Bogoljubow ' s
me with any leads, even those read original work. I have used these
ers with substantial libraries of various annotations extensively, as
Russian chess literature. Later, I you' ll soon notice by the numerous
had occasion to speak with 1M cases of 'cited in MCSG' that are
John Watson on this subject. Wat sprinkled throughout the next two
son, despite his extensive knowl chapters.
edge of Chigorin's games, was Let's turn our attention to those
unable to point to where this false games of Chigorin ' s that other
hood first surfaced. writers delight in using to demon
My opinion is that neither Chi strate his supposed preference for
gorin nor the major commentators knights.
Chigorin and the Knight Pair - The Traditional Case for Success 31
Steinitz - Chigorin become targets, as the king would

Havana World Championship not be available to defend them.
Match (6) 1 889 9 i.d2 i.cs
10 :gl
1 tLlf3 dS 2 d4 i.g4 3 c4 i.xf3 4 If White doesn ' t intend to cap
gxf 3 dxc4 !? S e4? ture at g7 , then this move is a mis
Steinitz later preferred 5 e3 . In take. Steinitz preferred 10 i.xc4
the second edition of ECO D, Cvetk (cited in MCSG).
ovic and V. Sokolov give ' 5 e3 ! ;t'. 10 ... tiJge7 (D)
S eS!
6 dxeS 'iUxdl+
7 c.t>xdl (D)
W
This is awarded an exclamation
B mark by Cvetkovic and V. Sokolov,
who misassess the position as
Both players have approached clearly better for Black. In fact,
the opening in what later genera White can improve at move six-
tions would consider their signa teen.
ture styles. Steinitz has given up 11 i.xc4
the right to castle for a pawn and It' s unclear whether Steinitz
the bishop pair, while Chi gorin has could have played 1 1 :xg7 ! ? tiJg6,
ceded the bishops - and a pawn - when White has two interesting
for the smoother development and continuations:
the better pawn structure. a) 1 2 e6 safeguards the rook
7 .. tiJc6 and bids to activate it along the
8 f4 :d8+! seventh rank. The drawback, of
It seems more natural to play course, is that it does nothing for
.
8 0-0-0+ but then f7 and g7 could
.. White's lagging development.
b) 1 2 .txc4 ! ? lb xf4 1 3 ':xf7 b22) 14 :xf4 lbxc4 is given by

(D) is critical: Chigorin (cited in MCSG). Bo
goljubow continues this with 1 5
'it>c2 ':g8 1 6 b3 i.d4 1 7 i.c3 .txc3
1 8 'it>xc3 ( 1 8 lbxc3 ':d2+ forces
mate) 1 8 . . . lb e5 (D), which he as
sesses as winning for Black.
bI) White should be better in

the exchange-down ending that
ari ses after 1 3 . . . lb d3 ( ' ! ' - Bogol
jubow, cited in MCSG) 14 i.xd3
W
'it>xf7 1 5 i.c4+ and 16 f4, because
the mass of passed pawns looks Again, this strikes me as ques
dangerous - Mayer. Bogoljubow tionable after the further 19 ':f5,
gave ' 1 3 . . lb d3 ! ' with no further
.
when White gains a tempo to con
comment or evaluation, so it's not tinue his development. It's true that
clear if he saw the exchange sacri Black has some active pieces, but
fice. we 're in an endgame and White's
b2) 1 3 . . . lb xe5 , with yet another connected passed pawns should be
subdivision: given their due.
b2 1 ) 1 4 lhc7 i.xf2 is clearly 11 lbg6
better for Black according to Stei 1 2 'it>c 1 i.xf2
nitz. This is debatable after the fur 13 ':g2 i.b6
ther 1 5 i.b5+ 'it>f8 1 6 ':xb7 - 14 lbc3 lbd4 (D)
Mayer. White has two extra pawns Black can stir up interesting
and his dark-squared bishop threat complications with 14 . . . ':xd2 ! ? 1 5
ens to become active on a variety of 'it>xd2 ( 1 5 ':xd2? .te3 is good for
diagonals, e.g., c l -h6, al -h8, and Black) 1 5 . . . lb xf4 16 lhg7 lb xe5
a3 -f8 . I don ' t see enough compen 17 i.b3 'it>f8, when it looks as if the
sation for two pawns. active black minor pieces will run
Chigo rin and the Knight Pair - The Traditional Case for Success 33
down the rook, but White emerges 21 .tb3 t'iJg6

with the better position after 1 8 22 e6 'it>e7 (D)
:xt7 +! t'iJxt7 1 9 .:n t'iJe5 20 llxf4+
- Mayer.
W The complications have re

solved themselves in Black' s fa
15 t'iJd5 t'iJf3 vour. White ' s passed e-pawn isn ' t
16 t'iJxb6? going anywhere against proper
Now the white position slides play, while Chigori n ' s knight i s
downhill. It's odd that Steinitz did superior to Steinitz's bishop.
not play to retain his bishops with 23 ':gl IIf2+
16 .ic3 ! (noted by both Steinitz 24 'it>e3 ':xh2
and Chigorin) 1 6 . . . t'iJxf4 17 ':xg7 25 ':g5
( 1 7 t'iJxf4? .te3+ should be a well It looks better to try the immedi
known theme by now) 17 . . . t'iJe6 - ate 25 ':f1 ! , but then Black obtains
Chigorin. Then Steinitz has at least a clear advantage with 25 . . . ':h3+ !
retained the bishop pair, but Black 26 'it>e2 t'iJe5 2 7 l:tf5 l::t x b3 (but
is ready to continue 1 8 . . . .id4 ! , certainly not 27 ... 'it>d6?? 28 l::t x e5 ! )
which neutralizes White's queen ' s 2 8 axb3 - Vasquez, who assesses
bishop and makes the doubled e the position as equal (cited in
p awn s an important feature of the MCSG) . However, the connected
position. kingside passed pawns should
16 t'iJxd2 cause White a great deal of trouble
17 l:txd2 axb6 after 28 . . . 'it>xe6 - Mayer.
18 l::txd8+ 'it>xd8 25 l:Ih3+
19 .ixt7 t'iJxf4 Black could obtain an even bet
20 'it>d2 ':f8 ter version of the previous note
with 25 .. .lbb2 26 lIf5 lIxb3+, but game against Emanuel Lasker at

it's unnecessary - Mayer. Hastings 1 895 . It's a fascinating
26 'it>d4 lIn struggle, with uneven play by both
Black also does well after sides, but Lasker missed at least
26 . . . c5+ 27 'it>c4 I:.e3 - Mayer. one win.
27 I:.bS 4Jf4
28 a4 hS (D) Em. Lasker - Chigorin
Hastings 1895
1 d4 dS 2 4Jf3 i.g4 3 c4 i.xf3 4

gxf3 4Jc6 S 4Jc3 e6 6 e3 i. b4 7
cxdS xdS (D)
w
Now the black h-pawn is able to
sail in, despite White's best efforts
at breaking through on the queen
side. w
29 as h4
30 axb6 c6 Chigorin typically recaptured
31 :fS 4Je2+ with a piece in such positions so as
32 'it>cs I:.xfS+ to retain the option of using d5 .
33 exfS h3 The drawback to this approach is
34 .ta4 h2 that it gives White a majority of
3S .txc6 bxc6 pawns in the centre, so Black's
36 b7 h1 middlegame strategy will revolve
37 b8'iV c1+ around restraining the white cen-
0-1 tre.
8 i.d2 i.xc3
Chigorin's strategic masterpiece 9 bxc3 4Jge7
in the handling of two knights ver This development of the knight
sus two bishops is undoubtedly his leaves B lack with the option of
later u sing his f-pawn to restrain or centre a bit shakier and prone to at
attack the white centre. The knight tack.
can also help indirectly protect the This type of position is conse
g -pawn from lth l - g l -xg7 , since quently difficult for both sides and
. . . tLlg6 and . . . <t>f8 would then trap presents problems that only strate
the white rook. gic play of a very high order will
Of course, White has two bish solve.
ops versus two knights; he also has 10 ltg1
two pawns to one in the centre, In the tournament book for
with his bishops' pawns supporting Hastings 1 895, Steinitz gives 10 e4
his centre pawns. On strictly classi 'iVh5 1 1 ltbl - threatening 12 ltb5
cal grounds, White must be better, - as ' much stronger' . This was
but Chigorin was always willing to tested in a consultation game Pills
play these positions out of his de bury and Blackburne vs Chigorin
fence. Why? and Schiffers sometime during or
For one thing, the white pawns immediately after Hastings 1 895.
are somewhat scattered, so Black The Anglo-American team won
might ultimately benefit from that. that game, but neither Pillsbury nor
Further, the white king will be Chi gorin must have been too im
slightly exposed wherever it finally pressed by the result, as the line
resides, as the pawn cover on both wasn't featured in any of their later
sides of the board has been dis Chi gorin Defence games.
rupted. The king might prove rea Another interesting continu
sonably safe in the centre, hiding ation is Romanovsky 's 10 'iVb3 ,
behind the mass of centre pawns, with the point that 1 O . . . 'iVxf3 1 1
but . . . e5 or . . . f7-f5-f4 could open ltg l opens the position for the
lines against it. Finally, while it's king's bishop, while 1O .. :iVxb3 1 1
true that White has the bishop pair, axb3 improves White's pawn struc
it's difficult for both of them to ture and gives him a better ending
achieve activity at the same time. If than he later obtains (cited in
they stay in their current formation, MCSG) .
then the king ' s bishop has activity, 10 'iWhs
but the queen' s bishop is stifled. 11 'iVb3 lbd8
On the other hand, the advance e3- 12 'iVbS+
e4 will open the c I -h6 diagonal for It seems mistaken to exchange
the qu een' s bishop, but it will also queens here, though White doesn't
s hu t down the bl -h7 and h l -a8 di lose his advantage by doing so.
agon als for the king's bishop. Fur Steinitz preferred 1 2 l':tg3 'iVxh2 1 3
ther, e3-e4 will make the white e4 'ih5 1 4 c4, when the bishops
and the centre pawns give White a majority and created options for
lot of play for his pawn. his queen's bishop.
12 xb5 1 6 e2 ?
13 .txb5+ e6 This is a mistake, as White
14 .t d3 lDg6 (D) should have leapt at the chance to
dissolve his doubled f-pawns with
1 6 [5 ! , while after the text, Black
should have leapt at the chance to
fix them with 1 6 .. .f5 ! , when Wat
son assesses the position as t.
16 l:te8?
17 l:tg3?
Yet another mistake. Levenfish
gives 17 f5 ! exf5 18 i. xf5 %:tc7 1 9
c4 a s clearly better for White (cited
in Watson). Watson comments that
"Black has no compensation for
W the active bishops and strong cen
tre."
15 f4 After 17 l:tg3 , Black should
Steinitz doesn 't go so far as to again have played 1 7 . . . f5 ! ; after
query the text, but he considers it Black's failure to do so, White
pointless unless White dissolves his should again have played 1 8 f5 ! .
doubled f-pawns with a later f4-f5 . 17 . e5 ?
H e prefers 1 5 e4, intending a later 18 %:tagl? e4!
c3-c4, while 15 . . . e5 1 6 dxe5 lDxe5 Black establishes a queenside
17 .te2 is favourable to White. majority and secures d5 as an out
15 0-0 post for a knight.
It seems more natural to block 19 .te2 f5! (D)
ade the doubled f-pawns with We finally arrive at the type of
1 5 . . . f5 . This is considered by John position that should have arisen
Watson in his opening monograph several moves back. Each side has
Queen 's Gambit: Chigorin Defence. his advantages and disadvantages.
He concludes that White 'keeps White is faced with the problem
control of the position ' after the that his bishops are both awful
further 1 6 h4 ! f7 1 7 h5 lDe7 1 8 pieces ; in one case because of
c4, intending 1 9 .tc3 and 20 a4. In black pawns, in the other case be
that case, White has clamped down cause of white pawns. However, it
on Black's potential queenside is always possible when one has
However, Levenfish points out

that 22 . . . a6 23 i.a4 b5 24 i.c2 al
lows White to place Black's queen
side majority under pressure by
means of 25 a4 and 26 litbl (cited
in MCSG) .
23 a4 ltJc6
24 libl ltd7
25 ltgg1 ltJge7
26 l:.b2 ltJd5
27 'itd2 lta5
W 28 lIgb1 b6
29 i.a3 g6
the bishop pair that they may later 30 l:.b5 lia6
become liberated. Black has the 31 i.el ltJd8
nice square d5 for a knight and a 32 lita1 ltJt7 (D)
queenside majority. Further, it is
possible that White 's centre pawns
will become weak when they even
tually advance, which they are al
most bound to do if his bishops are
ever to increase their scope.
Lasker drifts in the subsequent
play and never seems to come up
with a firm plan. Still, the difficult
nature of the position for both sides
is highlighted by the fact that
White will later miss plenty of
good opportunities even after be W
ing 'positionally busted' .
20 i.el ltt7 33 ltbb1 ltJd6?
21 i.a3 lIc6 Steinitz gives the immediate
22 i.c5 lta6 33 ... g5 34 fxg5 ltJxg5 35 l:.a3 ltJe4+
Chigorin manages to defend his (35 . . . ltJf3+ 36 'ite2 ltJxh2 3 7 f3 is
a-pawn, but at a high price; not good for White) 36 i.xe4 fxe4, and
only is his rook out of play, but its . . . ltg7 , when " . . . Black has a win
presence on the a-file makes it dif ning attack".
ficult for him to utilize his queen I'm not so convinced, as I fail to
side majority. see a definite win after 37 i.b2
:g7 3 8 'it>e2 l:.g2 39 lih l . Perhaps Chi gorin plays to open u p the
Black can then open a second front position in the hope of attacking
with 39 . . . bS ! ?, but White's possi the white pawns with his knights,
ble replies include 40 axbS , 40 as, but the text is now over-ambitious,
and 40 :ha l , with none of them as it exposes the black king and the
looking absolutely clear-cut. Black opening of the position should
can usually reduce play to an end benefit the bishops.
ing where he has knight and rook 36 'it> e2 gxf4
against a bad bishop and rook, but 37 e4! lLlf6
White will still have some counter 38 Jt.. xf4!
chances , particularly if he can Steinitz praises the text as supe
achieve piece activity. rior to 38 exfS exfS 39 Jt.. xfS :e7+
Bogoljubow gives a similar vari 40 'it>f2 lLld6, when the f-pawn can
ation that runs 3S Jt.. b 2 lLle4+ be defended by either 4 1 . ..lLldS or
(3s . . . lLlf3+? 36 'it>e2 lLlxh2 37 f3 is 4 1 . . .lLlhS .
once again good for White) 36 38 lLlh5
Jt.. xe4 fxe4 37 l:.gl + l:.g7 38 l:.xg7+ 39 Jt.. e3 f4
'it>xg7 39 l:.g l + 'it>f7 40 :g4 lLlf6, 40 Jt.. f2 lia5 (D)
when Black should win the ending
after grabbing the a-pawn (cited in
MCSG) . Yet it's unclear how he
makes progress if White isn't so
quick to unbalance the position by
surrendering the a-pawn, e.g., 39
'it>e2 or even 40 l:.a l , putting the
onus on Black as to how to convert
his advantage.
34 f3!
Lasker finally prepares the ad
vance of his central majority. It' s
true that the black knights, with the W
help of the c-pawn and f-pawn,
have done a good job of restraining 41 l:.g1+??
the centre and keeping the white Both players have done a good
bishop pair at bay, but White's cen job of playing a complicated posi
tre has been slowed down, not crip tion; this should be stressed, as
pled. most annotators have approached
34 lLlt7 this game solely by result and have
35 :a3 g5? ! credited Chigorin's play, while
lambasting Lasker's. In fact, White connected passed pawns should

could have obtained a winning po win (cited in MCSG) .
sition here with 41 eS ! , stealing d6 This wonderful refutation of 42
from a black knight and preparing eS makes Chigorin's failure to ad
to annex the c4-pawn, when I sim dress 4 1 eS ! all the more frustrat
ply don't see a defence for Black, ing. Perhaps the check at gl 'feels '
e .g. , 4 1 . . .ltJg7 42 lib4 ! ':'c7 43 so natural that he accepted it as a
i.bl ltJd8 44 .ta2 ltJc6 4S ':'xc4 bS given.
46 l:cS ':'xa4 47 ':'xa4 bxa4 48 dS ! 42 e5!
(D) wins for White - Mayer. 43 ':'ab1 ltJg7
44 ':'b4 Ilc7
45 .tb1 ltJe6
46 ':'d1 ltJed8 (D)
Unfortunately, Steinitz is com

pletely silent on the play from W
Black's 3 8th move to White's 47th,
while Chigorin and subsequent an 47 l:td2
notators also failed to spot 41 eS ! . The central tension helps Chi
41 f8 gorin defend his advanced c-pawn,
42 ':'aa1? for example, 47 .ta2 ltJc6 48 lhc4
Chigorin addressed 42 eS, when ltJd6 wins the exchange - Leven
he gives the pretty variation 42 . . . bS fish (cited in MCSG).
43 i. xh7 ltJxeS ! 44 %:tg8+ f7 4S For some reason, Bogoljubow
dxeS b4 ! 46 cxb4 IlxeS+ 47 fl gives 47 dS an exclamation mark;
li:Jf6, which he assesses as winning it's hard to believe that Chigorin's
for Black. Kan continues with 48 knights would have objected to
l'l g4 lL) xg4 49 fxg4 ':'dl + SO g2 the position becoming even more
:te2, when Black's active rooks and blockaded.
A more interesting try is Leven material with 53 ::'xe5 lLJxe5 54

fish's 47 dxe5 ! lLJc6 48 l:!.b5 l:!.xb5 .l:txe5 owing to 54 . . . l:tb3 ! - Mayer.
(48 .. Jha4 49 e6 is good for White, b) Even simpler is Steinitz's
as his e-pawn is now a factor) 49 52 ... .l:th6 53 .i.g5 l:!.xh2+ 54 'it>f1
axb5 lLJcxe5 50 l':td5 'it>e7 (cited in l:!.al 55 .i.xf4 l:!.xbl + and 56 ... l:!.h l +,
MCSG). The resulting position ap when " . . . Black's queenside will
pears much more resilient for win, especially as White 's c-pawn
White than the one that eventually will also fall soon."
arises in the game. His bishops 53 .i.d8 lLJd3+
aren't active and his e-pawn isn ' t 54 .i.xd3 cxd3 (D)
going anywhere, but his rook is ac
tive and the b-pawn does a good
job of crippling the black queen
side majority.
47 lLJc6
48 l':tb5 l:!.xa4
49 dxe5 lLJfxe5
50 .i.h4 l:!.g7
51 'it>f2 Ilg6
52 .l:tdd5 (D)
W
55 l:!.xd3??
Only now does the position be
come completely winning for
Black, as he is able to toss up a mat
ing net that will cost White a piece.
It's unclear how he wins after Ste
initz's 55 .i.c7 ! , with the point that
55 . . . Ilgg l 56 .i.xf4 gives the white
B
king shelter at e3 . It's true that
Black has the advantage after
52 l:!.a1? 55 . . . 11 a2+ 56 'it>f1 l:!.gg2 57 Ilxd3
Chigorin misses at least two Ilxh2 5 8 'it>g l , but even Steinitz
stronger lines of play: was unable to find a definite win
a) 52 ... l':ta3 ! ? targets the c for Black in this position.
pawn, when White cannot win 55 l':tag1
S6 :fS+ \t>e8
S7 .tgS :6xgS
0-1
The trouble Steinitz and Lasker

encountered must have suggested
to Chigorin's peers that the plan of
leaving the white king in the centre
was not the best in the position. Af
ter Hastings 1 895, the accepted ap
proach for White was to castle
queenside and attempt to use the W
bishops and open g-file to drum up
an attack against the black king. finds a number of files directly in
However, Chigorin demonstrated front of it flying open.
repeatedly that Black can generate 12 .td3 .txc3
an effective attack against the 13 'iWxc3 :b6!
white king once it is committed to This rook lift was another
the queenside. Chi gorin discovery that served to
bolster Black's attacking chances.
Pillsbury - Chigorin 14 \t>bl as
St Petersburg 1 895/6 15 :hgl lLlb4
16 .te4
1 d4 dS 2 c4 lLlc6 3 lLlf3 .tg4 4 e3 White 's bishops have managed
e6 5 lLlc3 .tb4 6 b3 i.xf3 7 gxf3 to work up more activity than they
lLlge7 8 i.d2 0-0 9 f4 :b8 10 0-0-0 typically do in positions of this type.
dxc4 11 .txc4 b5! (D) 16 lLledS
Pillsbury may have thought that 17 'i!Vcs 'iWa8
he was the only one with good at 18 .txb4 axb4
tacking chances in this opposite 19 i.xdS exdS
wing castling situation. In fact, as 20 xb4 (D)
Chigorin showed in this game and White has won a pawn at the
many others, Black can whip up a cost of exchanging both his bish
sUrprisingly strong attack by ad ops for the black knights. The rest
vancing his b-pawn and a-pawn. of the game is presented solely for
These serve as battering rams and the record, as it does not directly
secure the b4-square for a knight. It address the topic of this book.
may also prove possible later to 20 .. JIa6 21 b3 :b8 22 l':tgS
p lay . . . c7-c5 , when the white king c6 23 :dgl g6 24 fS b4 2S fxg6
Pillsbury - Chigorin
London 1 899
1 d4 dS 2 c4 lLlc6 3 lLlf3 .tg4 4 e3

e6 S lLlc3 .tb4 6 'iVb3 .txf3 7 gxf3
lLlge7 8 .td2 0-0 9 .td3 (D)
hxg6 26 WNd3 c;t>f8 27 11xg6 l:Ixa2

28 I1g8+ c;t>e7 29 l:Ixb8 WNa4 30
':b7+ c;t>e6 31 c;t>c1 ':al + 32 c;t>d2
':xgl 33 'ilic2 WNxc2+ 34 c;t>xc2 I1g2
3S ':xb4 11xh2 36 l':f.b7 11xf2+ 37
c;t>d3 fS 38 lIc7 c;t>d6 39 ftf7 cS 40 B
dxcS+ c;t>xcS 41 ':c7+ c;t>d6 42 ':c2
11f3 43 c;t>d4 ':h3 44 11f2 c;t>e6 4S An attempted improvement on
b4 ':h4+ 46 ':f4 ':hl 47 c;t>cS .:ct + the 9 f4 of Pillsbury-Chigorin, St
48 c;t>d4 ':bl 49 c;t>c3 ftel SO c;t>d2 Petersburg 1 895/6. It should be
':al liz_liz noted here that Pillsbury, unlike
An exciting game. Pillsbury was many of his contemporaries, was in
certainly better throughout the the habit of doing extensive open
complications, but White does not ing preparation, so it seems likely
appear to have an obvious win. that the text was an attempt at a
prepared improvement on 9 f4,
Both Chigorin and Pillsbury rather than a spur-of-the-moment
must have been reasonably satis decision.
fied with the course of their St. Pe 9 ':b8
tersburg 1 895/6 game, for they 10 cxdS lLlxdS
were to debate the same type of po 1 1 0-0-0
sition in another game a few years Hoffer, who wrote the tourna
later. On this occasion, Chigorin ment book for London 1 899, sug
showed that he had learned how to gests 1 1 0-0, but he overlooked that
cause the white king even greater Black can play 1 1 . . . WNg5+ 1 2 c;t> h l
problems. lLl xd4. However, Graham Burgess
poi nts out that White then has 1 3 16 l:.g3 a4 17 'iNdl lbcb4 18
exd4 'iUxd2 14 lbxd5 exd5 1 5 l:.fd l J.xb4 lbxb4 19 l:.hgl lIf7 20 i.bl
( 1 5 I:ad l 'i!t'f4 1 6 'i!t'xb4 'i!t'xf3+ 1 7 a3 21 bxa3 lbd5 22 'iNb3 b4 23
gl f5 1 8 l:.fe l l:.f6) 1 5 . . . 'iVxf2 1 6 axb4 l:.xb4 24 'iNd3 c5! 25 dxc5
'iWxb4 'iVxf3 + 17 'it>gl , when i t might 'iNa5 26 i.c2 'iVxa2 27 f3 :c4 28
be best for Black to take the perpet l:.lg2 l:.d7 29 c6 l:.xc6 30 'iUd4
ual check. 'iNa3+ 31 'it>d2 Ilxc2+ 32 'it>xc2
11 J.xc3 lbxe3+ 33 'it>bl litxd4 34 l:.xg7+
12 i.xc3 'it>f8 35 l:.g8+ 'it>e7 0-1
Now Hoffer suggests 1 2 bxc3 ,
with the idea 1 2 . . . b5 1 3 e4, when This game must surely be con
1 3 . . . b4 'would not affect [White] ' . sidered one of Chi gorin's most im
12 b5 pressive efforts with the knight
13 J.d2 l:.b6 pair. Like many of his successes
14 l:.dgl as featuring this material distribution,
15 f4 f5! (D) White's task was greatly compli
cated by concerns over king safety.
Somehow, very few of Chigorin's
opponents chose to castle into the
broken kingside, despite the fact
that a later 'it>g l -h l and l:.g l (and
l:.g2 if necessary) might have
proven a safer formation for the
king.
Bum - Chigorin
Berlin 1897
w 1 d4 d5 2 c4 lbc6 3 lbf3 .tg4 4 e3

e6 5 lbc3 J.b4 6 'iNb3 J.xf3 7 gxf3
Hoffer considers this position lbge7 8 i.d2 0-0 9 a3 J.xc3 10
winning for Black; the point is that i.xc3 l:.b8 1 1 l:.gl lbg6 12 0-0-0
Chigorin can defuse the pressure 'i!t'd6 13 l:tg3 dxc4 14 'iNxc4 b5 15
On the g-file with the economical 'iVc5 b4 16 axb4 lbxb4 (D)
. . . :f7 but White has no compara Another typically interesting Chi
ble method of safeguarding his gorin Defence. Black has opened
own king. Cvetkovic and V. Sok lines around the white king but
olov assess the position as clearly Burn is able to defuse the situation
better for Black. by walking his king toward the
Now Panov analyses both 2 1 b3

and 2 1 i.xb4 to clear advantages
for Black, but 2 1 i.e2 seems ten
able for White - Mayer.
Instead of this, Burn overlooked
Black 's threat and soon had to re
sign: 21 e2?? ':c6! 22 Wlia7 Wlic8!
(a trap closes around White's queen
and major material is now lost) 23
iLxb4 ':'xc4 24 i.c5 ':a8 25 'i!Vxa8
Wlixa8 26 d3 tbe5+! 27 e2 tbd7
w 0-1
kingside. Cvetkovic and V. Sok I 've deliberately focused pri

olov assess the position as +. marily on Chi gorin games with
17 i.c4 litb6 two knights versus two bishops,
18 ':dgl :tb8 but probably his best known game
19 d2? in the Chigorin Defence featured a
Panov gives 1 9 Wlixd6 cxd6 20 bishop and knight against two bish
d2, intending 2 1 i.d3 , 22 lIal ops . However, the ideas he used
and the advance of the h-pawn, as were similar enough to those used
unclear (cited in MCSG). in his efforts with two knights that
19 . Wlid7 it seems appropriate to present the
20 :al a6 (D) following game.
The immediate 20 . . . ':c6? loses
the exchange after 21 i.b5 . Teichmann - Chigorin
Cambridge Springs 1 904
1 d4 d5 2 c4 tbc6 3 tbf3 i.g4 4

cxd5 i.xf3 5 dxc6 (D)
This variation gained in popu
larity after it became clear that 5
gxf3 doesn't result in a simple ad
vantage for White.
5 iLxc6
6 tbc3 e6
7 i.f4
It was later establi shed that 7
w e4 i.b4 8 f3 f5 ! ? (8 . WHh4+ is the
. .
16 :hgl 'iVe7! (D)
modern preference) 9 Jt.. c4 ! repre W

sents the most dangerous approach.
7 lLJf6 The queen prepares to join the
8 e3 Jt.. b 4 attack on White's king. The rest is a
9 'iVb3 lLJd5 massacre : 17 :dfl 'iHa3+ 18 'it>d2
10 Jt.. g3 0-0 b4! 19 c4 Jt.. a4 20 'iHbl lLJc3 21 al
11 Jt.. d3 'iHg5! ':d8 22 g3 lLJe4+ 23 'it>e2 lLJc5 24
12 'iVc2 'iHbl lLJxd3 25 'iHxd3 'iHxa2+ 26
White simply loses material af 'it>f3 Jt.. c2 0-1
ter U O-O? Jt.. xc3 1 3 bxc3 lLJxe3 . No less a player than Botvinnik
12 . f5! called attention to Teichmann-Chi
Preparing to harass the g3- gorin, although many of Black's
bishop with ... f4, while also open ideas had been used by Chi gorin as
ing f7 as a defensive post for the early as the 1 890s.
king's rook.
13 Jt.. e5 lIn Verdict: The games in this chapter
14 0-0-0 Jt.. xc3 present the ' traditional case for
15 bxc3 b5! success' of Chigorin' s handling of
The e5-bishop looks impressive, the knight pair against the bishop
but it's the only white piece of pair. However, the astute reader has
which that can be said. In the undoubtedly noted by this time that
meantime, Black has a solid grip all of these games feature Chi go
on the light squares and is ready to rin 's own defence ( l d4 d5 2 c4
blast open the pawn shelter around lLJc6) or its close relative I d4 d5 2
White 's king with an opportune lLJf3 Jt.. g4. Chigorin's wins in this
.. . b4. chapter are impressive efforts and
most of them have become widely success or strategy in the handling

anthologized standards in chess lit of the knights. In particular, the
erature. Yet it seems to me that slightly airy position of the white
these games illustrate Chigorin's king and the defensive power of the
success in the handling of the un manoeuvre . . .f7-f5 and . . . : f8-f7
usual positions that arise out of his seems to have played a large part in
defence rather than any intrinsic Chigorin's successes.
4 The Rest of the Story -
Chigorin us ually lost with
t he Knight Pair
A database scan of Chigorin's knights could occupy. Finally, in

games reveals that he had twenty the games that Chigorin lost with
four efforts in which he fought the knight pair, he usually had a
with two knights versus two bish bad game very quickly, which sug
ops. His record with this material gests that many of the games where
distribution was quite poor; all told, he ' sought' the knight pair were
he was +5 =4 - 1 5 when he had the poody thought out.
knight pair. A percentage of 29% Curiously, he managed to fight
hardly suggests that Chigorin had back from bad positions in many of
devised a sound 'rebuttal ' to Ste these games and often had at least
initz' s preference for the bishops. one winning opportunity before ul
It's always possible that my data timately losing. Chigorin may have
base is incomplete and that I've been in bad form in these games (as
overlooked a game or two where evinced by his missing some easy
Chigorin won with the knight pair, tactical wins), which may also call
but even then, his score with the into question his decision to give
'advantage' of the knight pair was up his the bishops for knights .
probably no higher than 40% . One final point should be made
Several things leap out when before we tum to a consideration of
one looks at Chigorin's losing ef Chigorin's losses . The knight pair
forts with the knights. First, he usu came to figure in his games rela
ally did not give up the bishops to tively late in his career, when he
cause pawn weaknesses in his op was already in his forties. Claims
ponent's camp, as , say, Nimzo of any special 'insight' Chigorin
witsch did. Secondly, he almost may have had as to the value of the
never had good squares waiting for knight pair must take this into ac
his knights when he surrendered count, as it is very rare for a major
the bishops, i.e. , there were no im player to make such discoveries so
portant holes or outposts that the late in his career.
48 The Rest a/ the Story - Chigorin usually lost with the Knight Pair
Bum - Chigorin advance of the white h-pawn and

Vienna 1 898 target the weak f4-pawn.
13 0-0-0 'tIif6?
1 d4 tZJf6 2 c4 c5 3 d5 d6 4 tZJc3 g6 14 h4
5 e4 iLg7 6 f4 0-0 7 tZJf3 .tg4 8 h3 One can understand Burn's de
iLxf3 9 'tIixf3 tZJbd7 (D) cision to play for a straight attack
on the king but 1 4 eS ! was also
strong, e.g., 14 . . . dxeS 15 'tIixb7
l:td8 1 6 tZJe4 ! Wif7 1 7 xa7 , threat
ening 1 8 iLa5 +- Mayer.
-
14 ... t7
Chigorin could have tried to im
prove the placing of his queen's
knight with 1 4 ... tZJe5 but White
maintains a large advantage by 1 5
h3 tZJc6 1 6 h5 . Black's attack
simply lags behind White's due to
his failure to prepare . . . b5 .
W 15 d3 iLh6
16 g5 iLg7 (D)
Chigorin's decision to give up
his light- squared bishop for the
white knight can be seen as an ex
ample of 'changing the colour of a
bishop ' (see Chapter 9). Black's
play in the centre is based on the
dark squares, so the exchange of
the bishop is strategically justified.
10 g4!? e6
11 dxe6 fxe6
12 iLd2 tZJe8?
Unaccountably passive; perhaps
Chi gorin was concerned about a W
later e4-e5 or g4-g5. A modern mas
ter would probably have played 17 h5!
1 2 . . a6, paving the way for a later
. Burn shows good positional
. . . bS. Further, an eventual g4-g5 judgement in not going after the
might be answered in some cases backward d-pawn, e.g., 17 tZJb5 ? !
by . . . tZJh5 , which would stop the e5 ! 1 8 tZJxd6 tZJxd6 1 9 xd6 exf4
The Rest of the Story - Chigorin usually lost with the Knight Pair 49
20 'iVxf4 'ii'xf4 (20 . . . 'ii'e 7 ! is an al 21 cxd5 exd5

ternative, as it provides Black with 22 exd5 %;lCe8
go od compensation for the pawn 23 J.f3 b5
in the form of play on the dark 24 l':tdhl lDfS
s quares) 2 1 .txf4 l':txf4 22 l':txd7 25 .txh5 lDxh5
%;laf8 and 23 . . . l':t4f7 , when the su 26 l':txh5 %;le4! (D)
perior minor piece and the pres
ence of opposite-coloured bishops
gives Black good drawing chances
- Mayer.
17 gxh5
18 J.e2 .txc3!?
A n audacious decision. Chigorin
surrenders his remaining bishop
with the idea of controlling the h5-
square as long as possible. White's
rooks will then try to break through
on the h-file while Chigorin will
create counterplay by opening the W
centre with . . . d5. This plan smacks
of desperation but later events will 27 'iVh3?
show that it sets White unusual The further course of play sug
problems. gests that both players ended up in
19 'ii'xc3 lDg7 severe time pressure, perhaps as
20 ':h4 d5! (D) early as this phase of the game. The
text allows Black to defend suc
cessfully, but 27 g6 ! appears to be
very strong:
a) 27 . . . lDxg6? 28 l':txh7 'ii'x h7
29 l':txh7 xh7 30 'ii'h 3+ g8 3 1
f5 ':c4+ 3 2 b l (or 3 2 .tc3 lDe5
33 b3 ! +-) 32 . . . lDe5 (32 . . . lDe7 3 3
f6 ! is crushing) 33 'ii'g 3+ lDg4 3 4
.tf4 ! + - is a neat line-breaking
theme - Mayer.
b) 27 .. :i!t'xg6 is a tougher nut to
crack:
W b l ) 28 'iVxc5 is possible, when
White should win the ending after
50 The Rest of the Story - Chigorin usually lost with the Knight Pair
2s .. J;Ic4+ 29 'iVxc4 bxc4 30 ':g5 It's clear that the position at

xg5 ? 3 1 fxg5 ':dS 32 l:th4 ':xd5 move 27 is very complicated and it
33 ':xc4 ttJe6 34 lIa4. However, seems likely that Burn rejected 27
Black's defence can be improved g6 ! on the basis of variation 'b'
in this line with the immediate above (presumably he missed Bur
30 . . . lIdS ! 3 1 :hh5 <tif7 32 ':xg6 ges s ' s 30 ':xh7+ ! in line 'b2' ) .
hxg6 ! , when Black's better king However, the fact remains that
position and centralized pieces Chigorin is now able to stir up
make it difficult for White to win - complications that are almost fa
Mayer. vourable to him - and are defi
b2) 2 S f5 ! looks strong but I nitely favourable to him in the
didn't see a cold win after 2S ...'iVd6! event of one more mistake by
29 l':tg l + <tif7 , because I thought Burn.
that White would continue with 30 27 ... ':c4+ (D)
g7+ <tieS, e.g., 3 1 'iVb7 ':bS 32
'iVxa7 1Ie7 ! , when Black hits the d
pawn and hi s queen is ready to go
on a checking spree after 33 ':e l
lIxe 1 + 3 4 ..txe 1 'iVf4+ ! . However,
Graham Burgess pointed out 30
':xh7+ ! (D), when the black posi
tion is cracked.
28 bl ?
Natural and bad ! It's under
standable that Burn thought that
this was the right move, but he
should have braved 2S <tid 1 ! , with
the point that his king is safe from
B queen checks on the bl -h7 diago
nal. Black could then try 2S . . . ':d4
For example, 30 ... ttJxh7 3 1 g7+ but White gains a superior ending
<tieS 32 'iVxh7 or 30 . . . <tieS 3 1 'id3 with 29 g6! ':xd2+ 30 <tixd2 'iVxf4+
- Burgess. 3 1 'iVe3 ! - Mayer.
The Rest a/the Story - Chigorin usually lost with the Knight Pair 51
Other tries instead of 28 . . . I:.d4 could now have won with the
are possible but the looming g5-g6 straightforward 32 l:r.h8+ cj;; f7 33
looks like a winner with the heavy I:.h7+ t2J xh7 34 'iVd7 + ! Wie7 35
pieces tripled on the h-file. lIxh7+ - Mayer.
28 'iVxd5!
29 i.c3 lI e8!
30 g6 'iHe4+
31 cj;;a 1?? (D)
This allows mate in three but 3 1
cj;; c 1 'ii'xf4+ ! 3 2 cj;;b l 'iVe4+ and
33 .. . 'iHxg6 is good for Black, as he
has managed to keep his h-pawn in
place and has emerged with a ma
terial advantage.
32 'ii'xh 1+??
Another blunder - the last of the
game. Now it was Chigorin's turn
to panic in time pressure or he surely
would have found 32 . . . lhe5 ! , with
the possibilities :
a) 33 i.xe5 cj;; f7 ! 3 4 'iHh8 ! , with
the division:
B a l ) 34 . . . 'iVc2? plays for tactical
tricks but allows White to activate
31 hxg6?? his rook, e.g., 35 'iVf6+ cj;; e 8 36 b3 !
Proof, in my view, that both I:.e4 (36 ... 'iHe4 37 I:.d l is also win
players were in extreme time pres ning) 37 i.d6 and the possibility of
sure. The mate was to be had by 38 I:.h8+ should seal Black's fate.
3 1 . . . 'ii'e l + ! ! 32 I:.xe l I:.xe l + 3 3 a2) However, after 34 . . :d5 !
.txe l I:.c l#. Black is clearly better, as he seals
32 I:. e5? (D) the d-file against %;Ih l -d 1 and ties
Burn must have realized that the white rook to the back rank.
h e'd just missed a bullet, so he b) White could also try 33 fxe5
p lays to close the e-file against but then 33 . . . 'iVf5 ! is good for
the back-rank mate. In fact, White Black - Mayer.
33 'ii'xh1 b4 A n extraordinary sacrifice that

34 'ii'dS+ 1-0 nets Black a dangerous attack in re
Games such as Burn-Chigorin turn for a queen. Incidentally, the
show the gritty nature of nine ' ! ! ? ' symbol was Marshall 's per
teenth century chess. Burn handled sonal choice in Marshall 's Best
his initial advantage well but even Games of Chess.
a modern grandmaster might ap Marshall claims that "it would
plaud the deep and practical play have been more advantageous for
begun by Chigorin with the move me to play 1 3 . . . tZJh5 1 4 gxf4 l:.g2 !
1 8 . . . i.xc3 ! ? 1 5 f5 'YfIg7 1 6 I:.xh3 'YfIg4, etc." The
problem lies in the 'etc.' It looks to
Sometimes Chi gorin found him me as if Chigorin could then play
self in difficult endings as a result 1 7 i.xf7+ ! ? cJ;; x f7 1 8 'i!Vd3 ! , with
of a middlegame decision to ex the threat 1 9 I:.h4. Meanwhile, the
change his bishops for knights. white queen guards f3 directly and
h3 indirectly (not l 8 . . . 'iVxh3 ?? 1 9
Chigorin - Marshall tZJg5+) , while 1 8 . . . I:.xe2 1 9 cJ;; x e2
Vienna 1 903 (D) and now:
1 e4 eS 2 f4 exf4 3 Jt.. c4 dS 4 Jt.. xd5

'ih4+ 5 cJ;;n g5 6 g3 h6 7 tZJc3
tZJf6 8 d4 tZJc6 9 cJ;; g2 Jt.. d7 10 h4
I:.g8 1 1 tZJf3 gxh4 12 tZJe2 h3+ 13
cJ;;n (D)
a) 1 9 . . :iVg2+? 20 cJ;; d l keeps

White more than afloat due to the
continuing indirect protection of
the rook at h3 - Mayer.
B b) Graham Burgess suggests
19 . . . tZJb4 ! ?, which aims to exploit
13 fxg3! ! ? the clumsy interaction among the
wh ite pieces. After 20 'it'c4+ (20 White is embarrassed after Mar

'i'b3 + e8 is similar, when White shall's 1 9 h2? J.xh3 ! .
s hould play 2 1 I:.xh5 , because 2 1 19 ltxg3+
l:Lh4? 'i'g2+ 22 d l J.b5 ! is good 20 xg3 lbxd3 (D)
for B lack - Burgess) 20 . . . e8 2 1
l:Lxh5 xh5 ! (2 1 . . . 'ifg2+? 22 e3 !
tLlxc2+ 23 f4 is good for White,
e.g . 23 . . . lb xa1 24 lhh7 wins on
the spot - of course 24 . . 0-0-0 is
.
not legal ! ) , when White has com

pensation for his material - Bur
gess.
It seems as if 13 . . .fxg3 ! !? might
not only be Black's most spectacu
lar move in the position but also his
best.
14 J.xh6 w
White could also have played 14
lbxg3 g6 1 5 I:.gl (but not 1 5 J.f4? 21 cxd3?
lbh5 -+), which Marshall thinks Marshall gives 2 1 .txc6 J.xc6 22
may have been superior to the text. cxd3 as minimizing Black's advan
In the tournament book for Vienna tage.
1903 , Marco gives 15 I:.g 1 , but also 21 lbb4!
offers 15 J.g5 as a possibility. 22 :0
14 g2+ Black does very well after 22
15 gl J.xh6! J.xb7 I:.b8 23 J.d5 lbxd5 24 exd5
Threatening to mate with Ilxb2, which Marshall assesses as
16 . .. J.e3+ 17 h2 lbg4+ 18 xh3 winning for Black.
lbe5+ 19 h4 lb xf3+ 20 h5 Another possibility was 22 : h l
i. g4# - Marshall. lb xd5 23 I:.xh7 , with the threat 24
16 'it'd3 lbg4 I:.h8+, which both Marco and Mar
17 I:.xh3 J.e3+ shall analysed for some length
18 xg2 (Marshall demonstrated that Black
Both Marshall and Marco sug did very well with 23 . . . e7) . How
gest 1 8 xe3 lbxe3 , with the for ever, their analyses are moot, as
mer preferring 1 9 I:.xh7 and the both men overlooked the simple
latter giving 19 J.xc6. 23 . . . lbb6, which guards the a8-
18 lbf2+ rook and wins a piece.
19 lIg3 22 lbxd5
23 exd5 e7 (D) 27 l%h1 .trs

28 lLJg3 .t g6 (D)
W
W
Hair-raising complications have
resolved themselves into an ending 29 lLJe4?
that looks terrific for Black, since Marshall makes the apt com
he has the bishop pair and White ' s ment "White must remove one of
tripled d-pawns are a real fright. the terrible bishops." Oddly, nei
However, Black can ' t just win the ther he nor Marco mention that 29
d-pawns and subsequent possi lLJxg6+ ! , which does remove one
bilities suggest that proper defence of the terrors, is the best move:
by White might well have held the a) 29 . hxg6? 30 lLJe4 .tf6
. .
balance. (30 . . .f6 3 1 :h7+ is worse) 3 1 l:.c l !

Yet aside from his positional ad c6 (3 1 . . . l:.cB? 3 2 d6+ ! is the point)
vantage, Black probably had a big 32 dxc6 bxc6 33 lLJxf6 ! (the imme
psychological advantage here. Chi diate 33 l:.xc6? .txd4 favours
gorin has defended superbly but all Black) 33 . . . xf6 34 Ilxc6+ g7
'
that he has earned for his efforts is 35 Ilc7 l';IdB 36 Ilxa7 l:.xd4 37 e3
an ending that he probably wrote is good for White - Mayer.
off as ' lost' . Thus, while Marshall b) 29 :xg6 30 .l::tx h 7 elimi
. . .
will continue to play hard 'I have

- nates the outside passed h-pawn,
to win my won game' - there is the though the value of White 's extra
risk that Chigorin will slack off and pawn is unclear.
not make the most of his chances. 29 h6
24 lLJe5 l:.g8+ 30 lLJc5 b6
25 f3 .th3 31 lLJe4 f6
26 l:.d1 .t g5 32 lLJxg6+
32 lLJc6+ 'itd7 33 lLJxa7 ':a8 fa- 37 lLJxf6 J.xd4

vo urs Black - Marshall. 38 lLJg8+ 'itd7
32 ':xg6 39 lLJxh6 .txb2
33 'itg4 J.e3+ 40 'ite4 'itc6
34 'itf5 ':g1 41 d4 b5
35 ':xg1 42 lLJf5 b4
Marshall remarks that 35 l:.h3 43 lLJe7+ 'itd7
]1n + 36 'itg6 .tg5 is "also in 44 lLJd5 a5
Black's favour". 45 'itd3 'itc6
35 .txg1 (D) 46 'itc4 .tel
47 lLJe7+ 'itd7
"The king must get to e4 to men
ace the d-pawn." - Marshall.
48 lLJd5 'ite6
49 lLJb6 J.d2 (D)
W
36 d6+
Marshall awards this an excla
mation mark and claims a win for
Black after 36 lLJxf6 J.xd4 37 lLJg8+ W
'itd6:
a) 3 8 'ite4 J.g7 ! 39 d4 h5 "and 50 'itd3?
White can resign" - Marshall. This move passes without com
b) However, it is silly to leave ment by either Marshall or Marco.
the h-pawn on the board, which That's a pity, as I would have liked
su ggests 38 lLJxh6 'itxd5 39 b3 , to know how Black can win after
When White can play with the idea 50 d5+ ! 'ite5 5 1 'itb5 ! 'ite4 52 lLJc4
of the limited number of pawns re J.c3 53 lLJxa5 'itxd5 54 lLJc6 = -
maining to give himself decent Mayer.

drawing chances - Mayer. Note that White 's drawing pos
36 . cxd6 sibilities are based on the limited
number of pawns and the activity 'had their chance' only after the
his king and knight can achieve on bishops missed theirs.
the light squares, where White has
the advantage of an extra piece. In any event, an excellent anti
50 i.c3 dote to inappropriate romanticism
51 'it>c4 i.e1 concerning Chigorin' s knights is
52 llJd5 'it>f5 provided by the rest of the games in
53 llJb6 'it>e4 the chapter.
54 llJc8
Marco claims that White could Englisch - Chigorin
have drawn with 54 llJd5 but Mar London 1883
shall points out a win for Black
with 54 . . . 1.h4 ! 55 llJc7 1.f2 56
. 1 e4 c5 2 llJc3 llJc6 3 llJf3 e6 4 1.e2
llJe6 i.e3 57 d5 (or 57 'it>b5 1.d2 ! ) llJge7 5 d4 cxd4 6 llJxd4 llJg6 7
5 7 . . . 1.d2 -+. 0-0 i.e7 8 1.e3 0-0 9 f4 1.c5 10
54 d5+ 'it>h1 1.xd4 1 1 1.xd4 f5 12 i.c5
55 'it>c5 a4 ':f7 13 e5 b6 14 i.e3 1.b7 15 llJb5
56 llJd6+ 'it>d3 ':b8 16 llJd6 (D)
57 llJb5 a3
0-1
The excitement and missed op

portunities of games such as Bum
Chi gorin and Chi gorin-Marshall
might lead one to say, 'Ah, so Chi
gorin was on the right track in
playing for the knight pair.' The
problem with this view is that
Chigorin does not appear to have
'played for' the knight pair in these
games; he got ' stuck' with the B
knights as his best chance of de
fending a pair of really bad posi 16 l:tf8 17 llJxb7 l:txb7 18 1.f3

tions. (This is especially true in the Wie7 19 g3 l:tc7 20 "tIid2 lIfc8 21 c4

case of Chigorin-Marshall.) Fur llJf8 22 b3 llJd8 23 :tfd1 llJf7 24
thermore, while both sides 'missed a4 1Id8 25 a5 bxa5 26 "tIixaS d6 27
opportunities ' in these games, Chi exd6 ':xd6 28 1.xa7 ':xd1+ 29
gori n ' s opponents missed their i.xd1 e5 30 fxe5 llJxe5 31 1.d4
opportunities first, i.e., the knights llJc6 32 d5+ 'it>h8 33 1.c3 e8
34 i.f3 c8 35 ftel h6 36 d6
h7 37 i.dS lbg6 38 i.e6 'iVb7 39
i.xfS 4:Je7+ 40 i.e4 xb3 41 'iWeS
4:JdS 42 i.xg6+ 'it>xg6 43 e4+
'it>h5 44 'iVfS+ gS 4S h3+ 'it>g6 46
Ite6+ 'it>f7 47 f5+ 'it>g8 48 :e8#
(1-0)
Englisch-Chigorin was an espe
cially disheartening example of
just how badly Chi gorin could do
with the knights. The ' analysis
junkies' among you are probably w
disappointed that I ' ve given this
game without notes but where of his space advantage on the
could they have been inserted? queenside with the pawn lever c4-
Simply put, 'Black stunk up the c5 but then the white d-pawn is
joint' . The knights never created a weak. This isn't a problem in, say,
meaningful threat, the black pawn a King's Indian, as then White
structure was awful and the black would typically have a pawn at e4
army choked on a lack of space or to support the d-pawn.
counterplay. 12 l:tel fte8
This is undoubtedly the worst of 13 i.f4 lbb6
the games examined in this chapter 14 b3 .tf5
- and one of the worst in the entire 15 'iVd2 4:Jd7
book - but it's not an orphan. 16 %lac1 4:Jf8
17 l:te2 4:Jg6
Chigorin - Janowski 18 .tg3 h5!
Monte Carlo 1 901 19 h3 h4
20 .th2 'it>h7
1 e4 e5 2 lbf3 4:Jc6 3 c3 lbf6 4 d4 21 ftcel .tg5!
lLlxe4 S d5 lbb8 6 i.d3 lbc5 7 22 'iVdl
lLlxeS lbxd3+ 8 lbxd3 i.e7 9 0-0 White loses material after 22
0-0 10 c4 d6 1l lbc3 lbd7 (D) %lxe8? .txd2 -+.
Janowski is remembered as a 22 ... %lxe2
gre at connoisseur of the bishops, 23 'iVxe2 .th6
so it seems appropriate that he The white bishop at h2 is a par
shoul d have them in this game . ticularly poor piece, because it is
White's problem is that his natural 'biting on granite' against the pawn
plan is to play to make something chain c7-d6. The advance c4-c5 ,
designed to liberate it, would still 29 \tel t2Je5

serve primarily to weaken White' s 30 .txe5 dxe5 (D)
d-pawn.
24 t2Je4 Wld7
25 WIn ':e8
26 \to ? ! (D)
An odd move. It' s under
standable that Chigorin has no in
terest in the 'tactics ' of 26 t2J f6+?
gxf6 27 lIxe 8 , as both 27 . . . .txd3
and 27 .. :i'xe8 28 Wlxf5 Wle2 ! are
clearly better for Black. Still, it
can ' t help matters much to bring
the king toward the centre in a situ
ation where the centre isn ' t thor W
oughly closed.
31 c5?
For better or worse, Chigorin
had to play for a kingside attack
with 3 1 Wlh5 ! ? .txe4 32 ':xe4 c6
(32 ... g5 33 f4! carves out play on the
kingside and on the light squares)
33 t2J f4 ! ':d8 34 Wlg6+ \tg8 ! (this
pins the g-pawn but it also ensures
that captures at h6 don't occur with
check) 35 t2J e6 (35 t2Jh5 ! ? is worth
a look but how does White make
progress against 35 . . . i.g5 , which
B guards f6 and h4?) 35 . . . cxd5 36
cxd5 ':c8 37 ':g4 (37 ':xh4 'ii'x d5
26 a5 is very good for Black) 37 . . . f5 ! 3 8
27 a4 'ixf5 Wlxd5 (D), when i t appears
Chigorin prevents a later . . . a5- that White can hold the balance:
a4 but the white pawns are now a a) 39 ':d4 'ixb3 (39 . . . Wlxg2 40
major target in any endgames, as ':c4 ! - B urge s s , allow s the rook
they are fixed on the colour of the to find a safe square with gain of
light-squared bishop. tempo; then 40 . . . ':'xc4 4 1 Wlf8+
27 b6 \th7 42 Wlf5+ leads to a perpetua l
28 Ile2 f6 check on the light squares) 40
36 'iHc4 'iVd6 (D)
W
'iYxe5 'iWa2 leaves Black for choice W
- Mayer.
b) 39 tZJd4 Ilc l + (39 .. J;Id8 may 37 tZJdl
keep a slight edge for Black - Bur Trying to make room for the
gess) 40 'it>e2 'iWc5 ! , and now not rook with 37 tZJf1 is answered by
41 tZJe6? 'iHd6 when Black wins 37 . . . f7 and the d-pawn drops .
(Mayer), but 4 1 'i'e6+, when White Notice how strong the black bish
has at least a perpetual check - ops have become, while the white
Burgess. knights have been steadily re
Variations such as these always stricted and pushed back by the ad
remind me of Larsen's epigram vancing black kingside majority.
'Long variation, wrong variation' . 37 h5
It's certainly true that an analyst 38 tZJn
will often miss things in his study White loses after 38 f3 'iWg3+ 39
that two people playing a game of 'it>f1 to both 39 . . . exf3 and the even
chess will find. However, 3 1 c5 ? is stronger 39 . . . e3 ! 40 tZJb l 'iHh2 -+.
Worse than useless, as now the d 38 .txe2
pawn is fatally weak and the black 0-1
pieces soon get their shot at the
White king without having to worry Mieses - Chigorin
about any distracting counterplay. Karlsbad J 907
31 .tg6
32 c6 'Wie7 1 e4 e5 2 tZJc3 tZJc6 3 g3 tZJf6 4
33 tZJb2 f5! .tg2 .tc5 5 d3 d6 6 tZJa4 .tb6 7
34 tZJd2 'iHc5 tZJxb6 axb6 8 tZJe2 .tg4 9 f3 .te6
35 'iHd3 e4 10 f4 .tg4 1 1 h3 .txe2 12 'iVxe2
lLJd4 13 'if2 iVd7 14 .te3 lLJe6?

(better is the immediate 14 . . . lLJc6)
15 0-0 0-0 16 f5 lLJd8? (what's
wrong with 1 6 ... lLJc5?) 17 g4 lLJe8
18 g5 f6 19 h4 lLJc6 20 c3 (D)
24 lLJc7
25 l:Id2 iVe7
26 ':adl ':fd8
White has built up an impressive
B strike force on the d-file, as is
shown by the variation 26 . . . exd4 27
Chi gorin has played the opening cxd4 l:Ixa2 28 dxc5 dxc5 (28 . . .bxc5
in miserable fashion. Not only does 29 ':xd6 also favours White) 29
White have the bishop pair against l:Id7 'iVe5 30 g6 h6 3 1 ':xg7 ! , with
knights that have no good squares, a mating attack - Mayer.
but he also has a large space advan 27 'ig4 lLJb5
tage on the kings ide and chances of 28 a3 exd4
opening the centre with a later d3- 29 cxd4 c4
d4. The pawn structure on the It may have been better to elimi
kingside recalls a King 's Indian At nate one of the bishops by means of
tack. Chi gorin played this opening 29 . . . cxd4 30 .txd4 lLJxd4. How
quite well from the white side but ever, Black has the problem that his
as his play in this game demon b-pawns are very weak, so the end
strates , he was as clueless as his ings should be winning for White.
contemporaries when it came to Further, any ending is likely to oc
playing the black side. cur on an open board with pawns
20 lLJe7 on both sides, so White's bishop(s)
21 't\Vf3 rJilh8 would have a field day.
22 'ih5 lLJg8 30 ':cl d5
23 l:If2 c5 31 e5 fxe5
24 d4 (D) 32 dxe5 d4! ? (D)
C hi g orin tries for tactical com with 36 . . . WNxb2? is soon disposed

pli cations. These should favour of by 37 I:.c7 I:.d2 38 lIxh7+ !
White, but his decision is under 'it>xh7 39 I:.c7+ and White's mating
standable in view of 32 . . . WNxe5 3 3 attack arrives first - Mayer.
i.xb6, with a solid edge for White.
W
W
It should be noted that while
33 f6!? Black has succeeded in eliminating
Mieses was known a s a n attack one of White 's bishops, the other
ing player, so he has no qualms in will play a major role in the rest of
mixing it up from a superior posi the game. Meanwhile, the black
tion. It seems simpler to play 3 3 knight is doing little more than
i. xd4 ltJxd4 3 4 I:.xd4 'it'xe5 3 5 serving as the sole defender near
:'cxc4, when White ' s advantage the black king.
must be close to winning. 37 g6 hxg6
33 gxf6 38 xg6 I:.ad8?
Black is in difficulties after Chigorin could have rounded
3
3 . . . xe5 34 i.f4 'it'c5 3 5 fxg7+ off his defensive strategy with
rJi xg7 3 6 b4 ! , gaining control of 38 .. J I c8 ! ! , with the point that 39
the e5 -square for a bishop check - I:.xc8 WNd4+ gives him a perpetual
Mayer. check on the dark squares. White
34 i.xd4 ltJxd4 isn't obliged to capture the rook
35 I:.xd4 WNxe5 but then a single Black rook serves
36 I:.dxc4 I:.d2! (D) to tie down both White rooks, so
Chigorin realizes that his best Black's defensive chances are im
C han ce of defence lays in centrali proved greatly - Mayer.
z ation and activity. Pawn grabbing 39 I:.e4?
This game reminds me of Bum 39 litd1+

Chigorin, which we saw earlier in 40 J.f1? 'iVb5?
the chapter. In both cases, Chigo The continuing mistakes sug
rin's opponent mistakenly chose to gest time-pressure. Black is better
challenge a heavy piece on an open after the simple 40 . . .'iVxe4 4 1 xe4
centre file rather than to pursue his l:.xc l , as the bishop at fl can ' t es
own attack. cape from the plan of doubling on
Instead, Mieses could have set the eighth rank.
Black insoluble problems with 39 41 l:.cc4 d7
l:.g4 ! lId l + 40 lixd l l:.xd l + 41 42 liVh5+ liVh7
J.fl e3+ 42 'it>h2 'iVf2+ 43 J.g2 43 f3
'iVg l + 44 'it>h3 e3+ 45 l:.g3 e6+ Mieses really must have wanted
46 'it>h2 (D). to win by direct attack; the ending
after 43 xh7+ 'it>xh7 44 litc7+
should be winning for White, be
cause all the black pawns are weak
and White's pieces are far more ac
tive.
43 l:.1d7
44 l:.g4 l:.g7
45 l:.xg7 "iWxg7+
46 l:.g4 liVh6
47 e4 :f8
48 l:.g6 c1
49 'iVg4 (D)
B Here 49 'iVxb7 looks very good
for White, as it wins a pawn and
Now the safer king position threatens mate. Mieses's disdain
wins, e . g . , 46 . . . lid7 47 J.e4 ! +-, for grabbing material in this game
when a defence based on exchang suggests that he assumed that the
ing loses after 47 . . . f7 48 xf7 weak black pawns 'wouldn' t run
lIxf7 49 J. d5 , followed by mass away ' .
exchanges at g7/g8, as the outside 49 h6
passed pawn decides for White. 50 liVh5 'it>h7
Similarly, direct attack fails after 51 l:.g3 c5+
47 . . . l:.d2+ 48 'it>gl litd l + 49 'it>f2 52 xc5 bxc5
%Id2+ 50 'it>el +- and the white 53 J.d3+? !
king sends ' kindest regards' to his Still playing for attack ! It seems
colleague at h8 - Mayer. simpler to play 53 J.g2 b6 54 lib3
B W
l:b8 55 a4 + - . Note that the g2- i.xb7 b3 61 i.d5+ c4 62 h5

bishop serves to restrict not only xb2 63 a4 c3 64 i.e4 f5 65 i.d3
the rook - Black can never play c1 66 e3 lbg4+ 67 d4 f4 68
. . . l:a8 - but also the knight, i.e., xc3 1-0
55 . . . lbf5? 56 i.e4 wins a piece.
53 h8 Verdict: Neither Chigorin's record
54 I:.g6 I:.g8 with two knights vs two bishops
55 I:.xg8+ xg8 (D) nor the aesthetic quality of the re
But in the end, it really doesn ' t spective games suggest that he had
matter, a s White has reached a win worked out any systematic method
ning ending: his pieces are more of playing with knights against
active and he has an outside passed bishops. The onus is now on histo
pawn that will tie down at least one rians and researchers to provide any
black piece. The rest requires no proof that Chi gorin - or any of his
comment: 56 g2 f7 57 f3 major contemporaries - claimed
'itte6 58 f4 d5 59 i.e4+ c4 60 that he had.
5 The Pro b le m Knight
Minor pieces are particularly prone compared to 7 0-0 a6 8 d5 liJa5 9

to ending up in passive situations. liJd2 c5 , etc.
This is not at all uncommon; in
deed, the minor pieces start the
game in passive positions. However,
once the middlegame is reached,
one really does expect to have rea
sonably active minor pieces . They
occasionally need to be repositioned
as the pawn structure changes, but
it is typically feasible to find an ac
tive post for a minor piece that has
strayed into passivity.
Unfortunately, there are times
when a knight or bishop becomes B
irrevocably passive. I call such poor
unfortunates Problem Pieces. In 7 liJa5
this chapter, we examine Problem 8 liJd2 c5
Knights, while the next chapter Geller is also inaccurate, be
will feature the Problem Bishop. cause 8 . . c6 ! opens the c-file and
.
usually allows the knight to play to

Not surprisingly, a knight on the c4 later.
rim can end up as a Problem Instead, Oeller chooses to play
Knight. The following game is a 8 .. c5, which is a standard transpo
.
particularly striking example. sition from the Panno Variation to

the Yugoslav Variation. While it is
Botvinnik Geller
-
an accepted part of opening theory,
USSR Ch (Moscow) 1 952 it is critical for Black to remember
that the a5-knight is poorly placed
1 d4 liJf6 2 c4 g6 3 g3 JLg7 4 i.g2 strategically and only vigorous play
0-0 5 liJc3 d6 6 liJf3 liJc6 7 d5 (D) and tactics can justify its place-
It was discovered later that this ment.
move-order is slightly inaccurate 9 c2 a6
Th e Problem Knight 65
10 0-0 i.f5
White has more space in the
centre, so it's a mistake to provoke
the advance e2-e4, which he would
have played momentarily in any
event. Preferable is 1 O . . . .:b8 and
1 1 ...b5 .
It's now time to elaborate on
Black's knight on a5 . In the given
pawn structure, it is certainly badly
placed. Black needs to change the
pawn structure to justify the W
knight's placement on the rim. This
can be done by advancing the b White ' s knight heads for e3,
pawn, which serves to undermine where it will overprotect the im
the white centre. By spending two portant c4-pawn and will also have
tempi ( . . . i.f5 and . . . i.d7) before chances of leaping into an attack
making this advance, Geller gives on the kingside.
Botvinnik extra time to consolidate Black's problem is that he has
his space advantage, after which nothing to look forward to: White
the a5-knight really is a Problem has more space and can continue to
Knight. gain space with an eventual f2-f4
11 e4 i.d7 and e4-e5, while the knight at a5
12 b3 b5 has little chance for activity and
13 i.b2 bxc4 may even get trapped. under some
14 bxc4 ':b8 circumstances .
15 lIab1 lIb4 17 Wic7
16 a3 ':b8?(D) 18 ltJe3 ':b7
Geller lets slip his chance for 19 i.c3 ':fb8
any kind of fun. In Botvinnik 's Best 20 h3 ':xb1
Games 1 94 7- 1 970, Botvinnik sug 21 ':'xb1 ':xb1+
gests 16 .. Jhc4 1 7 ltJxc4 ltJxc4 1 8 22 ltJxb1 (D)
i. e l as better. Black' s knight is Botvinnik's piece arrangement
then active and his queenside play ensures that the a5-knight remains
should not be underestimated. a spectator. Inferior was 22 'ii'xbl
After the text, Botvinnik does a i.a4, when the knight might later
masterful job of locking out the play into b3 .
knight. Botvinnik suggests that Black's
17 ltJd1 ! only chance from here on "was to
66 The Problem Knight
attack: 27 liJe8 28 'iVc3 f6 29 e6

iLa4 30 iLe4 liJg7 (White wins a

piece after 30 . . . liJb3 3 1 'i'b2 aS
32 liJxb3 b6 { 32 ... e 1 33 d2 }
33 iLc2 - Botvinnik) 31 iLd3 liJh5
32 iLe2 liJg7 33 h4 liJe8 34 iLd3
liJg7 35 f5 liJb3 ("This loses a
piece, but it makes no real differ
ence." - Botvinnik) 36 'iVb2 a5 37
liJxb3 'iVe1 38 liJg2 d1 39 liJc1
gxf5 40 iLe2 WHd4 41 'it'xd4 1-0
B
This proved to be Botvinnik's
play . . . eS at the appropriate mo sole win from Geller. Excluding
ment so as to slow down the devel draws, Geller won their lifetime se
opment of White's breakthrough." ries 4- 1 .
22 6
23 liJd2 liJe8 Of course, the PannolYugoslav
24 iLxg7 liJxg7 is playable, but the failure to justify
25 'iith2 liJe8 the inferior placement of his knight
26 f4 liJf6 is comparable to Black 's failure to
27 e5 (D) justify his bad queen' s bishop in a
French Defence.
Another accepted approach in
which Black takes on a Problem
Knight at as is in certain lines of
the Chigorin Variation of the Ruy
Lopez, i.e . , 1 e4 eS 2 liJf3 liJc6 3
iLbS a6 4 iLa4 liJf6 S 0-0 iLe7 6 lite 1
bS 7 iLb3 d6 8 c3 0-0 9 h3 liJaS 1 0
iL c 2 cS 1 1 d 4 WHc7 1 2 liJbd2 liJc6
1 3 dS liJaS 14 b3 (D)
Now if Black is fortunate, he
will be able to activate his knight
B
through c4 or possibly by . . . c4,
. . . liJb7 and . . . liJcS . If he fails to do
The position is now a technical so, he will face many of the same
win, as there is no good method of problems as we saw in Botvinnik
stopping White's looming kingside Geller.
The Problem Knight 67
B W
Black's queen's knight can also U f3
become a Problem Knight in cer There was a time when White
tain lines of the Modem Benoni. In would play 1 1 f4 in this position,
the following game, Black does a but the text is now considered more
good job of solving the problem, logical, as it overprotects the key
but then goes astray and is out e4- square and frees White's c3 -
played in the tactical complica knight to take at b5 if necessary.
tions . 11 liJc7
12 a4 b6
Ivanchuk - Manor 13 liJc4 i.a6
Groningen, European Junior Ch Black's queen's bishop has few
1 986 active possibilities, so Manor pre
pares to exchange it for Ivanchuk's
1 d4 liJf6 2 c4 c5 3 liJf3 g6 4 liJc3 well-placed knight at c4.
i.g7 5 d5 0-0 6 e4 d6 7 i.e2 e6 8 0-0 14 i.g5 h6
exd5 9 cxd5 :e8 10 liJd2 liJa6 (D) 15 i.e3 i. xc4
Black's queen's knight heads for 16 i.xc4 a6
c7, where it will aid his queenside 17 'ifd2 h7
pawn majority in advancing. The 18 I:.abl :b8
drawback to this approach is that 19 b4
the knight will have few active pos In Informator43, Ivanchuk rec
sibilities from c7 ; in theory, it may ommends 1 9 Wld3 ! ? b5 20 axb5
eventually reach the juicy square at axb5 2 1 i.xb5 liJxb5 22 liJxb5
d4 via b5, but White would have to liJxd5 23 Wlxd5 (23 exd5 Wld7 ! is
play very inaccurately for that to clearly better for Black - Ivan
happen. chuk) 23 . . . I:.xb5 24 'ilixf7 t.
The text is intended to leave the skilful manoeuvring, Manor is able

black knight at c7 without any to improve the knight's placement.
good squares. Black succeeds in 23 11bd1 lLlce8! (D)
advancing his queenside pawns
and is even able to create a pro
tected passed pawn, but his queen's
knight will continue to be a Prob
lem Piece. Interestingly, Ivanchuk
marks 19 b4 as ' ? ! ' .
19 b5 (D)
W
It's taken some work, but Manor
has succeeded in rearranging his
knights so that they each have a
good square. Now the black king's
knight can play to d7 , which will
W free the f6-square for the queen's
knight.
20 J.e2 24 lIfe1 lLld7
This is the first new move of the 25 J.xg7 lLlxg7
game. Earlier efforts had seen the 26 f4 f5
alternative 20 axb5 , but it is in 27 J.f3 lIbe8?
White's best interests to keep the Manor maintains the central ten
queenside closed, because that sion a tad too long. Ivanchuk gives
makes it more difficult for Black to 27 .. .fxe4 ! as superior, when he as
queen his c-pawn. sesses both 28 i.xe4 1Ibe8 29 c2
20 c4 lLlf5 and 28 lLlxe4 WHf8 ! as +.
21 J.d4 lIe7! 28 e5! dxe5
22 a5 8! 29 d6 'Jlt7 (D)
Manor's play in this game is im White also wins after 29 . . . 11 e6
pressive. Ivanchuk has succeeded 30 J.d5 ! 'Jlxd6 (or 30 . . . 11 f6 3 1 J.c6
in leaving the black queen's knight +-) 3 1 J.g8+ +- Ivanchuk.
-
with no active possibilities, but by 30 J.b7!

35 f2+
36 'it>f1 (D)
'+-' - Ivanchuk. Black must

deal with the passed d-pawn, but he B
also faces the problem that his
fixed queenside pawns will be re 36 ltJf6
moved one after another by White. Or 36 ... ltJfS 37 xh8+ 'it>xh8 3 8
30 exf4 'it>xf2 ltJxd6+ 39 'it>g l - Ivanchuk,
Ivanchuk also considers the line when the connected passed pawns
30 . . . ltJe6 ! ? 3 1 fxeS ! ltJxeS 32 .txa6 carry the day for White.
ltJd3 33 .txbS 'iVd4+, and now in 37 d7! d8
stead of Ivanchuk' s 34 'it>h l , 34 It's too late for pyrotechnics ,
'ie3 ! is clearly winning, as White for example, 37 . . . ltJe6 ! ? 3 8 'iVeS !
offers a queen swap and hits the ltJg4 39 'iVxe6 ! ltJxh2+ 40 'it>e2
rook at e8 - Mayer. fl'iV+ 41 ':'xfl ':'xfl 42 'iVe7+ with
31 .txa6 ':'e3 a winning position for White -
32 ':'xe3 fxe3 Ivanchuk.
33 xe3 f4 38 xf2 ltJf5
34 d4 f3 39 6! ltJe3+
35 .txb5! 40 'it>g1 a8
Ivanchuk just goes ahead and Ivanchuk queries this and indi
takes everything that isn't nailed cates that it was a result of time
down. After 3S gxf3 a8 ! 36 .txbS pressure. However, White still wins
'it'xf3 , Ivanchuk assesses the posi easily even after the nominally bet
tion as unclear, although I suspect ter 40 . . . 'iVxb6 4 1 axb6 ltJxd 1 42
that White's fleet of passed pawns d8 ltJxc3 43 .txc4 ltJfdS 44 h3 -
still leaves him very much for Ivanchuk.
choice. 41 xe3 1-0
In the previous two games, the square at f4, while if the paw n
Problem Knight arose deliberately stays put at g5 , then White chops
in exchange for other possibili and the square g5 is weakened.
ties. It's also possible for the Prob 7 lZ'lf6
lem Knight to arise almost by 8 a3 d6
accident, as the following game 9 lZ'lgl .1i.. g7
shows. 10 WNc2 0-0
11 h3
Lucena - Browne Lucena finally brought his
Brasilia 1 981 knight back to gl anyhow, but he
can ' t bring himself to play 1 1 h4.
1 c4 e5 2 lZ'lc3 lZ'lc6 3 lZ'lf3 f5 4 d4 The text may allow White to play
e4 5 lZ'lg5 h6 6 lZ'lh3 g5 (D) g2-g4 in some cases, but in the
game it just never happens.
White hasn' t solved the problem
of his king' s knight. It's true that
it's back from the rim, but in the
meantime, Black has gained so
much space on the kingside that it
is difficult for White to find an ac
tive square for the knight.
11 lZ'le7
12 .1i.. d 2 c6
13 lZ'lge2 .1i.. e6
14 lZ'lcl (D)
W
White has played a provocative
opening in which he hopes that the
black kingside pawns will prove
overextended. On the downside,
his king's knight has ended up off
side in the process and getting it
back into play is difficult.
7 e3
It's probably best to play the
immediate 7 lZ'l g l , followed by 8
h4. If then Black pushes . . . g4, the B
white knight can find a good
A continuing side-effect of the positions are favourable for knights,

kni ght's original poor placement. then I don't know what will.
It's extraordinary that six tempi 20 tDg6
have been spent bringing the 21 libl f4
knight to c 1 . 22 tDc3 tDh4
14 e8 23 Ilg1 tDxg2+
15 .te2 ti Black also keeps a nice advan-
16 ':0 tage after 23 .. .f3 .
Lucena doesn' t dare castle king 24 Ilxg2 f3
side, as Brow ne has a ready-made 25 ':h2 fxe2
.
attack there with . . . f4. Still, f2 re 26 tD1xe2 d5
quires protection, so the rook slides 27 c5 tDh5
over to f1 . 28 'it'd1 c7 (D)
16 ':ac8
17 'it'a4 a6
18 c2
Another two moves are ex-
pended in vain ('it'c2-a4-c2).
18 'it'h8
19 b3 b5
20 tD3a2 (D)
w
Now Black crashes through on
the kingside and wins a couple of
pawns. The rest is given for the re
cord:
29 ':h1 ':xf2 30 'it'c1 Ilh2 31
'it'b2 ':xh3 32 lixh3 .txh3 33 ':h1
B .tg4 34 'it'd1 tDg3 35 ':gl tDxe2
36 tDxe2 .txe2 37 xe2 lU8 38
The white knights have to be ':0 ':xO 39 'it'xo d7 40 'it'c3
among the more pathetic in chess 'it'h7 41 'it'b4 'it'g6 42 'it'a5 'iVc8 43
history. If this game doesn't con 'it'b6 .tf6 44 'it'a7 h5 45 .ta5 h4
vince the reader that not all closed 46 .tel h3 47 .tg3 g4 48 .th2
'iVf3 49 'iVg1 g4 50 .tg3 'iHg2 51 now the knight at e8 is out of play

'iWxg2 hxg2 52 .tf2 .th4 53 i.g1 and will have difficulty getting
.tel 54 'it>xa6 g3 0-1 back into the game. The manoeu
vre . . . g6 and . . . lLl g7 is the most
Sometimes a Problem Knight likely future for the knight, but it is
arises when a knight is poorly also time consuming and weaken
placed and rather than its situation ing.
improving, it worsens. The follow I think it would be premature to
ing game is a good example. say that the knight at e8 is a Prob
lem Knight. However, it is clearly
Geller - Mikhalchishin on the verge of becoming one and
USSR Ch (Tbilisi) 1978 Black must be careful that it has a
future ahead of it.
1 e4 c5 2 lLlf3 d6 3 d4 cxd4 4 13 .te3 lLlc6
lLlxd4 lLlf6 5 lLlc3 a6 6 .te2 e6 7 14 'iHe2 g6
0-0 .te7 8 f4 0-0 9 'it>h1 'iVc7 10 a4 15 I:.ad1 dxe5?!
b6? ! (D) This is quite bad. Mikhalchishin
probably wanted to make sure that
White ' s e-pawn would provide a
target for later counterplay, but
now e4 remains available for the
white knight and f6 becomes a hole
on an open file. Instead, I s . . . lLlxd4
1 6 .txd4 i.xf3 and 1 7 . . . dS would
have left the position closed and
perhaps Black would have had
time to organize a defence.
16 fxe5 lLlg7
17 lLlxc6! .txc6
W 18 'iHc4! (D)
There's an old saying concern
This move is an inaccuracy for a ing the Sicilian : 'White wins the
variety of reasons. Superior was middlegame s ; Black wins the
1 O . lLlc6.
. .
endgames.' Here Geller is eager to
11 .tf3 i.b7 liquidate into an ending, because
12 e5! lLle8 he has a plan for dominating the
The weakness of d6 forces this knight and breaking into the black
move if Black wishes to avoid the position.
line-opening 1 2 . . . dxeS. However, 18 I:.fc8
B B
This leaves f7 exposed, but the tZJe4 ':xc2 27 tZJf6+ i.xf6 28 exf6
queen's rook must guard the a- ':xb2 29 fxg7 f5 30 ':f3 fxg4 31
pawn. l':tf4 l':tc8 32 Jt.. c5 1-0
19 Jt.. xc6 'it'xc6
20 'ii'xc6 l'hc6 Verdict: Problem Knights can arise
21 ':d7 .'I:le8 whenever a knight gets moored on
22 g4! (D) the edge of the board. Its owner
And that's it: Mikhalchishin should be alert to solving the prob
could already start packing his be lem if he can. Attention should also
longings, as his knight will never be paid to improving the position
be able to do anything useful. The of poorly placed knights that are on
game concluded: 22 ... h5 23 h3 their way to becoming Problem
hxg4 24 hxg4 b5 25 axb5 axb5 26 Knights.
6 The Pro b le m Bishop
Problem B ishops arise less fre Bronstein A. Zaitsev

-
quently than Problem Knights. I Berlin 1968

think this is a reflection of the way
the two pieces move. A knight, af 1 e4 c5 2 t2Jc3 e6 3 t2Jf3 t2Jc6 4
ter all, has its movements restricted Jt.. b 5 t2Jd4 5 Jt.. d3!? lL'lxf3+ 6 WUxf3
toward the side of the board. A Jt.. d6?! (D)
bishop does also, of course, but it
continues to reach a lot of squares.
Further, the bishop can usually
reach two diagonals ; coupled with
its rapid movement, this usually al
lows the bishop to escape from the
fate of Problem Piece. However,
this is not always possible, as we
shall see in this chapter. When a
bishop does end up as a problem
piece, it' s nearly always on or to
ward the side of the board.
Incidentally, I think a distinction w
is useful between Problem Bishop
and B ad Bishop . A Bad Bishop In The Sorce rer's Apprentice,
may or may not be a 'Problem' , Bronstein allows this move to pass
while a Problem Bishop can be, without meaningful comment, but
technically speaking, the Good it strikes me as a definite mistake.
Bishop - there are good 'bad' bish It's unclear what Zaitsev hoped to
ops and bad ' good' bishops. How accomplish; perhaps he wanted to
ever, this chapter should be read in show that he could match Bron
conjunction with the chapter on stein in 'eccentricity' . In any event,
bad bishops, as this will afford the the immediate 6 . . . e5 was more rea
reader the broadest insight into sonable, as was the flexible 6 . . . d6.
bishops that encounter troubles as 7 'iHe3! e5
they go about their normal busi Bronstein ' s last move made
ness. 7 . . . t2J e7 unattractive, since 8 t2J b5
The Problem Bishop 75
would follow. Black also wants to 12 Wixf6 lbxf6

avoid 7 . "ciJf6?? 8 e5 , so he resorts 13 a3! (D)
to the text. However, in Bronstein's
view, he should have continued
with 7 . . . b6 8 f4 ! .te7 ! 9 e5 d5 ,
when the position resembles an
odd French Defence.
8 lbb5 Wie7
Black loses after 8 . . . .te7 ? 9
'ii'g 3 ! a6 1 0 'Viiix g7 .tf6 1 1 lb d6+
<j;;e 7 1 2 lb f5+ - Bronstein.
9 b4! (D)
Black's queenside is a mess, so

Bronstein hurries to open lines
for an attack. Further, the black a
pawn will now be pinned (or about
to be pinned) for the remainder of
the game, so Zaitsev will have no
good way of evicting the knight
from b5.
B Zaitsev ' s king ' s bishop i s not
his only problem, for he must also
A wonderful move with over be careful that White doesn 't ac
tones of the Sicilian Wing Gambit. quire the dark-square bishop for a
Black's c-pawn is deflected from the knight, in which case his own dark
centre and the a3-f8 diagonal be squares will be very weak.
comes a potential factor in White's 13 d5
initiative. 14 f3 ':g8
9 . cxb4 15 g3 .th3
10 'ii'g3 .t b8 16 exd5 .tg2
Believe it or not, the bishop just 17 ':gl .txf3
made its last move. Bronstein 's 18 d6!
further play is a marvel of ingenu Ensuring that the bishop will re
ity and positional sophistication. main out of play and also setting up
11 'Viiixg7 Wif6 a potential lbb5-c7 manoeuvre. It
76 The Problem Bishop
has also become clear that Black 31 ':xeS ':xc4+! (D)

will find it difficult, if not impossi
ble, ever to bring his queen's rook
into play. An extra bishop and rook
is a lot to cede to Bronstein !
18 bxa3
19 lIn e4
20 J.e2 IlgS
21 c4! a6
22 Ilxa3! (D)
W
32 'it>dS!
It would be a horrible blunder to
play 32 'it> xc4??, as then 32 . .. axb5+
leaves White fighting for a draw.
32 ':cS+
33 'it>e6 lixeS+
34 'it>xeS 'it>d7
B 3S 'it>dS as
36 lita4 hS
Bronstein makes the pin on the 37 ':f4 1-0
a-pawn official. White is ready to queen the d
22 J.xe2 pawn and/or mate the black king.
23 'it>xe2 tLJd7 What a beautiful game ! Bron
24 d4! stein relates that German Grand
Either the d-pawn will support master Lothar Schmid called it a
c4-c5 or Zaitsev will have to acti "work of art" and said that it would
vate the white king. be "his favourite for many years to
24 exd3+ come".
2S 'it>xd3 :'cS Bronstein makes the curious
26 i.f4 f6 comment "White's main achieve
27 'it>d4 b6 ment was that the rook on a8 was
28 Ile1+ 'it>d8 not able to play a part in the game
29 lIe6 tLJeS at all." It seems to me that this puts
30 J.xeS! fxeS+ the cart before the horse, as the
ro ok would never have ended up

out of play if Bronstein hadn' t
managed to lock away the bishop.
Perhaps motivated by his expe

rience in the preceding game,
Bronstein played an even more im
pressive game a few years later. In
the Zaitsev game, the Problem
Bishop arose almost by chance. In
the following game, Bronstein
finds an astonishing pawn sacrifice B
to create a Problem Bishop.
1 1 g5 tZJd7
Bronstein - Beliavsky Black has an extra pawn and
Erevan 1 975 faces no immediate danger. On the
downside, he has some difficulty
1 e4 c6 2 d4 d5 3 tZJc3 dxe4 4 developing his kingside and his
tZJxe4 f5 5 tZJc5! ? b6 6 g4!? bishop is buried behind its own
White launches a plan o f attack pawns. Of course, Black can al
directed against Black's queen's ways liberate the bishop by return
bishop. The danger in this approach ing the pawn with . . . f4 at some
is that White's kingside pawns point, but then White is simply bet
could end up weak and exposed. ter, as he has a pawn in the centre
6 Jt.. g6 and pressure on the f-file. Beliav
7 f4 e6 sky is certainly to be commended
8 e2 Jt.. e7 for his refusal to accept an inferior
9 h4 h5 game without a fight.
10 f5! ! ? (D) 12 tZJb3 Wic7
This may be the single most 13 tZJh3 0-0-0
imaginative move in this book. 14 Jt.. f4 d6 (D)
Bronstein sacrifices a pawn so that 15 2!
the bishop can be turned into a Bronstein fights for control of
Problem Piece. It could well be the f4-square. A secret of the posi
that 10 f5 is unsound, but that re tion is that White can afford to go
ally doesn't matter: one must take into an ending, because a poorly
risks to play creative chess - and placed piece's relative importance
winning chess ! increases with fewer pieces on the
10 ... exf5 board.
with the bishop imprisoned at g6.

However, Beliavsky sees matters
differently and eventually pursues
an exchange of all the rooks.
22 c3 I:.heS
23 4Jc5 4JcS
24 4Jd3 4Jd6
25 i.b3 I:.e3
26 4Jf4 .l::tdeS
Black has gained some piece ac
tivity and it's difficult for White to
W make any progress without assis
tance from the opponent.
15 4Jf8 27 I:.hg1 I:.Se7
16 0-0-0 4Je6 2S l::tdfl 4Je4
17 i.xd6 Ibd6 29 .td1 r:i;d6
1S .tc4 4Je7 30 .tf3! (D)
19 4Jf4 4Jxf4
20 'ii'xf4 I:.ddS
21 'ifxc7+ r:i;xc7 (D)
Ambitious play ! White could

W capture at hS, but then Black would
be able to exchange off his Prob
The ending has arrived. It seems lem Bishop.
to me that Black should sit tight 30 ... c5
and avoid the exchange of rooks, as I ' m not sure why Beliavsky did
he really doesn' t want to get in this. His rooks are already active,
volved in a minor piece ending so it isn't to open files for them. It's
true that his king now reaches the 40 d3 c7

fourth rank, but it can ' t penetrate 41 d4 lbc8
further and White will now find it 42 b4 axb4
easier to create a passed pawn on 43 cxb4 lbe7
the queenside. 44 as f6
31 dxcS+ xcS Beliavsky was presumably loath
32 lbg2 lId3 to do this when rooks were on the
33 lbf4 (D) board, because it would have given
White a variety of targets.
45 gxf6 gxf6
46 cS Jt.. e8
47 bS b8
48 b6 1-0
Porreca - Bronstein
Belgrade 1 954
33 ':d8?!
In hindsight, it seems clear that
Black should have played 33 . . . lIe3,
when I don' t see how White can
make progress.
34 lIdl ! lIed7
35 llxd7 ':xd7 B
36 lIdl ! ':xdl+
37 xdl lbd6 There is a difference between a
The minor-piece ending has ar Problem Bishop and a bishop that
rived and White is very happy ! plays to an unusual square. In the
Black still hasn ' t solved the plight diagrammed position, Bronstein
of his bishop ; after White gains the was concerned that the natural
considerably more active king, it is 1 1 . . .e6 would invite a variety of
as if he is two pieces ahead. piece sacrifices against e6. He pro
38 c2 as duced a wonderful defence with . . .
39 a4 b6 1 1 ... Jt.. g8 ! !
Now f7 is guarded, e6 will be White to exchanging bishop for

well-defended and Black has time knight. In ECO C (First Edition),
to castle queenside. Porreca didn' t Keres gives ' 8 d4 ! ?' as an alterna
achieve anything special after 12 tive.
4Jd3 e6 1 3 i.f4 i.d6 14 i.xd6 8 ... h6
'iHxd6 1S 4JfS 'iHf8 16 'iHf3 0-0-0 9 Jt.. h 4?!
17 4Jg3 i.h7! This natural move is a mistake,
Black has solved his immediate as the bishop will soon run the risk
problems and went on to win the of getting stuck in a cul-de-sac of
ending eventually. pawns.
9 ... cS!
Problem bishops are sometimes Black must make sure that the
caused by pawn avalanches. The centre can't be opened if he 's going
manoeuvre in the following game to bury the bishop at g3 . Winter
is a useful one to know and is appli could now have played 10 Jt.. x f6,
cable more often than it is used. but he failed to realize the danger.
It was also acceptable to play 1 0
Winter - Capablanca 4J d2, which would ensure that the
Hastings 1919120 bi shop could get back into play
after a subsequent f2-f3 if Black
1 e4 eS 2 4Jf3 4Jc6 3 4Jc3 4Jf6 4 played along the same lines as the
i.bS i.b4 S 0-0 0-0 6 i.xc6 dxc6 7 game.
d3 i.d6 8 i.gS (D) 10 4JdS? gS!
11 4Jxf6+ 'iVxf6
12 i.g3 Jt.. g4! (D)
This is perhaps a bit dubious, as W

it either loses time or commits
Passing sentence on White's cxb3 l:xb5 26 1:.a4 1:.xb3 27 d4

bishop. After the looming ex ':b5 28 1:.c4 lib4 29 1hc6 ':xd4 0 1 -
change at f3 , the f2-pawn will be This game is undoubtedly a tri

unable to move, so the bishop will fle, but a very instructive one.
have no way back into the game.
Keres evaluates this position as It's also possible for the Prob
clearly better for Black. lem Bishop to be only one feature
13 h3 i.xf3 of a position. A later game of Capa
14 xf3 xf3 blanca's saw him bury an oppo
15 gxf3 f6! nent's bishop but still encounter
Useful prophylaxis against any difficulties.
subsequent d3-d4 breaks.
16 'it'g2 a5 Capablanca - Bogoljubow
17 a4 'it'f7 London 1 922
18 1:.hl 'it'e6
19 h4 litb8 (D) 1 e4 e5 2 liJf3 liJc6 3 i.b5 a6 4
Jt.a4 liJf6 5 0-0 Jt.e7 6 1:.el b5 7
Jt.b3 d6 8 c3 0-0 9 d4 exd4 (D)
If Black wishes to adopt the line
used by Bogoljubow in this game,
it is better to play 9 . . Jt.g4 10 Jt.e3
.
exd4, etc.
W
The sorry plight of the g3-
bishop restricts it from doing any
thing on the queenside, so Capa
takes play to the part of the board
where he effectively has an extra W
piece. The rest requires no real
comment: 10 cxd4 i.g4
20 hxg5 hxg5 21 b3 c6 22 ':a2 Fighting for control of the dark
b5 23 lIhal c4 24 axb5 cxb3 25 squares, as White ' s d4 is a hole.
White should now continue with

Lasker's I l lLlc3 ! , when Matano
vic considers 1 1 . . . .txf3 12 gxO
lLla5 ;!; best play in ECO C (First
Edition).
11 .te3 lLla5
12 .te2 lLle4
13 .tel e5
14 b3 lLla5
15 .tb2
Best is 1 5 dS ! t. The point is that
White wants to play d4-d5 , but it's B
not clear if his bishop should go to
b2 or stay on the c l -h6 diagonal. kingside pawn storm. Superior was
15 lLle6 2 1 .. . .txf3 , with an unclear game.
16 d5 lLlb4 It's possible that Bogoljubow re
17 lLlbd2 lLlxe2 alized the danger connected with
18 WNxe2 ':e8 2 1 . . . .thS ? ! , but believed that the
The positional jockeying has re white position would be compro
sulted in a pawn formation typical mised (see note to Black' s 26th
of the Modern Benoni. Black has a move).
queenside majority, while White 22 tD3d2!
has a central majority that may The bishop won ' t get a second
translate into a passed d-pawn or a chance to chop at O . White will
kingside attack. Matanovic consid now try to bury the bishop with his
ers the position equal. kingside pawns, but its inability to
19 WNd3 h6 escape means that Capa doesn ' t
20 lLlf1 lLld7 have to rush the advance.
21 h3 (D) 22 .tf6!?
Putting the question to the bish A fighting move. Black could
op. If White were to play 2 1 lLl3d2 have played 22 . . . f6 ! ?, ensuring the
(aiming to hem in the bishop by 22 bishop ' s retreat along the hS-e8 di
h3 , 23 g4, etc.), then Black would agonal, but weakens e6 and does
just move the d7-knight and pre nothing to address White's space
pare to retreat along the c8-h3 di advantage.
agonal. 23 .txf6 WNxf6
21 . .th5? ! 24 a4! e4!
This is probably bad, as the 25 bxe4 lLle5
bishop can now be harassed by a 26 WNe3 bxa4
27 f4! (D) and even blockade the white pawns

for some time to come. White's
space advantage on the kingside
indicates that he should attack
there, but that appears to involve
the double-edged g2-g4, h2-h4 and
g4-g5 . Does White want to expose
his king that way?
Further, Black can generate ac
tivity on the queenside. His knight,
incidentally, is a tremendous piece
and the front a-pawn is a dangerous
passed pawn rather than a weak
B ness.
So who's better? For many
Capablancajudges that the time years, I thought that White ' must
is right to start his kingside opera be' better, but I'm not sure any
tions . The black pawn at a4 will more. As we shall soon see, con
provide Black with a lot of coun crete lines always seem acceptable
terplay. or even superior for Black. And re
27 Wie7 member: Capa was at the height of
2S g4 .tg6 his powers at the time of this game;
29 f5 .th7 if he couldn' t find a convincing
The bishop is now buried. It can way of demonstrating White's ' ad
be activated later by . . . f7 f6, but the
- vantage' ; perhaps that's because
weakening of e6 will be serious. there isn ' t any.
This game is perhaps the most 30 tbg3 Wie5
difficult to understand of any in the 31 g2 nabS
book. A quick look suggests that 32 Ilabl (D)
White should be much better, be 32 f6
cause the bishop is buried alive and In The Immortal Games of Ca
White seems to have the better pablanca, Reinfeld gives 3 2 . . . l:tb2
pawn structure. But the 'quick as stronger, with the idea 3 3 1Ixb2
look' must be revised after a closer Wixb2 34 llb l 'iVc2 35 f3 tbb3
look. It's true that White 's pawns 'with strong pressure' .
are in one island, but their herky 33 tbf3 ':b2+
jerky advance has left plenty of 34 ':xb2 'ii'xb2+
holes, so Black's four pawns and 35 };;te2! 'ii'b3
dark-square control can restrain 36 tbd4! Wixe3
rook and knight, yields an unclear

ending.
37 Ibe3 l:tb8
38 lIc3 'it>ti
39 'it>f3 :b2
40 llJge2 .tg8 (D)
Tartakower and Du Mont indi
cate that the further 4 1 . . . 'it>eS
would equalize, so Capablanca
must strike.
Bogoljubow decides to chop

queens rather than to try his luck in
the complications of 36 . . . 'i!hc4. In
500 Master Games of Chess, Tar
takower and Du Mont give 37 llJe6
"with very strong pressure in the
centre". Curiously, Reinfeld con
tinues with 37 .. .libS 3 S llJxc5 dxc5
39 lId2 lIb3 (this seems silly, as
White will soon push the d-pawn) W
40 f2 +-, but credits the analysis
to Tartakower. This is apparently 41 llJe6! llJb3
from a second source of Tartak Or 4 1 . . .llJxe4 42 'it>xe4 l:txe2+
ower analysis, but I've been unable 43 'it>d4 l:td2+ 44 l:td3 +- - Tartak
to track down the primary source, ower and Du Mont.
so all references to 'Tartakower' 42 c5! dxc5
without Du Mont's name stem 43 llJxc5 llJd2+
from Reinfeld. It's possible that in 44 'it>f2 'it>e7?
some cases Reinfeld conflated his Reinfeld cites analysis by Tar
own analysis with the 'second Tar takower that runs 44 . . llJbl 45 1Ic4
.
takower' analysis. (according to Reinfeld, Tartakower

It might seem that the ending gives simply "45 llJxa4 llJxc3", but
should be ' an easy win' for White Golombek continues 46 llJxb2
on account of Black's bishop being llJxe4+ 47 'it>e3 llJd6 and "Black
locked out of play. In fact, the would still lose eventually after 4S
pas sed a-pawn, supported by the 'it>d4 'it>e7 49 llJf4 followed by
tDe6"; this variation highlights 'it>e7 and White should give perpet
the sorry state of Black's bishop) ual - Tartakower.
45 . . . a3 46 tD e6 ! (D), with the di 45 'it>e1 tDbl
vergence: 46 ':d3 a3?
Better was 46 . . . 'it>d6 47 tDxa4
':b4 48 tDac3 tDxc3 49 tDxc3 J.f7
50 'it>d2 g6 and "Black could still
put up a fight." - Tartakower and
Du Mont.
47 d6+ 'it>d8
48 tDd4 l::tb 6
49 tDde6+ i.xe6
50 fxe6 litb8 (D)
a) Tartakower believed that

Black loses after 46 . . . a2 47 d6. The
idea is clear: White tries either to
queen his d-pawn or to throw up a
mating net around the black king.
However, Black has sufficient time
to force a draw with 47 . . . 'it>e8 48
litc7 J.xe6 49 fxe6 ':b8 50 ':f7 ! W
(50 ':xg7 ? 'it>f8 -+) 50 . . . .:d8 5 1
l::te 7+ 'it>f8 5 2 ':f7+ 'it>g8 (or 51 e7+ 'it>e8
52 ... 'it>e8 53 l::te 7+ =) 53 e7 ':b8 54 52 tDxa6 1-0
d7 'it>xf7 55 d8 'iW litxd8 56 exd8'i; White ' s knight and d-pawn
a l 'i!t' 57 'iVd5+, with a perpetual make an impressive team. Even the
Mayer. It's possible that 46 . . . a2 47 a-pawn doesn' t save Black, e.g.,
d6 stems from Reinfeld and was in 52 ... a2 53 tDxb8 al'iV 54 d7+ and
terpolated in Tartakower's analysis White will soon mate.
in 'b' below, as it's difficult to be
lieve that a GM didn't find Black's There are cases where a Prob
forcing defence. lem Bishop is taken on willingly.
b) More straightforward is One must then work to free the
46 ... 'it>e7 47 ':c7+ 'it>d6 48 ':c6+ Problem Bishop or face long-term
problems. Reshevsky ultimately where it is. Another method of ad

loses the following game because vancing the e-pawn is 9 e3, 10 'iVe2
he doesn't improve the situation of and 1 1 e4, with the same piece ar
his Problem Bishop. Further, he rangement for White as that which
exchanges too many pieces and arises after his eleventh move.
finds - like Bogoljubow and Be 9 ... liJbd7
liavsky in prior examples - that a 10 e4 i.h7
Problem Piece becomes more and The black bishop is stifled by
more important the fewer the the white pawn chain c2-d3-e4.
pieces remaining on the board. Reshevsky 's method of develop
ment is acceptable in the event that
Korchnoi - Reshevsky White plays an early c2-c4, as then
Amsterdam, Candidates ' the white chain is weakened. Here
Match (4) 1 968 it is less appropriate. One situation
in which it could work out for
1 liJf3 liJf6 2 g3 d5 3 i.g2 c6 4 b3 Black is if he had his c-pawn at c5 ,
i.f5 5 .i.b2 e6 6 0-0 i.e7 7 d3 h6 S as he could then play for a later
liJbd2 0-0 (D) . . . c4, which would hack away at
the pawns that are containing the
bishop.
11 e2 as
12 a4 6
13 e5 lLJes
14 i.h3!
Korchnoi takes steps to slow
down or prevent a move of the f
pawn. Further, . . . i.f5 would now
result in doubled pawns.
14 liJc7
15 h1 :aeS
W 16 liJh4
Paving the way for the advance
9 e1 of the f-pawn and offering Black
A sophisticated move. Korchnoi the chance to play 1 6 . . . i.xh4,
would like to advance his e-pawn, when White gains dangerous play
but it requires support. The natural down the g-file.
approach would be to play 9 l:te l , 16 f6
but h e may later wish to advance 17 exf6 i.xf6
his f-pawn, so he leaves the rook 1S i.xf6 litxf6
19 f4 23 ':xfl+
Grabbing control of e5 and This plays into White's hands,
kee ping the black e-pawn as a tar as it leaves fewer pieces on the
get. board, thus making the plight of
19 'iUc5 the h7-bishop relatively more im
20 liJdf3 c3 portant.
21 .t g4 d4 (D) 24 ':xfl 'iUc5
25 'it'f2 ':f8
26 'iUxfS+ 'it'xfS
27 ':xf8+ xf8
The minor-piece ending is diffi
cult for Black, as he has yet to find
a way to liberate his Problem
Bishop.
28 liJf3 c5
29 liJd2 liJd5
30 liJc4 (D)
W
Black prepares . . . liJd5 , which
homes in on the weakened e3-
square.
22 liJe5 liJxe5
23 fxe5 (D)
30 liJb4?
In Candidates ' Matches 1 968,
Furman and Kirillov point out that
30 . . . e7 was better, with the idea
that 3 1 liJxa5 liJe3 32 1.f3 b6 is
equal. However, they also suggest
that White could play 3 1 .tf3 or
B bring his king to the defence of the
c-pawn, which in either case 're 41 e6 <tie7 42 <tif3 <tid6 43 <tie4 g5

tains winning chances ' . 44 g4 1-0
The rest of the game, while re
quiring some accuracy, doesn't re Verdict: While the Problem Bishop
quire any further comment: 31 does not occur as frequently as the
lDxaS lDxc2 32 lDxb7 c4 33 bxc4 Problem Knight, it's still common .
.txd3 34 lDcs .txc4 3S .txe6 It tends to occur when a bishop is
i.xe6 36 lDxe6+ <tie7 37 lDcs lDb4 on the side or toward the side of the
38 as lDc6 39 a6 <tid8 40 <tig2 g6 board.
7 The Over-rated Knight
One of the most difficult things in bishop. The knight, with its ability
all of chess is seeing a position as a to attack both colours, can fre
whole. It' s very easy for even a quently avoid this fate. Yet some
grandmaster to pick one feature of thing similar can happen and it is
a position, such as a well-placed inherent in the way a knight moves.
piece, and decide that it's the most Let's say that we get a knight to a
important thing on the entire board. good post where it can ' t be driven
It' s not uncommon to see an anno away. To do so, we may well have
tator point to a really well-placed to take several moves. If the knight
piece and explain how its owner is proves unable to attack anything
better or has a winning game be directly from its 'pretty position' ,
cause of it. Indeed, this will even then the time spent in getting it
be true in many cases. there may prove critical. Further,
However, it's also possible to because two knights may both
take a piece that looks wonderful want the same post, they can end
and decide that it's terrific, when in up tripping over each other; this is
fact it does little for one's position. Dvoretsky's concept of The Super
I call such pieces 'over-rated ' , as fluous Piece, which we shall have
they are literally given too much occasion to discuss in connection
credit. Over-rating can occur with with the following game.
any piece, but it seems to be most
common with knights or bishops. Larsen - Donner
A queen is so intrinsically power Beverwijk 1 960
ful that if it looks well-placed, it
probably is well-placed. Similarly, 1 g3 e5 2 it.g2 d5 3 lLlf3 .td6 4 0-0
a rook on an open file and the sev lLle7 5 c4 (D)
enth rank is usually doing exactly In Larsen 's Selected Games of
what you want it to do. Chess 1 948-69, Larsen queries this
One of the problems that an move and comments that it is ' not
over-rated bishop can encounter is very effective in this position' , pre
that it lacks meaningful objects of sumably because Black is able to
attack, e.g., the opponent's pawns support his centre with 5 . c6.
. .
are on the colour opposite to the 5 ... c6

his queenside to be torn up, thus

giving Larsen a source of potential
counterplay.
17 bxc6 bxc6
18 WHet a4
19 c5 .ta5
20 .tc3 ':e7
21 .txa5 ':xa5
22 lbfd2 lbed4
23 lbc4 l:.aa7
24 f4 (D)
B
6 d3 0-0
7 lbbd2 lbd7
8 e4 dxe4
9 lbxe4 .tc7
10 b3
" 1 0 d4 exd4 1 1 xd4 lbe5 is
good for Black" - Larsen. The
most important feature of the posi
tion then would be White's weak
nesses in the d-file.
10 ':e8 B
11 .tb2 lbf5
12 lle1 lbf8 Larsen makes an insightful
13 d2 f6 comment here:
14 %lad 1 "Of course Black has every rea
Larsen gives 1 4 d4 as equaliz son to be proud of the knight on d4,
ing, while he suggests that 14 b4 but exaggerated care for it now
immediately would possibly be leads him astray. The black position
better if White is playing for a win. must not be overestimated; also the
14 lbe6 white knights are full of pep.
15 b4 a5 "Black should play 24 . . . exf4,
16 b5 .tb6 but Donner thought that the cen
Larsen considers 16 . . . cxb5 wor tralized knight deserved to keep its
thy of ' serious attention' . With the solid pawn protection. During the
text, Donner plays to retain control rest of the game this knight does
of the d5-square, but he also allows not do very much."
The Over-rated Knight 91
One thing that's remarkable occupy the 'important high ground' ,

about this game is how well-placed and the others will turn out to be, so
all four knights appear to be. to speak, superfluous)."
Whose knights are actually better In Larsen-Donner, we see an ex
placed? A straightforward arit cellent illustration of The Super
hmetic approach suggests that fluous Piece, and it helps us to
Black' s are; after all, White ' s determine whose knights are really
knights are 'only ' o n the fourth better placed. Donner's knights
rank, while Black has a knight on look impressive, but only one of
the fourth rank and a somewhat them can occupy d4, while the
further advanced knight on the other has no special square that
fifth rank. beckons . Thus, Donner' s knight at
You may have noticed that the f5 is superfluous . By contrast,
black knights ' link up ' with each Larsen's knights both occupy good
other. This can sometimes be a squares in their own right and have
very useful arrangement, as it the potential of playing into the
means that one knight can replace hole at d6. At the moment, how
another if an exchange occurs. In ever, a white knight playing into d6
fact, Mark Dvoretsky has made a would make the other knight
useful discovery that he terms The largely superfluous, as it would
Superfluous Piece. Let's see how have no better square to play to
he described The Superfluous than its present post. In part, this is
Piece in his excellent Tra iningfor the reason behind Larsen' s sug
the Tournament Player: gesting 24 . . . exf4 here, as no new
"From Nimzowitsch ' s writing squares would open up for the
we know that pieces which are able white knights.
to cover a strategically important In conclusion, I think that we
square, making it possible to oc could say that each side ' s knights
cupy that square, are usually well are equally well-placed following
placed. White ' s 24th move. However, this
"In the fight for a given square evaluation will soon change.
players most often try to exchange 24 jte6?
these pieces off. But sometimes a 25 fxe5 fxe5
totally different strategy is adopted: 26 'iiith l !
if the square cannot be won back A subtle move. Larsen ensures
by means of exchanges, then one that the knight at d4 can never
may . . . forget about exchanges al check the white king.
together (after all, only one of the And now we can see how Don
opponent's pieces will be able to ner's mistake at move 24 has hurt
his position. His own knights still White menaces the a-pawn and
have the same possibilities as be also prepares to swing his queen
fore and the knight at f5 is still su over to the kingside.
perfluous . But the white knights It's interesting that the Superflu
have gained a number of possibili ous Knight that once stood at f5
ties, for example, the knight at c4 has found itself all the way back to
presses on the weak e5-pawn, f7 . No method of improving its
while the knight at e4 has gained placement arose when it was at f5
the possibility of using g5 as a stag and its placement subsequently de
ing ground for tactical operations cayed.
on the kingside and possible entry 32 a5
to the hole at e6. Consequently, if 33 'tlYh5 'iVc7
either white knight plays into d6, 34 Sl.h3 lIh6
the other knight will still retain Larsen gives this two question
possibilities of improving its cur marks and suggests 34 . . . 1:.e7 , when
rent placement and therefore avoid Black may still be able to defend
the fate of Dvoretsky 's Superflu the position.
ous Piece. 35 g4 1:. g6
26 Sl.d5 36 d1 a7
27 1:.fl 1:.e6 37 "iVb 1 ! tiJg5
28 lIf2 1:.17 38 "iVb6! (D)
29 1:.dfl Sl.xc4
30 dxc4 tiJh6
31 1:. x17 tiJx17
32 'iVd1 ! (D)
B
38 a8
39 tiJxg5 1:.xg5
B 40 c7 h6
41 1:.b1 1-0
Resignation is appropriate, as .te7 8 'iHf3 Wic7 9 0-0-0 liJbd7 10

4 1 . . .h7 42 .tg2 or 4 1 . . :it'a6 42 g4 bS U .txf6 gxf6! ? (D)
.tfl both give White decisive at
tacks. "What did the knight on d4
really accomplish?" - Larsen.
Larsen's comments on Donner's
'overestimating' the black position
give us additional insight into the
nature of over-rated pieces. In an
abstract case, a knight occupying a
hole or a bishop on a wide-open di
agonal are both excellent pieces.
The 'over-rating' is a human fail
ing, not a failure of the position or
the 'laws of chess strategy' . In w
deed, a good computer program
will typically not fall into the prob Fischer's choice; Black accepts
lem of 'over-rating' a piece, as its doubled f-pawns, but hopes to
evaluations are analytical in nature. demonstrate that his dark-square
Whether over-rating is a function control and central pawn mass are
of psychology (undue optimism, more important.
self-satisfaction) or aesthetics (the 12 fS liJeS
piece looks 'beautiful' ) , it seems to 13 3 O-O!
be peculiar to humans. White is playing for pressure on
e6 and a kingside attack. With his
There's an old line, generally at last move, Black reasons that his
tributed to Steinitz, that a knight at king will prove difficult to mate, as
e6 i s like a rusty nail in the knee. White has a relatively limited front
It's true that such a knight can in which to operate.
cause the defender a pain in the 14 liJce2 h8
knee, but it's also been known to 15 f4 l:.g8
happen that the knight is over Both sides are pursuing laudable
rated. goals; White has stepped up the
pressure on e6, while Black has
Kotkov - Belinkov taken steps to strengthen the de
USSR 1971 fence of his king.
16 .te2 .tb7!
1e4 cS 2 liJf3 d6 3 d4 cxd4 4 17 fxe6 fxe6 (D)
xd4 f6 S liJc3 a6 6 .tgS e6 7 f4 18 liJdxe6?
position is assessed as + by Gip

slis in ECO B, Second Edition, al
though I would be inclined to call it
unclear. As a side note, Gipslis did
not note White's possible improve
ment at move 1 8, thus giving the
casual reader the false impression
that Black is better in the lines aris
ing from 1 6 .te2.
20 ':'hfl ':'ac8
21 i.d3 .txd3
W 22 cxd3 b4
23 l2Jh5 b3! (D)
Kotkov captures with the wrong
knight ! He must have thought that
the f4-knight would later perform
good service from its present loca
tion, but he would have done better
with 1 8 l2Jfxe6 ! 'iYa5 1 9 l2Jf5 ! and
White is for choice - Gufeld in In
formator 1 1 .
lt's interesting that i n the above
variation, White gets a knight to e6
under much more favourable cir
cumstances than in the game con
tinuation. The secret is not the W
strength of the knight at e6, but the
fact that the other knight stays in White ' s knights hover menac
the centre and gains the valuable ingly close to the black king, but
f5- square. As Kotkov played, the there is no good way for them to
white knight on f4 is a Superfluous break down the defence. Mean
Piece, as it must support the e6- while, Black is pursuing a king at
knight and has no other really good tack of his own.
square to which it can play. 24 axb3 l:':tb8
18 ... a5 25 d4 (D)
19 'it>b1 .txe4 White doesn't care to get in
Black has the bishop pair and the volved in 25 l2Jxf6 .txf6 26 llxf6
e6-knight is not as impressive as ':'xb3 , although it is not clear who
it might look at first glance. This this favours.
There are also cases where one

side gets a knight to a central hole
on the sixth rank but lacks the
proper development (or piece place
ment) to do much with it. The fol
lowing game is a good illustration
of this.
Cheney - Mayer
Washington, D. C. 1990
B 1 e4 c5 2 4Jf3 d6 3 d4 cxd4 4
4Jxd4 4Jf6 5 4Jc3 g6 6 f4 i.g7!? 7
25 4Jc6 e5 (D)
26 :f5 .l::tb 5
27 d5?
Now Black is able to crash in
with his pieces. The correct con
tinuation was 27 ':xbS ! - defend
ers should simplify - when the
natural 27 . . . axbS , going for play
down the a-file, allows the pretty
28 4J xf6 ! .txf6 29 WNf3 , when
White is doing well - Mayer.
27 4Jb4
28 WNc3 WNa2+
29 c1 4Jxd5! B
30 4Jg5
Playing for a trick (30 . . . 4J xc3 ?? 7 4Jh5!
3 1 4J f7#) , but now Black wins There was a time when 6 ... .tg7 ! ?
massive material. However, 30 was considered practically a losing
.:tfxdS was no better, since Black move, as White's e-pawn push is
then has a winning attack after very disruptive to the black posi
30 . . . ':xb3 3 1 "iVc2 l:tgb8. tion. Indeed, the position after
30 iia1+ White ' s seventh move is known as
31 d2 WNxd1+ the 'Levenfish Trap ' and it i s cer
32 xd1 4Jxc3+ tainly a trap if Black falls into
33 bxc3 ':xg5 something like 7 . . . dxeS? 8 fxeS
0-1 and then 8 . . . 4J dS 9 i.bS+ ! , 8 . . . 4J g4
9 i.b5+ ! or 8 . . . lLlfd7 9 e6 ! . How Black must exercise caution, of

ever, after 7 . . . lLlhS ! , matters are not course, but the secret to this posi
so clear, because Black threatens tion is that White has difficulty
the trick 8 . . . lLlxf4 ! 9 i.xf4 dxeS finding anything useful to do with
-+. his queen' s bishop. Meanwhile,
8 i. bS+ i.d7 Black is crawling all over the light
9 e6 fxe6 squares and targeting both c3 and
10 lLlxe6 i.xc3+ c2.
11 bxc3 c8! 18 lilae1 i.fS
Before this finesse was discov 19 'iUd4 'iVcS! (D)
ered, Black had tried 1 1 . . :iVaS ?, Black's advantages show better
but then the reply 1 2 i.d2 ! is good in the ending.
for White, for example, 1 2 . . . i.xbS
13 c4.
After 1 1 . .. 'iic 8 ! , the knight at
e6 is a bit unstable and the white
c-pawns are suddenly under pres
sure. Of course, White has the bish
ops and the knight prevents Black
from castling, but things are very
obscure here. It could be that White
has the better chances in the mid
dlegame, but it's practically certain
that Black has the better endgame
chances. W
12 'iid3 lLlc6
13 lLlgS 20 'iixcs dxcS
Now the knight really hangs, so 21 c4?
White brings it back out. This is a mistake, after which
13 0-0 Black is able to overrun the white
14 0-0 lLlf6 queenside. Also mistaken was 2 1
IS i.d2 i.fS '1:.xe7 ? c4 ! , when the bishop i s im
16 'iie2 i.g4!? prisoned and soon to be trapped.
I t was also possible t o play However, after the superior 2 1
1 6 . . . i.xc2, but this was a last i.d3 ! , White would cover the c2-
round money game and I had no pawn and really threaten the e
desire to grab an isolated pawn that pawn.
I felt 'wasn' t going to run away' . 21 a6!
17 d3 lLlaS! 22 i.xaS axbS
23 .te3 b4 The white knight, which kept

24 .tal lUe8! hopping in and out of e6, never re
Black prevents the entry of the ally found a place to call home.
white rook before he cleans out the
queenside. As should be clear by now, a
25 ':e2 h6 knight can become over-rated
26 lLle6 l1xa2 when there are other features of a
27 lLlxe5 b6 position that are important. Indeed,
28 .txf6 bxe5 there are many opening variations
29 ':xe7 ':xe7 that include what could be called
30 .txe7 .txe2 (D) over-rated knights, for example,
the Dragon game we just looked at.
However, probably the most com
mon are those Sicilian Variations in
which Black plays an early . . . e7-eS
and allows a white knight to oc
cupy dS . The best known of these
lines is the Sveshnikov : I e4 cS 2
lLlf3 lLlc6 3 d4 cxd4 4 lLlxd4 lLlf6 S
lLlc3 eS 6 lLldbS d6. After the fur
ther 7 .tgS a6 8 .txf6 gxf6 9 lLla3
bS 10 lLldS fS , it is clear that the dS
knight is not the only thing in the
w position. Black has the bishop pair,
a central majority and the white
The ending is winning for Black knight on a3 might even be termed
and requires no further comment: a Problem Piece, as it's on the side
31 .txe5 b3 32 .td4 b2 33 .txb2 of the board and has no good
':xb2 34 ':f2 ':b1 + 35 ':fl 1:txfl + squares available for the foresee
36 xf1 .td3+ 37 Wf2 .txe4 38 g4 able future. This position gives us
f7 39 e3 .td5 40 d4 .tf3 41 some insight into how a knight can
g5 h5 42 e5 .tg4 43 d6 .tf5 44 be over-rated.
d5 e7 45 e5 d7 46 d5 e7 Another instance of an arguably
47 c5 b7 48 b5 .td3+ 49 c5 over-rated knight occurs in the
a6 50 e6 a5 51 e5 .tf5 52 King' s Indian after the sequence I
e4 b6 53 b4 e6 54 e4 d4 lLlf6 2 c4 g6 3 lLlc3 .tg7 4 e4 d6
d6 55 d4 e6 56 h4 .tb1 57 S .te2 0-0 6 lLlf3 eS 7 0-0 lLlc6 8 dS
e3 f5 58 f3 .te2 59 g3 lLle7 9 b4 (the B ayonet Attack)
e4 0-1 9 . lLlhS 1 0 g3 fS 1 1 lLlgs . White's
. .
kin g ' s knight angles for the gap Verdict: An over-rated knight is
ing hole at e6, but the cost is two objectively well placed but is sub
tempi and possible loss of the pawn jectively given too much impor
at e6. While this variation is very tance in assessing a position. The
theoretical nowadays and still very over-rated knight is most commOn
much in flux, one has to wonder if in dynamic positions where there
White 's knight isn't a tad over are other important factors that are
rated. not considered or given their due.
8 The Over-rated Bis hop
I n this chapter w e examine over

rated bishops. We are already fa- .
miliar with the concept of the
over-rated piece from the chapter
on over-rated knights. As we saw
there, an over-rated piece is one
that looks very impressive but
proves to be less important than
one might think. Further, there is
an element of human failing, since
the side with the over-rated bishop
usually is gUilty of placing too B
much value on an otherwise well
placed piece. but he has managed to reach a theo
The following example is par retical position that he knew little
ticularly striking, as White has not about at the time, while he had to
only an impressive-looking bishop assume that his opponent knew it
but also an impressive-looking inside out.
middlegame position. White ' s plan is both simple and
dangerous. Left to his own devices,
Lukovnikov - Polugaevsky he will mobilize his qualitative
Krasnodar 1 983 pawn majority by advancing his
g-pawn. The black king will have
1 e4 c5 2 liJf3 liJc6 3 d4 cxd4 4 difficulty surviving such an on
liJxd4 e6 5 liJc3 d6 6 f4 liJf6 7 i.e3 slaught, particularly as the c4-
i.e7 8 f3 e5 9 liJxc6 bxc6 10 f5 bishop and the queen can join in
l::tb8 1 1 0-0-0 a5 12 J.c4 (D) the attack with ease.
As Polugaevsky explained in his Black has trumps , of course,
excellent book The Sicilian Laby most noticeably in the half-open b
rinth Volume 1 , he unexpectedly file and his central majority. How
found himself in a difficult position ever, as Polugaevsky well knew,
while still in the opening. It's not White had won many games from
that his position is objectively bad, this position in the 1 970s when
100 The Over-rated Bishop
Black had relied solely on obvious to advance the pawn to h6) 1 7 . . . c5

counterplay. How can he, without 1 8 Jt.. g 5 ! c4 1 9 1.a4+ 'it>f8 20 Jt.. xf6
especially knowing the accumu Jt.. x f6 2 1 lhd6 Jt.. e 7 ! , with unclear
lated theory of the day, resist the play in Psakhis-Pritchett, Troon
white attack? 1 984.
12 h5! ! Psakhis ' s appraisal of 14 h3
' ! ! ' - Polugaevsky. shows the dynamic nature of chess.
' ! ? ' - Psakhis and Stetsko in In As we shall see, Polugaevsky con
formator 38. siders the similar move in this
"I am not afraid to attach two ex game a mistake of major propor
clamation marks to Black's reply, tions.
despite all its drawbacks. In a diffi 14 c5
cult psychological situation this 15 Jt.. c 1 c4
prophylactic operation enabled me 16 1.a4+ 'it>f8
to solve a problem of primary im 17 1.c6
portance: to prevent the white g The bishop plays around to the
pawn from storming forward." - hole at d5 , where it appears that it
Polugaevsky. will be very well-placed.
Of course, Black must worry 17 c5
now that his king will have prob 18 Jt.. d5 Jt.. d 8! (D)
lems, but it already had problems.
The genius of Polugaevsky' s move
is that it is both generally sound
Nimzowitsch certainly would have
liked it! - and it forces his oppo
nent to confront new problems at
the board.
13 Jt.. b 3 Jt.. a6
14 'it>b1
A lacklustre move. One gets the
impression that White was dis
armed once Polugaevsky deprived
him of his obvious plan. Perhaps W
Lukovnikov felt that he had time
for quiet moves that 'safeguard' his Hitting the nail on the head ! It is
king position. usually a good idea for a minor
A sharper approach is 14 h3 (' ! ' piece to occupy a hole, particularly
- Psakhis and Stetsko) 1 4.. h4 1 5 g4
. when it's in the centre. In the given
hxg3 1 6 WUxg3 1Ih6 17 h4 (intending case, however, the bishop at d5 is
The Over-rated Bishop 101
what I mean by 'over-rated' ; it position is safeguarded at the high

lo oks terrific, but it is difficult for cost of keeping the c3-knight and
White to do anything really useful c l -bishop nearby. Finally, Black's
with it. As Polugaevsky explains, endgame prospects should be bet
"The point is that Black has not in ter, as the pawns at e4 and h3 make
fact lost control of the d5- square, very nice targets.
which is being carefully guarded 23 a1
by the knight at f6, while the cen Polugaevsky points out that 23
tral bishop is essentially tying the .ixf7 fails to 23 . . . .ixc3 .
hands of its own rook and knight, 23 .icS! (D)
which cannot leave it in proud iso
lation."
With 1 8 . . . .id8, Black prepares
to place White's knight at c3 under
'constant surveillance ' by . . . i.a5 ,
when it may be possible to ex
change at c3 at an opportune mo
ment and then eliminate the
over-rated bishop by . . . tDxd5.
19 h3
Polugaevsky queries this move,
as it allows Black to damage the
white pawn structure. On strict po W
sitional grounds, White should pre
pare the advance of the g-pawn by 24 .id2?
g2-g3 , h2-h3 , and only then g3-g4. A mistake in a difficult position.
As mentioned above, two-time So Polugaevsky points out that 24
viet champion Lev Psakhis be .ixf7? .ixc3 25 .ig6 tDxe4 is good
lieves that the opening of the g-file for Black, so 24 ':gfl should have
in such positions can be worth been played.
while. 24 .ixf5!
19 h4 25 .ixti
20 g4 hxg3 Polugaevsky analyses two alter
21 'i!Vxg3 .ia5 natives, viz., 25 exf5 i.xc3 26 .ixc3
22 l:lhg1 l:lh7! lLlxd5 27 .ixe5 dxe5 28 xe5 'ifb6!
The black rook defends g7 while and 25 ':gfl .ig6 ! , which safe
remaining active on the half-open guards the black king and consoli
h-file. B lack's king position seems dates the extra pawn.
to be secure, while the white king 25 .ixc3
26 Jt.. xc3 4Jxe4 helped him win a game that might

27 f3 'iff2! have turned out poorly for him if
28 'ii'xf2 4Jxf2 he'd been a slave to convention.
29 l:Ixd6
Black wins the ending after 29 Since the middle of the century,
lI dfl rt;xf7 30 ':xf2 rt;e6 - Polu- the most popular approach to chess
gaevsky. has been that of seeking active
29 rt;xf7 counterplay. Indeed, most experi
30 .:n ':xh3 enced players probably prefer an
31 ':d2 4Je4 inferior position that has counter
32 lld7+ rt;e6 play to one that is just a shade
33 llfd1 (D) worse but offers no possibilities
of counterplay and allows the op
ponent to ' grind away ' . In the fol
lowing game, Geller gets his fair
share of counterplay, but Spassky's
well-timed defensive measures limit
the damage. When Geller fails to
take appropriate actions against
Spassky ' s pending attack, the re
sult is carnage for the black king.
Spassky - Geller
Sukhumi, Candidates '
B Match (6) 1 968
33 lIh1 1 e4 c5 2 4Jc3 d6 3 g3 4Jc6 4 .tg2

This wins, but 33 . . . Jt.. g4 is a g6 5 d3 Jt.. g7 6 f4 4Jf6 7 4Jf3 0-0 8
shade stronger, as it wins the ex 0-0 llb8 9 h3 b5 10 a3 a5 1 1 Jt.. e3
change - Mayer. b4 12 axb4 axb4 13 4Je2 Jt.. b 7 (D)
34 lIxh1 rt;xd7 Each side has 'his' respective
35 .txe5 Ilg8 side of the board: Black attacks on
0-1 the queenside while White is going
for the king. However, a certain
Games such as Lukovnikov amount of defence is also neces
Polugaevsky show the importance sary. The fourth game of the match
of original thought in chess. saw 14 d2, but Black's rook sub
Whether B lack's idea is good or sequently reached a2 and tied down
bad, it is certainly logical and the white queen to the defence of
1 8 . . . tt'lh5) 1 8 . . . :e8 is the typical

expression of Black's plan.
b) 17 'iHe 1 i.a6 slows down the
thematic transfer of the queen to
h4, as 1 8 WHh4? allows 1 8 . . . .txd3
or 1 8 . . . tt'lxe4.
17 'iHe1 WHa6
18 'iHf2
The immediate 1 8 WHh4 tt'lxe4 !
is good for Black.
In Mode rn Chess Brilliancies,
W Evans suggests that Black should
now continue 1 8 . . . tt'ld7 , which he
the c-pawn. Spassky won that assesses as equal.
game also with a nice kingside at 18 tt'la7
tack, but he came to the board for 19 f5 tt'lb5 (D)
his next White with a big improve
ment prepared.
14 b3! :a8
15 .l:Ic1 !
Spassky ' s manoeuvre ( 1 4 b3 !
and 1 5 lic l ! ) is designed to clear
the long diagonal of any targets.
The black king's bishop now looks
pretty, but there 's nothing to attack
on its diagonal. Further, White uses
the rook to defend c2, thus freeing
up his queen to play around to h4
and join in the kingside attack. W
15 li a2
16 g4 'iHa8 20 fxg6!
The closed centre is what gives This attack is 'anti-positional' in
White's pending attack its strength. that it captures away from the cen
Vasiukov, cited in Cafferty ' s Can tre, but Spas sky has calculated that
didates ' Matches 1 968, suggests the tactical chances gained from
1 6 . . . e6 ! , with the idea of gaining the open f-file make up for the
counterplay in the centre: drawback of exchanging an f-pawn
a) 1 7 f5 exf5 ! 1 8 exf5 ( 1 8 gxf5 for an h-pawn.
can be answered by 1 8 . . . :e8 or 20 ... hxg6
Soviet sources cited by Cafferty 25 tDxt7 ! lbe2

in Candidates ' Matches 1 968 The automatic 25 . . . xf7 allows
suggest 20 . . .fxg6 (aiming to defuse White to crash through after the
White 's play on the f-file) 2 1 tDf4 continuation 26 .th6 I:.g8 27 tDf4
.tc8 22 tDg5 , intending tDfe6, llxc2 (or 27 . . . d5 28 exd5 f5 29
which is assessed as clearly better tDe6 +- - Evans) 2 8 llfl with a
for White. winning attack - Soviet sources
21 tDg5! cited in Candidates ' Matches 1 968.
The white knight takes its post in The same Soviet sources indicate
preparation for the kingside siege. that 26 tDf4 also yields a winning
21 tDa3 attack.
22 'it'b4 lIe8 (D) 26 .th6 lbc1 +
White wins after 26 . . Jbe2 27
'it'xg7+ e8 28 tDg5 fxg5 29 .txg5
- Soviet sources.
27 tDxc1 xt7
Or 27 . . . .txh6 28 tDxh6 e8 29
tDg8 (or 29 'it'xg6+ d7 30 'ii'f7+
c6 3 1 e5+ wins - Evans) 29 ...f8
30 tDe7 +- - Soviet sources.
28 'iUxg7+ e8 (D)
Black loses after 28 . . . e6 29 g5
fxg5 30 .txg5 I:.e8 3 1 h4 or 3 1
iVxg6+ - Soviet sources.
W
23 llxf6 !
A standard sacrifice in this type
of position. The real point is re
vealed on move 25.
There's a fair helping of irony in
the fact that the black king's
bishop, which has already had its
prospects lessened by White's 14th
and 1 5th moves, must now consent
to being buried behind its own
pawns. w
23 exf6
24 \\h7+ f8 29 g5
White also wins after 29 e5 ! , hit

ting the bishop and threatening 30
e6 - Mayer.
29 fS
30 xg6+ 'it>d7
31 t7+ 'it>e6
32 exfS+ 1-0
The pawns are unstoppable after
32 ...'it>b6 33 'iVxb7+ 'iWxb7 34 Jt.. xb7
'it>xb7 35 f6.
Spassky's method of ' clearing w

the diagonal' is now an accepted
part of opening and middle game arrangement knights on b5 and c4,
theory in many different types of pawns on a4 and b3 . This will
positions. In particular, it is an ap dampen Black's queenside play, as
proach that can cause Black great the knights will protect things
difficulty in the Benko Gambit while the pawns will be shifted off
when White is successful in imple the colour of Black' s bishop. It's
menting it. true that b3 will be weak, but the
knight at b5 will shield it, while the
G. Kuzmin - Stein a-pawn will be a protected passed
USSR 1 972 pawn.
12 liJbd7
1 d4 liJf6 2 e4 eS 3 dS bS 4 exbS a6 13 a4 lUeS
S bxa6 g6 6 liJe3 .txa6 7 liJf3 Jt.. g7 Aiming to create play with ... c4,
S liJd2 d6 9 e4 a5 10 .txa6 xa6 but this chance never materializes.
1 1 e2 0-0 (D) 14 liJe4 liJeS
In the Benko Gambit, Black is 1S :a3
not particularly concerned if the This is a useful precaution
queens should go off, as his com against any tactics based on captur
pensation is positional, i.e., pres ing at b5 and then taking the loose
sure on the white queenside. His rook at a l . Further, the rook is
king' s bishop, supplemented by evacuated from the long diagonal,
the heavy pieces, typically plays an thus making Black's bishop less
important part in this pressure. powerful, as it will soon lack any
12 liJbS! targets. This position is evaluated
Kuzmin plays ambitiously. He as somewhat better for White by
plugs the b-file and aims for the Kotov in Informator 14.
106 The Over- rated Bishop
15 CDc7 be careful that Stein doesn' t man

16 CDxc7 l'iIxc7 age to break into the position with
17 .td2 CDb6 hi s rooks, as then the king could
18 b3 (D) prove a juicy target, particularly if
Black's bishop were to set up shop
at d4. Kotov gives a clear advan
tage to White in this position.
21 . c4
A typical Benko Gambit move
by which Black hopes to fracture
White's queenside pawns and re
new his pressure.
22 CDc6! cxb3
With White's knight now plug
ging the c-file, it would be foolish
to play 22 . . . c3 , as the pawn would
B eventually be surrounded and ab
sorbed, leaving White up two
While Black managed to ex healthy pawns.
change off the knight at b5, he has 23 ':xb3 f5
not been able to prevent White 24 as! CDd7
from completing the rest of his 25 ':el ! f8
'ideal arrangement' described in 26 CDb8 ! Ibel?!
the note to White ' s 1 2th move. Stein admits that his position is
Black retains practical chances, but hopeless and aims for salvation in a
he is being outplayed and must be piece-down ending. The alterna
careful that things don't get any tive was 26 . . . ':aa7 27 ':xc7 ':xc7
worse. 28 CD xd7+ ':xd7 29 .te3 +-, fol
18 lowed by pushing the a-pawn - Ko-
A mistake that allows White to tov.
gain time, simplify the position 27 CDxd7+ e8
and improve the placement of his 28 CDb8 ! (D)
pieces. This consolidates the extra
19 CDa5 ':c7 piece. Stein presumably hoped to
20 'ixa6 ':xa6 run White out of pawns, but he
21 e2! doesn't come close and the rest of
Kuzmin adds the advantage of the game requires no comment:
the better-placed king to his trophy 28 Iba5 29 .txa5 fxe4 30 CDc6
..
collection. Of course, he must still g5 31 .td2 ':c2 32 CDb4 ':c5 33

.txg5 ..t>f7 34 .td2 e6 35 dxe6+

..t>xe6 36 f3 d5 37 fxe4 dxe4 38
1:.h3 h5 39 1:.h4 1:.e5 40 lbc2 i.f6
41 1:.f4 1-0
Verdict: Like the over-rated

knight, the over-rated bishop is ob
jectively well placed but is subjec
tively misevaluated. The trait most
common to over-rated bishops is
that of being on an open diagonal
B with no accessible targets.
9 Changing the Colour of a
Bishop
B eginning chessplayers soon re A very odd move. It could be

alize that a knight has the poten that Topalov's second move is pre
tial to reach every square on the mature, as the central pawn struc
board, while a particular bishop ture isn ' t set yet. With the text,
has the potential to reach only half Korchnoi deliberately weakens d4
of the 64 squares. Indeed, this is the and practically begs Topalov to
biggest problem facing a single give him the bishop pair. Yet didn't
bishop . However, it's possible to Black's second move suggest that
use a bishop to influence events on he was willing to part with the
the colour opposite it - by pinning bishops even without such a major
or exchanging it off for a knight ! concession from White?
This process can be seen as 'chang 3 ... c5
ing the colour of a bishop ' . 4 lbc3 .txf3
This is also an odd move, be
Korchnoi - Topalov cause there was no good way for
Vienna 1 996 White to avoid this 'changing col
our' exchange. It seems better to
l lbf3 d6 2 c4 .tg4 3 e4?! (D) play 4 . . lbc6 and wait to see if
.
White would accommodate Black

further by playing h2-h3 at some
point.
5 'ixf3 lbc6
6 d3 g6
7 g3 i.. g7
8 i.. g2 e6!
A deployment seen frequently
in this type of position. Black' s
king's bishop will work o n the long
diagonal, while his king's knight
may later manoeuvre to the hole at
B d4.
Changing the Colour of a Bishop 109
9 0-0 lLlge7
10 .te3 0-0
11 'it'e2
The white queen clears the way
for the subsequent advance of the
f-pawn.
11 lIb8
12 d2 lLld4
13 h1 lLlec6
14 f4
It was also possible to exchange
off Black's king's bishop with 1 4 W
.th6, but Black's king position
would remain reasonably safe, 2S 4JgS .td4!? (D)
while White would then have one Topalov decides that the time is
less minor piece capable of fight right to ' finish off' Korchnoi with
ing for control of d4. tactics, but this move strikes me as
14 ... fS! too ambitious. Simply 25 . . d7
.
Another typical move; now (defending both e6 and h7) would

White 's queen 's bishop remains have kept Black's advantage.
locked behind the pawn at f4.
1S :ae1 'it'aS
16 .tg1 h8
17 h3 a6
18 g4 litbe8
Topalov centralizes his rooks
and waits to see how Korchnoi re
solves the central tension.
19 .th2 d8
20 exfS gxfS
21 gxfS lLlxfS (D)
White 's exchanges have opened
his bishop's diagonal and the e4- W
square, but Black now has the bet
ter pawn structure and chances of 26 4Jxh7!?
later playing a knight into g3 or h4. Korchnoi launches a n obscure
22 lLle4 lLlcd4 attack that takes all of Topalov's
23 d1 lLlh4 defensive skills to meet. Yet it
24 :g1 lLldfS seems to me that a better approach
110 Changing the Colour of a Bishop
was 26 tZJxe6 ! (26 .l::t g fl ? tZJxg2 such a simple trap. Instead, White
-+) 26 . . . l he6 27 I he6 i. xgl 28 should have proceeded more sim
'i!Vxg l litg8 29 ':e2 and White ply with 35 b4, when he can nibble
seems on the verge of unravelling - away at Black's king's pawn cover.
Mayer. 35 ':xe6
26 'it>xh7 36 i.xf5 ':xf5
27 iVh5+ 'it>g8 37 'i;xf5 'i;d7!
28 i.e4+ i.g7 0 1
-
29 l::txg7+! 'it>xg7 There ' s nothing better, as Topa

30 ':gl+ 'it>f6 lov threatens both the bishop and
31 i.g3 tZJg6 the win of the queen via . . . lle 1 +.
32 i.e1 'it>e7
33 l::txg6 'it>d7 (D) It should be noted that a bishop
pinning a knight also changes its
colour. The entire Winawer French
is built around this idea, which
serves to strengthen Black's play
on the light squares.
Diez del Corral - Petrosian

Palma de Mallorca 1 969
1 e4 e6 2 d4 d5 3 tZJc3 i.b4 4 e5 c5
5 a3 i.xc3+ 6 bxc3 'i!Vc7 7 'iVg4 f5
8 'iVg3 cxd4 9 cxd4 tZJe7 10 i.d2
W 0-0 1 1 i.d3 b6 12 tZJe2 ':f7 (D)
Topalov has managed to save his

king, but Korchnoi 's play is justi
fied, as he has succeeded in dis
rupting the flow of the game and
can look forward to practical com
pensation in the form of his bish-
ops.
34 'iVg4 'it>c8
35 ':xe6??
This i s tempting, but Korchnoi
was undoubtedly in his customary
time pressure to have fallen into W
A prophylactic defence of g7. In

Informato r 8, Petrosian indicates
that White should now respond
with 1 3 h4 ! . In ECO C (First Edi
tion), Ivkov assesses that position
as ;t.
13 0-0 .ta6
The point of Black's play be
comes clear. Petrosian has the bar
rier of pawns on the light squares
that is typical of the French De
fence, so he aims to exchange off W
his queen 's bishop. After the ex
change, Black will be left with two a) 1 9 'iHxg6 f4 20 i.c 1 lLJxd4
knights, i.e., pieces that can play on should be fine for Black - Mayer.
light squares, vs a knight and a b) 1 9 ltac l f4 20 .txf4 lLJxd4
dark-squared bishop. 2 1 ltfe l ;t - Petrosian.
14 h4 .txd3 19 'iHe8
15 cxd3!? 20 g3 ltc7
This takes away c 4 from Black 21 g2 f7
and it is also useful that the c-pawn 22 llhl (D)
is no longer backward on an open
file. However, Black now has a
queenside maj ority and can aim
to exploit the light-square weak
nesses at b3 and d3 .
15 lLJbc6
16 .te3 d7
17 lLJf4 lLJg6!?
Petrosian finds a fascinating
method of kingside defence. Also
feasible was 1 7 . . . ltc 8 , when Pet
rosian considers the position un
clear. B
18 lLJxg6 hxg6 (D)
19 'iHf4? ! 22 ... f8! !
This is a mistake, after which Few players have had a s good a
Petrosian has the advantage. There 'feel for the king' as Petrosian. His
were two alternatives: kingside is somewhat porous and
White ' s space advantage might crashes through on the kingside

start to mean something after the only to find that the chicken has
further h4-h5 . With the text-move, flown the coop. Although material
Black' s king starts a long journey is still even, Black's positional
away from the danger zone. trumps (better pawn structure and
23 'iHg5 'it>e8 better minor piece) are such that
24 l:.ac1 'it>d7 White must find something in the
25 h5 gxh5 middlegame, as endings are likely
26 ':'xh5 l:.g8 to be lost without much of a fight.
27 ':'h7 'it>c8 29 l:.h8 l:.xh8
28 '1i'h4 (D) 30 'iHxh8+ 'it>b7
31 'iVt'S l:Ic8
32 'iHd6 'iVe8
33 a4 l:Id8
34 'iHa3 'iHe7
Offering an exchange of queens,
the better to put the trumps to work
in the ending.
35 'iHc3 ':'c8
36 J.d2 g5
37 'iHc2 f4!
38 gxf4 gxf4
39 J.xf4 l:Ig8+
B 40 J.g3 liJxd4
41 'iHc3 liJe2
28 g6 42 'iHc6+ 'it>b8 (D)
Petrosian marks this as dubious
and instead offers '28 . . 'it>b7 ! ' with
.
out further elaboration. There was

also a long forcing sequence that
he undoubtedly saw but didn ' t
mention : 28 . g5 ! ? 2 9 l:.xf7 gxh4
. .
30 ':'xc7+ 'it>xc7 3 1 l:.h l (resolving

the tension) 3 1 .. .hxg3 32 fxg3 l:.g4
33 l:Ih4, when Black is better but it
is unclear whether he can win -
Mayer.
One thing that's impressive
about this game is the way White W
43 ':e1 Changing the colour of a bishop

White gives the knight a kick, as is usually done for long-term play
43 ':d l 'iWg7 ! is even worse. of a positional nature, but it can
43 4Jf4+ also be used to launch a quick at
44 f1 4Jxd3 tack.
45 ':b1 t7!
46 '*'d6+ b7 Kasparov Nunn -
47 e2 li e8 ! Lucerne Olympiad 1 982

Here and on the next couple of
moves the knight is safe from 1 d4 4Jf6 2 e4 e6 3 4Je3 e5 4 d5
xd3 , because then Black plays exd5 5 exd5 d6 6 e4 g6 7 f4 i.g7 8
. . . '*' h7+. J.b5+! 4Jfd7 (D)
48 as ':e2+ The variation chosen by Kaspar
49 f1 ov must be met with care, as both
Alternatives also lose, for exam 8 . . . i.d7 and 8 . . 4Jbd7 allow 9 e5 !
.
ple, 49 d 1 lia2 ! or 49 e3 4Jxf2 ! with great effect.

- Petrosian.
49 4Jxf2 (D)
W 9 a4!
There was a time when White
50 ':xb6+ invariably played 9 i.e2 or 9 i.d3
White hopes that a perpetual so as to get out of the way of
check opportunity will surface, but Black's queenside majority. How
Black is safe: ever, it was eventually determined
50 ... axb6 51 xb6+ e8 52 that the text is the most flexible
a6+ b8 53 b6+ b7 54 move, as White will nearly always
d6+ e7 0-1 want to restrict Black's b-pawn by
a4, while it is not yet clear which 1 2 ... i.xd7

square is best for the bishop. 13 fS!
9 4Ja6 The point of Kasparov 's play.
10 4Jf3 4Jb4 Normally this sort of advance
11 0-0 a6? (D) would be considered highly anti
Nunn puts the question to the positional, as it makes the e-pawn
bishop. Instead, 1 1 . . .0-0 1 2 lite I ! , backward and leaves e5 as a hole .
clearing fl for the bishop's retreat, However, in the given position,
favours White - Kasparov in The Nunn i s unable to steer a knight to
Test of Time . Nunn had earlier e5 , nor is he able to mount any
given this as unpleasant for Black quick pressure on the e-pawn. In
in The Benon ifor the Tournament the meantime, Kasparov is getting
Player, so the text was an attempt ready to storm the black king posi
at improving Black' s play, hoping tion.
to kick back the bishop before the 13 0-0
fl-square becomes available. Castling into the storm on the
kingside. Kasparov considers the
following alternatives :
a ) 1 3 . . . gxf5 14 i.g5 f6 (White
is also better after 14 . . . i.f6 1 5 i.f4
O-O ! 1 6 e5 ! - Kasparov) 1 5 i.f4
'fic7 1 6 4J d2 ! 0-0-0 (or 1 6 . . . 4J d3
17 i.xd6 ! 'fixd6 1 8 lbc4) 17 4Jc4
should make Black very unhappy.
b) 1 3 . . . c4 (aiming at . . . 4Jd3) 14
i.g5 'iVb6+ 1 5 'ith l 4J d3 1 6 f6
i.f8 1 7 a5 ! 4Jf2+ 1 8 1hf2 'ixf2 1 9
4J a4 +-, with a winning attack -
W Kasparov.
14 i:gS f6
12 i.xd7+! A nightmarishly ugly move to
Kasparov hits the nail on the have to make, but Kasparov com
head ! This exchange, which would ments that 'Black has no way of
fit into the chapter on 'The Unex opposing' White's attack after
pected Exchange ' equally well, al 1 4 . . . i.f6 1 5 'iVd2 . He adds that
lows White to shift the attack to the White keeps ' an enormous posi
dark squares, where he will now tional advantage' after the continu
have a three to two minor-piece ad ation 14 . . . i.d4+ 1 5 'ith l f6 1 6 i.h6
vantage . lte8 17 ltc 1 ! .
The exchange at move 1 2 has al "Or 19 . . 'iVe5 20 'iYg4 : f7 2 1

.
lowed White to develop an attack 4Jh6+" - Kasparov.

that would normally not be effec 20 4Jxe4 'it>h8
tive, as Black would use his addi Black also loses after 20 . . J l ae8
tional knight to control eS and f6. 21 'iVg4 'it>h8 (2 1 . . J:t xe4 22 4Jh6+ !)
15 .tf4 gxf5 22 4Jxc5 +- Kasparov.
-
Another very ugly move. Kas 21 4Jxc5 1-0

parov recommends 15 . . . 'iYe7 , al White wins after 2 1 . . : xd5 22
though he grants that White's game 'it'xd5 4Jxd5 23 4Je6 - Kasparov.
is extremely good after both 1 6 :tel
and 1 6 fxg6 hxg6 1 7 4J h4 'it>h7 1 8 The colour of a bishop can be
.tg3 . changed for relatively subtle rea
16 .txd6 .txa4 sons, as in the following game.
Kasparov remarks that ' things
were not essentially changed' by Suba - Farago
16 .. .l::t e 8 17 .txc5 fxe4 1 8 4J d4 Prague Zonal 1 985
4Jd3 19 4Jxe4! +-.
17 :xa4 'iVxd6 (D) 1 c4 4Jf6 2 4Jc3 e5 3 4Jf3 4Jc6 4 d3
d6 5 g3 g6 6 .tg2 .tg7 7 0-0 0-0 8
.tg5!? (D)
w
18 4Jh4! B
"The knight triumphantly estab
lishes itself at f5, where its value is White prepares to change the
immediately increased several times colour of his bishop by exchanging
over." - Kasparov. the knight at f6. Suba decides on
18 fxe4 this course even though Black
19 4Jf5 'iVd7 doesn ' t have any weaknesses on
the light squares, as his own play is 21 lLlce4! (D)

directed toward the central light
squares.
8 h6
9 ..txf6 xf6
10 l':tb1 a5
11 a3 lLld4
12 lLld2
The exchange 1 2 lLlxd4 ? ! exd4
is unfavourable for White, as Black
could subsequently develop pres
sure on the e-file. With the text,
Suba points his king's knight at the
e4-square; his other kni ght is al B
ready leaning on the d5-square.
12 c6 21 lta7?
This blocks the long diagonal An unfortunate mistake that
and takes away d5 from White, but leaves White significantly better.
it also presents a target for White's In lnformato r 39, Suba gives
queenside advance. 2 L..lLlxd3 ! 22 ltb7 'ii'd 8 23 lLlb3 (23
13 b4 axb4 lLlxd6? lLlxf2 ! - Mayer) 23 . . . i.c 8 !
14 axb4 WHd8 2 4 ltb8 lLlc5 (24 . . . WHc7? 2 5 ltxc8)
15 e3 lLle6 25 lLlbxc5 dxc5 26 lt8b6 ;t as best
16 'fIc2 'fIc7 play for both sides. However, in
17 b5 ..td7 this line Graham Burgess points
Black overprotects c6, since it is out that 24 .. .f5 ! is far more critical.
the key to his defence of the light The point is that Black is ready to
squares. The strategic battle in this secure his knight with . . . e4 if the
game is well-conducted by both white knight at e4 should move,
sides and White must be careful while 25 WH d3 fxe4 followed by
that Black doesn' t manage to dem 26 . . . WHc7 leaves the white rook at
onstrate that his bishops and space b8 in trouble. Maybe White should
advantage on the kingside are more try 26 ..txe4 c7 27 :b4, but then
important than White 's light 27 . . . c5 ! 2 8 :b5 i.a6 nets an ex
square play and queenside pres- change for what looks like inade
sure. quate compensation.
18 ltb4 lLlc5 22 lLlxc5 dxc5
19 l':ttb1 l':ta3 23 ltb6 f5
20 bxc6 bxc6 24 'iib 3 d6
25 .tf1! h7
Suba gives 25 . . . e4? 26 lld l !
exd3 27 lLIn as clearly better for
White.
26 lIh7 l':ta6
27 .th3!
Intending 28 lLIe4.
27 .tc8
28 l:.b8 lIa7
29 e4!
Stepping up the pressure on the
light squares and preparing to cre B
ate a target at fS.
29 lIat7 .txf8 41 'ifh5 (the immediate 4 1
30 'ifc2 .te6 'ifn i s also winning) 41 f4 (Suba

31 ':'xfS l:.xfS points out that White wins the end

32 exf5 gxf5 ing after 4 1 . .. e4 42 .txfS+ .txf5
Now the pawn at fS is attack 43 'iVxfS+ 'iHxfS 44 lL1xfS exd3 4S
able, but 32 . . . .txfS 33 lLIe4 is f1 ) 42 .txe6 'ixe6 43 'iff3 'iff6
clearly better for White - Suba. 44 g2 g8 45 'iVg4+ f7 46
33 lLIo .tf6 'iHd7+ iLe7 47 lL1f5 fxg3 48 fxg3!
34 l:.b7+ h8 1-0
35 'iHd2 .tg7 White simply chops everything
36 lLIh4 (D) at e7, brings his king to e4 and wins
Suba evaluates the position as due to his outside pawn on the
winning for White. Black's fS kingside.
pawn will eventually fall and even
eS is a target in some variations. Verdict: Bishops that change their
The rest of the game is conducted colour by pins or exchanges
well by White: 36 .tc8 37 l:.b3
against opposing knights can influ
h7 38 'ie2! (the queen heads for ence events on the colour complex
hS) 38 'iff6 39 l:.b8 .te6 40 ':'xfS
from which they are barred.
1 0 Increasing the Speed of
Your Knights
Every chess player knows that Knights are short-range pieces,

bishops like open positions, while which means that they are slow
knights prefer closed positions. In when they try to move across vast
general, this is true, but not always. expanses of the board. However, if
John Watson elucidated an excep one side surrenders a bishop for a
tion in his opening book Queen 's knight, it frequently results in a
Gambit: Chigorin Defence : gain of time, for example, the side
Given that a game is in the open that gains the bishops must first
ing or early middlegame stage (i. e. move a pawn to provoke the ex
that not too many pieces and change or the side that gives up the
pawns have been traded), it tends bishop may subsequently develop
to be desirable for he who has ac tactical threats that require time
qu ired the two bishops to immedi loss by the owner of the bishops.
ately close the position, andfor he The side that yields the bishops can
who gains the knights to immedi be seen as increasing the speed of
ately open it. his knights , as time hasn ' t been
Watson's point is that it usually spent on the bishops, e.g., retreat
takes a certain amount of time to ing in response to a pawn attack.
acquire the bishop pair in the open Let's proceed to some examples.
ing or early rniddlegame. Hence, the
side with the bishops is likely to fall Spassky - Petrosian
behind in development, in which Moscow, World Ch match (19)
case an opening of the game is the 1 969
last thing to be desired. Long-term,
there can be no doubt that the 1 e4 c5 2 tZJf3 d6 3 d4 cxd4 4
bishop pair will do better in an tZJxd4 tZJf6 5 tZJc3 a6 6 .tg5 tZJbd7
open position than in a closed posi 7 .tc4 WHa5 8 'id2 h6 (D)
tion, but the side with the bishop Putting the question to the
pair must first get developed. Thus, bishop . White decides to increase
what we might call 'Watson ' s the speed of his knights.
Rule' , which i s stated above. 9 .txf6! tZJxf6
Increasing the Speed of Your Knights 119
W W
10 0-0-0 e6 .i.b3 and 14 'it>bl ) . Black is cer

This move is the ' normal ' ap tainly behind in development and
proach, but the usual doesn' t al has had to play some passive moves
ways apply when things become to defend against White's inten
unusual. In Informator 7, Geller pre tions, but with enough delay, he
fers 1 O . . . e5 ! ? 1 1 lLlf5 .i.xf5 , while could hope to start putting his posi
he thinks that the text leaves Black tional trumps (bishop pair, centre
in a clearly inferior position. pawns, queenside initiative) to
11 lihe1 .i.e7 work. Spas sky has timed his attack
In ECO B (Second Edition), Gip well and now heads for the king.
slis considers 1 1 .. . .i.d7 12 f4 0-0-0 15 lLlxg4
1 3 .i.b3 lLleS 14 f5 , which he as This looks exceedingly danger
sesses as clearly better for White. ous, but Geller gives 1 5 . . . lLld7 1 6
12 f4 0-0 g 5 lLlc5 1 7 h4 a s clearly better for
13 .i.b3 lie8 White.
Petrosian prepares to overpro 16 'it'g2 lLlf6
tect e6, as he must always be ready In Spassky 's 1 00 Best Games,
to contend with f4-f5 , aiming to Cafferty mentions 16 . . . e5 17 lLlf5
soften up Black on the a2-gS di 'with strong threats ' .
agonal and open the f-file. 17 :gl i. d7
14 'it>b1 .i.fS (D) Black must be alert for White
15 g4! breaking with either f4-f5 or e4-e5 .
Spassky has completed his de Averbakh suggests 1 7 . . . 'it'c5 , but
velopment, centralized his pieces then Cafferty gives l S lLlf3 i.d7 1 9
and made some moves to get his e5 ! as very good for White.
house in order (for example, 1 3 18 f5 'it>h8
120 Increasing the Speed of Your Knights
Cafferty gives l S . . . e5 'or even 25 l:.g4 'iVh3 26 i.c4 'with a vari

l S . . . exf5 ' as considerably better ety of threats ' - Cafferty.
than the text. 21 e5! (D)
19 Jii dfl A clearance sacrifice to open the
GipsIis assesses this position as e4-square for the knight. Cafferty
clearly better for White. suggests that a slower approach
19 ... d8 with 2 1 liJf3 , threatening 22 e5 and
Yet another point where Petro 22 liJh4, also wins.
sian could have defended better.
Geller gives 19 . . . e5 20 liJe6 ! fxe6
2 1 fxe6 i.xe6 22 l:txf6 +- as hope
less, but he considers 1 9 .. :iVe5 ( ' ! 1 '
- Geller) t o be clearly better, but
not decisive for White. Cafferty
continues this with 20 liJf3 c5
(20 . . . f4 2 1 'iVh3) 2 1 h4, intend
ing 22 liJg5 or liJd l -e3 .
Cafferty independently consid
ers 19 . . . exf5 20 liJxf5 Jt.. xf5 2 1 l:.xf5
'iWdS (the immediate 2 1 . . . .:e5 ! ? al
lows 22 l:.xf6 ! l:.g5 23 xg5 hxg5 B
24 Jii x f7, which gives White a pow
erful initiative on the light squares 21 dxe5
- Mayer) 22 Jt.. x f7 ':e5 and con 22 liJe4 liJh5
cludes that it is superior to the text Black loses after 22 . . . exd4
move. (22 . . . liJxe4 23 l':txfS+) 23 liJxf6
20 fxe6 fxe6 (threatening 24 'iVg6) 23 . . . g5 24
Petrosian could have taken the h3 Jii e 7 25 ':xg5 Jt.. g7 26 1bg7 !
opportunity to remove a pair of cJ;; x g7 27 lIg 1 + and mates - Caf
minor pieces with 20 . . . Jt.. xe6 2 1 ferty.
liJxe6 fxe6, but then White has two 23 g6 exd4
good continuations : This goes down without a fight.
a) 22 liJe2, when White has A better try was 23 . . . liJf4, but then
strong pressure for his pawn - Caf White wins with 24 ':xf4 exf4 25
ferty. liJf3 (intending 26 liJe5 or 26
b) 22 e5 ( ' ! +-' - Geller) liJfg5) 25 . . . b6, and White could
22 . . . dxe5 23 liJe4 liJh5 24 'iVg6 ! solve the problem of his exposed
'iVh4 (24 . . . liJf4 25 ':xf4 exf4 26 c3 , rook with Geller's 26 ':g5 ! ! , with
intending 27 Jt.. c 2 +- - Cafferty) the threats 27 liJf6 and 27 liJe5 .
Increasing the Speed o/ Your Knights 121
24 lDg5 1-0 A principled decision to speed

Black loses his queen, since up the knights . White surrenders
24 . . . hxg5 25 xh5+ g8 26 "'f7+ the bishop, but now Black has
h8 27 :f3 mates. holes in his position at d4, d5, d6
and f6. White also gains a nice lead
Sometimes, the bishop exchange in development, which soon trans
is made in response to a piece at lates itself into a cru shing attack
tack on the bishop. after some further inaccuracies by
Black.
Petrosian - Korchnoi Petrosian ' s exchange is further
Curarao, Candidates ' justified by the fact that Korchnoi
Tournament 1 962 had to violate basic opening princi
ples to provoke the exchange, i.e.,
1 c4 c5 2 lDf3 lDf6 3 d4 cxd4 4 by moving the black queen early in
lDxd4 g6 5 lDc3 d5! ? 6 J. g5 dxc4 the opening.
7 e3 a5? (D) S exf6
9 J.xc4 i.b4!?
More typical would be 9 . . . J.g7.
The intention behind the text-move
would seem to be . . . J.xc3 , which
would prevent the c3-knight from
reaching the hole at d5 , but Korch
noi later blanches and allows the
knight to stay on the board.
10 l:tc1 a6
11 0-0 lDd7?
This should have resulted in se
rious problems. In ECO A (First
w Edition), Gipslis gives 1 1 . . .J.xc3 ;;!;
as superior.
Korchnoi has used an unusual 12 a3?
fifth move in the attempt to elimi A reciprocal mistake. It's sur
nate White's central space advan prising that Petrosian didn ' t jump
tage. In fact, he had reached this at the chance to play 1 2 lDd5 ! ,
position twice in the fifties ; both when White i s simply better.
times he played 7 . . . J.g7 and won, 12 ... i.e7?
so it's obscure why he felt the need This wasn ' t a good day for
to innovate here. Korchnoi. 1 2 . . . i.xc3 ;;!; was still in
S i.xf6! dicated.
13 b4! e5 19 liJdc7+ e7
Of course, 1 3 . . :iVxa3 ?? 1 4 liJd5 White also wins after 1 9 . . . .txc7
wins for White, as he threatens a 20 liJxc7+ d8 (seeking shelter
fork at c7 and also the trapping of behind his own pieces) 2 1 'iVe6 !
the queen with 15 ':a1 'iVb2 16 ':a2. +-, luring the rook to f8 after, e.g.,
14 f4! S (D) 2 1 . . .'iVb6 (or 2 1 . . .'iVxc7 22 :'fd 1 ! )
Black's queen must continue her 2 2 ':c6 'iVxc7 2 3 ':d I ! and the
fast, as 14 .. :xe3+?? 1 5 h 1 and black position cracks, for example,
1 6 :'f3 wins the queen. 23 . . . .:f8 24 ':xc7 xc7 25 'iVd6+
b7 (25 . . . d8 26 'iVxf8+) 26 l:.c 1
l:. a7 27 l:.c7+ a8 28 ':xd7 :'xd7
29 'iVxf8 - Mayer.
20 liJd4 f8
21 liJxaS 1-0
The bishop exchange for in

creased knight speed can also
come in response to tempi-losing
manoeuvring.
Polugaevsky - Gheorghiu
W Palma de Mallorca 1 972
15 .txf7+! 1 c4 c5 2 liJf3 liJf6 3 liJc3 d5 4

A nice sacrifice that cuts Black cxd5 liJxd5 5 g3 liJc6 6 J.g2 liJc7
open on the light squares. Gipslis 7 d3 e5 S liJd2! (D)
assesses the position as winning
for White.
15 xf'7
16 3+ eS
17 liJd5 .td6
1S liJe6 b5
This doesn't help Black, but nei
ther does 1 8 . . . f7 1 9 liJdc7 .txc7
20 liJxc7+ g7 2 1 e6 ! liJe5 22
e7+ liJf7 23 liJxa8 'iVxa8 24 ':c7
':f8 25 l:.d 1 ! and the twin threats of
26 l:.d8 ! and 26 l:. d6 are decisive -
Mayer. B
The text-move intends to shatter must first complete his develop

the black pawn structure with 9 ment.
..txc6+ ! bxc6 10 t2J c4, when prac 11 ..te7
tice indicates that White has a com 12 f4! exf4
fortable advantage. 13 ..txf4 t2Je6 (D)
Black has established a Ma The attempt to develop normally
roczy Bind, which i s defined by with 1 3 . . . 0-0 loses material to 1 4
his pawns at e5 and c5 in a posi ..txc7 Wixc7 1 5 t2J d5 - POlugaev
tion where the c-pawn has been sky.
swapped for the d-pawn. This
grants Black a space advantage
which, given time, he will use to
stifle White. The key phrase here is
' given time ' , for White's lead in
development (a temporary advan
tage) gives him a window of op
portunity in which to create
counterplay.
8 ... Wid7!?
Gheorghiu defends against the
positional threat 9 ..txc6+; he as
sumes that he will be able to w
straighten out his cumbersome de
velopment later by fianchettoing 14 as!
the queen' s bishop. In Grandmas "By giving up one of his bish
ter Performance, Polugaevsky in ops, White makes prompt use of
dicates that neither 8 . ..td7 nor
. . his lead in development. Moreover,
8 . . . f6 is significantly better than the he continues playing on both
text. wings, which in principle is always
9 0-0 b6 promising for the side with the in
10 t2Jc4 f6 itiative." - Polugaevsky.
11 a4! Gheorghiu has lost time with
This discourages Black from multiple moves of his knight -
playing a later . . . ..tb7 , because he four, to be exact - so the ensuing
would have to be worried about the exchange at f4 costs him two
a-pawn playing down to a6. It's tempi, as the bishop has only
true that the text weakens b4, but moved once. Polugaevsky ' s ap
Black will never have time to ex proach circumvents Black's build
ploit this in the present game, as he up, which was slow but intended
to consolidate a space advantage. "when Black is bound hand and

By surrendering the bishop pair, foot." Indeed, this looks winning;
Polugaevsky gains a large enough Black can't play 20 ... 0-0 21 ':xc6 ! ,
lead in development that Black 's but the threat i n the meantime i s 2 1
space advantage is meaningless, l:.e4, winning material.
since White ' s knights and other 18 l'Ie4+! (D)
pieces prove able to gain active
play.
14 tbxf4
15 l:.xf4 ':b8
16 axb6 axb6
17 tbd5
"Increasing the tempo of the of
fensive, and not allowing the oppo
nent a respite. After all, White must
constantly remember: if Black had,
for example, castled, he would be
perfectly all right." - Polugaevsky.
It should be stressed again that B
Gheorghiu is playing for various
long-term (or 'static' ) advantages, The rook is now effectively on
such as space, the better pawn an open file, despite the fact that it
structure and the bishop pair (as would be on a closed file if it were
suming that the bishops are likely at e l .
to prove better than the knights) . 18
rJ;;f7
Polugaevsky is using the short Polugaevsky gives the line
term ( 'dynamic' ) advantages in 18 . . . rJ;;f8 19 tbf4 'Wic7 20 ':e6 tbd4
herent in his lead in development 2 1 tbd5 +-, but I don' t see a win
and greater piece activity. Given after 2 1 . . :'f7 ! , when 22 ':d6 Jt.. g4 !
time to consolidate, Black will creates counterplay and 22 ':e4
probably be beUer, since White's calls White's 20th move into ques
short-term advantages will evapo tion - Mayer.
rate. Thus, it is Polugaevsky's task 19 'fib3! tbd4
to convert his short-term advan It was better to play 1 9 . . . rJ;;f8, al
tages into some sort of long-term though Polugaevsky comments
advantage. that Black would be 'under severe
17 Jt.. d8 pressure' after 20 'iVb5 .
POlugaevsky gives 17 . . . b5 1 8 20 ':xd4 cxd4 (D)
tbcb6 'iVd6 1 9 tbxc8 ':xc8 20 l:.a6 21 tbb4
Increasing the Speed o/ Your Knights 125
after, for example, (22 liJc6 I:.b7

23 I:.aS) 23 ... .tc7 24 I:.xcS+ 'iVxc8
2S a3 + 'it>f7 26 .tdS+ +- or
23 . . . .te7 24 liJ4eS ! fxeS 2S liJxeS
'iVeS 26 I:.xcS .tdS and now 27
.txb7 wins the house, while Bur
gess points out the sadistic 27
'iVb4+ I:.e7 2S .tc6.
22 liJc6 c7
23 liJxb8 'ilixb8
24 5 c7
W Polugaevsky points out that
24 . . . .tb7 loses to 2S .txb7 xb7
Polugaevsky gives this move an 26 liJd6 e7 27 'iVfS+ 'it>h6 2S
exclamation mark and comments 'iVh3+ 'it>gS 29 h4+ 'it>g6 30
that 2 1 liJdxb6 'it>g6 or 2 1 . . . 'it>eS 'ilig4+ 'it>h6 3 I liJfS#.
would ' sharply reduce' the tempo 25 'iVd5 (D)
of White's attack.
In fact, it appears that he missed
a pretty win with the thematic 2 1
.th3 ! , when the black position col
lapses :
a) 2 1 . . :iUxh3? 22 liJf4 h6
(other squares drop the queen at
least to a discovered check) 23
liJd6++ mating.
b) 2 1 . . . c6 22 liJb4 'iic 7 23
.txcs l'hcs 24 lba6 ! +-, forcing the
queen off the rook at cS - Mayer.
21
'it>g6 B
Gheorghiu hopes to escape the
centre and later drop the king back 25 . I:. e8
to h7 . Polugaevsky gives the vari White ' s position is winning, as
ation 2 1 . . .'it>eS 22 liJc6 I:.b7 23 I:.aS the black king is in extreme danger
+-, but as 22 .tc6 wins the queen, and the knight is ready to play to
it seems likely that he meant d6. Polugaevsky examines the al
2 1 . . .'it>fS. ternatives :
Even assuming he meant a ) 2S . . .'i1HcS 26 :taS + - wins
2 1 . . .'it>fS , White is still winning material.
b) 25 . . . i.e7 26 i.f3 ! threatens temporary increase in piece activ

27 'iVh5#. Black is lost in all vari ity, but that the side with the bish
ations: ops will eventually manage to
bl) 26 . . .f5 27 lLle5+ 'it>f6 28 neutralize the knights' speed, after
'ilif7+ 'it>xe5 29 'iVxg7+ +-. which the bishops may become the
b2) 26 . . . h5 27 :f1 ! i.c5 28 dominant positional feature.
i.e4+ 'it>h6 29 ':'f5 ! i.xf5 30 'iVxf5
'ilif7 3 1 lLle5 'iVe8 32 WNf4+ mates. A. Anderssen - L. Paulsen
b3) 26 ... 'iVc5 is relatively best, Vienna 1 873
but White wins the ending after 27
lLlxb6 ! . 1 e4 e5 2 lLlf3 d6 3 d4 exd4 4
26 lLld6 l:.e5 'iVxd4! ? lLlc6 5 Jt.. b5 Jt.. d7 6 Jt.. xc6
White wins after 26 . . . :xe2 27 Jt.. xc6 7 Jt.. g5 (D)
Jt.. e4+ ':'xe4 28 xe4+ f5 29 WNe8+
+- - Polugaevsky.
27 i.e4+ 'it>h5
28 g8! f5
29 lLlrl fxe4
30 WNxh7+ 'it>g4
31 :n :f5
Or 3 1 . . . g5 32 'it>g2, mating with
h2-h3 - Polugaevsky.
32 'ilt'g6+ :g5
33 WNxe4+ 'it>h5
34 7+
An acceptable way to mate , B
however 34 WN h4+ 'it>g6 3 5 lLlh8#!
would have been more fitting in Anderssen has adopted an ap
my view. proach that allows him to maxi
.
34 'it>g4 mize his development at the cost of
35 WNh4# ( 1-0) the bishop pair. Further, he has a
Polugaevsky's play in this game nice space advantage in the centre,
was a model of dynamic play and as his centre pawn is on the fourth
increasing Knight Speed. rank, while Black's centre pawn is
only on the third rank.
One thing that the player with 7 lLlf6
the speedy knights must be care 8 lLlc3 Jt.. e7
ful about is the risk that surrender 9 0-0-0 0-0
ing the bishops will lead to a 10 :he1 :e8
One difficulty Black faces is that for the knights) . Anderssen' s next
his heavy pieces have trouble oper move is particularly good; he sim
ating in the centre. By comparison, plifies and ensures that Paulsen
White ' s heavy pieces are placed cannot subsequently deflect the
comfortably on the e- and d-files. white rook from the d-file.
This is a result of White's previously 16 ':xe8! .txe8
mentioned space advantage. 17 llJd2!
11 bl .td7 Anderssen's reputation today is
12 J.xf6! J.xf6 as a dangerous attacker and imagi
13 e5 J.e7 native tactician. As this game
14 llJd5 J.f8 shows, he could also play very so
15 exd6 (D) phisticated positional chess. With
the text, he starts his knight on a
journey to c3 , where it will supple
ment the knight at dS.
17 .tc6
18 llJe4 f5
19 llJec3 d7
20 a3 f7
21 h3 a6
22 g4 .l:te8
23 f4 ':e6
24 g5! (D)
B
15 cxd6
Black just loses a pawn after
l S . . . .txd6 1 6 llJxc7 ! .
The use o f tactics has allowed
Anderssen to carve out weaknesses
in the black centre, particularly at
d6 and dS . This early game is a
wonderful example of transform
ing a dynamic advantage (the lead
in development resulting from B
surrendering the bishop pair) into
a static advantage (better pawn Paulsen has avoided a pawn
structure and good centre squares storm for the moment, but now
White controls f6 and his threat of 34 liJef6+! gxf6

advancing the h-pawn down to h6 3S liJxf6+ 'it>f7
will soon make matters worse for 36 I:.xh7+ i. g7
Black. 37 I:.xg7+! 'it>xg7
24 bS 38 liJxe8+ 'it>f8
2S h4 I:.e8 39 'it'xfS+ i.xfS
26 'iVd3 l%.b8 40 liJxd6 1-0
27 hS as Unfortunately, I don' t have ac
28 b4! cess to primary sources with re
Another strong move. Anderssen spect to this game. In Masters of
prevents any speculative attacks the Chessboard, Reti indicates that
based on the pawn sacrifice . . . b5- Black resigned here, but some da
b4, while also guaranteeing that the tabase sources suggest that Black
pawn at b5 will be a fixed weak played on for a while with 40 . . . i.d7
ness in any endgame that arises. 4 1 liJe4 'it>g7 42 liJg3 'it>f7 43 f5
28 axb4 'it>e7 44 'it>c l 'it>d6 45 g6 i.e8 46
29 axb4 xhS 'it>d2 'it>d5 47 'it>d3 'it>e5 48 'it>e3
30 'ilixfS 'V$f7 'it>d5 49 'it>f4 'it>c4 50 liJe4 'it>xb4 5 1
31 'it'd3 i.d7 liJd6 i.c6 5 2 f6 'it>c3 5 3 liJxb5+
Paulsen prevented the white h i.xb5 54 f7 before resigning.
pawn from coming down to h6, but
he has had to yield a half-open h It is rather unusual for Black to
file and give White a potentially adopt a strategy by which he yields
mobile f-pawn. the bishops for increased knight
32 liJe4 fS speed. This isn't surprising, since
33 I:.h1 I:.e8 (D) knight speed is a dynamic feature
of a position rooted in time (tempo).
Just as there are very few reason
able opening gambits for Black,
there are few cases where he yields
the bishops for the specific reason
of increased knight speed. Many of
the cases where Black does play
for increased knight speed arise out
of the Nimzo-Indian, where Black
has an intrinsic possibility of gain
ing a lead in development, as he is
busy developing his kingside, while
w White's early development is on the
queenside. As a result, White can In Informator 49, de Firmian

never castle kingside before move and Fedorowicz give both 1 0
seven, which presents a built-in dxe6? i.xe6 and 1 0 e4 exdS a s fa-
window of opportunity for Black vourable for Black.
to catch White's king in the centre. 10 ltJa4
11 'iWb3 bS
Miles - de Firmian 12 e4 a6
Manila Interzonal 1 990 13 ltJe2 0-0
14 .tgS h6
1 d4 ltJf6 2 c4 e6 3 ltJc3 i.b4 4 IS .th4
'iWc2 cS S dxcS ltJa6 6 a3 .txc3+ 7 Black's lead in development and
WHxc3 ltJxcs 8 f3 (D) active piece play give him compen
White can also try 8 b4 ltJce4 9 sation for the pawn after I S .txf6
WHd4 dS 10 cS ! ?, when the knight at 'ixf6.
e4 is short on squares. This vari IS exdS
ation is extremely sharp and was 16 eS!? l:.e8 ! (D)
particularly topical at the time of
Miles-de Firmian.
B Black has such a large lead in

development that he is able to offer
8 dS a speculative piece sacrifice with
9 cxdS b6! ? the intent of catching the white
10 b4 king in the centre. The alternative
In New Ideas in the Nimzo-In was 16 . . . gS 1 7 .tf2, which is as
dian Defence, Kosten indicates sessed as clearly better for White
that 10 .tgS ! ? is probably superior by de Firmian and Fedorowicz.
to the text. 17 f4
Awarded an exclamation mark has other options, e.g., 20 . . . a5 ! ? or

by Ftacnik in New in Chess no. 6, 20 . . . WN b6 ! ?, among other tries. In
1 990, while de Firmian and Fe any event, his take on the position
dorowicz mark it with ' ! ?' . It's was that White ' s exposed king po
clear that 17 exf6 d4 gives Black a sition should give Black enduring
strong attack in return for his piece, compensation for his piece and it
but it is not clear if it is adequate: was clear that he favoured Black ' s
a) de Firmian and Fedorowicz practical chances.
consider only 18 l:.d l J.e6 19 'iVbl The critical continuation is
( 1 9 'iVd3 ? tZJb2 20 'ilixd4 tZJxd 1 clearly 1 7 exf6; to date, I don ' t
wins for Black - de Firmian and think the published analysis o n i t i s
Fedorowicz) 1 9 . . . J.c4 gives Black exhaustive, perhaps since d e Fir
a strong (possibly winning) attack mian and Fedorowicz felt that this
for the piece. position could arise again in one of
b) Ftacnik examines 1 8 'iVd3 their games.
tZJb2 1 9 'iVc2 tZJc4, and now rather 17 ... g5
than Ftacnik's 20 l:.d l tZJe3, which 18 J.f2
he assesses as giving Black a de Firmian and Fedorowicz as
strong initiative, White can try 20 sess both 18 fxg5 tZJe4 and 18 exf6
O-O-O ! ? (D), with the point that 'iWxf6 1 9 :d 1 gxh4 20 <it>f2 as
20 . . . tZJe3 2 1 l:.xd4 ! favours White clearly better for Black.
- Mayer. 18 tZJe4
19 J.d4 J.e6
20 'iVf3 :c8
This move passes without com
ment from de Firmian and Fedoro
wicz, but Ftacnik gives it a dubious
mark. Instead, he prefers 20 . . .f5 2 1
exf6 g4 2 2 'iVe3 tZJxf6 2 3 f5 J.xf5 ,
but now rather than 24 WNxh6 <it>f7
(unclear - Ftacnik), it seems to me
that White can win a piece with 24
WNf4 ! , when the exposed position
of White ' s king is offset by the
B even more airy position of the
black king.
However, when I mentioned the 21 f5!
above possibility to Fedorowicz Miles goes for broke. de Firmian
in 1 997, he pointed out that Black and Fedorowicz give 2 1 tZJg3 gxf4
(2 1 . . .f5 22 .id3 is unclear - Ftac Black, as he's able to penetrate on

nik) 22 'ii'xf4 "i!Vg5 23 "i!Vxg5+ hxg5 the e-file with his queen.
24 .id3 liJac3 25 0-0 liJxg3 26 Ftacnik cites analysis by Miles
hxg3 liJe4 27 .ixe4 as equal. that sees Black take a different ap
21
... .td7 (D) proach in the above variation, i.e.,
24 ... ':'c2 25 .ih3 g4 (25 . . Jle7 26
0-0 +- finally sees the king castle
into safety) 26 .ixg4 ':'xe2+ 27
'iUxe2 liJec3 28 'ifxe6 fxe6 29
.ixe6+ 'it>h 7 30 llf1 "i!Ve8 3 1 llf7 +
'it>g6 32 llf6+ 'it>h7 (D) .
W
22 liJg3?
A mistake that allows a very
imaginative sacrifice. The position
is extremely complicated, but ap
parently equal with best play: w
a) 22 "i!Vh5 .ixf5 23 'iUxh6 ':'e6
- Ftacnik. At this point, Ftacnik claims that
b) 22 e6 with the idea of shut Miles gives ' 3 3 'it>f1 , intending 34
ting the e-file: ':'f1 ' (perhaps a typo for 34 l':te 1 or
b I ) 22 . . . .ixe6 was apparently 34 ':'f7+ which are indeed both
de Firmian's intention, when he strong threats) . Of course, this is
expected 23 fxe6 ':'xe6 24 g3 liJac3 doubly garbled, since 3 3 g4 ! grabs
25 .th3 liJd2, which is assessed as control of the h5-square and threat
unclear by de Firmian and Fedoro ens to mate with 34 l':tf7+ 'it>g6 35
wicz. A likely continuation is 26 1:. g7# .
'it>xd2 ':'xe2+ 27 'it>d3 , when White However, when I spoke with Fe
intends to seal the c-file with 2 8 dorowicz years after the game, he
.ic5. It's also possible to try 27 was of the opinion that de Firmian
'iUxe2? ! liJxe2 28 'it>xe2, but then probably would have played the
28 . . . 'ii'e 8+ ! looks favourable for drawing line in 'b2 ' below.
b2) Black can achieve an equal

position with 22 . . .fxe6 23 'iWh5 (23
f6 %:tf8 ! is clearly better for Black -
Ftacnik) 23 . . . e5 (forced) 24 g6+,
when de Firmian and Fedorowicz
indicate that White should take the
perpetual check. Ftacnik analyses
further with 24 . . . 'it>f8 25 xh6+
'it>g8, but now his claim that 26 h4
is unclear is proved wrong by
26 . . . .:c6 ! - Mayer.
c) 22 h4 ! ? is the really interest B
ing line:
c 1 ) 22 . . . gxh4? 23 f6 ! is practi 22 ':xe5!! (D)
cally winning for Black - Ftacnik.
c2) 22 .. :Wic7 ( ' ! ?' - Ftacnik;
also possible is 22 . . . .:c2 ! ?, which
Fedorowicz considers playable - I
agree) 23 hxg5 Wlc2 24 ':d 1 lDac3
25 lDxc3 lDxc3 26 .td3 ( ' ! ' -
Ftacnik; also possible is 26 gxh6
.txf5 - Mayer) 26 . . . Wlxd 1 + 27
'iWxd 1 lDxd 1 2 8 'it>xd 1 %:tc4 (Black
loses after 28 . . . hxg5 29 e6 fxe6 30
':h8+ 'it>f7 3 1 %:th7+ - Ftacnik) 29
i.xc4 dxc4 30 g4 ':d8 (inferior is
30 . . . i.c6 3 1 ':xh6 :d8 32 l':td6 - W
Ftacnik), and now Ftacnik gives 3 1
:xh6 .txf5 as somewhat better for de Firmian uncorks a wonderful
White. However, 3 1 .tb6! (D) looks sacrifice to complement his earlier
stronger here. aggressive play. The skin of centre
Then 3 1 .. . .tc6+ 32 .txd8 .txh 1 pawns is torn away and the white
33 'it>e2 is good for White - Mayer. king is exposed to whatever the
As one can see from these vari black pieces can inflict upon him.
ations, the position is very compli 23 i. e2
cated and Miles lost in part because de Firmian and Fedorowicz only
he played for the win and in part give 23 .txe5 e8, which they as
because he failed to play the criti sess as clearly better for Black.
cal 22 h4 ! ? Ftacnik continues the analysis and
confirms the view that Black is do

ing very well:
a) 24 l:Id l 'it'xe5 25 'ie3 (25
xe4 dxe4 transposes to the next
note) 25 . . . l1c3 and then 26 .td3
tLJb2 or 26 lid3 l:.xd3 27 .txd3
Wia l + makes Black happy - Ftac
nik.
b) 24 tLJxe4 'it'xe5 25 lidl dxe4
(25 . . . tLJc3 is also very good for
Black according to Ftacnik) 26 'ii'e 3
.txf5 is clearly better for B lack - W
Ftacnik.
23 'it'e8 31 .txf5 'ii'e3+
24 tLJh5 32 .td3 e2+
Black still does very well after 33 'it>c2 .tf5
24 .txe5 Wixe5 25 l:.d l l:.c3 or 24 34 11ad1 'ii'e5
l:Id l tLJac3 25 l:.c 1 tLJxg3 26 .txe5 35 'it>d2
tLJgxe2 - Ftacnik. Black also wins after 35 .txf5
24 l:.xf5 'ic3+ 36 'it>bl Wib3+ 37 'it>al 'ii'x a3+
25 'it'e3 l:.c3! 3 8 'it>bl 'it'xb4+ 39 'it>a l 'ii'a 3+ 40
Ftacnik analyses the active 'it>bl tLJc3+ 4 1 'it>c2 xd l 42 ':'xd l
25 . . . l:.c2 to an unclear position. 'it'c5+ - de Firmian and Fedorowicz.
26 .txc3 tLJaxc3 (D) 35 2+
27 .tg4 36 'it>e3 .tg4
White loses after 27 11 f1 (27 37 l:.d2 'it'd4#
'it'd4 f6) 27 . . . l:Ie5 28 11c l tLJxe2 29 (0-1)
'it'xe2 tLJc3 30 tLJf6+ 'it>h8 3 1 tLJxe8
l'he2# - Ftacnik. Verdict: Increasing the speed of
27 d4! your knights is a dynamic opera
28 'ixd4 tLJc5+! tion in which a bishop or both bish
29 'it>d2 tLJb3+ ops are exchanged against knights
30 'it>xc3 xd4 so as gain time.
1 1 The Bad Bishop
Bishops must always contend with matter of some concern. On the

their own pawns getting in their other hand, if it is on the bl -h7 di
way. Having at least one bishop agonal or if it managed to ex
cramped or otherwise hindered is change itself via a6 or by playing to
usually unavoidable, as pawns do g4, then Black may have a very
tend to get set into bishop-blocking comfortable game. At the least, he
pawn chains. When a bishop is suf won't have to worry about a bad
ficiently hampered by its own pawns bishop .
that one takes special note of it in Some openings, e.g. , the French
evaluating a position, it earns the or the King' s Indian, lend them
label ' bad bishop ' . However, there selves to the topic of bad bishops in
is a great distinction to be made be a particularly instructive manner.
tween bad bishops that are 'outside These are openings in which Black
the pawn chain' and those that are practically always has a bad bishop
'behind the pawn chain' . to worry about. Sometimes the
bishop is not a very important fac
tor, while at other times it is obvi
ous to even the weakest or most
inexperienced player that the bad
bishop is a disgrace.
The following game could never
be played among modern grand
masters, but the extreme plight of
Black's queen's bishop is still a
common sight among amateur
players today.
A look at the pawn structure in Tarrasch - von Gottschall

the above diagram doesn' t tell us Dresden 1 892
much about Black's queen's bishop.
We must also know where it is lo 1 e4 e6 2 d4 d5 3 tbd2 tbf6 4 e5
cated; at c8, b7 or d7, it will be be tbfd7 5 .lid3 c5 6 c3 tbc6 7 tbe2
hind the pawn-chain and will be a b6 8 tbf3 .lie7? (D)
The Bad Bishop 135
The correct approach is S . . . cxd4 ! 13 a3 'Wic7?

9 cxd4 f6 ! , when Black has an open Under no circumstances should
file for counterplay and is about to White have been allowed to play
open another. b2-b4, as he will now be in a posi
tion to neutralize any potential
black counterplay in that sector.
Correct was 1 3 . . . c4, followed even
tually by a queenside pawn ad
vance.
14 b4 b6
15 lLlh3 a5
16 b5 lLld8
17 a4 .tb7 (D)
W
von Gottschall prefers to de
velop his kingside and castle, but
this approach, so suited to an open
game, is inappropriate here, as it
leaves Black cramped and without
active chances.
9 0-0 0-0
10 lLlf4! W
With Black eschewing the break
. . . f6, White owns the dark squares Black is strategically lost, be
on the kingside. Here he's getting cause White can build up for a
ready to play his knight to h5 as an breakthrough on the kingside and
aid to a kingside attack. Black can neither prevent it nor
10 cJth8 distract White with counterplay.
11 .tc2 I:.g8 Note that, in addition to Black' s
12 'iiVd3 g6 awful queen's bishop, h e is also
Now Black's kingside dark saddled with a problem knight at
squares are a mess. It was better to dS and even the knight at d7 has
play 1 2 . . . lLlfS, which at least does trouble getting activated.
not create any permanent weak 18 .td2 ':c8
nesses. 19 I:.fc1 cJtg7
136 The Bad Bishop
20 'iUe3 'it>h8 32 .tel

21 i.d3 c4 33 f4
With the queenside closed, there Clears the third rank for a rook
is absolutely no chance of distract lift and may also allow a later
ing White. If there were a way to breakthrough based on g2-g4 and
seal the kingside, then Black might f4-f5 .
be able to angle for a blockaded 33 'it>g8
position that would be difficult for 34 l1e3 f5
White to crack, but as matters stand, 35 exf6 lLlxf6 (D)
White will always be able to keep
the kingside at least semi-open.
22 i.c2 lLlfS
23 'ilVe2 'it>g7
24 'iUe3 'it>h8
25 ':el 'iUd7
26 'it>hl l1c7
27 'ii'h 6 'ii'e8
28 lbfg5 ':g7
29 f3! (D)
W
B lack has finally achieved the
exchange of his f-pawn for the e
pawn, but it comes 25 moves too
late. White is able to build up his
kingside attack and Black can do
little but watch in horror: 36 lbe5
.td6 37 ':f3 'iUe8 38 i.d2 .txe5 39
fxe5 lbg4 40 'iUh4 h5 41 h3 lbh6
B 42 g4 hxg4 43 hxg4 lLldf7 44 'it>g2
lbf5 45 gxf5 gxf5 46 l1g3 lbxg5 47
Preparing to bring the knight to .txg5 'iUf7 48 l:hl 1-0
g4, where it will intensify White 's It is fitting that the variation em
pressure on the dark squares. ployed by Tarrasch in this game (3
29 .tc8 lLld2) bears his name today. He
30 lbf2 lLld7 won many games with it, several of
31 lbg4 'iUg8 which were similar to this game.
The Bad Bishop 137
von Gottschall was hampered by 8 d5 e7

the fact that the methods of solving 9 d2 c5
Black' s problems in the position Is this move a mistake? It's hard
after White ' s move eight are rela to say. In general, Black should
tively sophisticated and had yet to avoid creating a ' Full Benoni '
be worked out systematically. A pawn structure (c4, d5, e4 vs c5 ,
modern master sits down at the d6, e5) if the white king has not
board well-versed in how to create castled kingside, as White 's king
counterplay, but sometimes these may then stay in the centre or castle
methods only make the situation queenside, followed by a pawn
worse. storm on the kingside. However, in
the given situation, White's king
Polugaevsky Am. Rodriguez
- was already relatively safe in the
Toluca Interzonal 1 982 closed centre, so he was preparing
c4-c5 , when his queenside attack is
1 d4 f6 2 c4 g6 3 c3 j"g7 4 e4 considerably faster than it is in the
d6 5 .te2 0-0 6 f3 e5 7 j"e3 main line of the Classical King's
c6?! (D) Indian (7 0-0 lLlc6 8 d5, etc.).
10 g4
Polugaevsky plays to restrain
Black's natural counterplay. In
most King' s Indians, the advance
. . . f5 will be played at some point or
another, with the goal of gaining
space on the kingside. By playing
g2-g4, Polugaevsky makes it clear
that he intends to swap twice on f5
in reply to that plan; this will offer
White attacking chances based on
the g -file, the b 1 -h 7 diagonal and
W the use of the e4-square.
10 ... xg4? !
Already an inaccuracy, as White Rodriguez plays what he hopes
gains greater flexibility in his de will prove to be a freeing combina
ployment in light of the fact that he tion. In Grandmaster Performance,
has yet to commit his king. Further, Polugaevsky relates that he had
with the centre closed, White' s never seen the move played in this
king may b e able to stay in the cen position, so it came as quite a
tre in more or less perfect safety. shock to him.
138 The Bad Bishop
Here we see an instance of an im 13 hxg4 a6

perfect freeing manoeuvre back 14 a4
firing on the originator. There can Polugaevsky makes a very in
be little doubt that Rodriguez is a structive comment:
stronger and better-informed player "At first sight, this seems to be
than von Gottschall, but the end excessive prophylaxis, but ... What
result in this case is 'a little knowl is of primary importance in posi
edge is a dangerous thing' , since tions of this type is the overall stra
Black ends up saddled with an ex tegic plan, and not some move
tremely bad king's bishop. taken in isolation."
11 i..xg4 1'5 (D) White ' s idea is to squeeze the
opponent over the entire board, and
then to exploit his spatial advan
tage and the bad bishop at g7.
14 . i.. d7
15 f3 lbc8?
Polugaevsky gives I S . . . 'iVaS ,
preventing the further advance of
White's a-pawn, as the only move.
16 g5! (D)
W
12 h3! fxg4
Black has a choice of continu
ations that give him an inferior
game:
a) 12 .. .f4 1 3 i.. xcS dxcS 14 i.. xc8
'it'xc8 I S 'i; g4 is clearly better for
White thanks to Black's bad bishop
- Polugaevsky. B
b) 1 2 . . . hS 1 3 i.. x hS gxhS 1 4
i.. gS wins the h-pawn - Polugaev This space gain stops . . . i.. f6 and
sky. . . . i.. gS, aiming either to activate or
c) 1 2 . . . b6 1 3 f3 leads to es sen to exchange off the bad bishop.
tially the same type of position as 16 :b8
the game - Polugaevsky. 17 a5! b5
The Bad Bishop 139
18 axb6 Ibb6 24 WNc7

19 b3 IU4! (D) 25 tbdl !
The knight heads for d3 , where
it will support the pawn advances
b3-b4 and f3-f4.
25 tbb6
26 t2Jf2 tbd7
27 tbd3 lIa8
28 WNc3 1.f8
29 ':gal Ilb6
White wins the a-pawn after
29 . . . tbb8 30 b4 cxb4 3 1 tbxb4
1.b5 32 d3 ! - Polugaevsky.
30 b4 'it'c8
W Polugaevsky indicates that the
only try was the exchange sacri
20 ':gl ! :ti fice 30 . . . cxb4 3 1 i.xb6 WNxb6 32
Both sides showed good judge tbxb4, though he concludes that it
ment concerning the manoeuvre would be inadequate after the fur
. . . 1If4. Black, of course, hoped that ther 32 . . :ilid8 33 "iVe3 tbc5 34 tbd3
White would win the exchange at tbxd3 35 xd3 i.e7 36 ':xa6 Ilxa6
the cost of activating the dark 37 ':xa6 i.xg5 38 'it'b6 +-.
square bishop, while White refused 31 bxc5 tbxc5
to be distracted from his positional 32 tbxc5 dxc5 (D)
build-up.
Polugaevsky notes that 20 . . . 1.f8
(intending 2 1 . . . i.e7) allows White
to play 2 1 e2 i.e7 22 1.xf4 exf4
23 e5 ! and the square e4 is cleared
for the knight. Instead, he recom
mends 20 . . . .:h4 as the move that
offers Black the most chance of
counterplay.
21 e2 1.e8
22 l':r.a3 :a7
23 "iVaI libb7
24 :a5 W
White loses his queen after 24
1:ha6?? :xa6 25 WNxa6 :a7 . 33 ':b1 ! lhbl
140 The Bad Bishop
34 t2Jxbl .td6 idea is to attack on the kingside and

35 t2Jd2 l:. a7 this is what White will do if Black
36 'iVa3 'iib 3 doesn't play . . . hS . However, if
37 lba6 'iVg2+ Black does stop the further ad
38 d3 1-0 vance of the h-pawn, then White
still retains attacking chances, but
So far, we have seen an example can also continue with strict posi
where the side with the bad bishop tional methods, as having the pawn
made no effort to rectify the prob at hS gives Black certain problems.
lem and another example where his 9 h5
cleverness backfired on him. In the 10 t2Jcl e5
following example, he does noth 11 d5 t2Jd4
ing unusual, but makes sure to play 12 t2Jb3 t2Jxb3
actively. Yet in the end, his bishop 13 axb3 c5 (D)
stays locked up and White wins a
very pretty game.
Petursson - W. Watson
Palma de Mal/orca 1 989
1 d4 t2Jf6 2 c4 g6 3 t2Jc3 .tg7 4 e4

d6 5 f3 0-0 6 .te3 t2Jc6 7 t2Jge2 a6
8 WHd2 l:.b8 9 h4! (D)
W
Black couldn't maintain his
knight at d4, so he had to swap it
off. Yet even with a doubled b
pawn, White is still better on the
queenside, so Watson plays to shut
down the sector by playing . . . cS .
The downside is that his king ' s
B bishop is now bad.
It should be noted that the most
This finesse gives White a wide important factor injudging whether
range of possibilities. The basic a bishop is good or bad is to take a
The Bad Bishop 141
look at which pawns cannot be

shifted from the colour they stand
on. For instance, here Black has
three pawns fixed on dark squares,
so despite the fact that he has fewer
pawns on the dark squares than the
light squares, his king's bishop is
the bad bishop, as there is no good
way to liberate it.
14 .te2 liJe8
IS g4 fS
16 gxfS gxfS B
17 exfS .txfS
The main feature of the position Petursson examines several al
is now clear. White has opened the ternatives :
g-file, the bl -h7 diagonal and the a ) 2 2 . . . 1:.f4 23 :tg l ! is clearly
e4-square. If Black is unable to ad better for White.
vance his e-pawn, then his king ' s b) 22 . . . .tf6 23 O-O-O ! , intend
bishop will stay bad and h e will ing 24 1:.dg l , is also clearly better
also face problems on both sides of for White.
the board. c) The best course is 22 . . . ':'f7 !
18 liJe4 .txe4 23 O-O ! (the issue is now over con
In Infonnator 48, Petursson gives trol of the f-file, so 23 O-O-O? is an
1 8 ...liJf6 19 .td3 ;t as an alternative. inaccuracy) 23 . . . 1:.bf8 24 .txg4
The problem with the text is that hxg4 25 lixf7 ':'xf7 26 ':'f1 ;t is
it leaves Black exposed on the light better for White, as he has the plan
squares and practically ensures that 27 ':'xf7 , followed by 'it>g l -g2-g3,
his king's bishop will stay buried. when the pawn at g4 is weak -
Perhaps he will be able to counter Petursson.
attack the e-pawn at some point, 23 O-O!
but White should be able to provide Petursson might also be able to
it with sufficient protection. get away with 23 1:.xa6, but he pre
19 fxe4 liJf6 fers to castle and leave pawn-grab
20 .tf3 V!lVd7 bing for later.
Or 20 . . . V!lVb6 2 1 'i!Vd3 liJg4 22 23 bxc4
.td2 ;t - Petursson. 24 bxc4 1:.b3
21 e2 liJg4 2S 1:.xa6
22 .tgS! (D) The immediate 25 .txg4? is met
22 bS? ! by 25 . . ..l::t g 3+.
142 The Bad Bishop
25
7 (D) 31 'it>g3 lile2 (D)
W W
26 ':a2 32 'it>xg4 Ilxe4+
Solid and good enough to leave 33 'it>f5
White clearly better. But he also Petursson's king is able to stroll
had an interesting alternative avail in on the weakened light squares,
able in 26 .txg4 ! ?, with the point undetected by Black's dark-square
that 26 . . . xa6?? loses to the reply bishop.
27 i.e6+. White also does well af 33 ':xc4
ter 26 . . . I:.g3+ 27 'it>h2, with an ex 34 'it>g6 I:.b4
tra pawn and better bishop, so that 35 h5 I:.b7
leaves 26 . . . hxg4 27 ':xf8 + ! (27 36 h6 i.f8
I hd6? l::t g 3+ 28 'it>h2 :ff3 ! gives 37 lilhl I:.d7
Black a winning attack) 27 . . . .txf8 38 lilbl c4
28 'iix g4, and White mates if 39 I:.b8 I:.c7
28 . . . xa6? by 29 .tf6+ 'it>f7 30 40 .tf6 c3
e6+ 'it>g6 3 1 h5+ ! , etc. - Mayer. 41 i.g7 1-0
26 I:.b8 A cute mating attack in a re
27 'it>g2 I:.xb2 duced ending. White 's extremely
28 lilxb2 xb2 active king deserves special recog
29 .txg4! nition.
Petursson assesses this position The moral of Petursson-Watson
as winning. Inferior was 29 'ii'xb2? is that some bishops are bad and
l hb2+ 30 'it>g3 l1b3 - Petursson. even the best efforts may not be
29 hxg4 enough to solve the problem they
30 "ii'xb2 I:.xb2+ pose.
The Bad Bishop 143
Sometimes a bishop that is well Shirov ' s queenside activity is

placed and unobstructed by its own sufficient to distract White from
pawns 'goes bad' . In the following any kingside build-up he might
game, one of the world's youngest have intended. Furthermore, it
grandmasters has a bishop that turns out that the bishop doesn' t
looks particularly impressive end have a stable square o n the a2-g8
up as bad as any in this book. diagonal. Perhaps White 's appar
ent opening advantage was just
Leko - Shirov that: apparent.
Tilburg 1 996 18 .ta4
19 lIal bxc3
1 e4 e5 2 lLlf3 lLlc6 3 .tb5 a6 4 20 bxc3 lIb2
i.a4 lLlf6 5 0-0 .tc5 6 c3 b5 7 .tb3 21 lLlfl lIxf2
d6 8 a4 .tg4 9 d3 I:.b8 10 axb5 22 'iVxf2 lLlb7 ! (D)
axb5 11 h3 .th5 12 .te3 .txe3 13 Continuing to harass the white
fxe3 .txf3 14 'iVxf3 0-0 (D) bishop.
Leko has emerged from the
opening with an acceptable posi
tion; indeed, the open f-file and his
well-placed bishop would seem to
give him the tools for forging a real
advantage.
w
23 .tc6 lLlc5
24 'iVc2 'ib8
25 I:.bl 'iVa7
26 d4 lLle6
w 27 lLld2 I:.b8
28 .td5 I:.xbl+
15 lLld2 b4 29 WUxbl lLld8 (D)
16 lIf2 I:.a8 An interesting ending has
17 I:.bl lLla5 arisen. Queen and bishop vs queen
144 The Bad Bishop
w B
and knight is generally considered 38 Jt.. d3 g4
favourable for the side with the 39 hxg4 Wld7
knight (see Chapter 1 7) . It' s un 40 f1 Wlxg4
clear whether the addition of a pair 41 Wle2 Wlih4
of knights should be enough to 42 Wlif2 1+
change this assessment in the gen 43 e2 Wlia1
eral case; it certainly doesn't prove 44 Wlie1 Wlia2+
enough in this particular game. 45 Wlid2 Wle6 (D)
30 f2 g6 Leko survived the first charge of
31 lOf3 c6 the black queen, but he is far from
32 .tc4 g7 solving his problems, as the queen
33 .td3 (D) can try to penetrate on both the
The bishop retreats to a defen kingside and the queenside.
sive role, which it will play for the
rest of the game. White's centre is
fragile and Leko is trying not to be
forced into any further weakening
advances.
33 lOe6
34 'iWb2 Wlic7
35 .tc2 h6
36 e2 lOg5!
37 lOxg5 hxg5
Black ' s doubled g-pawn is an
asset, as it can be used to open up
territory on the kingside. w
The Bad Bishop 145
46 'iHc2 c5
47 dxc5
Now any pretence of White ' s
pawns being mobile is gone. Leko
was presumably afraid of the tactic
47 . . . cxd4 48 exd4 exd4 49 cxd4
liJxe4 ! .
47 dxc5
48 'it'e1 WHc6
49 c4 'tWb6
50 'iVc3 liJd7
The knight will eventually reach w
d6, where it will press on the weak
pawns at e4 and c4. It might have pawn would provide him with ade
been nice if the knight didn ' t have quate counterplay. The rest of the
to delay its trip to d6, but Black ' s game is instructive, but doesn't re
structural and positional advan quire comment: 64 'it'f6 65 'iHa5

tages are so huge that he doesn ' t liJxe4 66 .txe4 'iHxe4 67 'iHxc5
have to rush. 'ii'xg2+ 68 'it'd3 'iHn+ 69 'it'd2
51 'it'e2 'ii'e6 'iHf2+ 70 'it'd3 'iHf5+ 71 'it'c3 g5 72
52 .tc2 'iHg4+ 'iHb6+ cJ;g7 73 c5 g4 74 c6 g3 75 c7
53 'it'n 'iHg3 g2 76 e4 WHn 77 'it'd2 g1 WH 78
54 'iHe1 2 'iHxg1 + 'iHxg1 79 c8'ii' 'iHd4+ 80
55 'iHf2 liJf6 'it'e2 'ii'xe4+ 81 'it'f2 'iHf4+ 82 'it'e2
56 'it'e2 '+lVh1 e4 83 'iHc3+ 'it'g6 84 'iHc6+ 'it'g5 85
57 'iHn 'iWh4 'iHc5+ 'it'g4 86 'iHc8+ 'it'g3 87 'ii'c5
58 WHf2 2 'it'g2 88 'iHd5 WHf3+ 89 'it'e1 WHe3+
59 'iHf3 liJe8 0-1
60 .td3 liJd6 (D)
The knight has arrived and will There are several cases in open
add greater meaning to the tacking ing theory where the player with a
of the black queen. particularly bad bishop plays for its
61 'iHn exchange. The most typical case is
62 'iHf2 'iHe7 seen in French Defences, e.g., 1 e4
63 WHn 'tWb7 e6 2 d4 d5 3 liJc3 i.b4 4 e5 b6, in
64 'iHa1 tending 5 . . . .ta6. One might also
Leko elects to take play into a see this in a Modem/King 's Indian
pure queen and pawn ending, pos type of position, for example, 1 d4
sibly because he thought that his c- g6 2 c4 d6 3 e4 e5 4 d5 .th6 ! ?,
146 The Bad Bishop
when Black hopes that the holes on 9 f4 c6

the kingside won ' t miss the dark 10 h3 lLle6
squared bishop. 1 1 f5 lLld4
However, it can also be wrong Now we can see the result of
to swap a bad bishop, as sometimes Scherbakov 's loss of time. Instead
it is performing other valuable de of playing . . . lLlc6 and . . . lLld4, he
fensive duties. The following game has tossed away two entire tempi,
is a good illustration of this. which can be seen as White's king
side space advantage (f2-f4-f5) .
Botvinnik V. Scherbakov
- 12 g4 lLle8
USSR Ch (Moscow) 1 955 13 i.e3 a6
14 'iVd2 b5 (D)
1 c4 lLlf6 2 lLlc3 g6 3 g3 i.g7 4
i.g2 0-0 5 e4 d6 6 lLlge2 e5 7 0-0
lLlbd7 (D)
w
15 i.g5!
w Botvinnik teases Black. The im
mediate 1 5 i.h6 was also possible,
There is nothing wrong with this but then W h ite wouldn' t have any
move, but given that Scherbakov immediate way to penetrate on the
uses a time-consuming method of kingside, as the black king would
playing it to d4, he would have reach g7 . With the text, Botvinnik
done better to play 7 . . . lLlc6. offers an attractive-looking trap.
8 d3 lLlc5 15 i.f6?
Heading for e6. A superior ap Into which Black falls ! It seems
proach is the immediate 8 . c6, in
. . natural to exchange off the king's
tending to initiate queenside play bishop, but now his king position is
with 9 . . . a6 and 1 O . . . b5 . impossible to hold together. It's
The Bad Bishop 147
understandable that he didn't want

to play the superior 1 5 . . . f6, when
his king's bishop is buried, random
queen moves allow 1 6 f6, and
1 5 . . . liJf6 walks into a pin, but the
text allows White a decisive attack.
Further, 1 5 . . . b6 would threaten a
discovery and Black could then
follow up with 1 6 . . . liJf6, when the
knight isn't pinned.
16 i.xf6 liJxf6
17 liJxd4 w
Ensuring that the centre will be
closed before attacking on the 21 g5 liJh5
flank. 22 liJf4 liJg7
17 exd4 23 f6 1-0
18 liJe2 "iVb6
19 cxb5 Verdict: B ad bishops are frequent
This also is intended to keep the occurrences. Their importance var
centre closed before White decen ies from position to position, but a
tralizes his queen. bad bishop inside its own pawn
19 axb5 (D) chain is always a matter for con
20 6! i.d7 cern.
1 2 The Sacrifice for Active
Bis hops
Few attacking forces are as power bishops with his remaining bishop
ful as a pair of active bishops. Two (assuming he has one)? Secondly,
bishops placed optimally on an can the opponent's knight reach a
open board will attack 26 squares, stable post that will serve to blunt
which is nearly as many as the the bishop pair? Finally, does a
queen's 27 squares on an open concrete analysis of the position
board. Of course, a pair of bishops arising after the pawn sacrifice
is usually considered to be worth seem to justify such a measure?
six 'points ' , but common sense If the answer to the last question
suggests that a pair of active bish is 'yes ' , then by all means sacrifice
op s must be accorded a value of the pawn. However, if the answer
almost nine under some circum to the third question is ' maybe ' ,
stances . It should come as no sur then the pawn sacrifice should
pri se that it often proves fruitful to probably be rejected if either of the
sacrifice a pawn - or even an ex first two questions can be an
change - to activate a bishop pair. swered in the affirmative.
It is difficult to draw up any The difficulties posed to the de
'general rules ' for when one should fender by such a sacrifice are dem
sacrifice a pawn to activate a onstrated by the following game, in
bishop pair. In many situations, a which a former world champion
pair of bishops will be active 'in missed several superior defensive
their own right' , but a pawn disad continuations.
vantage can still prove to be a pawn
disadvantage. These decisions are Botvinnik Euwe
-
cases that must be considered on an Moscow World Ch

individual basis . The player with (Match Tournament) 1 948
the bishop pair should probably
consider three factors before de 1 d4 d5 2 c4 e6 3 lbf3 lbf6 4 lbc3
ciding to sacrifice material. c6 5 e3 lbbd7 6 .td3 .tb4 7 a3
First, is it possible for the op .ta5 8 c2 Wie7 9 .td2 dxc4 10
ponent to neutralize one of the .txc4 e5 11 0-0 0-0 12 :ae1 ! (D)
The Sacrifice for Active Bishops 149
gives 14 . . . iLd6 as superior in Bot

vinnik 's Best Games 1 947-1970.
1S iLa2! t2Jf6
16 'iib4 e4 (D)
B
Botvinnik's decision on how to
place his rooks shows a penetrating
insight into the position. The ten
sion of the centre pawns will soon w
resolve itself in one of three ways.
It could be that Black will capture 17 t2JeS!
at d4, in which case Botvinnik in Botvinnik sacrifices a pawn so
tends to recapture with his pawn, as to activate his otherwise bad
gaining a tempo on the queen. An bishop. White now has the oppor
other approach is for White to ex tunity to whip up a strong kingside
change at e5, in which case he will attack if the pawn is accepted, as
have a kingside pawn majority that his bishops and queen are well
is ably supported by the rooks at e l placed and his king's rook will
and fl . Finally, Euwe may choose soon enter the attack with f2-f3 .
to advance . . . e4, in which case 17 ... iLxeS
White can play f2-f3 to undermine Botvinnik writes that "Black
Black's strong centre point. Of certainly should not have ac
course, this will leave the white e cepted" the pawn sacrifice. He rec
pawn backward on an open file, so ommends 1 7 . . . iLe6 1 8 iLbl iLd5,
the e l -rook will guard it and can when 1 9 f3 I:.fe8 allows Black to
also help enforce the push e3-e4. defend his centre and keep the
12 iLc7 white bishops contained.
13 t2Je4! t2Jxe4 18 dxeS 'ii'xeS
14 'iVxe4 as? ! 19 iLc3 'ii'e7
Subsequent events show this to 20 f3 (D)
be a pointless move. Botvinnik 20 t2JdS?!
150 The Sacrifice for Active Bishops
B W
Botvinnik considered this the b) 20 . . . i.e6 2 1 i.bl lb d5 22
decisive mistake. It's true that WUxe4 f5 and 23 . . . lbxc3 - Botvin
Black must return his pawn one nik. However, 22 WUxe7 ! lbxe7 23
way or another, but now White's i.xe4 looks better, since it retains
bishops and better development the bishop pair and gives White a
give him an undisputed advantage. nice ending - Mayer. However, an
Black has two superior continu analyst who is not specifically
ations: credited (possibly Botvinnik in an
a) 20...exf3 !? 2 1 i.bl l:te8 ! (Bot other set of notes, but I am unaware
vinnik only considered the losing of a corroborating source) on the
2 1 .. .h6? 22 l'hf3 lbd5 23 litg3 ! +-) ChessB ase 'Mainbase' now gives
22 i.xf6 (22 l':txf3 lbe4 ! is fine for 23 . . . i.b3 as equal. It's true that
Black) 22 . . . 'iVxf6 23 WUxh7+ f8 White then can't use the d-file very
24 l':txf3 'ii' h 6 ! (not 24 .. :xb2? 25 easily, but Black will have to prove
i. g6 i.e6 26 l':tb1 ! and 27 l':txb7 is that he can neutralize the bishops
winning) with the following posi and this line strikes me as superior
tion (D): to Botvinnik's (original?) variation.
25 'iVxh6 (25 c2 i.e6 holds to 21 'ilt'xe7 lbxe7
gether) 25 . . . gxh6 and the ending 22 fxe4 (D)
should be tenable, particularly in 22 .. b6
view of Black's queenside majority Black's problem in this queen
and more active king - Mayer. In less middlegame is that the white
fact, this ending should be better pieces are so much more active
for Black than the one that arises in than the black pieces. Botvinnik
the next note, as he has managed to considers two alternatives but finds
eliminate White's bishop pair. both wanting:
24 i. a6
25 lIf2 j.,b5
26 e5 ltJe7 (D)
B
a) 22 . . . j.,e6 23 j.,xe6 fxe6 24
lIxf8+ xf8 25 ':'n + g8 26 lld I
+- - Botvinnik. The rook's pene W
tration on the d-file is decisive.
b) 22 . . . j.,g4 23 IIf4 j.,h5 24 g4 27 e4
i.g6, with the division: Botvinnik prefers to pile up the
b I ) 25 h4 h5 26 h2 h7 27 pressure and limit the knight's ac
':'gi f6 28 gxh5 i.xh5 29 e5 ! 'with cess to the centre.
a strong attack' - Botvinnik. 27 ':'d7 should also be favour
b2) 25 ':'d i ':'ad8 26 ':'xd8 ':'xd8 able for White, when 27 . . . ltJd5 al
27 j.,xa5, when Keres believes that lows 28 e6. Note that this is better
B lack has some counterplay after than 28 j.,xd5 cxd5 26 ':'xd5, when
27 . . . ':'d I + 28 f2 ':'c I (cited in the opposite-coloured bishops may
Botvinnik but apparently from the save Black, or 28 e4 ltJxc3 29 bxc3,
tournament book) . Botvinnik dis when the f7 -pawn will fall, but
misses this on the grounds that White ' s own pawns have become
White has an extra pawn and the weak.
two bishops. 27 c5
23 ':'dl 28 e6! f6
Note, however, that 23 llxf7 ? 29 lIxb6 j.,c6
.'I:lxf7 24 ll n ltJd5 ! favours Black. 30 lbc6! lDxc6
23 ltJg6 31 e7+ ':'17
24 l':td6! 32 j.,d5 1-0
The passive placing of Black's Understandable, as 32 . . . l':tc8 3 3
pieces makes it very difficult to put e8'i+ ! l':txe8 3 4 j.,xc6 and 35 j.,d5
up a sustained defence. nets a piece.
The following game impressed sacrifice, like so many made for the
me tremendously as a youngster; sake of active bishops, should be
indeed, it still does. Tal's imagina classified as 'speculative' .
tion seems leagues beyond that of 13 dxc5
practically any other chess player, Hecht really is constrained in his
but Black was not without his re response. He should take the pawn,
sources. because otherwise 14 cxd6 will un
double the pawn free of charge, but
Tal - Hecht 1 3 . . . bxc5 1 4 dS would leave his
Varna Olympiad 1 962 queenside light squares weak and
the bishop at b7 would be quite out
1 d4 tbf6 2 c4 e6 3 tbf3 b6 4 tbc3 of play.
.tb4 5 .ig5 .tb7 6 e3 h6 7 .th4 14 dxe5 Wlixe5
.txc3+ 8 bxc3 d6 9 tbd2 tbbd7 10 15 a4+ c6? !
f3 Wlie7 11 e4 e5 12 .td3 tbf8 13 This move must be considered
c5!? (D) questionable, since it weakens the
d6- square. In Mikhail Tal Games I
1 949- 1 962 , the Bulgarian series of
Tal's complete games, the response
IS ...tb6d7 is examined, when White
has the choice of 1 6 ':c 1 Wlie6 1 7
0-0 .tc6, which i s assessed as un
clear, and 16 WHc2, when the indi
cation is that White has ' attack' ,
presumably i n view of his bishop
pair and kingside majority.
Another idea, apparently more
logical, is I S . . . tbSd7, preparing
B castling and intending to meet 1 6
Wlic2 with 1 6 . . . tbdS ! - Mayer. In
Having the bishops and having Study Chess with Mikha il Tal, Tal
active bishops aren't the same & Koblenc suggest that 1 6 .txf6 is
thing. With this pawn sacrifice - best, when "Black would then have
which is typical in this type of posi to lose a few tempi before cas
tion - Tal aims to open the a2-f7 di tling." The downside of this ap
agonal and clears the c4-square. proach, however, is that White
White certainly has practical com cedes the bishop pair and may still
pensation for his material deficit. It have to be concerned about double
should be stressed here that this attacks based on a later .. :ilixc3 .
The four-volume Bulgarian se

ries of Tal' s complete games fea
tures annotations to all of Tal 's
games. It's an impressive work (vol
ume three has just been released as
I write this), but for unknown rea
sons, none of the masters or grand
masters who worked on the games
are credited as annotators for par
ticular games. This inexplicable
approach makes it impossible to at
tach a name to any original analyti B
cal work included in the series, so I
have chosen to designate any unat a) 20 'ifxh4 WNxe5 - Bulgarian
tributed analysis from it as 'Bul Series . This is good for Black, be
garian Series' . cause 2 1 tbxb7 WNe3+ snares the
16 0-0 tbg6 bishop.
Black lands in major trouble af b) 20 :ae l ( ' ! ?' - Bulgarian Se
ter 1 6 . . .'ifxc3 ? 1 7 tbc4 ! , with the ries) 20 . . . 'ifd5 2 1 ttJxb7 'ilIxd3 22
options: exf6, when White has compensa
a) 17 . . . b5 18 tbd6+ 'it>d7 1 9 tion for his material but Black may
tbxb5 ! cxb5 20 .txb5+, i s assessed be able to defend - Mayer. How
as winning by the Bulgarian Se ever, in a practical game against
ries, e.g., 20 . . . 'it>e7 2 1 e5 , with a Tal, it would be hard to like Black's
powerful initiative. chances at all.
b) 1 7 .. .'it'xd3 1 8 }::t fd l b5 1 9 19 exf6! bxa4?
lixd3 bxa4 20 ttJd6+ 'it>e7 2 1 tbf5+ Hecht makes a serious mistake
(Tal & Koblenc), is crushing, as g7 in accepting the queen sacrifice.
hangs. An attitude of ' show me' is often
17 tbc4 'ife6 useful for the defender, but 1 9 . . . 0-0 !
Black loses after 17 . . . b5 1 8 20 :ae l ! 'ifxe l ! offered the best
tbxe5 bxa4 1 9 tbxg6 fxg6 20 e5 - chance at refutation, e.g., 2 1 }::t xe l
Tal (cited in the Bulgarian Series). bxa4 22 .txg6 fxg6 23 lie7 was
18 e5! (D) given by Tal as winning, as he ap
18 ... b5 parently only considered 23 . . . :f7?
Black raises the ante. The Bul 24 tbd6 ! (cited in the Bulgarian Se
garian Series offers some interest ries).
ing alternatives after 1 8 . . . tbxh4 ! ? In fact, Lilienthal (also cited in
1 9 tbd6+ 'it>f8: the Bulgarian Series) shows that
154 The Sacrifice/or Active Bishops
23 . . . g5 ! (D) results in extraordi 29 f4 ! ) 29 fxg4 I:.ad8 30 h4 gxh4

nary complications: 3 1 g5 when the addition of the g
pawn to the attack leaves Black
scrambling to find a defence. The
bind that White achieves in this
line is based in part on the presence
of opposite-coloured bishops, for
while he is down an entire rook, he
has a substantial advantage on the
dark squares.
20 fxg7 I:.g8 (D)
W
a) 24 1hb7 gxh4 25 I:.xg7+ h8
26 ':g6 l::t f7 blockades the danger
ous f-pawn.
b) 24 ':xg7 + h8 25 J.g3
(White plays to maintain his passed
pawn, which is eliminated after both
25 lLle5 ':xf6 and 25 J.xg5 hxg5
26 l::t x b7 lhf6) 25 . . . J.a6 leaves W
White at an important junction:
b I ) 26 lLld2 I:.xf6 and the white 21 J.f5 ! !
initiative is finally grinding down, A n extraordinary move that re
e.g., 27 lic7 I:.e6 28 lLle4 ':b8 ! or stricts the movements of the black
27 lil d7 ':f5 28 lLle4 J.c4 - Lilien king, which is already in a box be
thal. cause of the pawn at g7 . The coor
b2) 26 .te5 ! appears strong af dination between the various white
ter 26 . . . .txc4 27 f7 (as given by minor pieces is really impressive.
Lilienthal) . My first thought was 21 lLlxh4
that 27 . . . h5 ! ? is good, as it clears Black has a number of options,
the h6- square for the black king. but White is doing well by this
However, Graham Burgess points time:
out that 28 g4 ! continues the com a) 2 1 . . .'iVxc4 22 :tae l + 'iVe6 23
plications by fixing the g5-pawn as I:.xe6+ ! fxe6 24 J.xg6+ d7 25
a target, e.g., 28 . . . hxg4 (or 28 . . . h4 I:.d 1 + "followed by .tf6 and the
The Sacrifice /or Active Bishops 155
white pawn on g7 cripples Black" c2) 24 . . . lZ:lxh4 25 liJxc5+ 'it>c7

- Tal & Koblenc. 26 liJxa6+ 'it>b6 27 liJb4 Ilxg7 28
b) 2 1 . . :it'xf5 22 liJd6+ 'it>d7 23 g3 is clearly better for White - Lil
liJxf5 (the knight guards the h4- ienthal.
bishop) 23 . . . liJxh4 24 ':ad l + 'it>c7 Perhaps the nicest part of Tal's
25 liJxh4 ':xg7 26 :t fe l was given play in this game is the manner in
as a winning ending by Tal, but the which so many of the variations re
Bulgarian Series points out that sult simply in a superior ending
25 . . . iLc8 ! is an improvement, be which must still be won.
cause the knight can ' t return to 22 iLxe6 iLa6
play through f5. However, they sug White is also doing well after
gest 24 liJxh4 ! , when White will 22 ... fxe6 23 liJd6+.
either transpose to Tal's analysis, 23 liJd6+ 'it>e7 (D)
for example, 24 . . . ':xg7 25 Ilad l +,
or bring his knight into play via
f5 .
c) 2 1 . . .iLa6 22 i.xe6 fxe6 23
liJd6+ 'it>d7 24 liJe4 ! (D), with the
division:
W
24 iLc4!
This move, by which the hang
ing knight at d6 guards the bishop
at c4, is an amusing echo of the
B knight's guarding of the bishop at
h4 in variation 'b' in the note to
c l ) 24 ... iLxfl 25 liJf6+ 'it>c7 26 Black's 2 1 st move.
iLg3+ e5 27 liJxg8 l1xg8 28 'it>xfl 24 ... ':xg7
Ilxg7 29 Ilel is assessed as winning 25 g3 'it>xd6?
for White by the Bulgarian Series; Now Black's problems increase,
I suspect that this analysis stems as he finds himself with an inferior
from Tal. knight against a strong bishop. Tal
suggests 2S . . . i- xc4 26 lLlxc4 I:.d8 ,

which he assesses as clearly better
for White.
26 i-xa6 lLlfS?
This makes life easier for White,
as he now gains control of the b
file. 26 . . . I:.b8 was superior.
27 lIab1 f6
28 I:.fd1 + <tie7
29 :e1+ d6
30 <tin (D)
B
3S ::'c8?!
Tal indicates that Black should
have played to get as many pawns
off the board as possible with
3S . . . hS 36 h3 hxg4 37 hxg4 fS .
36 l':txa4 I:.xc3
37 I:.a6+ <tics
38 I:.xf6 hS
39 h3 hxg4
40 hxg4 I:.h7
B
41 gS I:.hS
42 I:.fS I:.c2+
30 c4?! 43 <tig3 <tic4
Losing ground. It was better to 44 :teeS d4
shore up the knight with 30 . . . hS . 4S g6 I:.h1
31 g4 lLle7 46 ::'cS+ <tid3
32 l':tb7 I:.ag8 47 I:.xc2 <tixc2
33 i-xc4 lLldS 48 <tif4 ::'g1
34 i-xdS 49 I:.gS 1-0
Tal takes play into a double rook White wins the king and pawn
ending. The black knight has fi ending that arises after 49 . . . I:.xgS .
nally reached a good square, so the
decision to surrender the bishop The bishop pair can be particu
isn't controversial. larly effective in situations with
34 cxdS opposite-wing castling. The fol
3S ::'b4 (D) lowing gem is one of a thousand
similar games that Black has won

in the Sicilian.
Shiyanovsky - Taimanov
USSR Ch (Ere van) 1 962
1 e4 c5 2 t2Jf3 t2Jc6 3 d4 cxd4 4

t2Jxd4 e6 5 t2Jc3 a6 6 i.e3 t2Jf6 7 a3
Wlic7 8 f4 d6 9 "i!t'f3 e5 10 t2Jxc6
bxc6 11 fxe5 t2Jg4! (D)
Taimanov prevents the white

bishop from coming to c4, where it
would pressure f7 and could also
serve to defend the queenside.
15 i.g2 .:tb8
16 0-0-0
"White no longer has time for
any other defences of the b2-pawn."
- Taimanov.
W 16 0-0
17 .:thel i.e5
As Taimanov comments on a 18 WHc5
similar situation in Taimanov 's Se Shiyanovsky prevents 1 8 . . . WHa5 ,
lected Games, "No Sicilian player which would step up the pressure
would begrudge giving up a pawn on the white king position. White's
for White's dark-square bishop." problem is that his king is men
12 exd6 aced, but he has no attack of his
Taimanov points out that pre own, despite the opposite-wing
serving the bishop by 1 2 i.f4 gives castling. Further, White 's remain
Black a very nice game after ing bishop is quite bad and his knight
1 2 . . . dxe5 1 3 i.g3 i.c5 , so White must stay put if he values his king.
might as well grab material and try The present position is a good
to weather the storm. example of lOOking at the specifics
12 i.xd6 of a position and not at the gener
13 g3 t2Jxe3 alities. For instance, in the Dragon
14 Wlixe3 i.e6! (D) Variation, White is usually careful
not to exchange his dark-squared 24 WNa5

bishop for a knight, as then he has 25 e5 i.a3+
no good way to oppose Black's fi 26 b1
anchettoed dark-squared bishop. On 26 d2, Black can win by
Of course, the present game isn't a either 26 . . . i.b4 or 26 . . . I:.d7+ - Tai-
Dragon, but the bishop at e5 is on manov.
the long diagonal and so might as 26 i.xb3! (D)
well be a Dragon bishop.
18 :b7
19 lite3 I:.tb8
20 b3 h5
A clever move that simultane
ously creates ' luft' and introduces
the positional threat ... h4.
21 i.fl i.f6! (D)
27 cxb3
Or 27 i.b5 i.f8 28 "*,h4 i.xa4 -
Taimanov.
27 l1xb3+
28 c2
Black also wins after 28 a2
W i.f8 - Taimanov.
28 l1xc3+!
22 a4 29 lhc3 I:.b2+
Taimanov indicates that 22 WNxh5 30 d3 WNd5+
i.e7 doesn' t help White, but that 31 e3 'ii'x d1
22 tZJb l might have increased his 0-1
defensive chances.
22 i.e7 The sacrifice for the active bish
23 WNxh5 g6 ops frequently unleashes both
24 "ilih6 bishops, as we saw in the last three
"Or 24 "*,f3 i.g5 with a mount games. However, it is also a motif
ing attack." - Taimanov. to free up one bishop in particular.
In the following game, Black's bishop. Moreover, the long dark di

king' s bishop becomes a power agonal is opened up.
house and the fact that White has 10 exd4
no good means of opposing it is the 11 'iUxd4 c6!
secret to Black' s pawn sacrifice. Naturally Smyslov wishes to
By comparison, Black's queen's open the position; not only does he
bishop doesn't stir until move 19 have the bishops, but the white
and then only to give itself up as king is still in the centre.
part of a queen trap. 12 dxc6 bxc6
13 0-0 tZJd5
Wade Smyslov
- 14 'iUc4 tZJf4! (D)
Havana 1 965
1 e4 tZJf6 2 tZJc3 d6 3 d4 g6 4 J.e2

J.g7 5 h4 h5 6 tZJh3 tZJc6 7 tZJg5 0-0
8 J.e3 e5 9 d5 tZJd4! 10 J.xd4 (D)
W
Black's knight takes up an active
post and creates possibilities of
forking on e2. Smyslov now gives
B 1 5 J.f3 lIb8 16 gabl ( 1 6 b3 'iUa5)
1 6 . . . ltb6 1 7 ltfd l i.a6 1 8 'ilia4
Wade accepts the pawn sacri J.e5 1 9 g3 'iVf6 ! as giving Black a
fice, as otherwise the d4-knight is ' strong initiative' . Wade plays an
well-placed. This type of sacrifice even worse approach.
to gain the bishop pair is typical of 15 'iUxc6? J.xc3
a variety of fianchetto defences, 16 J.f3 i.xb2
since it is now next to impossible 17 'iUxa8 i.xal
for White to pierce the black king 18 Ibal '1i'b6!
position, because he has no way Everything becomes clear: the
to eliminate Black' s dark-square white queen is in a trap.
19 e5 d5 The motif of sacrificing material

20 g3 to obtain a pair of active bishops
Smyslov notes that 20 J.xd5 does not always have to be an at
.te6 wins a piece. tacking one. In the following game,
20 J.g4 Gligoric uses an exchange sacrifice
21 WNxf8+ \t>xf8 to create complications that are not
22 .txg4 hxg4 necessarily in his favour, but that
23 gxf4 4! (D) offer practical defensive chances in
an otherwise difficult position.
Hort - Gligoric
Amsterdam 1 970
1 d4 tOf6 2 e4 g6 3 tOe3 .tg7 4 e4

d6 5 f3 0-0 6 J.e3 b6? ! 7 J.d3 a6
8 tOge2 e5 9 e5! tOfd7 10 exd6
exd6 1 1 WNd2 tOe6 12 J.e4! J.b7
13 O-O-O! tOf6 14 J.xe6 .txe6 15
.tg5 (D)
w
Smyslov goes after the weak
ened white pawns in such a manner
that the rook can ' t get into the
black position.
24 \t>g2 WNxf4
25 lte1 \t>e8!
Black minimizes the impact of
any e5-e6 attacks.
26 ':e3 g3
27 ':xg3 'iWxh4 B
28 e3 \t>e7
29 \t>gl 'iff4 Hort has played the opening in a
30 tOh7 a5 very aggressive manner, with no
31 tOf6 WNxe5 concern about his king possibl y
0-1 falling under an attack. As his re
The king and pawn ending is a ward, he gets to set up a nasty pin
win after 32 lte3 \t>xf6, etc. on the black knight and has the
The Sa crifice for Active Bishops 161
threat 1 6 dxc5 dxc5 1 7 WHf4 Wlie7 Many of the Classicists of the

1 8 .l::t d 6, as given by Levy in Gli late nineteenth and early twentieth
goric 's Best Games 1 945- 1970. century claimed that two bishops
15 :cS versus rook and knight were
16 d5 i.d7 equivalent. I don 't believe that this
17 tZJg3 lIeS! is the case, i.e., I think the extra
Black must take action against material will usually win out, but
the looming 1 8 Wlif4 and 19 tZJce4, this view continues to influence
which will step up the pressure on chess thinking to some extent even
the pinned knight. The text-move today.
envisions an exchange sacrifice 19 lIxg5!
against the powerful g5-bishop, af 20 'iUxg5 b5! (D)
ter which Black will have two bish
ops and practical chances. This
type of exchange sacrifice was later
to become common in the Soltis
Variation of the Yugoslav Attack
Sicilian Dragon, but this example
predates the numerous Dragons
that followed in its wake.
Hort has the advantage, but it
should be stressed that Black also
has his trumps: the two bishops ; a
powerful king's bishop; and he can
open files for his attack on the king w
more quickly than White can.
1S 'i!t'f4 :e5 Black already has the bishop
19 tZJce4 pair but hi s material deficit limits
Gligoric suggested after the the amount of time he can spend on
game that 1 9 h4 ! is stronger, when slow build-ups. The text aims to
he intended to play 19 . . . h5 20 coordinate the rook with the king's
tZJce4 lIxg5 21 Wlixg5 tZJe8 , offer bishop so as to generate an attack
ing to play an exchange-down end on the king. Razuvaev assesses the
ing where his bishop pair give him position as already somewhat bet
some practical drawing chances. It ter for Black.
should be noted that in ECO E 21 tZJxd6
(First Edition) , Razuvaev assesses Levy prefers 2 1 :he l , on the
the position after 19 h4 ! as clearly reasoning that White should not be
better for White. worried here about acquiring more
162 The Sacrifice jor Active Bishops
material. There is certainly sense to

this view in light of the white
rook' s failure to get developed in
subsequent play.
21 ':'b8
Gligoric decides to ' go for
broke ' by remaining down the ex
change and a pawn. He could also
have played 2 1 . . . WNf8 ! ? 22 tZJxc8
.ih6 23 tZJe7+ (Levy analyses 23
h4?? to a supposed advantage for
White but he overlooked 23 ... .ixc8 W
-+) 23 .. . ' h8 24 WNxh6 (24 h4 is
also possible) 24 .. :i1Hxh6+ 25 'it>bl 28 WNe1 'iVa3
bxc4 - Mayer. Black has restored 29 tZJxc3 .if5
nominal material equality but he' s Gligoric could play 29 . . . l':txb2+
running out o f attacking pieces and at once but he prefers to activate his
the white d-pawn could become a bishop and threaten mate in one.
factor. 30 'it>a1 ':'xb2!
22 tZJge4 h6 31 ':'xb2 i.xc3 (D)
23 'i!t'e3 tZJxe4
24 tZJxe4 bxc4
This pawn will prove dangerous
throughout the rest of the game, as
it can help open up the white king
position or support black pieces
near the white king.
25 l':td2 WNa5 (D)
26 'it>b1
Levy gives 26 WNa3 ! .ixb2+ 27
WNxb2 l':txb2 28 l':txb2 "with an un
clear position, possibly favouring
White". In fact, after 28 . . . WNa3 ! 29 W
'it>bl Wid3+, Black is considerably
better, as his queen is active, the 32 WNc1?
king ' s rook is still unmoved and Levy doesn' t comment on this
the c4-pawn is dangerous - Mayer. move but it strikes me as an obvi
26 c3 ous error. White should have played
27 ':'c2 .id4 to keep his queen active with 3 2
'ii'e S+ cJ;; g7 33 'ilibS c4 34 :c l , 36 d7! .txd7 (D,

with the idea of establishing a
blockade at c2 once the black pawn
reaches c3 - Mayer.
Black's second exchange sacri
fice makes a pleasing impression.
If one remembers my suggestion
at the beginning of the chapter that
a pair of active bishops can be ac
corded a value of almost nine
points in some circumstances, then
it should be clear that Black effec
tively has near material equality in
much of the play that follows. W
32 c4
33 d6 37 g4?
Levy relates that many com Levy asks the rhetorical ques
mentators gave 33 g4 as White's tion "How does Black make pro
best, but he dismisses it on the gress after 37 I hd7 ! 'iVxd7?" and I
grounds of 33 ... .td7 34 cJ;;b l .txb2 agree with him. The problem, of
35 'iVxb2 'iVxf3 , when Black is do course, is that he no longer has any
ing quite well, as 36 I:.e l c3 deacti way to raise the blockade at c2, so a
vates the white queen. However, it draw is a reasonable result. How
seems to me that White has better ever, Hort was in severe time pres
in 34 l:t f1 ! ?, with the idea of hold sure, so it's understandable that he
ing together the kingside pawns. missed his best chance.
Another idea is 33 ... .txb2+ 34 37 .te6
'iixb2 'iVxf3 35 :e l .td7 (35 ... c3 ?? 38 I:.et .tb3
36 'i!fbS+) 36 'iVd4! , when the white Teasing White. Levy claims that
pieces have gained some much 3 S . . . .txa2 ! is a blunder, but he
needed activity, since 36 . . . 'iVxg4?? missed 39 I ha2 c2+ 40 'iVb2 'ilif4 !
loses the queen to 37 l:teS+ - (instead of Levy 's 40 . . . .txb2+?,
Mayer. which he considers equal) and
33 .tf6. 4 1 ...c 1'ii'+ will mate - Mayer.
34 I:.dl c3 39 :ee2
35 :c2 The power of the bishops is il
Levy points out the pretty 3 5 lustrated by 39 I:.e4 .txc2 ! 40
l'IbS+ cJ;; g7 36 'iVxa3 c2+. I:.xa4 .txa4 4 1 a3 .tb3 ! and White
35 'iVa4 can resign - Mayer.
39 J.xa2 Verdict: The sacrifice for active

40 xh6 bishops is difficult to study in a
Or 40 l:Ie4 J.c4+ 4 1 'it>b1 b3+ systematic manner. However, a
42 'it>a1 i.d3 , winning - Mayer. pair of active bishops is frequently
40 J.c4+ adequate compensation for a pawn
41 'it>bl i.xe2 - or even the exchange - in a mid
0-1 dlegame position.
1 3 The Unexpected Excha nge
One of the basic principles of chess Bronstein - Boleslavsky

strategy is to exchange one's less Moscow, Candidates ' Play-Off
active pieces for the opponent's Match (14) 1 950
more active pieces, while avoiding
the trade of one's more active 1 d4 liJf6 2 c4 e6 3 liJc3 i.b4 4
pieces. When the pieces in ques liJf3 d6 5 WHb3 as 6 g3 liJc6 7 i.g2
tion are the same, it is generally liJe4 8 0-0 i.xc3 9 bxc3 0-0 10
simple to realize whether one should liJel f5 1 1 f3 liJf6 12 a4 'Wie7 13
exchange or avoid the exchange, as c5! d5 (D)
the pieces have the same charac Black's pawn structure would be
teristics. If the opponent's piece is fractured after 1 3 . . . dxc5? 14 i.a3 ,
more effective or has better pros as 1 4 . . . liJd7 1 5 liJd3 regains the
pects than ours, we exchange; oth pawn with advantage.
erwise, not.
However, bishops and knights
are a more difficult matter to judge,
as the pieces are accorded the
same material value (three points)
while having very different charac
teristics . One of the most difficult
things to see is when an apparently
well-placed minor piece should be
exchanged for an apparently less
active one that has different charac
teristics. Such manoeuvres might
best be remembered under the W
heading of 'the unexpected ex
change' . Since such unexpected 14 i.g5 h6
exchanges appear to defy 'com Black runs into trouble after
mon sense' , it is fruitless to discuss 14 . . . e5 ! ? 1 5 e4 ! : 1 5 . . . exd4 1 6 exd5
them as general cases , so let's or 1 5 .. .fxe4 1 6 fxe4, with the threat
move directly to concrete exam 17 exd5 - Mayer.
ples. 15 i.xf6 'iWxf6
166 The Unexpected Exchange
16 lUd3 b6! 'built-in' counterplay along the b

B ronstein has created a position fi le. As Bronstein remarked, "It
where he has the better minor was not easy to find a weakness in
pieces . If Boleslavsky were to sit Boleslavsky's superb classical
quietly, Bronstein would soon step strategy."
up the pressure with e2-e4, aiming Why was Bronstein able to find
to open up the position. the correct continuation when most
17 exb6 ':b8 players - even many grandmasters
18 "i!t'a3 ':xb6 - would not have? I believe there
19 f4 are two reasons. First of all, he was
B ronstein decides to seal the a strategist who did not hesitate to
centre on account of the activity play moves that might 'look wrong' ,
soon to be displayed by the bishop but that i n fact met the needs o f a
at a6. specific position. Yet more impor
19 J.a6 tant, I think, is the fact that he re
20 lUeS "Wie7 (D) fused to 'lull himself to sleep' with
pretty words. It's an easy matter to
try to turn chess into a ' war of
words ' rather than a war of moves.
The words sound so right: 'His
bishop is bad, as there are many
pawns fixed on the light squares.
Meanwhile, my knight is a tower
of strength, as it occupies a hole
where it can live for life.' Such an
approach works in many positions
but this strictly verbal approach
will always fail to find the excep
w tions to the general case.
21 ... ':xa6
21 lUxa6! ! Black can't ease his defensive
The knight, which looked like a problems by swapping queens with
much better piece than the 'bad' 2 1 . . :it'xa3 22 ':xa3 ':'xa6, as then
bishop at a6, willingly exchanges White secures a solid advantage
itself. However, Black intended to with 23 ':b l llb6 24 ':b5 ':fb8 25
place his bishop at c4, in which ':'c5 r:J;; f7 26 e3 r:J;;e7 27 c4, when
case the white bishop would be "it is obvious that the strong pawn
locked out of play by the pawn at on d5 has become very weak." -
d5 . Moreover, Black would have Bronstein.
The Unexpected Exchange 167
One of the main features of the I:.ac8 35 I:. 1 b3 I:.xh2 36 I:.e3+ d7

knight for bishop exchange is that 37 xd5, when the position is ' a
White will be able to activate his clear draw' - Bronstein.
bishop with a later c3-c4. 29 d3 d7
22 'iWc5! 30 e3 e7
Bronstein follows up with an 31 d2 ':c8
other insightful move. Now a 32 iof3 g6
queen exchange would leave White 33 .te2 h5
with a pawn at c5 that might seem 34 h3 I:.g8
weak. Indeed, this is certainly what 35 I:.b7 I:.gc8 (D)
'pretty words' would tell one. In
fact, the pawn would rob a black
rook of the b6-square and would
prove next to impossible to attack.
One of the central tenets of modem
chess is 'a weakness that cannot be
attacked is not a weakness' and that
would certainly be the case with
the white pawn at c5 .
22 I:.b8
23 fttb1 'iUxc5
Bronstein suggests that the sim
plifying 23 . . . I:.xb 1 + 24 ':xb 1 WUxc5 W
25 dxc5 f8 would have left Black
with an easier defence. But what's 36 g4
the rush, as after 24 dxc5 in the White opens a second front, as
game, Black could still reach this Black has done well in defending
variation with 24 ... I:.xb 1 +, etc. against an attack based solely on
24 dxc5 f8 ? ! the b-file.
25 l':tb5 I:.aa8 36 hxg4
26 f2 e7 37 hxg4 4Jd8
27 e3 I:.g8 The black knight has the prob
28 I:.ab1 lem that it can ' t get past the third
White builds up on the b-file. rank. Boleslavsky brings it over to
28 . I:.gh8 the kingside, where it defends f7
Better was 28 . . . g5 , beginning an and shields the second rank against
active defence, e . g . , 29 fxg5 hxg5 a white rook at h7, but this sort of
30 c4 4Jb4 3 1 cxd5 4J xd5+ 32 passive defence is not a knight's
.txd5 exd5 33 d4 I:.h8 34 ':b7 great dream in life.
38 1:.7b2 tiJf7 at g6 - but this doesn ' t look to me

39 g5 any worse than what happens in the
The g-pawn is fixed on a light game ' free of charge' after Black
square, while the e-pawn will also engineers . . . e6-eS .
be held in place if Black doesn ' t 40 dxc4
watch out. 41 .tfJ 1:.a7
39 d7 (D) 42 c3 c6
The knight will now be unable
to get any further than e6. Even
this unremarkable post will be dif
ficult to reach, since it will require
that Black play . . . eS , . . . tiJd8, and
. . . tiJe6. Further, this approach will
open the a2-g8 diagonal for the
white bishop, which may be able to
attack the black g-pawn from n in
some circumstances.
43 1:.h2 e7
44 .t e2 f8
w 45 .txc4 1:.eS
46 1:.b6 (D)
40 c4 Culminating the attack begun
This is 'positionally decisive ' , with 2 1 tiJxa6 ! ! . The white pieces
according to Bronstein. The bishop have achieved maximum power
will now gain greater activity and and are ready to start collecting the
have the opportunity in many vari weak black pawns.
ations to get at the weak black
pawns. If Black were now to de
fend with 40 . . . c6, then Bronstein
offers the fantastic pawn sacrifice
4 1 cxd5 exdS 42 e4, which he
awards two exclamation marks, as
it opens 'useful diagonals for the
bishop ' . It's true that 42 . . . fxe4 43
.tg4+ wins the exchange but it
seems to me that Black should have
tried 42 . . . dxe4 43 .tc4. The black
position is then very loose - there
is an especially conspicuous target B
46 l::tc7 fact that the f-pawn must finally

47 :a6 cJ;;g7 start to run, after which it is rounded
48 l::txa5 e5 up by the white rook with no real
Boleslavsky hopes for activity diminution of White's winning ad
but B ronstein demonstrates that it vantage) 64 cJ;; e7 65 l::txf4 cJ;; e6 66

is too little, too late. :f6+ cJ;;d5 67 cJ;;b4 1-0

49 l::th 3 ltd7
50 .tb3 exf4 There are chess games that can
51 exf4 lte4 be regarded as 'companion pieces'
52 :a8 l::txf4 (D) to each other. Bronstein-Boleslav
sky paves the way for a greater un
derstanding of a later game.
Fischer - Petrosian
Buenos Aires,
Candidates ' Match (7) 1 971
1 e4 c5 2 liJf3 e6 3 d4 cxd4 4 liJxd4

a6 5 Jt.. d3 liJc6 6 liJxc6 bxc6 7 0-0
d5 8 c4 liJf6 9 cxd5 cxd5 10 exd5
exd5 1l liJc3! .te7 12 'ilia4+! (D)
W
53 .txti!
This second 'unexpected ex
change ' , which results in a winning
rook and pawn ending, is an amus
ing echo to the earlier exchange of
knight for bishop. Note that Bron
stein judges the resulting position
on merits; he refuses to lull him
self to sleep with 'pretty words ' .
The game concluded 53 :xf7 54..
B
l::t ah8 l::tf3+ 55 ':xf3 cJ;; x h8 56 as
':a7 57 cJ;; b 4 cJ;;g7 58 :a3 :a6 59 12 iVd7? !
:d3 cJ;;f 7 60 %ld6 cJ;; g7 61 l::td7 + Petrosian offers a speculative
cJ;; g8 62 cJ;;a 4 cJ;;f8 63 l::tb 7 f4 64 exchange sacrifice, since 1 2 .td7
. . .
:b4 (the key to White's win is the allows White two continuations:
1 70 The Unexpected Exchange
a) 1 3 'iVc2 ! gives Black a wide 14 "iVxaS 0-0, when Speelman calls

choice: the position 'extremely messy' .
a l ) 1 3 . . 0-0 14 .tgS d4 is con
. Lilienthal analyses further with
sidered favourable to White but 1 5 "iVaS d4 1 6 lbxb5 .tb7, when
playable for Black by both Poluga Black has threats of 1 7 . . . lIaS and
evsky (cited in Bernard Cafferty 's 1 7 . . . .txg2, which will open up the
Candidates 'Matches 1971 , as are king for a perpetual on the light
all the other Soviet annotators squares, e.g. , I S 'it>xg2 "iVg4+.
quoted on this game) and Speelman Fischer will have none of such
in Best Chess Games 1 970-80. nonsense; instead, he takes play
a2) 1 3 ... .te6 14 .tgS h6 1 S .txf6 into an extremely favourable end
.txf6 1 6 "iVa4+ 'it>fS "and although ing.
White stands better Black can put 13 'ii'xa4
up a stubborn defence" - Poluga 14 lbxa4 .te6
evsky. 15 .te3 0-0
a3) 1 3 . . . d4 ! ? 14 lbe4 ( 1 4 lbe2 Petrosian must consent to the
is better - Polugaevsky) 14 . . . lbxe4 exchange of his good bishop, since
1 5 .txe4 l':.cS and 1 6 . . . .tbS - Polu 1 5 . . . lb d7 1 6 f4 ! g6 1 7 .td4 ( 1 7
gaevsky. .tc2 i s good according to Lilien
b) 1 3 'iVd4 ( ' ! ' - Botvinnik; Lil thal) 17 . . . 0-0 I S :ac l is very good
ienthal) looks even better, as White for White, as given by both Aver
places his queen in a central loca bakh and Speelman.
tion and uses it on the dark squares. 16 .tc5!
As every beginner is taught, the Fischer has a certain type of po
queen combines the movements of sitional advantage in mind and he
the rook and bishop. In fact, it is won' t be distracted by anything
more accurate to say that the queen else:
combines the movements of the a) 1 6 lbb6 :abS 17 .txa6 .tdS
rook and the potential movements I S lb a4 d4 'gives Black the initia
of both bishops. In the position at tive' - Botvinnik.
hand, Black is weak on the dark b) 1 6 lbc5 as ! 17 .td4 .txc5 I S
squares, so it makes perfect sense .txc5 - Botvinnik. This variation
to put the queen to work on the favours White but Botvinnik points
dark squares. out that it involves tempo loss ( 1 7
13 l':.el ! .td4 and then I S .txc5) and allows
Fischer cuts to the heart of the Black to sneak in 1 6 . . . as ! , when
position, as he did so often in his the a-pawn is not as weak as in the
best games. Petrosian was hoping game and serves to restrain White's
for the materialistic 13 .tbS? ! axbS queenside majority.
16 :fe8 The key lies in my remark that

17 i.xe7 l:txe7 the bishop 'had few chances for
18 b4! 'it>f8 activity ' . Does the bishop need to
19 ttJc5 .tc8 be active to be a useful piece? No,
20 f3! it doesn 't: "For although it was
This takes away e4 from the 'bad' , the bishop was holding to
black knight and clears a path to gether the black position. After its
the centre for the white king. exchange, the white rooks can
20 ':ea7? show their paces in a way which
This was Black's last chance to was not possible before." - Speel
set up a defence and he misses it. man.
Botvinnik suggests 20 . . . .:xe l + 2 1 Polugaevsky adds the enlighten
l:lxe 1 tt:JeS 22 'it>f2 ttJc7 23 'it>e3 'it>e7 ing remark that Fischer often used
24 'it>d4+ 'it>d6 "and the worst is be the method of 'transforming one
hind Black once is knight is on c7". advantage into another' . Imagine
21 lIe5! .td7 (D) for the moment that chess is mathe
matics and that at least to some ex
tent chess positions are maths
problems. There can be no doubt
that Fischer would have main
tained a solid advantage even if he
hadn't exchanged his knight for the
bishop, but in doing so, he is effec
tively ' simplifying the equation'
and making the solution to the
problem that much easier.
22 lIxd7
23 ':cl lId6
W This cedes the seventh rank to
the white rook but the immediate
22 ttJxd7+! ! threat was 24 llc6. Fischer is now
A sensational move, even for a able to reach an overwhelming
player acquainted with Bronstein position, as his rooks achieve tre
Boleslavsky. It's clear that the mendous activity, Black's knight
white knight was a tremendous remains inactive and the white
piece, while the black bishop had bishop has its choice of targets,
few chances for activity. Then why most notably a6 and dS .
did Fischer so willingly exchange 24 ':c7 ttJd7
his ' superior' minor piece? 25 ':e2 g6 (D)
"Black is in a sort of zugzwang. 34 :ce8 ! as winning for White.

2S . . . aS would be met by 26 bS and Further, he points out that the po si
2S . . . lLJb6 allows the other white tion is a kind of zugzwang , as a
rook to occupy the seventh rank." - move by the rook at a8 allows 3 1
Polugaevsky. : a7, while 30. . J:tdS 3 1 :e6 is also
winning.
31 :ee7 lLJd5
32 :f7+ e8
33 :b7 lLJxb4
34 J.c4 1-0
White will soon mate w ith 3 S
:g7 or 3S :h7.
Cafferty relates that Fischer
considered this his best game of the
Petrosian match.
There seem to be fewer exam

W ples of the unexpected exchange of
bishop for knight. This is probably
26 f2 h5 a function of the pieces: it may take
In Botvinnik's view, Black's last a knight many moves to reach a re
chance was to 'try for some coun ally excellent square, whi le the
terplay ' with 26 . . . :b8 27 a3 as 28 bishop that it eventually elimi nates
bS a4. may have moved only once or
27 f4 h4 twice, as we saw in the above ex
Both Speelman and Polugaev amples. Similarly, a well- placed
sky suggest 27 . . . lLJb6 28 :ee7 :f6, bishop may not even have moved
when the latter remarks that Black in a game, as bishops are long
can hope to 'complicate matters' . range pieces and consequently can
28 f3 f5 be ' developed' , i.e., accomplishing
29 e3 d4+ really important things, even when
White also wins after the some they sit unmoved upon their origi
what superior 29 . . . lLJf6 30 d4 nal squares.
lLJe4 3 1 :ec2, which is evaluated
as winning by both Speelman and E. Jimenez Larsen
-
POlugaevsky. Palma de Mallorca 1 967

30 d2 lLJb6
Speelman gives 30 . . as 3 1 bxaS
. 1 e4 lLJf6 2 e5 lLJd5 3 lLJf3 d6 4 d4
l:has 32 :c8+ g7 33 J.c4 f6 dxe5 5 lLJxe5 g6 6 J.c4 J.e6 7 i.b3
The Unexpected Exchange 1 73
i.g7 8 0-0 0-0 9 e2 a5 10 lLlc3 c6

11 lLlxd5? (D)
B
centre but the black bishop was not
B poorly placed and would seem to
be an excellent defender. However,
"I do not understand such Larsen has a very powerful con
moves ! " - Larsen in Larsen 's Se tinuation in mind.
lected Games of Chess 1 948-69. 15 dxe5 d4!
The text allows Black to develop "Advantage for Black ! He has
his knight to c6 and also gives him the initiative in the centre and both
play against White's queenside. the white king 's pawn and his
11 ... cxd5 pawns on the queenside are weak."
12 a4? - Larsen.
This simply weakens White's It may surprise the inexperi
queenside further. Larsen gives 12 enced player to learn that White's e
c3 as better, although he remarks pawn is weak, while Larsen makes
that Black has "an excellent posi- no comment concerning the possi
tion". ble weakness of Black's pending
12 lLlc6 doubled e-pawns. Modern grand
13 c3 6 masters have demonstrated re
14 i.a2 (D) peatedly that 'a weakness is not a
14 St.xe5!! weakness' if it cannot be exploited
"Jimenez was very surprised - and that is the case here. Further
and afterwards expressed hi s admi more, the twin exchanges of minor
ration." - Larsen. pieces will leave White with an un
The text is an extremely unex remarkable bishop against a knight
pected exchange. It's true that the that has a rosy future, as it applies
white knight is well placed in the pressure to the e-pawn and has
chances of obtaining even greater As Larsen points out, White

activity. would suffer if he lost the e-pawn
16 .th6 because of the central posts that
Jimenez decides to pursue some Black's pieces would gain as a re
vague chances against the black sult.
king. It might have been better to 19 I:.f8
go after the e6-pawn immediately 20 g3
with 1 6 .txe6 fxe6 17 'iVc4 ! ? ( 1 7 A really ugly move. Larsen
.th6 is pointless now, because speculates that White played it so
Black replies 1 7 . . . I:. f5) 1 7 . . . ltJxe5 as to keep the d2- square under ob
18 "ilixd4 xd4 19 cxd4 ltJc6, when servation in the event that Black
Black has the better minor piece decides to push his d-pawn.
and it's likely that the d-pawn will 20 I:.f5
prove weaker than the e-pawns - 21 I:.adl?
Mayer. White misses an interesting
But even so, it's likely that this chance here. Larsen considers 2 1
position would dissolve to a rook h4 'relatively best' , but then adds
and pawn ending in which Black has 2 1 . ..d3 22 'iVd2 'iVd8 ! , with the
a weak e-pawn(s) , so there would point that the e-pawn drops after 23
certainly be defensive chances for I:. ad 1 ? ltJxe5 24 .txe5 I:.fxe5 25
White. I:.xe5 I:.xe5 26 Wixd3 ??, when
16 I:. fd8 Black has the choice of winning a
17 .txe6 fxe6 rook by 26 . . . I:.d5 or 26 . . . l':te l +.
18 I:.fel I:.d5 But what if White plays good
19 .tf4 (D) moves? For starters, he might try
for activity with 23 I:.e3 ! ltJxe5 24
b3 ! . Now that the knight has been
deprived of forks, 25 l':tae 1 has be
come a legitimate threat, while 25
c4 may also be annoying - Mayer.
Note also that the d-pawn is going
nowhere for now.
21 'iVb3!
22 h4 Wixa4
23 Wie4 'iVb3
24 cxd4 'ixb2
25 I:.bl "ilixd4
B 26 I:.xb7 Ildxe5! (D)
27 'iVbl
W W
Forced, as Black wins after 27 Bronstein - Petrosian
i.xe5 WHxf2+ 28 h l ':xe5 - Amsterdam, Candidates '
Larsen. Tournament 1 956
27 ':xe1+
28 xe1 WHd5
29 c1 lbd4
30 ':b8+ ':f8
31 Ilxf8+ xf8
32 St.h6+ e8
33 WHc8+ d8
34 c4 tiJf5 (D)
The game has settled down into
a lost ending for White. It con
cluded 35 St.f4 f7 36 St.e5 'ilid1 +
37 h2 'iUd2 38 WHc5 a4 39 St.c3
WHc2 40 g1 'id1 + 41 h2 'id5 W
42 'iUb4 "'f3 43 g1 a3! 0-1
White is out of luck, as 44 'i!Vxa3 Black's centre pawns have been
tiJd4 ! wins the bishop or gives forced into a formation that is gen
mate - Larsen. erally not considered favourable in
the King's Indian. Despite a pro
There can also come a time tected passed centre pawn, Black
when the 'unexpected' exchange has the problem that White can
loses its surprise value, as the par eventually place a knight at e3 and
ticular manoeuvre becomes an ac undermine the fS-e4 pawn chain
cepted part of middlegame theory. with g2-g4. This may also result in
1 76 The Unexpected Exchange
White gaining the initiative on the chances of operating on the queen

kingside. side with a later b2-b4. Further, the
16 i.e3 h6 white c-pawns may prove a prob
17 lbh3 i.xc3! (D) lem, as in a Nimzo-Indian. Finally,
it is critical that Black controls d4,
so White never has any means of
getting his dark- squared bishop on
the long diagonal.
In the subsequent play, Bron
stein proved unable to demonstrate
any real advantage, as Petrosian
concentrated his forces in the cen
tre and on the kingside: 18 bxc3
lbf6 19 a4 h8 20 lbf2 ltg8 21
h1 'ie8 22 ltg1 "ii'g6 23 "ii'd 2
i.d7 24 g3 l::t ae8 25 a5 lte7 26
W ltab1 i.c8 27 lIg2 l::teg7 28 l::tb g1
lbce8 29 h3 h5 liz_liz
This exchange, now widely ac
cepted in this type of King' s In Verdict: The Unexpected Exchange
dian position, appears to have been of a well-placed minor piece for
played for the first time in this another type of minor piece is
game. Black parts with his impres something that must always be re
sive bishop, but eliminates White's membered as a possibility.
1 4 S hattered Pawn Positions
The most common case of a bishop Further, we are only considering

being given up for a knight is when those cases where one side or the
a doubled pawn is created. In and other has pawns that are weak and
of themselves, simple doubled susceptible to attack.
pawn positions are too common for One thing that many players
us to consider. Further, a doubled don't realize about weakened pawn
pawn may not even be a real weak structures is that they also carry
ness, e.g., the cluster a7-b7-c7-c6 square weaknesses in their wake.
that arises in a variety of Ruy Indeed, as Nimzowitsch demon
Lopezes has drawbacks, but the strates in the following game, it is
susceptibility of the pawns to di much easier to undouble a pawn
rect attack is rarely one of them. than to restore pawn protection to a
However, there are many cases hole.
where a doubled pawn is very
weak, for example, the cluster c3- Mattison - Nirnzowitsch
c4-d4 that White takes on in a vari Karlsbad 1929
ety of lines in the Nimzo-Indian
Defence. These ' shattered pawn' 1 d4 tDf6 2 c4 e6 3 tDc3 .tb4 4
positions can be compared to gam tDf3 i.xc3+ ! ? 5 bxc3 d6 6 'iVc2
bits, i.e., the player with the lousy 'Wie7 7 .ta3? ! (D)
pawn structure has dynamic com
pensation (the bishops), but time is
not on his side, as his shattered
pawns will result in his defeat if he
proves unable to make something
of his dynamic advantage. Such
positions are difficult to treat in a
systematic fashion, as having shat
tered pawns is always unfavour
able in general. This chapter is the
longest in the book, but the reader
should realize that we won't come
close to exhausting the subject. B
178 Shattered Pawn Positions
Mattison hopes to undouble his numerous weak points that the

c-pawn by playing 8 c5 , but this black knights can attack.
can be prevented, after which his 18 .tel a6! (D)
bishop will prove misplaced.
7 c5
8 g3?
In general, the bishop should be
developed to d3 , as then it defends
the weakened c4-pawn and makes
it easier to expand in the centre
with e4. Further, it might be possi
ble to translate the bishop ' s place
ment into a later kingside attack.
8 b6
9 .ig2 .tb7
10 0-0 0-0 W
11 liJh4 .txg2
12 <t>xg2? Fully in accordance with the
Now White's knight remains out principle that one should open
of play. A superior approach was lines in the part of the board where
1 2 liJxg2, followed by 1 3 e4 and 14 one has the advantage.
liJe3 , when the knight defends c4 19 bxa6 Iha6
and could obtain active chances 20 dxc5 bxc5
from its central post. White ' s pawns, which were one
12 'i!Vb7+ body as recently as move 1 8, are
13 <t>gl WNa6! now separated into three islands.
14 'i!Vb3 liJc6 The holes in White's pawn struc
15 Ilfdl liJa5 ture caused by the initial doubling
16 'i!Vb5 WUxb5 of the c-pawns remained after the
17 cxb5 liJc4 pawns were undoubled and opti
Here we see an important dis cally free of weakness. The c3-
covery of Nimzowitsch 's. Earlier, pawn is now 'weak' to the naked eye,
White had a weak pawn that was as it is isolated, but it was weak
subject to attack at c4. Mattison even when White had pawns at b5
managed to undouble it, i.e., make and d4, as the square was a hole.
it 'healthy' again, but the weakness 21 liJg2 liJd5
of the square c4 is left behind. 22 lId3 :fa8
Thus, while White ' s pawns are 23 e4 liJe5!
now back in one island, he has 0-1
Shattered Pawn Positions 1 79
White loses three pawns after 24 One refinement that White can
l':t d 1 lbxc3 25 lIn l1 xa2 26 I ha2 try is 7 i.e3 lbfd7 8 :t c l ! , which
lbf3+ ! 27 'it>h l l1 xa2, when the e takes prophylactic action against a
pawn will also go. possible exchange at c3 . However,
this line was still relatively unex
Nimzowitsch's games in the plored in 1 973 and the dangers as
Nimzo-Indian are very influential. sociated with an exchange at c3
Sometimes the ideas even pop up were not yet known.
in other openings. 7 lbfd7
8 i.e3 .i.xf3
Cuellar - Tal 9 .i.xf3 lbc6
Leningrad Interzonal 1 973 Black now has three minor
pieces that can potentially attack
1 lbf3 d6 2 d4 lbf6 3 c4 g6 4 lbc3 the hole at d4, while White has
.i.g7 5 e4 0-0 6 .i.e2 .i.g4 (D) only two.
10 d5 lba5!
11 .i.e2 .i.xc3! (D)
This is an instance of Changing W

the Colour of a Bishop, as Black is
prepared to exchange his light This might be seen as an exam
squared bishop for the knight at f3 , ple of an Unexpected Exchange, as
which guards the dark-square hole Black trades off his apparently
at d4. This sensible move has been powerful fianchettoed bishop. How
used by at least four world champi ever, in return, Black inflicts dou
ons (Botvinnik, Petrosian, Spas sky bled c-pawns that will be prone to
and Tal). attack and also secures c5 as a per
7 0-0 manent square for his knights.
12 bxc3 e5! 19 g3 :f7

13 dxe6 20 .tg2 :efS
A difficult decision, for while it 21 .th3 rbh8
makes sense to 'open the game for 22 l::te2 e5!
the bishops' , their scope is not re 23 f5 g5!
ally increased by this exchange, Both sides have set their sights
while Black regains the e5-square on the centre. White's failure to
for manoeuvring. Yet alternatives open the game has dire repercus
result in the position remaining sions for him, for while he has
closed, which would certainly de gained a protected passed pawn at
light the knights. f5 , his king 's bishop has become
13 fxe6 even worse and the pawn at e4 is
14 f4 Wlie7 now a fixed weakness.
15 'ii'a4 b6 24 .tg2 :g8
16 llae1 lbb7 25 h3 h6
17 .tf3 lbbc5 (D) 26 Wlidl lbf6
27 .txc5 (D)
W
B
White faces a dilemma in what
follows. It's true that he has two The game enters a new phase.
bishops versus two knights, but the White manages to double the black
fact that the knights have a good c-pawns, but he is now left in a
square on the colour of White 's good knight versus bad bishop
good bishop will tempt an ex game.
change at c5, after which White ' s 27 dxc5
remaining bishop is extremely bad. 28 :d2 lbe8
18 'ifc2 l::tae8 29 :d8 IlgfS
Shattered Pawn Positions 181
30 ':a8 a5 sets about mobilizing his queen

31 .tf3 lbd6 side majority ; the fact that both
32 'iWd5 lba8 sides have broken queenside pawns
33 xa8+ l:U8 allows him to create a passed
34 c6 ':f6 pawn.
White made an effort to break 45 b5
into the black position via the d 46 cxb5 cxb5
file, but the net result has been fur 47 d3 'iVc6
ther simplification. 48 h4
35 'iWd5 \t>g7 Cuellar tries to stir up some ac
36 :d1 ':f8 tion on the kingside. Black is too
37 i.h5 liId8 well placed in the centre for White
38 WNc6 \t>f6 to make progress there, e.g., 48
39 \t>g2 WNd7 "it'd5 ! ? xd5 49 ':xd5 lbb7 50
40 "it'd5 WNe7 ':xd8 lb xd8 5 1 .te8 b4, with the
41 "it'c6 "it'd7 division:
42 WNd5 c6 a) 52 cxb4? cxb4 53 .tx a4 bxa3
43 'iWd3 'iNc7 54 .tb3 lbc6 and the black a-pawn
44 WNe2 a4 (D) nets a piece - Mayer.
b) 52 .txa4 bxa3 53 .tb3 is bet
ter, as the black knight doesn' t
have access to b4 and d4 - Burgess.
48 lbf7
49 hxg5+ hxg5
50 "it'e3 Ilxd1
51 .txd1 (D)
W
45 a3
The a-pawn will subsequently
prove difficult to defend, but the al
ternative was to allow yet another
pawn to be fixed on a light square
by a subsequent . . . a4-a3 . Now Tal B
White 's chances of disturbing It's also possible for shattered

the flow of the game have dwin pawns to be a problem in more
dled to practically zero. It's true open positions. In the following
that the presence of queens may al game, one can always point to
Iow a swindle, but it also allows Tal open diagonals for the bishops, but
to pursue the advantages of Capa they never seem to reach them.
blanca's Ending (see Chapter 1 7)
while maintaining the safety net of Petrosian - A. Nielsen
a winning minor-piece ending. The Copenhagen 1 960
conclusion, while interesting, did
not really offer White any practical 1 d4 f5 2 .tg5 g6 3 tZJd2 .tg7 4 e3
chances of saving his position: tZJf6 5 e3 d6 6 tZJgf3 tZJe6? ! 7 'i!t'b3!
51 ...tZJd6 52 .te2 b4 53 ..t>h3 b3 54 (D)
.tb1 e4 55 ..t>g4 tZJf7 56 ..t>f3 'iHd6
57 't\Vc1 ..t>g7 58 ..t>e2 'iHh6 59 'iHg1
'i!t'h5+ 60 ..t>e3 tZJd6 61 ..t>d2 'i!t'f3
62 'ife5? ! (better practical chances
were offered by 62 'i'a7+, since
62 . . . ..t> f6 63 'i!t' b6 forces Black to
take perpetual check due to the
threat 64 't\Vc7+, while 62 . . . ..t> fS 63
't\VbS+ is also a mistake; however,
after 62 . . . tZJf7 63 'i'g l g4 ! , White
is lost, as his pawns can't be held in
the long run) 62 'iHg2+ 63 ..t>c1

b2+ 64 ..t>d1 'iHf1+ 65 ..t>d2 tZJf7 B

(Black also wins with 6S . . . 'i!t'xbl ,
when it appears that his king can A good ' nagging move ' . White
escape perpetual check; however, delays Black's castling and tempts
Tal 's pragmatic decision is under him to alter the pawn structure with
standable, as one must always be . . . d6-dS . Given the further course
cautious about a mistake near the of the game, it seems that Black
end of a long game) 66 .ta2 'iHd3+ would have done better to castle at
67 ..t>e1 'iHxe3+ 68 ..t>f1 f3+ 69 move six .
..t>g1 'iixg3+ 70 ..t>h1 'iHe1 + 71 ..t>h2 7 h6? !
d2+ 72 ..t>h3 g4+ 73 ..t>xg4 'i!t'f4+ It's mistaken to chase after the
74 ..t>h3 tZJg5+ 75 ..t>g2 'iWxe4+ 76 bishop pair when the black king is
..t>h2 'iVe2+ 77 ..t>h1 b1+ 78 .txb1 not in a position to castle. More
xb1+ 79 ..t>h2 tZJf3+ 0-1 prudent was 7 . . . e6 ! ?, when Black
can meet 8 e4 with 8 . fxe4 9 ltJxe4

. . knights. However, Petrosian is able
h6. to demonstrate in what follows that
8 .txf6 i.xf6 the black position is very poor al
9 e4 ready, and quite possibly lost.
Petrosian plays to open the posi 13 'iHa4!
tion against the uncastled king. This apparently crude attack has
This manoeuvre, in conjunction a more subtle point: White will ma
with the exchange of bishop for noeuvre his knight to the most ad
knight, also has the effect of speed vanced square it can reach, i.e., c5 .
ing up the white knights . It' s also A strategist of lesser ability might
important that the black king 's have been satisfied with the c4-
bishop 'bites on granite' in the form square.
of the pawn chain b2-c3 -d4. 13 'iHd6
9 ... e5 14 ltJb3 .td7
10 .tb5! f8 15 I:.d1 'Wie7
Black faces the fact that he won't 16 ltJc5 i.e8
be able to castle by normal means, 17 b4 g7
as 1 O ... i.d7 1 1 d5 wins material. Black's bishops haven' t man
11 i.xc6 bxc6 aged to gain any meaningful ac
12 dxe5 dxe5 (D) tiv,i ty, thus his position must be
regarded as lost in view of his
wrecked queenside pawns and the
general looseness in his kingside.
18 0-0 :f8
19 'iHa6 fxe4 (D)
W
Black has obtained two bishops
vs two knights at the cost of great
damage to his pawn structure, but
he has not had to yield any obvi W
ously good square to the white
20 lDd2! many defences, e.g., the King's In

The knight at c5 is already dian or certain Sicilians, it is com
placed optimally, so Petrosian mon for Black to sacrifice a pawn
strives to improve the placing of its if he can obtain a dark-squared
colleague. bishop for a knight. It should really
20 e3 (D) come as no surprise that players are
also willing to accept a broken
pawn structure, assuming that they
get an unopposed bishop with a
bright future ahead of it.
Reichenbach Sosonko
-
Mannheim 1 975
1 d4 tbf6 2 i.gS tbe4 3 i.h4 gS! ? 4

f3!? gxh4 5 fxe4 (D)
21 lDde4! exf2+
22 l'hf2
Threatening to win the bishop
by doubling rooks on the f-file.
22 i.gS
23 l'hfS xf8
24 lDxgS hxgS
25 'iWb7 1-0
Black has no interest in watch B
ing the white a-pawn march down
the board. An excellent, if one White has given up the bishop
sided, demonstration of knights vs pair, but has a pretty centre and has
bishops in a shattered pawn setting. left Black with a degraded king
side. In return, Black has the
The side with the shattered bishop pair and the white position
pawns doesn' t always lose. Some is exposed to attack on the dark
bishops (generally fianchettoed squares. Black's mission now is to
kings' bishops) are so active that a disrupt the white centre and open a
premium is placed on them. In line for his queen for dark-square
operations. The move S . . . cS is seen 14 e3 .td7

frequently, but Black has another 15 .te2 0-0-0
logical approach available. 16 0-0-0 'fIie7 (D)
5 e5!?
6 lbO!?
This position is little explored,
so it's hard to be certain what's
White ' s best move. 6 e3 looks
reasonable, but then 6 .. :iVgS and
7 . . . .th6 cau ses him problems .
Another way o f propping u p the
centre is 6 c3, but this has the draw
back of taking away the queen 's
knight's most natural square.
6 exd4
7 'ixd4 :g8 W
8 lbc3
Also possible is S WNeS+ WNe7 9 The opening is over and Black
xc7 lb a6, and now 1 0 WieS looks has every reason to be happy. His
better for White. Instead, Keitling king 's bishop is a powerhouse, his
haus-Knaak, Bundesliga 1 99 1 con h-pawn is holding down the white
tinued 1 0 c4 bS 1 1 WNdS (there's g-pawn and it's difficult for White
no obvious objection to 1 1 WixbS , to draw up a good plan. It's true
as 1 l ...:bS 12 Wic4 stops 1 2 ... :xb2) that he could play liJfS at some
l l ... lbc7 12 WNd3 .tb7 1 3 lbc3 .th6 point, but Black would just chop it
14 lbd4 :gS , and here I S lbfS ! + off and still have the weakness at
would have won the exchange. e3 to attack.
However, Black can look for im 17 :dfl .th6
provements with 9 . . . lbc6 or even 18 h3?
S . . . .te7. How can this be right? Now the
8 lbc6 g-pawn is never able to move and
9 WNd2 d6 the black rook gains a nice square
10 lbd5 .tg7 to use in pressing on e3 .
1 1 c3 liJ e5 18 :g3
12 liJd4 c6 19 :f2 liJg6
13 liJf4 g5 20 .to lbe5
Black doesn' t accomplish any 21 .te2 .te6
thing special after 1 3 . . . liJc4 1 4 Sosonko has no objection to an
WNc 1 Wib6 1 5 e 3 - Mayer. exchange at e6, as that would allow
him to add a centre pawn and use Perhaps the secret of this game
the f-file for kingside operations. is that White also had shattered
22 cJtb1 llJd7 pawns. They don't look significant
23 .to llJc5 in the position after White's fifth
24 'ii'c2 ':'e8 move, but practice suggests that
25 llJfxe6 fxe6 (D) the pawn weaknesses are severe for
both sides.
There are cases where a player

already has the bishop pair, per
haps with some degradation of the
pawn structure, and further sacri
fices pawn structure for the sake of
piece activity. This is not as com
mon as one wQuld think. Perhaps
the explanation is that the addi
tional degradation of the pawn
structure usually allows the other
W side to do something nice, e.g., win
material or break in to the position.
White eliminated the bishop pair,
but his reward is a distinctly infe Smyslov - Botvinnik
rior opposite-coloured bishop mid Moscow World Ch Match (20) 1 957
dlegame.
26 ':'e1 cJtb8 1 e4 e6 2 d4 d5 3 llJc3 J.b4 4 e5 c5
27 ':'d2 llJd7 5 a3 J.xc3+ 6 bxc3 'iVc7 7 'iVg4 f6
28 'iVb3 llJe5 8 llJf3 llJc6 9 g3 'iVf7? (D)
29 .th5 J.xe3 In World Chess Championship
30 ':'xe3 ':'xe3 1 957, Golombek suggests the su
31 J.xe8 xe8 perior 9 . .cxd4 10 cxd4 'iVf7 .
.
32 xe6 g6! 10 dxc5!?

33 'iVxg6 hxg6 "With two bishops, White natu
Play has come down to a rook rally wants to open the game. He
and knight ending where Sosonko now obtains a tripled, but never
is able to demonstrate that White 's theless extra pawn. He also gains
multiple weaknesses (e4 and g2) the possibility of developing his
are decisive: 34 llJe6 llJc4 35 :f2 bishop at d3 without having to fear
a5 36 llJg5 a4 37 cJtc2 ':'g3 38 llJf3 the blockading advance . . . c4." -
llJe3+ 39 cJtd3 llJxg2 0-1 Smyslov in 125 Selected Games.
primary source is My Best Games

of Chess 1 935- 1 957, while 'Smys
lov 2' will be used for material
from the later 1 25 Selected Games.
12 lLlxe5 lLlxe5
13 'i!Vxe5 0-0
14 0-0 lLlc6
15 'i!Vg3 e5
16 i.e3
Golombek points out that 1 6
i.b5 e4 1 7 i.xc6 bxc6 1 8 i.e3 fa
W vours White, but is drawish due to
the opposite-coloured bishops.
There are a few other side-ef Botvinnik has managed to oc
fects of accepting the tripled c cupy the centre with his pawns,
pawns. The d4-square is cleared but while they look pretty, they
and the fourth rank is opened for can't really be used for active op
use by White 's heavy pieces. Fur erations. Indeed, they may even
ther, Golombek points out that the become a target, as Smyslov later
pawn at c5 serves to blunt Black's demonstrates.
play on the c-file. 16 ... i.f5
While this type of idea had been Smyslov considers the immedi
played before, e.g., by Panov in the ate 16 . . . i.e6 better.
1 930s, there can be little doubt that 17 llabl ! i.xd3
the present game did a great deal to This eliminates a bishop but
popularize it in a variety of French goes a long way to making the
Winawers. pawn at c5 into a healthy extra
10 . . . lLlge7 pawn. However, the threat was 1 8
1 1 i.d3 fxe5 i.xf5 and Botvinnik apparently
Smyslov considers this ex didn't care for Tolush's suggestion
change too simplistic. In My Best of 17 ... l:.ab8 (cited in Golombek).
Games of Chess 1 935- 1 957, he 18 cxd3 l:.ae8
considers 1 1 ... lLlg6 better, but is of Black completes his develop
the view that 12 exf6 favours ment, but Smyslov takes this ap
White . In 1 25 Selected Games, he proach apart with ease. Golombek
mentions 1 1 . . .i.d7 1 2 l::tb l 0-0-0, suggests 1 8 . . . d4 1 9 cxd4 exd4 20
but terms it ' not altogether sound' . i.d2 :ad8 as superior.
Further citations of Smyslov will 19 f4! (D)
use 'Smyslov l ' to indicate that the 19 ... 'Wic7
B W
Now simplification favourable \t>f7 29 \t>f2 b6 30 ':'b1 \t>e6 31
to White occurs. Smyslov consid ':'b5 d4 32 c4 I,xc5 33 Jt.. h 2 :f7+
ers the alternative 1 9 . . . e4 better, 34 \t>e2 J:1e7 35 :xc5 \t>d7+ 36 \t>d2
when his annotations offer two dif ':'e6 37 ':'g5 g6 38 ':'d5+ \t>c8 39
ferent continuations : Jt.. g 1 ':'f6 40 Jt.. xd4 lbxd4 41 lbd4
a) 20 d4 lba5 2 1 f5 lbc4 22 Jt.. f4 ':'f2+ 42 \t>c3 1-0
favours White but 'Black is not
without counterplay' - Smyslov 1 . One player who noticed Smys
b) 2 0 dxe4 ':'xe4 2 1 f5 ':'fe8 22 lov 's idea was Spassky. The next
Jt.. g5 lb e5 23 h3, and a subsequent game, while not a weighty strug
f5-f6 will give White a strong at gle, demonstrates some of the king
tack - Smyslov 2. side fury that White can develop by
c) Golombek offers a third idea means of the surprising exchange
in 20 f5 exd3 2 1 f6 ! , intending 22 dxc5 .
fxg7 or 22 .th6.
20 fxe5 ':'xfl+ Spassky - Nagaizev
21 ':'xfl 'iUxe5 Leningrad 1 967
22 'iUxe5 lbxe5 (D)
Smyslov has whittled down the 1 e4 e6 2 d4 d5 3 lbc3 .tb4 4 e5 c5
black centre and taken play into a 5 a3 Jt.. a5 6 'iUg4 lbe7 7 dxc5! ? (D)
pawn-up ending. While some care Of course, White's c-pawns
is still required before White con were not even doubled yet, but the
verts the extra pawn, no further text is an open invitation to triple
comment is necessary: 23 ':'d1 them. Instead, 7 'ixg7 or 7 b4 take
\t>f7 24 h3 lbc6 25 Jt.. f4 ':'e7 26 play into a different type of com
Jt.. d 6 ltd7 27 llfl+ \t>e6 28 ':'e1 + plexity.
B B
7 .txc3+ 21 .txe5 'iWt7
8 bxc3 0-0 22 Wlig4
9 .td3 lbd7 The queen returns to the king
10 lbf3 f5 side. Notice the way in which
11 exf6 lbxf6 White ' s queen (and later his rook)
12 'tib4 .td7 uses the open fourth rank as a
For the moment, White has an means of rapidly shifting from one
extra pawn. His real advantage, side of the board to another.
however, is to be found on the dark 22 'ifd7
squares, where his pieces have ex 23 llb4! (D)
cellent possibilities.
13 0-0 lbf5
14 4 .tc6
15 lbe5 Wlic7
16 1Ie1 llae8
17 .tf4 (D)
White's play is a model of Nim
zowitsch' s ideas on restraint and
blockade. Black is two pawns up in
the centre, but Spassky's pieces oc
cupy good central squares and it is
hard for Black to counter this.
17 lbh5 B
18 .tg5 lbf6
19 I:.abl lbd7 Here comes the rook ! White
20 .tf4 lbxe5 now has a winning kingside attack,
as the force he accumulates over - which is why the Samisch isn ' t
whelms the black defences. seen much nowadays - a talented
23 l:tf7 attacker can still make the bishops
24 5 g6 sing.
25 l:tg4 liJg7
26 6 i.b5 Lilienthal - Smyslov
27 i.xg6! Piirnu 1 947
The most destructive finishing
blow, though it should be noted 1 d4 liJf6 2 c4 e6 3 liJc3 i.b4 4 e3
that 27 l:t h4 was also good enough c5 5 a3 i.xc3+ 6 bxc3 liJc6 7 i.d3
to win. b6 8 liJe2 d6 9 e4 liJd7 (D)
27 hxg6
28 l:th4 'it>f8
29 i.d6+ IUe7
30 'ilixg6 1-0
Spas sky later showed his affin

ity for this method of treating the
French Winawer in his 1 978 Can
didates' match with Korchnoi. Three
games saw the extraordinary 1 e4
e6 2 d4 dS 3 liJc3 i.b4 4 eS cS S a3
i. xc3+ 6 bxc3 liJe7 7 liJf3 i.d7 8
dxcS ! ? Although Spas sky scored w
poorly in these games, by the third
attempt he had refined play to the This manoeuvre prevents White
point of 8 .. :ifc7 9 i. d3 i.a4 1 0 from pinning the knight and may
Ilb l ! ! , when the white rook can allow the black f-pawn to join in
later swing into action via b4, in a the struggle for the centre.
similar manner to what we just saw 10 0-0 e5
in the game vs Nagaizev. 11 f4!?
Lilienthal begins immediate ac
Another setting in which the tion on the kingside. The danger is
player with the bishops takes on that his position isn ' t developed
shattered pawns is in the Samisch enough for it to work, but a greater
Variation of the Nimzo-Indian ( 1 danger lurks in waiting too long to
d4 liJf6 2 c4 e6 3 liJc3 i. b4 4 a3 ! ? take action. The player with shat
or 4 e3 and 5 a3 ! ?) . While Black tered pawns is usually better off
usually fares well in these positions thinking of his position as being a
gambit, Le., time is against him and his pieces will display greater ac
he must create his chances before tivity than their black counterparts,
his inferior pawn structure can be thus distracting Black from the at
taken apart. tack on the c4-pawn.
11 e7 A sharper approach for White
12 fxe5 dxe5 was 1 8 :f5, with the point that
13 dxe5! ? 1 8 . . . lb a5 is answered by 19 i.g5 ! ,
A surprising move, a s i t leaves when 1 9 . . . lb axc4 2 0 lbxf6+ or
White's pawn structure a wreck. It 19 . . . lbexc4 20 e5 ! both give White
was also possible to play 1 3 d5 lbaS an initiative.
1 4 lb g3 , angling for f5 . The draw
back to this approach is that Black
can play 14 . . . g6 and re-route his
queen's knight to d6, where it will
be a very well placed blockader.
13 ... lbdxe5
14 lbf4 i.g4
Lilienthal's 'argument' in this po
sition is that the knight outpost at
d5 and the open f-file make up for
the sorry shape of his pawn struc
ture. With the text, Smyslov pur
sues simplification, which makes B
sense in view of his superior pawn
structure. However, given that his 18 0-0
light-square bishop is much better 19 i.xe5 lbxe5
than White's, this whole exchang 20 a4! lbc6
ing manoeuvre is probably a mis 21 WNg4 'it>h8
take. 22 a5! lbxa5
15 .te2 i.xe2 Lilienthal has conjured up play
16 WNxe2 f6 on both sides of the board. 22 ...bxaS
17 lbd5 WNf7 would keep control of e5 , but the
The queen serves to protect the black queenside pawns would be
king, but it is also exposed to tac targets after 23 %lib l or 23 "ii'g 3, in
tics on the f-file. Better was tending to play to d6 or e3 . Still,
1 7 . . :it'b7 . this looks like a better way to play
18 i.f4!? (D) than the text.
Lilienthal intends a simplifying 23 e5 (D)
aprroach by which he hopes that 23 ... f5
The alternative was the equally

miserable 33 . . .'ifxd5 34 cxd5 .
B
24 'iUe2 I:.ae8
25 I:.ae1 I:.e6 w
Smyslov prepares to sacrifice an
exchange to establish a blockade at After the text, the rest requires
e6 . The alternative was 25 . . . lLlc6 no comment: 34 I:.xe5 'ifxe5 35
26 e6 'ifg6 27 e7 , and eventually I:.xe5 I:.xe5 36 'it'd6+ I:.e6 37 'iff8+
lLld5-c7 will win the exchange at e5 38 'iUb8+ f6 39 'iUxa7 g5
e8. 40 'iUd7 Ile4 41 d8+ h6 42 h4
26 lLlf4 I:.fe8 g7 43 'iUc7+ h6 44 'it'xb6 10
27 lLlxe6 'it'xe6
28 'iUd3 g6 Verdict: Not all doubled pawns are
29 I:.f4 lLlc6 subject to direct attack. Shattered
30 'it'd5 lLlxe5 pawns are isolated or doubled
31 f1 ! pawns which are at risk from direct
Lilienthal steps away from attack. In general, the weaknesses
3 1 . . . lLlf3+ and prepares to work on of at least some of the squares oc
the pinned knight with his heavy cupied by the shattered pawns are
pieces. permanent. Shattered pawns are al
31 g7 ways a liability, though inventive
32 I:.f2 f6 piece play can sometimes make up
33 I:.fe2 h5 (D) for their possession.
1 5 The Ruy Lopez Ending
One of the most interesting end Benjamin - Brooks

game battles that arises between USA 1 991
bishop and knight is when one side
has accepted a crippled pawn ma
jority by allowing the opponent to
exchange bishop for knight. This
ending occurs regularly in connec
tion with the Ruy Lopez Exchange
Variation ( 1 e4 e5 2 t2J f3 t2Jc6 3
.tb5 a6 4 .txc6 dxc6), when White
eventually play s d2-d4 and gains
a 4-3 kingside pawn majority.
Black's queenside majority is crip
pled, i.e., unable to produce a
passed pawn on its own if handled W
properly by White. Thus, Black is
in the position of having to do 11 t2Jf3 0-0-0
something before it's too late. Ide 12 :el :e8
ally, this means using the bishops 13 e5 i.d8
to generate sufficient counterplay 14 .td2 t2Je7
in the middle game. However, if an 15 t2Jg5 :ef8
ending arises, White will aim for 16 e6 (D)
his ' dream position' : a king and 16 f6
pawn ending where he has retained Now White has a passed e-pawn,
his superior pawn structure. but Brooks was understandably re
Normally Black has two bishops luctant to play 1 6 . . . fxe6 17 t2J xe6
vs bishop and knight, or else : f7 1 8 .tc3 , with strong pressure
bishop vs knight. Here we see a for White.
case of what can happen to Black if 17 t2Jn :hg8
he is unfortunate enough to part 18 :adl l':te8
with the bishop pair, keep same 19 g4 t2Jg6
coloured bishops on the board and 20 cJtg2 t2Jf8
fail to straighten out his pawns . 21 t2Jxd8 cJtxd8
194 The Ruy Lopez Ending
B W
22 i.b4+ 'it>c8 28 llde1 'it>d8
23 J.xf8 ':gxf8 29 a4 Ilg7??
Benj amin has taken play into a An unfortunate blunder that robs
double rook ending where he is ef us of seeing Benjamin win the end
fectively a pawn up. Double rook ing. After 29 . . . 'it>e8 the game would
endings are notorious for the coun continue for quite some time.
terplay they offer the defender - af 30 g5 1-0
ter all, they're a rook ending times
two - but here Black's rooks are Brynell - Geller
passively placed and the only ques Berlin 1 991
tion is how White will break in.
24 h3 ':e7
25 f4 f5
26 ':e5 g6 (D)
White attempts to raise the
blockade at f5 while Black seeks to
maintain it. Following the natural
26 . . . fxg4 27 hxg4 1'hf4, White has
2 8 11f5 ! , exploiting Black's unfor
tunately placed king. Then 28 ... ':xf5
(28 . . . g5 29 ':xf4 gxf4 30 ':e l ) 29
gxf5 wins for White due to his ac
tive king and Black's exposed king W
side pawns (29 . . . g6 30 f6 ':xe6 30
f7 llf6 fails to 31 ':f1 ). It helps the side with the bishops
27 'it>g3 11f6 if he's tactically talented. In this
The Ruy Lopez Ending 195
position, Black still has his bish 21 ttJxg6

ops, but has had to take on an iso Interesting tactics arise after 2 1
lated e-pawn. This may serve as a tDd7 . when Black has a choice to
target for White, but it also offers make:
its advantages to Black in the form a) 2 1 . . .:xf2? looks natural, but
of play on the f-file. Further, White White does well in the complica
won't find it as easy to create a tions of 22 tD xc5 tDxc5 23 'it' xf2
passed pawn on the kingside, as J.xd3 24 J.a3 ! +-. winning a piece
Black can make it difficult for him - Mayer.
to play f2-f4. b) 2 1 . . .':f3 ! hits the white
14 h3 J.h5 pawns hard. e.g., 22 tD xc5 tDxc5
15 tDn J.b4 23 ':c3 tDe4 24 Ilxc6 tDxf2 ! . when
16 ':ed1 tDxe4 25 'it' g2 is answered by 25 . . . .te4 !
17 g4? ! and Black has a clear advantage
More in line with White ' s strat Mayer.
egy was 1 7 ':xd8+ ':xd8 1 8 ':xd8+ This latter line bolsters the view
'it' xd8 19 tD xe5 , when White can that White should have removed all
try to grind Black in the minor of the rooks earlier.
piece ending - Mayer. 21 hxg6
17 .tg6 22 J.d4 J.xd4
18 tDxe5?! 23 ':xd4 tDc3? !
It was still best to remove all the Black goes after the wrong pawn.
rooks starting with 1 8 ':xd8+. Instead. 23 . . . tD xf2. hitting the h
18 l::txd3 pawn and threatening 24 . . . .:f3 .
19 ':xd3 J.c5! was stronger - Mayer.
20 tDe3 l::tf8 ! (D) 24 'it'g2 tDxa2 (D)
w W
2S f4? Kindermann - Vojska

Only now is Black winning. Katerini 1 992
White had a pretty drawing idea
with 25 lle4 ! (threatening both 26 1 e4 e5 2 ltJf3 ltJc6 3 i.bS ltJf6 4 0-0
lile6 and 26 11e7) 25 .. .'d7 26 ft d4+ ! ltJxe4 S d4 ltJd6 6 i.xc6 dxc6 7
=. White just checks back and forth dxeS ltJfS 8 'ii'xd8+ 'it>xd8 (D)
between e4 and d4, while any
black move away from the centre
allows a rook invasion - Mayer.
2S as
26 fS gxfS 27 ltJxfS ':'f7 28 h4
ltJb4 29 c4 g6 30 ltJe3 ':'d7 31 ':'e4
ltJd3 32 hS gxhS 33 gxhS ':'h7 34
':'d4 ltJcs 3S cxbS ltJxb3 36 ':'d3
ltJcs 37 ':'c3 llxhS 38 bxc6 'it>b8
(D)
W
This intriguing position is a case
where Black deliberately pursues a
'Lopez Ending' . However, there
are two maj or differences as com
pared to the normal Lopez Ending.
First, the white pawn is at e5 rather
than e4. This means that a certain
amount of flexibility is lost from
W his pawn structure, e.g., Black can
aim to set up a light square block
The ending is a win for Black, ade against the white kingside ma
assuming that White can ' t run him jority and his pieces can use the d5-
out of pawns. The rest requires no and f5-squares. However, a more
comment: 39 'it>f3 'it>a7 40 'it>g4 important factor is that Black has
l1gS+ 41 'it>xgS ltJe4+ 42 'it>fS ltJxc3 lost the right to castle. It's true that
43 'it>e6 a4 44 'it>d7 ltJbS 4S ltJdS queens are off the board and Black
a3 46 ltJb4 'it>b6 47 ltJdS+ 'it>c5 48 almost never loses this position as a
ltJxc7 ltJxc7 49 'it>xc7 a2 50 'it>b7 result of a quick attack on the king,
al'ii' 0-1 but he has trouble connecting his
rooks. He may aim to exchange a

pair of rooks or all of the rooks at
d8, but this has the drawback that
play becomes ever more simpli
fied.
9 lbc3 .te6
White could now play 10 lb g5 ,
when the black bishop at e6 will be
exchanged off. Then all question of
this ending having any interest
from the standpoint of ' bishops
vs pawn majority ' disappears and B
Black is either hoping that he can
achieve active play through cen A general restraining move di
tralization or may even just push rected against White 's kingside
wood in the hopes of holding a majority.
slightly inferior game. 18 ':fel
10 h3 e8 Or 1 8 lb xd5 ! ? ':xd5 1 9 lbc4 -
11 b3 .tb4 Kindermann.
12 lbe2! 18 i.c6
White avoids the exchange 19 e6 f6
. . . .txc3 , as then Black would have 20 lbc4 g5
created opposite-coloured bishops. 21 ':xd8+ xd8
Further, he can now play c2-c4, 22 lbd3 g4
when d5 will be taken away from 23 hxg4 hxg4
the black pieces. 24 lbxc5 i.xc5
12 .td5 25 .txc5 (D)
13 lbd2 (D) An opposite-coloured bishop
13 .. c5? position finally arises, but hardly
Black takes control of d4 but on the terms that Black desired.
now his king's bishop gets pushed White is a pawn up and has more
out of play. In Informato r55, Kin active pieces, so it is difficult for
dermann suggests 13 .. J;t d8 14 c3 Black to put up genuine resistance.
.tc5 and 13 . . . .txd2 14 .txd2 c5 as 25 lbh4
superior continuations. 26 lbe3 f5
14 c3 .ta5 27 ':dl+ c8
15 i.a3 lId8 28 lbd5 i.xd5
16 ':adl .tb6 Now Black will win back the
17 lbf4 h5 pawn, but the cost is that he takes
game requires no comment: 34

IU8 I:.xh4 35 lbf5 lIh7 36 I:.g5
I:.d7 37 I:.xg4 I:.dl+ 38 <J;;; h 2 I:.d2
39 I:.g7+ <J;;; b 6 40 I:.f7 I:.xa2 41 g4
<J;;; e 5 42 g5 <J;;; d6 43 g6 lIa5 44 f4
1-0
Karpov - Browne
San Antonio 1972
B
play into a bishop vs knight ending
(see the chapter on 'The Grindable
Ending' ) .
29 I:.xd5 I:.e8
30 .te3! ':'xe6
31 .tg5 e6
32 lId8+ <J;;;e7
33 .txh4 I:.h6 (D)
W
White is doing well in the dia
grammed position, as he has the
more flexible pawn structure. Kar
pov decides to step up the pressure
with . . .
17 d4! I:.fd8
18 lIedl
Threatening to win the ex
change with 1 9 dxc5 dxc5 20
t:tJ e7+. Browne elects to exchange
W at d4, but now Karpov will have a
4-3 queen.side majority, while
Surprisingly, White has no way Black's kingside majority is crip
of holding on to his extra piece, but pled.
he can go after B lack' s exposed 18 exd4
kingside pawns. The rest of the 19 exd4 <J;;;f8 (D)
of importance, despite its impres

sive- looking diagonal . It ' s truly
an 'over-rated bishop' (see Chapter
8).
26 !:iJxb4 a5 (D)
W
20 c5!
I trust that you 've heard the ex
pression ' Passed pawns must be
pushed' ?
20 . !:iJa7 W
21 !:iJe3!
Avoiding 2 1 !:iJ xb4? ! J.xg2 22 27 !:iJd5
!:iJ xa6 J.f3 , when Black's bishops Karpov maintains a large advan
give him defensive chances . In tage on the queenside, because his
stead, Karpov plays to exchange knights do a better job of attacking
off Black's queen's bishop, which the a-pawn than Browne 's bishop
is providing defence against the and knight do of defending it.
further advance of the c-pawn. 27 ':c6
21 J.xg2 28 !:iJe3 l:.c5
22 'it>xg2 dxc5 29 !:iJf4 J.h6
23 dxc5 l:.xd2 30 lId5 l:.xd5
24 l:.xd2 l:.c8 31 !:iJfxd5 J.xe3 (D)
25 !:iJd5 ':xc5 Black feels compelled to elimi
Browne eliminates the passed nate a knight and hope to draw a
pawn. Inferior is 2S . . . as 26 !:iJb6 technically lost ending. However,
':c7 27 ':d8+ 'it>e7 28 l:.g8 +-, as this proves impossible, as Karpov
White threatens the bishop and a displays his usual precision: 32
fork at dS. !:iJxe3 'it>e7 33 'it>f3 !:iJc6 34 !:iJc4
One problem Browne has in 'it>e6 35 'it>e3 'it>d5 36 a3 'it>e6 37
what follows is that his bishop re 'it>d3 'it>d5 38 f3 h6 39 'it>c3 h5 40
ally can't attack or defend anything 'it>d3 f6 41 f4 g5 42 !:iJe3+ 'it>e6 43
'it>xa5 'it>e4 49 b4 'it>f3 50 b5 'it>xg2

51 b6 CiJf8 52 'it>b5 CiJd7 53 a4
CiJxb6 54 'it>xb6 'it>f3 55 a5 'it>xf4
56 a6 'it>e3 57 a7 f4 58 a8'iH f3 59
'iHe8+ 1-0
Verdict: The Ruy Lopez ending,

which features bishops vs a supe
rior pawn majority, is generally
favourable for the side with the su
perior pawn structure. However,
W tough defence or inventive tactical
play can sometimes save such po
h4 gxh4 44 gxh4 CiJe7 45 'it>c4 sitions, or even turn them around
CiJg6 46 CiJg2 'it>d6 47 'it>b5 'it>d5 48 by 'uncrippling' the majority.
1 6 The Grind a b le End ing -
Rook and Bishop vs Rook
and Kni ght
The endgame of rook and bishop ending offers the defender too
vs rook and knight is one that arises much counterplay? Go into the mi
relatively frequently; explorations nor piece ending. The minor piece
of databases suggest that it occurs ending is too easy a draw? Break
in between 5% and 1 0% of all mas him in the rook and pawn ending.
ter games. Better yet, keep the remaining
' Grinding ' is the art of taking pieces on the board and grind him
favourable positions (frequently down!
endings, but also certain types of Above all, the 'grindable end
middlegame advantages) and win ing' is one more option for the
ning them. These aren' t 'winning player trying to win and one more
games ' of themselves, but games hurdle to clear for the player trying
that require some effort on the part to draw.
of the would-be winner. Fi scher is regarded as one of
Reference works on the ending the greatest proponents of the
typically do not pay any special at grindable ending. Indeed, Leonard
tention to rook and bishop vs rook B arden apparently referred to this
and knight. Indeed, Fine makes the material distribution as 'the Fis
point that the principles of such cher Ending' in one of his endgame
endings really don't differ from the books, though I have not seen the
underlying endings of rook vs rook book in question.
and bishop vs knight. This may be
true in terms of the principles , the Fischer Tal
-
general evaluations, etc . , but it isn't Curafao, Candidates '

entirely correct in a practical sense. Tournament 1 962
With a rook and minor piece on the (see diagram on the next page)
board for each side, the aggressor
has greater flexibility than in a sim Fischer is down a pawn but he has
pler ending. The rook and pawn active pieces. In particular, his king
202 The Grindable Ending - Rook and Bishop vs Rook and Knight
Playing to keep the tension,

which gives White more opportu
nities of realizing his advantage .
The Bulgarian Series assesses the
minor-piece ending that arises af
ter 3 3 I hc6+ bxc6 34 i.xg7 d7
as equal.
33
tDf6?
This allows White to gain a strong
initiative on the kingside. The Bul
garian Series gives 33 . . . I:.c5 34
w .ixg7 d7 as unclear.
34 f4! (D)
is better placed than its black coun
terpart.
30 I:.c1 I:. c7
In the Bulgarian Series volume
Mikhail Tal Games 1 949- 1 962 , the
editors give 30 . . . lId3+ 3 1 f4 d7
32 xf5 tD xa2 33 I:.e l as unclear.
Such an approach has the advan
tage that the black rook remains ac
tive.
31 i.f4 l:.c6
32 i.e5 tDd5
Recentralizing the knight. B
White does well after 32 . . . tD xa2?
33 ':xc6+ bxc6 34 f4! c5 35 xf5 34 g6
c4 3 6 .ixg7 c3 37 g4 c2 3 8 i.b2 The Bulgarian Series indicates
hxg4 39 h5 +- - Tal (cited in The that both 34 . . . lIc2 35 .id4 ! and
Bulgarian Series). This variation 34 . . . lI c4+ 35 i.d4! are good for
shows some of the characteristics White.
of a bishop versus knight with 35 f3 tDd7
pawns on both sides of the board. 36 i.d6 I:.c2
The knight has difficulty dealing 37 g3 I:.e2
with a passed h-pawn, while the White wins after 37 . . . lI xa2? 3 8
bishop likes the fact that there are g5 ! :f2 39 I:.d3, intending 40
pawns on both sides of the board. xg6 +- - Bulgarian Series.
33 I:.d1 ! Fischer's initiative on the kingside
The Grindable Ending - Rook and Bishop vs Rook and Knight 203
threatens to create menacing passed 4S ':b2 f4 (D)

pawns. The reason why lines based
on .. . l ha2 aren't dangerous is that
the black queenside pawns are not
advanced and would have a diffi
cult job advancing in the face of the
white rook and bishop.
38 gS l:.e6
39 i.f4 4Jf8
40 l:.d6 as (D)
w
46 i.xf4
Tal also considers 46 i.xf8 and
47 g4 to be winning for White. How
ever, one can understand Fischer's
desire to keep the better minor
piece for as long as possible.
46 1%fS+
w 47 h6 bS
48 i.d6 b4
41 h6! 49 g4! (D)
Tal analyses 4 1 ':xe6 4Jxe6+ 42
xg6 4Jxf4+ to a draw.
41 ':e2
Back again in the hopes of some
activity. White wins after 4 1 . . .b5
42 g7 l:.xd6 43 i.xd6 4Je6+ 44
xg6 b4 45 xf5 +- - Bulgarian
Series .
42 ':d2 %;Ie7
43 i.d6 l:.h7+
44 gS ':17
4J
Tal gives 44 . . . d7 ? 45 xg6
l:.h8 46 g7 ':e8 47 f7 ! as good B
for White.
Initiating the decisive break 61 l::tg7+ e8

through. 62 d6 b3
49 l':Ixf3 63 a3! 1-0
50 g5 tZJe6
White also wins after the con Fischer Taimanov
-
tinuation 50 . . . d7 5 1 J.xf8 ':xf8 Palma de Mallorca Interzonal

52 xg6 - Tal. 1970
51 xg6
52 J.e5 ':e3
53 f5 tZJfS
54 ':g2 ':f3+
55 J.f4 d7
56 g6 (D)
B
Here we have an additional pair
of rooks on the board. The white
rooks are more active, so the addi
tional material benefits him. Black
B once again faces the problem that
he has a knight vs a bishop on an
56 tZJe6 open board that has pawns on both
Or 56 . . . tZJxg6 57 ':xg6 ':f2 5 8 sides. Notice that if the rooks were
a3 ! and White wins - Bulgarian all off the board, Black would be
Series. able to play for a defence based on
57 g7 ':xf4+ the dark squares, while a pure rook
White's new-born queen shines (or two rook) ending would allow
after Tal's 57 tZJxf4 58 g8'iV tZJxg2+
. . . him to play for rook activity.
59 e4 :f4+ 60 e5 :f2 61 WNd5+ 37 g6
e8 62 WNa8+ f7 63 WNa7+. 38 h4 tZJh7
58 e5 ':fS 39 J.g4 tZJf6
59 gxfS 'iV tZJxfS 40 i.f3 ':d7
60 d5 a4 41 ':b5 ':d4 (D)
White is doing well despite his

pawn deficit, because the a-pawn
will prove difficult for Black to
handle.
46 g7
47 as ':e8
48 ':c1 lle5
49 ':a1 l:Z.e7
50 f2 t2Je8
51 a6 ':a7 (D)
42 c5!
There are always tactics in an
ending and Fischer was usually
alert to them. He doesn't mind that
his h-pawn now disappears with
check, because he's engineering a
breakthrough on the queenside.
42 ':xh4+
43 g1 ':b4
44 ':xb4 axb4 w
45 llc4 bxc5
46 l:Z.xc5 (D) Now the black pieces are tied
down, so all that remains is for
White to activate his king.
52 e3 t2Jc7
53 .tb7 t2Je6
54 l:Z.a5 f6
55 d3 e7
56 c4 d6
57 ':d5+ c7
58 b5 1-0
The black rook is effectively
locked out of play, so White has the
luxury of cleaning Black out on the
B kingside and then breaking the pas
sive defence on the queenside.
Petrosian - Lilienthal
USSR Ch (Moscow) 1 949
W
36 J.d3 lle7
W 37 ':ae1 lUe8
38 J.c1 lbc3 (D)
When two bishops are coupled Simplification helps the side
with a rook or rooks, the problems with extra material.
facing the knight(s) can be ex
tremely difficult. In this position,
Petrosian ' s light- squared bi shop
is momentarily stifled, but he can
free it up and weaken the black
central position.
29 h3! lbf6
30 g4 fxg4
31 hxg4
Now White has a passed f-pawn
and Black has significant weak
nesses in the e-pawn and h-pawn.
31 ':c7 W
32 cJtf2 h6
33 ':h1 e3+ (D) 39 ':xe7 ':xe7
Lilienthal hopes for counterplay 40 as b6
on the e-file, but it won' t prove 41 axb6 axb6
enough, particularly as White is al 42 J.d2 lbe2
ready well centralized. Now the knight can be trapped.
34 J.xe3 lbe4+ 43 c3 b5
35 cJtg2 lbt7 44 cJtf3 1-0
The bishop pair allows their White's kingside space advan

owner to ' gang up ' on critical sec tage makes it easier to activate his
tors of the board, frequently at long king.
distance. Here Black has the prob 29 'it>e7
lem that his a-pawn is exposed and 30 ':a5 lDc8
White has the advantage of three 31 ':e5+!
long-range pieces to one. Preparing to highlight one of the
advantages that bishops have over
Botvinnik Langeweg
- knights : they're usually easier to
Hamburg 1 965 exchange.
31 'it>d8
32 .tb5 ':e7
33 ':xe7 'it>xe7
34 .txe8 'it>xe8
35 'it>f4 'it>d7
36 .tc5 1-0
The knight will be captured if it
ever moves ; all resulting king and
pawn endings are lost.
Alekhine - Reti
New York 1924
W
25 ':c5 e4
White meets 2S . . . ':e7 with 26
':'c6.
26 .te2 h6
27 h3
"In view of the opponent's lack
of useful moves, White patiently
waits, so as to start active play in
the most advantageous situation.
The aim of both sides is to bring up
their kings." - Botvinnik in Botvin w
nik 's Best Games 1 947- 1 970.
27 lDfe8 The bishop pair can even create
28 'it>g2 'it>f8 dangerous attacks on the king in
29 'it>g3 the ending, assuming that at least
one pair of rooks is still on the

board. In this position, the absence
of rooks would make the draw rela
tively easy to secure for Black, as
all the pawns are on one side of the
board. As it is, Alekhine is able to
exploit his piece superiority on the
kingside in a surprisingly swift
manner.
37 fS! I:.a6
Black must defend his third
rank, because otherwise 38 hxg6+ w
hxg6 39 I:.d6 will win a second
pawn and establish a winning ad 43 l:.d7+ g8
vantage. However, the black king 44 J.xg6 1-0
position is still unstable and Alekh Saving the rook with 44 . . . l:.h8
ine is able to exploit the bind with allows White to advance the h
some unusual tactics. pawn with decisive effect.
38 h4! hS
39 g4! l:.aS Verdict: The grindable ending is
Black must try to prevent the still subject to typical methods of
white h-pawn from reaching h5. endgame evaluation, but it gener
40 fxg6+ fxg6 ally favours the side with the
41 gxhS I:.xhS (D) bishop. The addition of another set
42 J.gS! of bishops or a pair of rooks is fre
Trapping the rook and threaten quently extremely favourable for
ing to win it with 43 J.f3 . the side that has the bishop vs the
42 i.c3 knight.
1 7 Capa b l a nca's Theore m
"Wi + ttJ is better than "Wi + i..
i n the End i ng
Capablanca's contention that queen take play to a part of the board that
and knight are superior to queen the knight can't reach in time to de
and bishop in the ending is very in fend. A more interesting question
sightful. The positions with which is whether queen and knight is a su
he was concerned were too com perior material combination than
plex for his contention to be any queen and bishop in a middlegame
thing more than intuitive and setting. In some positions it is and
anecdotal, but he latched on to sev in some positions it isn ' t, but the
eral important factors. His reasons size of the problem is such that one
for favouring the queen and knight will probably have to wait for a su
are their ability to work together percomputer such as 'Son of Deep
smoothly and create a greater Blue' to address it before mere
number of threats than the queen mortals know the answer. Finally,
and bishop. The bishop, of course, it is well established that rook and
could encounter situations where knight is inferior to rook and
defence was required on the colour bishop in the ending, so it would
opposite it, in which case it would not be all that surprising if the side
be of no help. with the bishop were to be fa
However, one can come up with voured in a queen and knight ver
counter-proposals as to why Capa sus queen and bishop ending.
blanca could have been wrong. For In the long run, it would be nice
instance, the queen can move like a if Capablanca's contention could
bishop, so one would think that be tested by examining a very large
they could ' link up ' together on a database of grandmaster games to
diagonal and cause damage to the see if this ending could be shown to
side with the knight. The bishop is favour definitely one side or the
a long-range piece, so while it's other. To my knowledge, a reliable
true that it can't defend all the database of this type does not yet
squares a knight can attack, it can exist and there are a number of
210 Capablanca 's Theorem - 'iV+lLl is better than 'ii' + ..t in the Ending
problems that will need to be ad could use one piece to contain

dressed before it becomes feasible Black's e-pawn and the other to at
to construct such a database. In the tack the queenside.
meantime, we must be satisfied 24 lL'la4 'iVg5
with our human ability to general 25 Wif4 'iWh5
ize and abstract great masses of in 26 lL'lc3
formation. Gligoric has managed to repulse
the first attempt to penetrate on the
Najdorf - Gligoric dark squares by threatening his
Saltsjobaden lZ 1 948 own queen invasion. However, it's
difficult for his bishop to partici
pate in attacks, so it seems likely
that sooner or later White will be in
a position to penetrate while his
queen or knight serves to shelter
his own king.
26 h7
27 h4 g8
28 f2 Wie8
29 lL'le2 Wic8 (D)
W
Here we have a particularly in
teresting situation, as there are six
pawns apiece still on the board. Of
course, Gligoric probably wished
that there were fewer pawns, as his
bishop is bad and he has weak dark
squares that can be attacked by Na
jdorf's queen and knight. The pres
ence of protected passed pawns W
complicates matters for both sides,
because it means that a number of 30 el
endings will contain counterplay The blocked centre allows Naj
chances. Still, one must assume dorf to centralize his king. This is a
that the exchange of queens would big advantage for any pure minor
make White's job easier, since he piece endings that might arise, but
Capabianca 's Theorem - 'iIi'+tZJ is better than 'iIi'+i. in the Ending 211
he must be careful not to slip up 41 tDd1 'iVe6

and allow Black to launch an attack 42 tDe3 'iVb6 (D)
on the king.
30 'iVe8
31 'iVc1 'afg6
32 'iVc3 'iVg4
33 'iVb3 b5
This creates further dark -square
and queenside weaknesses, but
something like this was inevitable
unless Gligoric wished to pursue
purely passive defence with his
queen.
34 tDf4 i.f7
35 'ilie3 (D) w
43 tDf5
The knight has finally managed
to break into the black position.
White's queen now looks to join up
with it, so Gligoric tries one last
stab at counterplay against White's
king.
43 b4
44 a4 b3
45 'iVe3 'iVb4+ (D)
B
35 g5? !
This takes away f4 from White's
knight but the pawn at g5 is now a
target and the dark squares near the
black king are also opened up.
36 hxg5 hxg5
37 tDe2 cJ;g7
38 tDc3 .te6
39 'iVf2 cJ;g8
40 cJ;d2 i.f7 w
212 Capablanca 's Theorem - 'it'+tD is better than 'it'+ in the Ending
46 e2 'ic4+ 33 g1 tZJe3+
47 e1 'iVb4+ 34 h2 tZJg4+
48 f1 The king is driven up to the third
The white king has escaped dan rank. In a typical ending, White
ger; unfortunately for Black, his would have no objection to this,
own king is not so lucky. but with queens on the board, the
48 f8 guiding principles of the middle
49 'iVxg5 e8 game are more apt than those of the
50 tZJd6+ d7 ending.
51 Wig7 1-0 35 g3 h7!
This forces White to resolve the
Manakova - A. Kuzmin tension of the pin.
Alushta 1 994 36 'ixf8?
This allows the black queen ac
cess to e l , but after 36 .txfS 'ixc3+
37 f3 tZJe3 Black is better - Mayer.
The rest is butchery: 36 'ixc3+

37 f3 'ie1+ 38 f4 'id2+ 39 g3
tZJe3 40 'iVe7 'ixg2+ 41 4 tZJxd5+
0-1
Silman - Moskalenko
Pardubice 1 994
The combination of queen and

knight is in its element in attacking
situations. In this position, both
kings are endangered, but Black's
king will be able to escape to a col
our opposite that of the white
bishop, while White will discover
that the dark squares around his
king are difficult to defend. w
30 h4 f5
31 e8 tZJxd2 As we know, bishops really
32 xd6 tZJf1 + don ' t like closed positions, but
Capablanca 's Theorem - 'tIi+tZJ is better than 'tIi+i. in the Ending 213
there are few types of position that A typical method of attacking in

illustrate this as well as queen and queen and knight versus queen and
knight versus queen and bishop bishop endings. White has one less
endings. Here White's bishop has piece that can influence events on
become a Problem Bishop, as it the dark squares , so he must sit by
lacks active possibilities in light of helplessly as his pawns are snipped
Black' s kingside space advantage. off and his king opened up : 32
Further, the pawns at d7 and g7 b5 tZJxe3 33 e2 h6 34 a4
serve to shut down Black's second 'iVxf4 35 gl g3+ 36 h1 f4 37
rank, while White has no compara as tDxf1 38 'iVxf1 e3 39 a6 f3 40 a7
ble shielding of his second rank. e2 41 c1+ g5 0-1
Consequently, the black queen is
able to influence events on both the Boleslavsky - Panov
queenside and kingside. USSR Ch (Moscow) 1 940
27 i.f1 h7
28 d6 4
Now White must worry about
his d-pawn and his king position.
Unfortunately, even 29 f2
doesn't help, as then 29 . . .'iVd2+ 30
i.e2 h4 ! 3 1 gxh4 tZJd5 cracks
White's game.
29 'Wic7 'Wie1
30 'iVxb6 'iVxg3+
31 h1 tZJd5 (D)
Despite the reduced material,

White has a substantial advantage.
Black's pawns are exposed and he's
open on the light squares. Boleslav
sky turns his attention to an attack
on those weakened light squares.
43 tZJe5! g8
44 h2 e7
45 h3!
w Boleslavsky shifts his king off
the colour of the black bishop and
214 Capab/anca 's Th eorem - 'it'+tLl is better than 'it'+.t in the Ending
ensures that a capture at f4 can

never be met by a pin.
45 J.f6
46 tLlg6 'iVd6
47 tLlxf4 (D)
Angling for d6, which will make

targets out of both f7 and e4. One
thing that White must be careful
about here is a premature exchange
B of knight for bishop, as Black 's
practical defensive chances would
Now that White has an extra be high in any pure queen and
pawn, he can play for the exchange pawn ending.
of queens and win the minor-piece 53 i.e7
ending with ease. The game con 54 t2Ja5 J.b4
cluded: 47 J.d4 48 f3 a5 49
55 t2Je6 i.e3 (D)
'iVe6+ 'iUxe6+ 50 tLlxe6 i.f2 51
'it>g4 'it>f7 52 'it>f5 h5 53 tLlf4 h4 54
tLld3 1-0
Kupreiehik - Gavrikov
USSR Ch (Frunze) 1 981
Here the black position is more dif

ficult than one might think, as
White can play to pressure both the
e-pawn and the black king posi
tion.
51 t2Ja5 'fWb4 W
52 'iUd5 'iUxa4
53 t2Jxb7 56 h5!
Capablanca 's Theorem - 'ti'+tD is better than 'ti'+ in the Ending 215
One suspects tht Black should 68 tZJg7 + 1-0

have found time for . . . h6-hS , but White will win a piece after
where could he have played it? 68 . . .'.ti'd6 69 tZJe8+ and 70 'iWxdS+.
56 gxh5
A tough decision. Now fS is However, typical methods of
available for the knight, but other evaluating a position continue to
wise weaknesses would have ap apply to queen and knight versus
peared at g6 and e6. queen and bishop endings. For in
57 tZJe7+ rJ;;g7 stance, the side with extra material
58 tZJf5+ rJ;;g6 is generally always considerably
59 tZJh4+ rJ;;g7 better. Similarly, much better placed
60 'i!Vxh5 d7 pieces (or a better pawn structure)
61 tZJf5+ rJ;;f6 typically yield their owner a good
62 'iVxh6+ (D) game.
Karpov - Anand
Linares 1 991
B
62 rJ;;e5
Of course, the knight couldn't W
be taken, as then 63 'iVh3+ would
have won the queen. Now the black In this position, Anand is con
king is exposed and the e-pawn siderably better for a variety of rea
will prove undefendable. sons. The most noticeable feature
63 g4! f6 of the position is that he has the
64 8 Jt.. a5 better pawn structure. Karpov's
65 'ii'b 8 + Jt.. c7 pawns are all fixed and the situ
66 2+ rJ;;e6 ation is particularly bleak for him
67 3+ d5? on the kingside, where two black
216 Capablanca 's Theorem - 'tIi+tLl is better than 'tIi+i. in the Ending
pawns hold three white pawns.

Against passive play, Anand will
simply move his bishop to d4 or c3
and then load up the diagonal by
playing his queen to f6. Karpov
played to prevent this.
37 i.b2 .txb2
38 'ifxb2 'ii'f6!
But Anand's advantage isn ' t so
easy to dissipate. Even without the
bishop pair, he is able to challenge
Karpov for the long diagonal. If it W
should be ceded to him, he will
quickly play his queen into White's important is the fact that White's
position and use his bishop to team king is exposed while Black's king
up on the fixed b-pawn. Karpov has a wall of pawns that he can use
preferred to try his luck in a pure for shelter. Further, in endings with
minor-piece ending, presumably queens, the side with the further
on the basis of the blocked position advanced passed pawn usually has
making it difficult for Black's king the advantage. Black has a pro
to penetrate. However, Anand won tected passed pawn on the fourth;
easily through opening the position White does also, but it will be iso
with . . . d5 and using zugzwang to lated the moment the g-pawn is
break the white defence: 39 Wixf6+ lost. Further, the black king is near
xf6 40 gl .tb1 41 lbn .te2 42 White ' s passed pawn, while the
lbd2 e6 43 f2 dS 44 exdS+ white king is distant from the black
d5 45 e3 i.d1 46 d3 i.xb3 passed pawn.
0-1 36 'ie7+ f8
37 ,*,b2 g7
It's also possible for the side 38 tLlb4 'in
with the bishop to do well because Yermolinsky prefers to pursue
of having better attacking chances. the white king before settling for
capturing the g-pawn.
Hjartarson - Yermolinsky 39 e2 i.h6
Erevan Olympiad 1 996 Here we see the bishop in its tra
ditional glory. Even though it had
White has an extra pawn, but his g to retreat, it continues to influence
pawn is impossible to defend, so events on the other side of the
this isn ' t really important. What is board.
Capablanca 's Theorem - 'it'+lLl is better than 'it'+ in the Ending 217
40 a4 c1+ Other factors, such as superior

41 'it>d3 'iVd2+ centralization and a space advan-
42 'it>e4 g2+ tage, can also help the side with the
43 'it>d3 d2+ bishop win.
44 'it>e4 'i!t'g2+
45 'it>d3 (D) Ruban - Ehlvest
Novosibirsk 1 993
B
W
45 xg3+
Yermolinsky managed to cap Here White has a pretty-looking
ture the g-pawn with check and his knight and a protected passed b
initiative continues. pawn, but Black's passed pawn is
46 'it>c2 f2+ further advanced and the b-pawn
47 'it>b1 'i!t'e1+ isn't going anywhere anytime soon.
48 'it>a2 e2+ 36 d2 'iYd5
49 'it>b1 d1+ 37 f3 f5!
One thing about this ending that Ehlvest takes control of the e4-
is not immediately apparent is the square and hints at the possibility
extent to which the white queen is that he might later launch a king
out of play. It looks well-placed, side pawn-storm.
but it really does nothing to men 38 'it>f2 d3
ace the black king or defend its 39 'it>e1 'iYd4
own. 40 'it>f1 f4
50 'it>a2 j"c1 This gives the e4-square back to
51 'it>b1 bxa4 the white knight, but it serves to fix
52 bxa4 j"a3+ the pawn at g2 and seizes space in
0-1 the vicinity of the white king.
218 Capablanca 's Theorem - 'if' +ltJ is better than 'if' + i. in the Ending
41 ltJe4 h7 44 'iHxf4 'iHb2+ 45 e3 . .te8 46

42 g3 'iHal+ xd3 a3 47 ltJc3 a2 48 ltJxa2
43 f2 'iHxa3 (D) 'iHxa2 49 d4 'iHa7+ 50 d5 'iVb7+
51 c4 .tc6 52 'iVf5+ g6 53 'iHe5
.txf3 54 b5 .tg2 55 h4 .tfl + 56
b4 h5 57 'iVc5 .td3 58 b6 .te4
59 'iVd6 'iHg7 60 c4 'iVf7+ 61
c5 'iHf2+ 62 c4 'iHc2+ 0-1
Verdict: Capablanca's intuitive in

sight into the advantage of queen
and knight vs queen and bishop in
the ending is correct. While gen
eral positional methods of evalu
ation are still important, the
W attacking force of the queen and
knight can be very fierce, particu
The a-pawn will soon win a larly in blocked positions or those
piece. The rest needs no comment: featuring fewer pawns.
Index of G a mes and
Pa rt- Games
Numbe rs refer to pages.

The second-named player is White if his name appears in bold. Otherwise
the first-named player is White.
ALEKHINE - Reti 207 CHI GORIN & SCHIFFERS - Pillsbury

ANAND - Karpov 2 1 5 & B1ackburne 35
ANDERSSEN, A . - Paulsen, L. 1 26 CUELLAR - Tal 1 79
B ELIAVSKY - Bronstein 77 DE FIRM IAN - Miles 129
B ELINKOV - Kotkov 93 DIEZ DEL CORRAL - Petrosian 1 1 0
BENJAMIN - Brooks 1 93 DONNER - Larsen 89
BOGOLJUBOW - Capablanca 81 DVOIRYS - Malaniuk 20
BOLESLAVS KY - Bronstein 1 65; EHLVEST - Ruban 2 1 7
Panov 213 EM. LASKER - C higorin 34
BOTVINNIK - Euwe 148; Geller 64; ENGLISCH Chigorin 56; Steinitz
-
Langeweg 207; Scherbakov, V. 17

146; Smyslov 1 86 EUWE - Botvinnik 148
B RONSTEIN- Beliavsky 77; FARAGO - Suba 1 15
Boleslavsky 165; Petrosian 1 75 ; FISCHER - Petrosian 1 69; Taimanov
Porreca 79; Zaitsev, A. 74 204
BROOKS - Benjamin 1 93 FISCHER - Tal 201
BROWNE - Karpov 1 98; Lucena 70 GAVRIKOV Kupreichik 214
-
BRYNELL - Geller 1 94 GELLER Botvinnik 64; Brynell

-
B URN - Chigorin 43, 48 1 94; Mikhalchishin 72; Spassky

CAPABLANCA - Bogoljubow 81 ; 1 02
Winter 80 GHEORGHIU - Polugaevsky 122
CHENEY - Mayer 95 GLIGORIC - Hort 1 60; Najdorf
CHIGORIN Burn 43, 48;
- 210
Em. Lasker 34; Englisch 56; HECHT - Tal 152
Janowski 57; Marshall 52 ; HJARTARSON Yermolinsky 2 1 6 -
Mieses 59; Pillsbury 41 , 42 ; HORT Gligoric 1 60

-
Steinitz 3 1 ; Teichmann 44 IVANCHUK Manor 67 -

220 Index a/ Games and Part-Games
JANOWSKI - Chigorin 57 PANOV Boleslavsky 2 1 3

-
JIMENEZ,E. - Larsen, B. 1 72 PAULSEN, L. - Anderssen, A. 1 2 6

KARPOV Anand 215; Browne 1 98
- PETROSIAN - Bronstein 1 75; Diez
KASPAROV - Nunn 1 1 3 del Corral 1 1 0; Fischer 1 69;
KEITLINGHAUS - Knaak 1 85 Korchnoi 121; Lilienthal 206;
KINDERMANN Vojska 1 96
- Nielsen, A. 1 82 ; Spassky 1 1 8
KNAAK - Keitlinghaus 1 85 PETURSSON Watson, W. 140
-
KORCHNOI - Petrosian 1 2 1 ; PILLSBURY - Chigorin 4 1 , 42

Reshevsky 86; Spassky 1 90; PILLSBURY & BLACKBURNE -
Topalov 1 08 Chigorin & Schiffers 35
KOTKOV - Belinkov 93 POLUGAEVSKY - Gheorghiu 122;
KUPREICHIK - Gavrikov 214 Lukovnikov 99; Rodriguez, Am.
KUZMIN, A. - Manakova 212 137
KUZMIN, G. - Stein 1 05 PORRECA - Bronstein 79
LANGEWEG - Botvinnik 207 PRITCHETT - Psakhis 1 00
LARSEN Donner 89; Jimenez, E.
- PSAKHIS - Pritchett 100
1 72 REICHENBACH - Sosonko 1 84
LEKO - Shirov 143 RESHEVSKY - Korchnoi 86
LILIENTHAL - Petrosian 206; RETI - Alekhine 207
Smyslov 1 90 RICHTER, B. - Tarrasch 24
LUCENA Browne 70
- RODRIGUEZ, AM . - Polugaevsky
LUKOVNIKOV Polugaevsky 99
- 137
MACKENZIE - Steinitz 20 ROSENTHAL - Steinitz 12, 21
MALANIUK - Dvoirys 20 RUBAN - Ehlvest 217
MANAKovA - Kuzmin, A. 2 1 2 SCHERBAKOV, V. - Botvinnik 146
MANOR - Ivanchuk 67 SELLMAN - Steinitz 20
MARSHALL - Chigorin 52 SHIROV - Leko 143
MATTISON - Nimzowitsch 1 77 - Taimanov 157
S HIYANOVSKY
MAYER - Cheney 95 SILMAN- Moskalenko 2 12
MIESES Chi gorin 59
- SMYSLOV Botvinnik 1 86;
-
MIKHALCHISHIN - Geller 72 Lilienthal 1 90; Wade 1 59

MILES - de Firrnian 129 SOSONKO - Reichenbach 1 84
MOSKALENKO - Silman 212 S PASSKY - Geller 1 02; Korchnoi
NAGAIZEV - Spassky 1 88 190; N agaizev 188; Petrosian 1 1 8
NAJDORF - Gligoric 210 STEIN - Kuzmin, G . 1 05
NIELSEN, A. Petrosian 1 82
- STEINITZ - Chigorin 31 ; Englisch
NIMZOWITSCH - Mattison 1 77 1 7; Mackenzie 20; Rosenthal
N UNN - Kasparov 1 13 1 2 , 2 1 ; Sellman 20
Index of Games and Part-Games 221
SUBA - Farago 115 TOPALOV Korchnoi 1 08

-
TAIMANOV - Fischer 204 ; Shiyan VOJSKA Kindermann 1 96

-
ovsky 1 5 7 VON GOTTSCHALL Tarrasch - 134

TAL Cuellar 1 79 ; Fischer 201 ;
- WADE - Smyslov 159
Hecht 1 52 WATSON , W. - Petursson 140
TARRASCH - Richter, B. 24; von WINTER - Capablanca 80
Gottschall 1 34 YERMOLINSKY - Hjartarson 2 1 6
TEICHMANN - Chi gorin 44 ZAITSEV, A. Bronstein
- 74
I ndex of Openings
Numbers refer to pages.
1 g3 89
1 tiJ f3 d6 1 08
Alekhine, Kengis 1 72
Benko Gambit 1 05
Caro-Kann, Classical 77, 79
Dutch, 2 ,t g5 182
English, . . . e5 70, 1 1 5
English, Botvinnik 146
English, Rubinstein 122
English, Symmetrical 121
Four Knights 80
French, Tarrasch 134
French, Winawer 1 1 0, 145, 1 86, 1
King's Gambit 52
King's Indian, Classical 97, 1 75
King's Indian, Fianchetto 64
King's Indian, Gligoric 137
King's Indian, Samisch 140, 1 60
King's Indian, Simagin 1 79
Modern 145
Modern Benoni, Classical 67
Modern Benoni, Taimanov 113
Nimzo-Indian, 4 tiJf3 1 77
Nimzo-Indian, 5 'i!t'b3 165
Nimzo-Indian, Classical 129
Nimzo-Indian, Deferred Sarnisch 1 90
Nimzo-Indian, Samisch 1 90
Old Benoni 48
Philidor 126
Pirc, 5 h4 159
Index of Openings 223
Ponziani 57
Queen 's Gambit, Chigorin 34, 3 41, 42, 43, 44
Queen's Gambit, Semi-Slav 148
Queen's Pawn, 2 . . . .tg4 31
Queen 'slNimzo-Indian 152
R6ti (without c4) 86
Sicilian, Classical 99, 1 00
Sicilian, Closed 1 02
Sicilian , Dragon 95, 161
Sicilian, Kan 1 69
Sicilian, Najdorf 93, 1 18
Sicilian, Rossolimo 74
Sicilian, Scheveningen 72
Sicilian, Sveshnikov 97
Sicilian, Taimanov 56, 157
Spanish (Ruy Lopez), 3 . . . g6 1 7, 20
Spanish (Ruy Lopez), 9 d4 81
Spanish (Ruy Lopez), Berlin 24, 1 96
Spanish (Ruy Lopez), Chi gorin 66
Spanish (Ruy Lopez), Exchange 1 93, 1 94
Spanish (Ruy Lopez), Mller 143
Three Knights 12, 2 1
Trompowsky, 2 . . . lZJe4 1 84, 1 85
Vienna, 3 g3 59
Index of End ings
Numbers refer to pages.

All endings in the main games in the book are included, provided that,
apart from kings and pawns, neither side has more than two pieces. In
cases where both sides have one bishop, it is indicated whether the bishops
are same-colour or opposite-colour.
Pawn 20, 1 96, 200

tfj v !, 32, 200
tfj v tfj 88, 1 96, 1 99
iL v !, 24, 97
iL v iL (same) 72
iL v tfj 1 9, 32, 5 63, 12& 20 214, 2 1 6
.t+tfj v iL+tfj (same) 79, 87
Il v Il 42, 1 5 16 1 8& 1 98
2':v2': 156, 169, 1 94
: v tfj 204
':+iL v ': 97
It +tfj v 1':[ +tfj 85, 1 86, 1 95
It+iL v ':+iL (same) 142
l':t+iL v Il+iL (opposite) 97
It+.t v Il+tfj 1 9, 31, 5 62, 7 1 88, 1 9& 201, 205
"VJIi v !' 200
'iV v 'iV 1 45, 213
'iV v It 1 92
v It+tfj 160
'iV+tfj v 'iV 32, 1 82
'iV+.t v 'iV 24, 2 1 8
'iV+.t v 'iV+iL (same) 71
'it'+tfj v 'it'+iL 1 1 7, 128, 144, 1 75, 1 81, 2 1 0,
2 1 2, 2 1 2, 2 1 3, 214, 216 (2), 2 1 7
42

Bishop v. Knight - The Verdict PDF

Uploaded by

Copyright:

Available Formats

Bishop v. Knight - The Verdict PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bishop v. Knight - The Verdict PDF

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What topics does the book cover regarding bishops and knights?

What topics does the book cover regarding bishops and knights?

Bis hop versus Knig ht:

B. T. Batsford Ltd, London

British Library Cataloguing-in-Publication Data.

All rights reserved. No part of this book may be

Edited by Graham Burgess and typeset by Petra Nunn for

To my mother; Gloria Mayer

A BATSFORD CHESS BOOK

1 Some Characteristics of Knights and Bishops 9

It will probably be helpful to the two-fold. First, chess really does

cases of minor-piece superiority (or made a remark along the lines of

No book is written in a vacuum, so this book can hardly be an exception.

It's useful to consider some charac The knight is a short-range

The knight at al has only two

It's harder to speak of general

As every beginner knows, the

If there is some reason that the

The bishop pair make up for

Steinitz is commonly credited with Board, Reti points to the following

younger masters as Charousek, Las 8 WHd2 t2J e5 9 e2 d5 ! would lead

to move pawns when applying the 24 i.b7

and 32 f6; 30 tZJf4! ? is also interest These variations pass without

sic example of the Steinitzian Re eliminating White's spatial advan

38 ltJe2 2) 9 h3 c6 10 J.d3 d5 1 1 exd5

same position arose in Blackburne of the games at London 1 883. For

another game Rosenthal-Steinitz, 14 .ib3?

I S . . . cS? 1 6 f4 ! favours White.

37 'ixh7+ e6 and Englisch-Steinitz, was not lost

1 e4 e5 2 liJf3 liJc6 3 .tb5lLJf6 4 0-0

White uses a pin to win a piece.

'it'gl 'ie3+ 41 'it'n 'iVf4+ 42 'it'e2

will allow Black to queen a pawn. 22 ':e2 b5

30 ':5e2?? Momentarily preventing the ad

A sterling example of the Ste degree of simplification has oc

in Russian and most of his writings on his play believed that he fa

Steinitz - Chigorin become targets, as the king would

b) 1 2 .txc4 ! ? lb xf4 1 3 ':xf7 b22) 14 :xf4 lbxc4 is given by

bI) White should be better in

down the rook, but White emerges 21 .tb3 t'iJg6

W The complications have re

with 25 .. .lbb2 26 lIf5 lIxb3+, but game against Emanuel Lasker at

1 d4 dS 2 4Jf3 i.g4 3 c4 i.xf3 4

However, Levenfish points out

lambasting Lasker's. In fact, White connected passed pawns should

Unfortunately, Steinitz is com

A more interesting try is Leven material with 53 ::'xe5 lLJxe5 54

The trouble Steinitz and Lasker

1 d4 dS 2 c4 lLlc6 3 lLlf3 .tg4 4 e3

hxg6 26 WNd3 c;t>f8 27 11xg6 l:Ixa2

w 1 d4 d5 2 c4 lbc6 3 lbf3 .tg4 4 e3

Now Panov analyses both 2 1 b3

kingside. Cvetkovic and V. Sok I 've deliberately focused pri

1 d4 d5 2 c4 tbc6 3 tbf3 i.g4 4

16 :hgl 'iVe7! (D)

modern preference) 9 Jt.. c4 ! repre W

most of them have become widely success or strategy in the handling