Winning Endgame Strategy

Alexander Beliavsky,
Adrian Mikhalchishin

B.T. Batsford Ltd, London

First published in 2000

Alexander Beliavsky, Adrian Mikhalchishin 2000
I SBN 0 7134 8446 2
British Library Cataloguing-in-Publication D ata.
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.
Printed in Great Britain by
Creative Print and Design (Wales), Ebbw Vale
for the publishers,
B.T. Batsford Ltd,
9 Blenheim Court,
Brewery R oad,
London N7 9NT
A member of the

ChKalis

Group plc

A BA'J: SFORD CHESS BOOK

Contents
Page
Introduction

I.

Pawn Endings

2.

Kn ight Endings

34

3.

R ook Endings

56

4.

Shouldering: the struggle of the kings

1 09

5.

Complex Endings

1 12

6.

Defence in the Ending

131

7.

The isolated pawn in the Ending

136

8.

R ook and bishop against rook and knight

141

9.

R ook and two pawns against rook and knight 1 50

10.

Two minor pieces against a rook

153

1 1.

R ook against knight and pawn

172

12.

Queen Endings

1 80

Solutions to Exercises

19 1

Introduction
For the authors the most import
ant thing is-will readers study their
book? It was very pleasant for us
when Winning Endgame Technique
( in fact the first part of the present
book) was deeply studied by grand
masters Boris Gelfand and Ognjen
Cvitan. Then grandmasters Suat
Atalik and Alex Yermolinsky also
studied the book for a month
( looking for mistakes!) and found
much of interest for themselves.
Particular thanks to grandmaster
Alexei Kuzmin, trainer of the Qatar
team, who discovered many import
ant and interesting corrections in
pawn endings. We still cannot guar
antee there are no mistakes-but
then again, generally speaking, any
one who considers themselves free
fr om error makes more mistakes
than others.
The present book is a direct con
tinuation of Winning Endgame
Technique and to some extent the
second part in the sense that many

themes, such as queen, knight and

complex endings, were not dealt
with in the previous book. However
in some cases we have devoted a
little more attention to methods of
play in definite types of position,
rather than concrete cases, which is
a fault, on the whole, of all books on
the endgame. For example the
Encyclopaedia of Chess Endings
gives replies only to some ques
tions. We have tried to find more
examples in which play conforms
with general principles so that
readers can begin to apply these
methods in concrete practical situ
ations. Of course, correlation of
methods of play in typical positions
and techniques is not always uni
form, but the authors present their
vision of practical endgame prob
lems and are a long way from ex
hausting this theme.
It remains to thank our friend,
master Oleg Stetsko, for help with
the selection of practical examples.

Pawn Endings
Shirov-Timman
Wijk aan Zee, 1 996

Guru Grigoriev was right!

Readers might know that the

greatest connoisseur of pawn studies
was the Soviet master Nikolai
Grigoriev, who achieved the im
possible in his complicated pawn
studies. In fact he once said:
"Grandmasters do not like pawn
endings because they simply don't
understand them". The statement
sounds paradoxical but the thought
is topical even today. It seems that
pawn endings are simple but the
number of mistakes grandmasters
have made over the years has in no
way diminished. Since the publica
tion of our first book quite a few in
teresting endgames have been
played and quite a few instructive
mistakes made.
The most characteristic and sur
prising mistakes are ...
Premature resignation
of a game

Yes, even today, this happens at

grandmaster level.
In the following position Timman

resigned and both players thought

this to be in order, and appropriate

comments appeared in ChessBase...

But White does not win after

2 h4 xc6 3 f5! 'iitd 6 4 f6,
when there are two ways to draw:
1 ...d6

(a) 4 . . .d7 5 f3 e8 6 'ite4

rti c7 7 d5 e8 ! 8 rtic6 (8 rti e6
rtif8!) 8 . ..d8, and he does not win
the c7 pawn;
(b) 4 . . .c6 5 rtif3 d7 6 rtie4
@d8 ! , and the king holds the
squares corresponding to a5, c5,
e5-d7; a4, b4, c4, d4, e4-d8, e8;
a6-e6.
Another surprising resignation
followed in a game played two
years later...

8 Pawn Endings

Maciej a-Grabarczyk
Poland, 1 998

Krumpachnik-Polak
Slovenia, 1 985

Although after 1. ..bxa5+ 2 xa5

c5 3 b5 d6 it is a simple draw.

Here Black resigned, reckoning

that he would lose both of his
doubled pawns, but after 1...d7 2

It is also possible to resign be

cause of an incorrect calculation of
an arising pawn ending.
Svidler-Lobron
Erevan, 1 996

Here Black resigned, reckoning

on the variation 1...g2 2 @c5

hl='ir' 3 .!:l.xhl xhl 4 d5 g2 5

'itie5, but he did not calculate to the
end. 5 ... 'itit3 6 xf5 e3 and Black

captures the c3 pawn.

xe5 e7 3 g6 f8! 4 xe6 g7

5 f5 h6! 6 f6 it all ends in a

well known stalemate.

N atapov-Schuravlov

Moscow, 1994

Black resigned, forgetting that

there was a last chance of saving the
game by stalemate aft er 1. c6 2
.

axb5+ xb5 3 'it>d5 a4!

Pawn Endings 9

Yudasin-Osnos
Leningrad, 1 987

After an arduous defence and

transposition to a pawn ending
White had the impression that an in
teresting chance had cropped up,
and he took it.
1 'it>fS??

But he did not reckon on the

simple
l ...b4! 2 axb4+

Losing simply is 2 a4 'it>d4 3 'it>g6

'it>c3 and the black pawn is faster
than the white one.

After playing 1 'it>f2 Yudasin of

fered a draw, saying that this well
known position is in the books and
is drawn! His experienced oppo
nent, and international master, was
shocked by his own ' ignorance' and
accepted it there and then. However
it ought to be well-known to every
one that after l ... 'it>e4 2 'it>e2 f4 3
@f2 f3 4 @fl the triangulation
e5-f5-e4 is winning. It just shows
don 't believe everybody ! !
Erroneous play in pawn endings

Another type of very common

mistake-even strong grandmasters
have many technical shortcomings.

2 ...'it>xb4 3 'it>g6 'it>xb3 4 'it>xh6 a5

5 'it>xgS a4 6 'it>h6 a3 7 g5 a2 8 g6
al='i' 9 g7 11Yf6+ and he had to
resign.

Correct was the more logical l

'it>e5 a5 2 'it>e4 a4 3 bxa4 bxa4 4
'it>e5 'it>c4 5 'it>e4 'it>b3 6 'it>d3 'it>xa3
7 'it>c3 'it>a2 8 'it>c2 a3 and here the
extra move 9 h3 ! makes a draw.
Another way to the draw is 5 'it>d3
'it>d5 6 h3!.
A lack of understanding of basic
principles of play in pawn endings
Markovic-I van ovic
Vrnjacka Banja, 1 998

Hector-Speelman
Roskilde, 1 998

Because of his weakened structure

on the queen's flank (doubled
pawns) White has the inferior posi
tion, and the right way now was

1 0 Pawn Endings
I ... hxg4 2 'it>xg4 'it>e5 3 Wg5 d5 ! 4
cxd5 b 5! 5 axb5 a4 6 'it>g6 a3 7 d6!
'it>xd6 8 @xg7 a2 etc. The only
chance of a draw would be the diffi
cult 4 d4+! @xd4 5 'it>g6 @es 6
cxd5 @f4 (or 6 ... b5 7 d6! cxd6 8
axb5 a4 9 b6=) 7 c4 We4 8 @xg7
@xf5 9 @fl We5 10 @e7, and a
draw. But Black played

Schandorff-Speelman
Roskilde, 1 998

1...g5+? 2 fxg6 h4??

It was still a draw after 2 ...hxg4 3

'it>xg4 @xg 6.

3 @n wxg6 4 @g2 @f6 5 Wh3

@g5 6 d4 c6 7 c3
And Black resigned since after

7 ... d5 winning is 8 c5 b5 9 axb5 a4

10 bxc6, and in the resulting queen
ending White has too many pawns.
Ehlvest-Shirov
Vienna, 1 996

1 b4?? 'it1f6 2 'it>h4 g6! and White

resigned since after 3 'it>g 3 g5 4 f4

exf3 5 'it1xf3 'it>e5 6 @fl We4 7 @e2

f5 ! 8 gxf5 'it>xf5 thanks to the dis
tant passed pawn Black wins easi ly.
Correct was I @g3 ! b4 (or I . . a5 2
a4 bxa4 3 bxa4 g6 4 Wh3 'it>f6 5
@h4 @e6 6 'it1g5 'it>e5 7 'it>h6 Wf6 8
g5+! @f5 9 Wg7 with a draw) 2
@h3 f5 3 gxf5 @xf5 4 @h4! g5+ 5
'it>g3 @f6 6 'it>h3 @g7 7 'it>g3 '.t>h6 8
@h3 @h5 9 'it>g3 a6 I 0 'it>h3 g4+ 1 1
'it>g3 @g5 1 2 'it>g2 @h4 1 3 @h2 g 3+
14 fxg3+ @g4 15 'it>g2 as 16 fl
@h3 1 7 g4!, and a draw. Black
lacks a single tempo move with the
a-pawn.
.

Plaskett-Rowson
Scotland, 1 998

A simple position. White only

needs to meet . . . 'it>d5 with @e3 and
there is no problem. However
Ehlvest played carelessly.
1 @f4?? and after 1 ... 'it1e6 2 We3
@d5 3 'it>d3 f4! 4 gxf4 h4 5 @e3 h3
6 gxh3 gxh3 7 @n @xd4 8 'it1g3
@e4 he had to resign.

In the following position we have

another example of an incorrect
pawn move.

Pawn Endings 1 1

Black evaluated the position poor

ly and sacrificed a pawn.
1.. g4?? 2 fxg4 lt>g5 3 'itif3 e5 4

Ostenstad-Kuzmin
Biel, 1 990

c4 e4+ 5 lt>xe4 lt>xg4 6 b4 Wxh5 7

f5! lt>h4 8 c5 a6 9 a4 h5 10 b5
axb5 11 axb5 lt>g3 12 c6! bxc6 13
b6!, and Black resigned. After the

correct 'passive' defence l ...\t>g7! 2

c4 lt>h6 3 b4 lt>xh5 4 c5 lt>g6 5 'it>e5
h5 6 b5 Wfl 7 'itid6 g4 he would
easily achieve a draw. But it looked
all so simple-going directly for the
h5 pawn.
Greenfeld-Golod
Israel, 1 998

Only one winning move is left:

1 .. .lt>e3 ! 2 lt>c4 @d2 3 lt>b5 'it>c3 4
Wxb6 lt>b4, with a typical finish, but
there followed 1...lt>d5?? and after
2 'it>d3! draw.

Ignorance of typical methods of

struggle. Typical breakthroughs.
Adams-Lutz
Wijk aan Zee, 1 995

Correct play was the simple l ...f5 !

2 Wxa7 (after 2 gxf5 exf5 3 lt>xa7,
simply 3 ... g4 4 a4 f4 5 a5 g3) 2. . . e5
3 a4 e4 4 lt>b6 fxg4 (4 ... f4! ?-+) 5
a5 g3 ! 6 fxg3 e3 7 a6 e2 8 a7 e l ='l'W
9 a8=it' 4+ 1 0 @c7 'l'Wd6+! with
a transposition to a single, but easily
winning, pawn ending.
In the game Black decided to
'press' the white king.
1 ...lt>d6? 2 @xa7 'itic6 but here
followed 3 '1t>b8 ! ! f5 4 a4 fxg4 (or
4 ... f4 5 a5 lt>b5 6 @b7 ! 'itxa5 7
'itic6!) 5 a5 lt>b5 6 'it>b7! (Reti' s
typical manoeuvre) 6...'it>xa5 7 'it>c6
Wb4 8 lt>d6 lt>c4 9 @xe6 @d3 with
a draw.

1 h5??

Correct was 1 bxa4 bxa4 2 g4 ! g6

3 gxf5 gxf5 4 lt>d3 ! h5 5 c3 ! 'it>c5 6
c4 e4+ 7 fxe4 fxe4 8 lt>xe4 @xc4 9
'it>e3 'it>b3 1 0 lt>d3 lt>xa3 1 1 'it>c3
with a draw.
l...b4! White resigned.

J 2 Pawn Endings

Overlooking stalemating defences

Hellers-Eingorn
Debrecen, 1992

1 h4?

It is terribly dangerous to move

the 'wrong' pawn-necessary was 1
@c3 'it>d5 2 'it>d3 h5 3 b4! axb4
(3 ... a4 4 e4+ @d6 5 h4 e5 6 'it>e3
exf4+ 7 'it>xf4 'it>e6 8 'it>e3 g5 9 f4=)
4 axb4 e5 5 e4+ @e6 6 'it>e3 exf4+ 7
'it>xf4 g5+ (if 7 ... h4, then 8 'it>e3 D.
f4) 8 'it>e3 We5 (8 . . .f5 9 f4 g4 1 0 e5)
9 h4 ! 'it>e6 1 0 f4 with a draw.
1...'it>d5 2 b4?

Here easily winning is l . . . l:!xa4 2

i.xf6 (2 c6 fxe5 3 c7 l!d4+! 4 'it>c5
di) 2 . . .a2 3 c6 l:!xg2 4 c7 l:!c2.
However Hellers saw that in the
pawn ending he could force the win
of a pawn and decided that this was
sufficient.

Another incorrect advance-it

was necessary to play 2 e4+ @d6 3
@e3 ! (3 'it>d4? e5+ 4 fxe5 fxe5+ 5
@d3 'it>e6 6 Wc3 h5 D. g5 - +) 3 . . .e5
4 f5 ! gxf5 5 exf5 'it>d5 6 'it>d3 a4 7
bxa4 bxa4 8 h5 h6 9 'it>c3 e4 IO
fxe4+ xe4 1 1 'it>b4, and, though
Black has an extra pawn in the
queen ending, White has chances of
a draw.
2... axb4 3 axb4

1. .. J:Ixc5+? 2 'it>xc5 fxe5 3 'it>d5

Wg6 4 'it>xe5 'it>g5 5 'it>e4 h5 6 We5
Wh4! 7 'it>xf4 Drawn.

The Black king is stalemated!

Erroneous pawn advances

Indeed, this is a great mistake to

make in pawn endings.
Karpov-Kasparov
Las Palmas, 1 996

3 ... h6?

Now comes a mistake from the

other side-after 3 ... e5! 4 e4+ We6
5 'it>e3 d6! (5 ... 'it>f7? 6 f5 ! =) 6 f5
(if 6 fxe5+, then 6 . . .fxe5 7 'it>t2 e6
8 'it>g2 'it>f6 9 'it>g3 h6, and zug
zwang) 6 ... gxf5 7 exf5 'it>d5 8 'it>d3
h5 9 'it>e3 'it>c4 10 'it>e4 @xb4 1 1
'it>d5 'it>c3 1 2 'it>e6 b4 1 3 'it>xf6 b3 1 4
'it>g7 b2 15 f6 b 1 ='iV 1 6 f7 'i/Vb7 1 7
'it>g8 'iVxf3, winning.
4 e4+ 'it>d6 5 We3 e5 6 fxe5+ fxe5
7 'it>f2 e6 8 'it>g2 !

Pawn Endings 1 3

Weak was 8 'it>g3 c;f;>f6 9 'it>g4 h5+

IO c;f;>g3 g5, and Black wins, while
now on 8.. .f6 follows 9 c;f;>g3 h5
10 f4, and a draw.

6.. . d5??

Correct was to force e2-e3 by

6 . . . 'it>e4 ! 7 e3 d5 8 e2 i;;>c5 ! 9
'it>d2 'it>d6!, so that on 1 0 'it>d3 @d5

Matlak-Tseshkovsky
Lubniewice, 1 995

1 f4? ?

Any other move wins-simplest

was 1 c;f;>M 'it>f4 2 h5 'it>e3 3 g6
'it>xe2 4 f4 'it>e3 5 f5 .
1...'it>e4 2 c;f;>g4 c;f;>d4!

White had reckoned only on

2 ...'it>e3 3 f5 with a win, but now
on 3 c;f;>f5 follows 3 ... 'it>e3 !, and after
White moves this position is drawn!

. . .it will be White 's move-after

Black moves his position is lost,
since he is forced to allow the White
king to e5, and ... i;;>c5 loses after e4.

7 'it>e3 'it>c4 8 i;t>d2 'it>d4 9 e3+

'it>c4 10 i;;>c2! ,
and Black resigned since White

achieves the above-mentioned posi

tion with Black to move.
Drasko-VratonjiC
Ulcinj, 1 997

3 'it>h5 'it>e4 !

Again Black will not 'buy'

3 ... i;t>e3? because of 4 f5 winning
for White.
4 i;;>g4 i;t>d4 s wo rs 6 'it>f2

1 . .. f5 !

A colossal defensive resource,

though it was possible to hold the
position even by simple defence,
e.g. l...'it>e6 2 i;;>e4 (the assessment
is not changed by 2 f4 f5 3 g5 h5=)

Pawn Endings

14

2. . f5+ (the only move) 3 gxf5+

.

'it>d6 4 f6 'it>c5 5 'it>e5 a5 6 a4

'it>c6!=; or 2 c5 bxc5+ (losing is
2 . b5 3 f4 a5 4 h4, 5 h5, 6 g 5! + -) 3
'it>xc5 f5 4 gxf5+ 'it>xf5 5 b5 'it>f4
6 'it>a6 'it>f3 7 'it>xa7 'it>xf2= .
.

2 gxf5 f6! 3 a4?

White loses the thread of the

gam.e, starting to play on his weak
flank. He should reconcile himself
to a drawn result after 3 h4 'it>c6 4
h5 'it>d6 5 'it>d3 'it>c5 6 'it>c3 a6 7 a4
a5 8 f3=.
3...'it>c6 4 f4??

A very serious mistake. Without

need White himself gives up the op
portunity of exploiting the tempo
and hands back the move to his op
ponent in a position of mutual zug
zwang. It was still not too late to
lead the game to a draw by 4 h4 h5
5 'it>d3 'it>c5 6 c3 a6 6 f3 a5 7 f4.

4 'it>d6 5 'it>d3 'it>c5 6 'it>c3 h5 7

'it'b3 h4 + 8 'itc3 h3 9 b3 a6 10
a5
..

The assessment of the position is

not changed by 10 c3 a5 - + .

10 ... bxa5 11 'it'a4 'it>xc4 12 'it>xa5

'it'd4 13 'it>xa6 'it'e4 14 'it>b5 'it'xf5
White resigned.
Sulipa-Gricak
Lvov, 1 995

A pawn ending which is quite

difficult to evaluate. To gain victory
Black needs to solve the problem on
the queenside, where his opponent
has the chance to create a passed
pawn. This is achieved by I ...'it>d7!
2 h4 'it>c7 3 'it>f3 'it>c8 4 'it>e4 'it>b7 5
'it>f3 a6 - + . However in the game
followed

1... g5? 2 @f3 'it'd5 3 c6 'it>d6 4

e4??

The decisive mistake. The signifi

cance of a tempo in pawn endings is
far higher than in any other. White
should immediately exploit his op
ponent's mistake and himself break
up Black's pawn chain. This is done
by 4 h4! gxh4 5 'it>g4 a6 6 bxa6
'it>xc6 7 'it>xh4 'it'b6 8 'iitg4 'it>xa6 9
h4 + -.

4 ...a6 5 bxa6 'it>xc6 6 'it>f3 'itb6 7

h4 gxh4 8 g4 'it'xa6 9 xh4 'it>b6
10 'it'g4 'it>c6 11 h4 'it'd6 White
resigned.

As a matter of fact, in their first

book, the authors made several inac
curacies, which were corrected by
grandmaster Alexei Kuzmin.
Kupreichik-Mikhalchishin
Lvov, 1 988

Pawn Endings 1 5

1 @dS
...

In their first book the authors

placed a question mark against this
move, pointing out a 'direct' path to
a draw: l ... \t>f6 2 <;;>f2 <;;>g6 3 @f3
h5 4 gxh5+ 'it>xh5 5 'it>e4 <;;>g4 6 f5
g5 7 <;;>xd4 <;;>xf5 8 'it>c5 'it>e5 9
b6 i;t>d6 etc. However they did not
reckon on one finesse to which
A.Kuzmin drew attention: 3 f5+! (3
<;;>e2 h5 4 f5+ 'it>h6=) 3 ... \t>g5? (3
i;t>f6, returning to the basic vari
ation) 4 'it>e2 h5 5 f6 ! 'it>xf6 6 gxh5
\t>g5 7 i;t>d3 'it>xh5 8 'it>xd4 \t>g5 9
c;t>c5 i;t>f6 1 0 i;t>b6 'it>e7 1 1 <;;>xa6
<;;>d7 12 <;;>b7+-. Thus Black's at
tempt to simplify the position at
once is mistaken.
2 'it>f2 <;;>e4
A dubious move. Better, missed
by the authors, is 2 ... a5! 3 <;;>f3 (or 3
'it>e2 'it>e4 4 f5 'it>e5) 3 ... \t>c4 4 'it>e2
'it>c5 (but not 4 ... 'it>d5? 5 'itid3 'it>c5 6
g5 hxg5 7 fxg5 'it>d5 8 g6+ -) 5 i;t>d3
(5 'it>d2 'it>d6 6 'it>c2 @e6 7 c;t>b3
i;t>d5=) 5 ... i;t>d5 6 f5 'it>e5 7 <;;>c4
'it>e4 8 f6 d3 9 'it>c3 <;;>e3 lO fl d2 1 1
f8='ii' d l='fl' 12 'fl'xh6+ <;;>f3, and in
the resulting queen ending Black
achieves a draw without trouble.
3 rs \ties 4 'it>e2?

Much stronger is 4 a5, but the

authors mistakenly assumed that
this was easily winning, giving the
variation 4 ... 'it>d5 (4 ... h5? 5 gxh5
\t>xf5 6 <;;>e2 'it>g5 7 i;t>d3 l:>.
@xd4-c5-b6xa6-b7+ -)
5
'it>e2
'it>e5(?) 6 '>t>d3 d5 7 f6 'it>e6 8
W xd4 'itxf6 9 @c5 @g5 1 0 'it>b6
xg4 1 1 <;;>xa6 h5 12 i;t>b6 h4 13 a6
tc. Far more tenacious in the opin
ion of A.Kuzmin was 5 ... 'it>d6 6
d3 (6 d2 'iitd 5 7 c2 i;t>d6 8
b3 i>d5=) 6 ...e5 7 'it>c4 \t>e4 8
f6 d3 9 <;;>c3 'it>e3 10 fl d2 1 1 f8='ii'
dl='ii 1 2 'ti'xh6+ f3 .

In this queen ending White has

some chances of a win but upon
correct defence it should probably
be drawn. For example: 1 3 g5 'fl'el+
1 4 'it>c4 'it>g4, and i t i s not easy for
White to improve his position.
4 hS?
.

Better is 4 ... a5 ! (6. 5 'it>d3 'it>d5 6

f6 'iite 6) 5 @f3 i;t>d5 6 \t>f4 'it>c4 7 f6
d3 8 fl d2 9 f8='fl' d l =ii=.

S gxhS 'it>xfS 6 d3 'it>gS 7 'it>xd4

xhS 8 \tics Black resigned.

Now it is clear that best was an

immediate l .. . a5! 2 'it>f2 'it>d5 3 @f3
<;;>c4 4 'it>e2 'it>c3 5 'it>d l <;;>c4 6 'it>d2
'itc5 7 'it>c2 'it>c4 8 f5 i;t>d5 with a
draw, since on 1 ...'it>d5 could follow
2 a5! (the same as on l . ..f6 2 'it>f2 !
'iitg6 3 f5+ i;t>f6 !-4 a5!).
Vaganian-Portisch
Ti/burg, 1 992

1 6 Pawn Endings

A.Kuzmin also drew attention to

this pawn ending. We present his
more accurate analysis.
"The authors assessed this ending
in White's favour, giving 1 Wd2
Wc5(?-A.K.) 2 'it>c3 g5 (2 . . .Wb5 3
g3 g6 4 h3 h5 5 g4 h4 6 g5 'it>c5 7
a3 'it>b5 8 Wd4 i;t>a4 9 Wxc4 Wxa3
10 f5 a5 1 1 fxg6 fxg6 12 e5+ -) 3
g3 (?-A.K.) 3 . .. g4 4 a3 h5 5 a4 a5 6
e5 'it>d5 7 f5 ! Wxe5 8 Wxc4 Wxf5 9
Wb5 +-. But the variation leaves a
strange impression. Firstly let's look
at the final position.

'it>c5 4 a3 h5 5 a4 a5 6 e5 'it>d5 7
'it>d2 (of course White is not obliged
to give up the pawn by the move 7
f5?) 7 ... Wd4 8 Wc2, sailing into a
drawn harbour."
It remains only to see how the
game ended (comments by the
authors of the book).
1 h4? i;t>c5?

After l .. .h5 ! White is in no posi

tion to create a passed pawn: 2 d2
Wc5 3 'it>c3 'it>b5 4 a3 Wc5 5 a4 a5 6
g3 g6 7 e5 Wd5 8 'it>d2 'it>d4 9 c2
c3 1 0 'it>d l i;t>d3- + .
2 h5 f6?

Making it easy for his opponent to

create a passed pawn: 2 ... Wd4 3 e5
'it>c3 4 f5 'it>b2 5 e6 fxe6 6 fxe6 c3 7
e7 c2 8 e8='ii c l ='ii+
.

3 We3 a5?

Once again after 3 ... c3 4 Wd3 c2

5 Wxc2 d4 6 i;t>b3 'it>xe4 7 'it>a4
'it>xf4 8 'it>a5 'it>g3 9 Wxa6 f5 10 a4
'it>g2 Black has the better queen
ending.
4 a3 a4 5 e5 ! fxe5 6 fxe5 d5 7
e6 xe6 8 'it>d4 Black resigned.

After 9 ... 'it>e6 (9 . . . 'it>g6!?) 1 0

'it>xa5 f5 i t is not Black, but White
who needs to think about saving the
game: 1 1 Wb6 (1 1 Wb4 f4 1 2 a5
fxg3 1 3 hxg3 'it>d6- + ; 1 1 b5 f4 1 2
gxf4 h4 1 3 a 5 g 3 14 hxg3 h3 !- +)
l l . ..f4 1 2 a5 fxg3 ( 1 2 ... f3 ! ?) 1 3
hxg3 h4 14 a 6 hxg3 1 5 a 7 g 2 1 6
a8='ilk g l ='fl'+ with a theoretically
won queen ending.
Secondly, after 2 . . .g5 White can
also win directly by 3 fxg5 hxg5 4
a3 f6 5 a4 a5 6 h3 d6 7 'it>xc4 e5
8 'it>b5 xe4 9 xa5.
Thirdly, on 1 d2 correct is an
immediate l ...g5 D. 2 g3 g4 3 c3

Poor technique in transposing

to a pawn ending

It is surprising but true that mod

ern grandmasters experience great
difficulties in transposing to a pawn
ending. There are two factors-a
poor knowledge of pawn endings
and inability to switch from the fac
tors of evaluation and play of com
plicated endings to the factors and
methods typical for pawn endings.
This is a very great problem even
for high-level grandmasters and
here the authors simply must also
name themselves.

Pawn Endings 1 7

Beliavsky-Sveshnikov
Novi Sad, 1 9 79

Black should correctly transfer to

a pawn ending and this is achieved
by l ... .l:!a4! 2 a3, and the position of
the pawn on a3 gives Black the
possibility of more quickly creating
a passed pawn. But in the game
there followed . ..

1....l:!xd4+ 2 .!id3 .l:!xd3+ 3 xd3

f6 4 i;(.c4 a6 S fS!

It is necessary to clear a path for

the king, otherwise the march of his
h2 pawn will lead to nothing.

How is it possible to lose here?

very simply-by transposing to a
pawn ending.
1 'ii'e 2?? 1 @3, with a draw.
1. i;i>c3! Taking on e2 indeed leads
to a draw. 2 i;(fl 11Vxe2+ 3 i;i>xe2
'it>c2, and he had to resign since the
king goes over to the white pawn on
g3 while White's during this time
goes to g6 after which Black wins
by the well-known zugzwang . . . .
'it>g4 ! .
.

s ... 'it>d6 6 fxe6 xe6 7 'it>d4 @rs

8 'it>dS 'it>gS 9 'ii>e 6! fS 10 e3 'it>g4
11 h3+! <;;>gs 12 h4+! i;i>g6 13 hS+
<;;>gs 14 h6 g6 1s h7 'it>xh7 16
'itixfS 'itig7 17 'itie6 Black resigned.
Adams-Lautier
Ti/burg, 1996

Kramnik-Lautier
Belgrade, 1 995

White has an extra pawn and can

win the position as he pleases. As he
pleases? In the game followed .. .
1 h4? l:!.e6+!, and the pawn end
ing with an extra pawn was drawn.

2 .Iles @f6 3 .l:!xe6+ xe6 4 'it>d4

'it>d6 S i;i>c3 i;i>c7 6 b4 cxb4+ 7
xb4 c6, and a draw.

Correct was 1 c5 ! or 1 .!lb5.

18 Pawn Endings

011-Benj amin
New York, 1 995

2 'll xe6 lies 3 lies?

Both players "drift" commented

A.Alekhine. After 3 'it>e2 .l:Ixc6 4
lixc6 bxc6 5 b4 ! 'it>e7 6 'it>d3 'it>d6 7
'it>d4 White has a typically winning
position.
3 bxe6?
..

But here Black does not exploit

his chance-better was 3 .. . lixc6 4
.l:Ixc6 (after 4 lixd5 lic2 and Black's
active rook compensates for the loss
of a pawn) 4 . . .bxc6 5 b4 'it>e7 6 We2
c,t>d6 7 'it>d3 c5 8 bxc5+ 'it>xc5 9
@c3 a5 reaching an equal position.
Neither side can lose the rook
ending. But White decides to try to
win the pawn ending.
1 lid4? .l:!xd4 2 exd4 c,t>g6 3 'it>f3

@rs 4
gS+ 7
'it>dS!
resign

4 'it>e2 i;t>e7 S 'it>d3 'it>d6 6 :!.as

lia8 7 'it>d4 fS 8 b4 lib8 9 a3 lia8

'it>e3 i;t>e6 s 'it>e4 fS+ 6 'it>f4

'it>e3 @dS 8 f3 'it>e4 9 b3+
10 h3 hS, and he had to

1 l .. .g4.

since on 1 1 c,t>d3 follows

However, we should not think that

classical players from the past
handled analogous situations better.
Flohr-Vidmar
Nottingham, 1 93 6

10 e4!

A seemingly illogical move but

Black has only one weakness on a6
and White exchanges his weak e3
pawn, activates his king and rook
along the fifth rank and then sets
about creating weaknesses for the
opponent on the kingside.

10 fxe4 1 1 fxe4 dxe4 12 @xe4

a7 13 i;t>f4 h6 14 h4 ! @e6 lS 'it>g4
.l:ta8 16 hS gS
.

1...'ll e 6?

After the natural 1 . ..c,t>e7 Black

has every chance of holding the
slightly inferior isolated pawn.

Or 1 6 ...gxh5+ 17 'it>xh5 . l:!g8 1 8

g4+ -.
17 g3 .l:Ia7 18 'it>f3
Now the king transfers to the
other flank.
18 ... .l:IaS 19 'it>e4 l:!.a7 20 .l:!eS+!

Here Black has a choice: to allow

the rook to e8 or the king to f5.

Pawn Endings 1 9

20 ... @d6 21 .l:te8 c5 22 .l:Id8+

<Jtc6

After 22. . . rt;c7 23 .l:Ih8 it is time

to resign.

23 c8+ 'it>b6 24 .l:txc5 Black

resigned.

Again typical-only this wins.

19 rt;b6 h2 20 c7 Wd7 White

resigned .
Anastasian-Romanishin
Moscow, 1 994

We also certainly come across

transitions which are backed up by
splendid calculation.
Cruz-Seirawan
Moscow, 1 994

1 ...tt::lc3! 2 i.xc3 dxc3 3 a4 c;t>b4

4 a5 Wxa5 5 Wxc3 'it>b5 6 'it>d3
Wb4!

Black meets the white king's

roundabout route with one of his
own!
1 ....l:Ia5 +!

The best solution, after l ....l:th5 2

'it>b4 J::rxh2 3 c5 and White has
strong compensation for the pawn.

2 rt;b3 .l:txa2 3 'it>xa2 'it>d6 4 'it>b3

rt;cs 5 c;t>c3 e5

7 e3

Or 7 rt;e4 rt;c3 8 Wxe5 rt;d2,

catching up with the white pawns.

7 ...rt;b3 8 exf4 exf4 9 'it>e4 'it>c2

10 '\t>f5 rt;d3 11 'it>xg5 We3 12 'it>h4
Wxf3 13 g5 'it>e2 D rawn.

At first sight it seems that White

has the advantage because of his
passed pawn, but the most important
factor in this ending is the paralysed
white pawn chain on the kingside.

Kuzmin-Petrosian
Moscow, 1 979

6 rt;d3 f5 7 'it>c3 e4 8 'it>b3 h6! 9

'ii>c3 h5 10 'it>b3 f4!

Now follows a typical, well

calculated breakthrough.
11 gxf4 e3! 1 2 fxe3 h4 13 f5
d6!

It is this which had to be foreseen

-bad would have been 1 3 ...g3? 1 4
hxg3 hxg3 1 5 f6 rt;d6 1 6 c5+.
1 4 'it>b4 \ties 15 c5 <bxf5 16 c6
e6 1 7 'it>c5 g3 1 8 hxg3 h3 !

1 c5!

20 Pawn Endings

The transfer to a pawn ending is

forced, since after 1 ...bxc5 2 'it>d3
and 3 \tc4 Black stands badly.

1.. .!lxc5 2 J:Ixc5 bxc5 3 'it'd3 e5 4

\tc4 f5 5 \txc5 h5 6 b4 axb4 7
@xb4 f4 8 a5 e4 9 \tc3 e3 10 c;(d3 !
Black resigned.
.

Ap. analogous idea was not taken

into account by Black when trans
posing to a pawn ending in the
following game ...

capture on a3 the king on c3 can do

nothing since Black will have a
tempo after ... a5 ; h3 a4; h4 @a2;
\tc2 a3 ! ) 6 ... a5 7 \tc4 f4 8 @d4 f3 9
\t e3 <;txc5 10 @xf3 \tc4.
Ivanchuk-Kasimdzhanov
Elista, 1 998

Finkel-Mikhalchishin
Belgrade, 1 998

Here 1 l:td7 is quite simply win

ning. But Ivanchuk was reckoning
on the pawn ending.
1 l:txe6 'iVxe6 2 'iVxe6 fxe6 3
J:ixf8+ <;txf8 4 <;tg2 and all of a

sudden

Winning easily is 1 . ..\te5 ! 2 l:l'.c3

.!lc7 and 3 ... <;td5, but Black wants to
take the bull by the horns at once by
1 ...@d5 ??

White is frightened by the pawn

ending and after. .
.

2 c;(f4 l:tc7

. . .he could quietly resign.

Meanwhile after 2 :id 1 + \tc6 3
l:txd7 <;txd7 4 'it'd3 ! 'it>c6 (4 ... @e6 5
'it>d4 changes practically nothing) 5
<;td4 a6 (after 5 ... f4 6 'it>e4 <;txc5 7
@xf4 'it>b4 8 \t e5 \ta3 9 @f6 'it>xa2
1 0 cJitg7 'it>b 1 1 1 h4 ! a5 1 2 h5 an
endgame with an extra p awn for
White is reached. Therefore Black
rightly plays 7 ... @d5 8 \te3 'it>e5 9
h4! a6 1 0 a3, with a draw) 6 a3!
(bad is 6 'it>c4 f4! , and the white
king does not get to g7, while after a

4...a5 ! !

The only move-bad was 4 ...cJite7

5 c;(f3 @d7 6 @f4 <;tc6 7 <;tg5 cJitd5
8 c;(f6 a5 because of 9 b3 ! with a
win for White.
5 @f3 @f7 6 cJite4
Nothing is gained by 6 cJitf4:
6 ... a4 ! 7 g4 hxg4 8 \txg4 \tg8 ! !,
creating the distant opposition. And
if 6 b3, then 6 . . .g5 7 \te3 @g6 8
'it>e4 'it>h6 9 'it>d4 \tg6 l 0 'it>c5 c;(f5
1 1 hxg5 <;txg5, with a draw.
6 ... g5 7 c;(d4 <;tg6!

We must mention the erroneous

ness of the exchange on h4, e.g.
7 ... gxh4? 8 gxh4 \tg6 9 \t e4 a4 1 0
\tf4 @h6 1 1 <;t e3 @g6 12 cJite4 ! ,
triangulating, and Black is forced to
move away his king to h6, which
gives White the opportunity of
entering via f4.

Pawn Endings 21

8 @cs gxh4 9 gxh4 c,t>fs 10 'it>b5

wxe5 11 'it>xa5 c,t>d4 and he had to
agree a draw.
Pawn endings in the creative work
of Robert Fischer

Each of the great champions has,

besides his own style, also his own
methods of play in the various
stages of the chess game.
Thus, upon his ' taste' depends the
ar ising various structures and ac
cording to his 'taste' his tr eatment
of them. The transfer to a pawn end
ing is one of the instruments for
realisation of an advantage or a
method of defence. In F ischer's case
this ending is met more frequently
than any other champion, and the
authors were interested in how the
great Fischer played them at these
moments. The examples show quite
a broad spectrum of quality of play.
Fische r-Letelie r
Mar de! Plata, 1 959

White clearly has the better king

but the asymmetrical structure gives
Black the possibility to reply to the
creation of a passed pawn on the
king's flank with the creation of a
passed pawn of his own on the op
posite side. F ischer continued to
play for a win.

1 a4?

In principle, a serious mistake,

though White's plan is understan
dable-he places a pawn on a5 and,
by sacrificing a pawn on the king
side, breaks through with his king to
the pawn on a6. However it is not
possible to win this position.
E.Mednis in his book How to beat
Bobby Fischer assessed the position
as a draw and did not criticise the
move in the game. Nevertheless
White has a path to victor y and it
consists of the move 1 g4! . In reply
Black has the following contin
uations:
a) l .. .Wd6 2 f5 gxf5 3 c,t>xf5 c4 4
bxc4 bxc4 5 'it>e4 c3 6 c,t>d3 @e5 7
c,t>xc3 c,t>f4 8 'it>b4 'it>xg4 9 @a5 Wf5
10 'it>xa6 c,t>e6 11 a4, and White is
victorious;
b) L.. c,t>d6 2 f5 g5 3 a4! c,t>c6
(3 ... b4 4 a5 'it>c6 5 f6+-) 4 axb5+
axb5 5 'it>e5 +-;
c) l ... a5 2 a4 b4 3 'it>d3 ! (Why
not 3 f5+ gxf5 4 gxf5+ 'it>d6 5 f6?
Because of 6 ... c4! 7 bxc4 e6! , and
for the time being White must think
about saving himself) 3 . .. 'it>d5

4 g5 ! 'it>e6 5 'it>c4 c,t>f5 6 'it>xc5

'it>xf4 7 'it>b5 'it>xg5 8 'it>xa5 '\t>f4 9
'it>xb4 g5 10 a5, and White reaches a
queen ending with a b-pawn-and
every chance of a win. This

22 Pawn Endings

variation was found by A.Bely

avsky. But we return to the game:
1. .'it>d6 2 f5??
Mednis passed over this move in
silence. Fischer, apparently, noticed
that as a result of the correct 2 g3
'it>e6 3 a5 'it>d6 4 g4 'it>e6 5 f5+ gxf5
6 gxf5 'itf6 7 c,t>d5 c4 8 bxc4 bxc4 9
'it>xc4 c,t>xf5 the Black king hurries
back to c8.
.

Here is another example.

Gligoric-Fischer
Candidates (t), Belgrade 1 959

2 gxf5+ 3 'it>xf5
.

Here, leading to a draw is 1.. ..l:!h5 !

2 l:!c5 l:!xc5, and whichever way
White retakes there follows 3 . . . 'it>c8
with a draw. But Fischer mistakenly
played...
1 . l:!h8? 2 c,t>xb5?
..

3 'it>d5?
..

(!)Mednis. Why not 3 ... c4! , and

White must resign at once.
4 g4 'it>d4 5 g5?

5 axb5 leads at once to a draw.

5 ...c4! 6 bxc4 b4 7 c5??

After 7 g6 White reaches a slight

ly inferior queen ending but with
chances of a draw. Now however
it's all simple-7... b3, and it is time
for White to resign.
The pawn ending, apparently so
simple, is difficult in that on the
transfer to it the players must calcu
late the changes that take place in
the game by comparison with other
aspects of the ending.

Now Gligoric met mistake with

mistake, wrongly transferring to a
pawn ending. Winning was 2 c7+!
'it>d6 3 c6+ 'it>d7 4 'it>xb5 J:ib8+ 5
J:ib6 J:ih8 6 Il:b7+ 'itic8 7 'it>a6 nh6+
8 'it>a7.

2 llb8+ 3 'it>a4 J::ra8+ 4 'it>b3

nc8! 5 nxc8 'it>xc8 6 'it>c4 'it>b8!
...

Gligoric did not reckon in his

previous calculations that White
could not maintain the opposition,
therefore it's a draw.
With time Fischer managed to
correct his shortcomings and his
transfer to the endgame became
immaculate.

Pawn Endings 23

Lombardy-Fischer
USA (ch), 1 960/61

We also come across ignorance in

transferring to a pawn ending by
Fischer's rivals.
Fischer-Bisguier
USA (ch), 1 959

R eturning the exchange to win a

pawn was also a recipe of Capablan
ca himself.

1...lixc3+! 2 bxc3 .l:IxeS+ 3 'it>d2

xel 4 c;t>xel 'it>dS S 'it>d2 'it>c4 6 hS
b6 7 'it>c2 gS ! 8 h6 f4 9 g4 as 1 0
bxaS bxaS 1 1 'it>b2 a 4 12 Wa3
@xc3 13 'it>xa4 'it>d4 14 'it>b4 'it>e3
and White resigned .
Fischer-Larsen
Candidates (m) Denver, 1 9 71

Black should suffer a little in the

rook ending by 1 ....l:Ig5 2 l:id4 b5 3
'it>e2 'it>c5 4 @f3 .l:!g8 5 'it>f4 f8+
with the idea of breaking through on
the second rank with the rook. But
Bisguier decides to transfer to a
pawn ending in which it requires in
credibly accurate play to achieve a
draw.
1. ..Ii.dS? ! 2 'it>e2 .l:Ixd2+

No help is 2 ... b5 3 xd5 Wxd5 4

'ite3 a5 5 g5 b4 6 g6 We6 7 'it>d4

+ -.
3 'it>xd2 @dS 4 'ite3 @es S 'it>O!

Fischer 'waits' for the weakening

of Black's pawn structure-there is
no win by 5 g5 'it>f5 6 'it>d4 'it>xg5 7
'it>xc4 Wf6 etc.
S ... aS

Here White won easily after

1 a4 Wf8 2 .lii.c 3! .txc3 3 Wxc3

c;t>e7 4 'it>d4 'it>d6 S aS f6 6 a6 'itc6
7 a7 'it>b7 8 'it>dS h4 9 'ite6 and
Black resigned.

If 5 . . .'it>f6 6 'it>f4 'it>g6 7 @e5 @g5

White goes not for the c4 pawn but
for the pawn on b7-8 'it>d6! 'it>xg4
9 'it>c7 @f3 1 0 'it>xb7 'it>e2 1 1 'it>xa7
@d3 1 2 'itb6 'it>c2 1 3 'itc5, and
wms.
6 'it>e3 a4

On 6 ...b5 winning is 7 g5 b4 8 g6
'it>f6 9 'itd4.

24 Pawn Endings

7 g5 @f5 8 'it>d4 'it>xg5 9 'it>xc4

@f4 10 Wb4 'it>e3 1 1 'it>xa4 c,t>d2

.
,
q
-
'
w8
-8
-8

Off at last!
20...'it>c2

20 ... 'it>d3 21 'it>b5 +-.

21 b4 Black resigned.
Rossolimo-Fischer
USA (ch), 1 962/63

. . .
DL.
.
.

L,
m m ... m m

1 2 'it>b3

The position appears very simple,

but in fact it is just the opposite as
testified by the mistakes made in
past analysis by authors even in the
Encyclopaedia of Chess Endings.
Fischer avoided 1 2 c4 because of
1 2 ... Wc2 1 3 b4 'it>c3 14 c,t>b5 'it>d4
1 5 c5 'it>c3 1 6 'it>a5 'it>c4 1 7 'it>a4
Wd4 1 8 'it>b3 'it>d5 1 9 'itc3 b6 ! with
a draw.
12 ... 'it>d3 13 c4 Wd2 14 Wa4

Also no help is 14 c5 'it>d3 1 5

'it>b4 'it>d4 1 6 @b5 c,t>d3 1 7 'it>a5
'it>c4 ! .
14 ... 'it>c2 15 'it>a3 !

Again 1 5 b4 draw, as pointed out

above.
15 ...'it>d3 16 'it>b3 b6?

It is incomprehensible why Bis

guier did not give Fischer the
chance 'to show' the win after
1 6 ...'it>d2.
17 c,t>b4 c,t>c2 18 'it>a3

Clearly not 1 8 Wb5? because of

1 8 ... Wb3 ! , but in the Encyclopaedia
of Chess Endings is given the vari
ation 1 8 b3 ? 'it>b2 19 'it>a4 'it>c3 20
'it>a3 'it>d3 21 'it>b2 Wd4 22 'it>c2
b5 !=.
1 8...'it>d3 19 Wb3 'it>d2

1 9 ... 'it>d4 20 'it>b4 wins.

20 'it>a4!

Black clearly has the more active

king but the White g5 pawn paral
yses Black's flank and it is imposs
ible to win.
1 ...a4 2 'it>e2

The Encyclopaedia recommends

2 b3 ! a3 3 h3 b4 4 h4, and now
Black must play 4 ... d5 5 exd5 'it>xd5
6 c,t>d3 'it>c5 7 'it>e4 Wd6 with a
draw.
2 ...a3 ! ?

The last try.

3 bxa3 Wc3 4 a4 ! bxa4 5 'it>e3

Wb2 6 @d2 'it>xa2 7 Wc2 a3 8 h3
And a draw.
Some conclusions

In the earliest stage of his career

Robert Fischer had technical prob
lems in playing pawn endings, and
particularly when transposing to
them. But by the beginning of the
60s he had already managed to
eradicate these shortcomings and no
longer had technical problems with
pawn endgames.

Pawn Endings 25

Exercises:
Pawn Endings
1

How can White win the endgame?

What plan must White adopt

to realise his advantage?

How do you assess the position

with White to move?

What is correct l jxg4 or J...hxg4?

..

2 6 Pawn Endings

How does Black save the endgame?

What is correct: J... @e5 or 'iPg5?

Assess the outcome of playing

the typical 1 g4?

What is correct: J ...g4 or l . rtJf6?

.

10

Should White resign ?

What is the simplest way

for White to win ?

Pawn Endings 2 7

11

14

How does White win ?

Can White hold the position ?

12

15

Can Black save himself?

White to move and win.

13

16

What is right J . @e6 or J ... c6 ?

Point out the right move for White.

28 Pawn Endings

17

20

White wins.

How does White make a draw?

18

21

What is correct: l . . e4, 1 g4

or l . . . @g8?

What is correct:
l . . . @g7 or l . . rJ;h5 ?

19

22

What is the right move for Black?

What is correct: l . a 5 o r l ... axb5?

...

Pawn Endings 2 9

23

25

Find the right continuation

for Black.

How should White continue?

24

26

What is correct: 1 c4 or 1 b4?

Wha t is correct:

l @xb4 or l rJilc4?
. ..

...

30 Pawn Endings

Exercises:
Transposition to a Pawn Ending
1

Can Black transpose to

a pawn ending?

Demonstrate a plan to realise

the extra pawn.

Can White transpose to

a pawn ending?

How do you assess th e endgame

after 1 c5?

Pawn Endings 3 1

In whose favour is the pawn ending

after I b5 ?

Which should Black prefer,

the rook or the pawn ending?

Is it worth White avoiding the

transposition to a pawn ending?

Can Black transpose to

a pawn ending?

10

...

Can

0,d3 be played?

In what way can White transpose to

a pawn ending?

32 Pawn Endings

11

14

Is it possible to transpose to
a pawn ending by I g5 ?

Find the right con tinuation

for Black.

12

15

Find the right continuation

for Black

How can White save the game?

13

16

Can Black transpose

to a pawn ending?

Find a winning plan for Black.

...

Pawn Endings 33

17

20

Can White transpose to

a pawn ending?

How can Black hold a draw?

18

21

Find the right continuation

for Black.

How does White transpose to

a winning pawn endgame?

19

22

How should Black realise

his extra pawn?

Find the right continuation

for Black.

Knight Endings

Due to their small radius of action

knights and pawns have very defi
nite features. M.Botvinnik described
these features well by the very com
prehensive and deep statement:
"Knight endings are pawn endings".
Indeed the method of playing knight
endings is very close to those with
pawns. Here also a space advantage,
active position of the king, pawn
structure are all significant. An
important role is played by the
presence of outside passed pawns.
Often the zugzwang motive is ex
ploited in the game. However, apart
from the general principles char
acteristic for p laying pawn endings,
play in knight endings is compli
cated by the possibility of pursuing
a knight with an enemy king. Apart
from this, knight endings with
limited pawn material have specifi c
features
associated
the
with
possibility for the weaker side of
sacrificing the knight to liquidate
the opponent's pawns and thereby
achieve a draw. This feature creates
significant difficulties in the
realisation of a positional and
material advantage.
Let us look at a few examples
where there is a compact pawn
mass.
Upon the central placement of
pawns the pieces can have quite a
broad range of activity. In the first
position White has a positional ad
vantage due to the active position of

his king and the presence for Black

of a weak backward pawn on b6.
Fedotov-Arkhipov
Moscow, 1 9 78

However the fact that there are

only a few pawns on the board com
plicates White's task.
An insufficiently deep penetration
into the features of the position
leads to Black's defeat, practically
without a struggle. However,
analysis shows that Black has very
considerable defensive resources,
and to achieve a win White has to
demonstrate very accurate p lay
throughout.
Let us look at the possible course
of events. The first part of White's
plan consists of the creation of a
passed pawn and exploitation of the
weakness of the b6 pawn.
1 f5 exf5 2 liJxf5 ctJd4 3 ctJd6

White carefully controls the b5

square, preventing the move

Knight Endings 35

.. .b6-b5, e.g. 3 tfJg7? @c6 4 e6

tll xe6 5 tfJxe6 b5, and the white
pawns are exchanged. Also no good
is 3 lll xd4 cxd4 4 rJ;fJ d3 5 e6+
rJ;; c7 6 e7 d2 7 e8='ili' d l =WV, with a
draw. Black, for his part, is obliged
to control the e6 square. Now, for
example, 3 ...c6 with the idea
...b6-b5 , does not work because of 4
e6! rJi;xd6 5 e7, and the pawn cannot
be stopped.
3 ttJe6 4 tfJb5 tfJd4 5 tfJc3 tfJe6
..

.
,
.
&,,,,, M%ft..
%ft.'.
J
.
.
.
.
...;
. . .

....

Black's possibilities. Now his best

counterchance lies in the formation
of a passed c-pawn, since giving up
the b6 pawn without a murmur leads
quickly to defeat. And so . . .
6 ... b5 ! 7 cxb5 Clld 4 8 Cll c3

The realisation of the extra pawn

requires accuracy. Black's hopes are
linked to the win of the b-pawn and
the subsequent sacrifice of the
knight for the e-pawn. Since his
king is already close to the b-pawn,
any advance of this pawn must be
backed up by accurate calculation.
The second part of White's plan
lies in the win of the c-pawn while
holding on to his own pawns. Let us
see whether Black can hold the de
fence by the active advance 8 ... c4 or
the passive move 8 ... tfJe6.
I. 8 . . .c4

6 ltJd5

And so the first part of the plan is

completed: White wins the b6 pawn,
since the opponent's pieces are
forced to control the passed e5
pawn. Also possible is 6 tlJa4 How
ever it is interesting that, with the
knight on a4, tactical resources
suddenly appear for Black which
requ ire precise attention from his
opponent. F or examp le: 6 . . . b5 7
cxb5 tfJc7. This trappy move was
not possible with the knight on d5.
Now, on the natural 8 tfJxc5+ fol
ws 8... 'it>e8 ! with a draw after 9 b6
"Z.l d5+ . However after 8 tfJc3 White
can return to the channels of the
main var iati on, looked at below.
lso possible of course is 8 b6
"Z.l d5+ 9 'ii>fl c4 1 0 e6+ @c6 1 1
ctJc3 ! tfJ xc3
1 2 e7 and White
should win. But
7 ...tfJ 7 is not obli
gatory and is given to illustrate

On c4 the pawn supported the ma

noeuvres of the white knight but
now Black is left without a base on
d4 for his own knight which is also
deprived of its defence. This creates
a motive to carry out the threat of a
breakthrough with the king to the c4
pawn, in combination with thr eats to
advance the passed pawn.
But the direct approach of the
king to the c-pawn does not achieve
its objective. For example: 9 c;i;>g5

36 Knight Endings

Wc7 IO Wf4 Wb6 1 1 @e4 liJxb5 is not successful. For example: 9

@f5 liJd4+ l 0 @e4 @e6 1 1 b6 liJc6
with a draw.
The solution lies in the creation of 12 b7 @d7 and ... @c7. Therefore
a zugzwang position where Black White first endeavours to provoke
simultaneously controls the e6 and the advance of the pawn to c4, aim
rs squares, i.e. in this position it is ing to obtain the position in the
necessary to hand over the move to previous diagram. The direct 9 liJe4
Black. This is achieved by the does not achieve its objective be
endgame king manoeuvre, 'trian cause of 9 .. liJc7! , and he has to go
gulation', characteristic of pawn and back, 1 0 liJc3, so as not to allow a
indeed also of knight endgames. In draw in the variations l 0 liJxc5+
the present case, along the g6, II @e8! 1 1 b6 liJd5+ or 10 b6 liJd5+.
and f6 squares. This method is ac It becomes clear that the manoeuvre
tively exploited also later on.
liJe4 is good when the king is situ
9 wn liJe6 1 0 @g6 liJd4 1 1 @f6
ated on fl and cannot immediately
The objective is achieved! Now in be attacked. The solution to the
the position shown in the previous position is again achieved by the
diagram it is already Black to move. triangulating
king
manoeuvre
I l...liJe6 12 @rs liJc5 13 liJd5 g6-fl-f6 .
liJd3 14 e6+ @d6 15 e7 Wd7 1 6
9 'iti'g6 We7
We4
After 9 . . liJd8, preventing the
The c4 pawn is doomed.
move to fl, the king breaks through
Instead of l O. . .liJd4 there is no to the c-pawn: 1 0 @rs @e7 1 1 @e4
salvation in 1 0 ... @e7, on which We6 12 liJd5 followed by liJf4+ and
possible are both 1 1 @rs liJc5 12 'it>d5.
tlld 5+ @If 1 3 b6 @e8 14 e6, and 1 1
The continuation 9 .. . c4 l 0 @rs!
liJd5+ We8 ( 1 l .. .Wd8 1 2 @fl liJc7 leads to a position looked at under
13 liJc3 @d7 14 b6) 12 @f6 liJd4 1 3 the analysis of the first variation.
b6.
The continuation 9 . . liJd4 1 0 @fl
And so 8 ...c4 loses. But what if liJe6 1 1 liJe4 is looked at below (see
the pawn is left on c5?
moves 1 1 - 1 3) .
.

II. 8 ...liJe6

1 0 liJd5+ 'iti'd7

On 1 0 ... We8 winning is 1 1 b6

liJd8 1 2 'iit f6 'iii' d 7 1 3 e6+.
11 @t7 liJd4
Losing at once is l 1 ... liJd8 1 2
@f6 liJe6 1 3 liJb6+.
12 liJc3 liJe6 13 liJe4 liJd8+

Also here the immediate break

through with the king to the c-pawn

The position arising after 1 3 ...c4

14 liJc3 liJd4 1 5 @f6 is examined in
the commentary to the continuation
8 .... c4. Also losing is 1 3 ... liJd4 be
cause of 14 liJxc5+ @c7 1 5 liJe6+.
14 Wf6 c4
Black already cannot hold on to
the pawn on c5: since in the event of
14...liJe6 winning is 15 b6 liJd8 1 6
liJxc5+ @c6 1 7 @e7.

Knight Endings 3 7

15 lLic3 lLie6 16 'it>f5 , and White

wins by continuing as in the analy
sis to Variation I.
The difficulties also encountered
by the defending side with a flank
pawn fonnation are well demon
strated by the following example.

lLixg7 However he does not notice

this possibility.
6...tt:::lf4? 7 'it>e5 lLig6+ 8 'it>f5 gxf6
9 gxf6

Kremenetsky-Razuvaev
Moscow, 1 981

How can he fight against the

passed f6 pawn? Black cannot allow
its further advance to fl since, with
out support from the king, the
knight will not be able to cope with
it, e.g. 9 . . tt:::lf8 1 0 fl 'it>h5 1 1 lLic5
'it>h6 1 2 'it>f6 ll'ih7+ 1 3 'it>e7 Wg7 14
lLid7. The transfer of the knight to
this key point ends the struggle,
since the f8 square is indefensible.
This is a typical knight manoeuvre
with a pawn one square from queen
ing. And so the knight is forced to
move to a passive position on h8. In
contrast to the variation given after
Black' s 3 rd move, Black's pieces
now have no room for manoeuvre.
9. lLih8 10 tLld6 'it>h5 1 1 i;t>e6
.

White has a material and posi

tional advantage, but his chances are
based on the fonnation of a passed
pawn which is inevitably linked to
exchanges and gives Black defens
ive resources.
1 'it>e3 tt:Jd5+ 2 i;t>d4 lLif4 3 g5+

Wh5
More natural looks 3 ... fxg5 4
hxg5 'it>h5 5 i;t>e5 'it>g4, creating de
fensive possibilities after 6 g6 lLid3+
7 i;t>e6 tt:Jf4+, and also upon 6 f6
gxf6 7 gxf6 lLig6+ 8 We6 tLlf4+ 9
'it>d6 'it>f5 1 0 fl lbg6.
4 lLixf6+ 'iitxh4

After 4... gxf6 5 gxf6 the white

p awns are very dangerous.
5 lLie4 tLlh3 6 f6?

On 6 g6 'it>h5 7 'it>e5 White main

tains the advantage, but now Black
can momentarily exploit his own
drawing resource, linked to a knight
sacrifice to liquidate the pawns:
6 . . . lLixg5 ! 7 fxg7 lLie6+ and 8 . . .

..

g6 12 'it>e7

38 Knight Endings

The black king has two squares at

its disposal: h5 and h7. Black played
12 'it>hS
Why did he reject 12 . . . 'it>h7? Let
us look at a possible continuation of
the struggle. 1 2... h7 1 3 CiJfl
CiJg6+ 14 e8 _g 8 1 5 CiJh6+ 'it>h7
16 CiJg4 CiJh8 17 tll e 5 @g8 1 8 'it>e7,
and there arises a well-known posi
tion from Cheron, presented in
Y.Averbakh's monograph Compre
hensive Chess Endings (2nd edition,
1 980). This position is won in the
following way: 1 8 ... h7 1 9 @f8
'it>h6 20 'it>g8 'it>g5 2 1 g7 f5 22
CiJd7 CiJg6 23 fl 'it>g5 24 CiJe5 CiJf4
25 'it>g8 CiJe6 26 CiJD+ and 27 CiJd4.
Instead of 14 ...g8, there is no
saving himself either by 1 4 .. .'Llf4,
or 1 4 ... CiJM. For example: 14 ... CiJf4
14 CiJe5 CiJd5 1 6 fl CiJf6+ 17 'it>e7
CiJg8 1 8 'it>e6 'it>g7 19 CiJd7; or
14 ... CiJh4 1 5 CiJg5+ 'it>h6 16 fl CiJg6
17 CiJf3 'it>g7 18 CiJe5 CiJf8 1 9 'it>e7
ttJh7 20 CiJd7 'it>h8 2 1 CiJf8 (2 1 CiJf6
ttJf8! ) 2 1 ...tLlg5 22 CiJg6+.
After 12 ... h5 ! the king at the
necessary moment can attack the
pawn from the rear. This method of
defence is presented in the same
book by Y.Averbakh in examples
470 and 47 1 . For example:
(a) 13 CiJfl CiJg6+ 1 4 @d6 (also
insufficient is 14 'it>e8 because of
14 ...'it>g4 1 5 CiJd6 CiJe5 !=) 14 ... 'it>g4
1 5 CiJe5+ @f5 1 6 f7 f6=;
(b) 1 3 'it>d7 'it>g5 1 4 e6 @g6.
In the game followed:
13 'it>e6 'it>g6 14 'it>e7 'it>h5 15
...

We6 'it>g6 16 tt:irs 'it>gs

Great accuracy is required of

Black. Simpler is 1 6 ... 'it>h7, since
White does not manage to reach
Cheron 's position . For exampl e:
(a) 17 tllh4 'it>_g8 1 8 We7 CiJfl 1 9
ti:J f3 CiJh6 2 0 t'zJe5 CiJf5, defends
successfully;

(b) 17 CiJe7 h6 18 CiJc6 @g5 1 9

CiJe5 @f4 20 CiJ fl CiJg6 2 1 ttJd6
ttJh8 ! 22 e7 CiJg6+ 23 @fl CiJe5+
24 g7 'it>g4! (it is important to
leave the file where the promotion
of the pawn will take place!) 25
CiJc4 CiJc6 26 'it>g8 CiJd8.
17 CiJd4 !

The knight heads fo r the key e5

square!
17 @h6 18 CiJf3 tt:::lg6
On 1 8 ... 'it>h7 1 9 CiJe5 @g8 20 @e7
arises the above-mentioned Cheron
position.
..

19 'it>d7 'Llh8 20 CiJe5 'it>g5 2 1

e6

21. 'it>h5?
Black cannot stand the tension
and makes a mistake. With the only
move 2 1 .. .'it>f4 ! he holds the posi
tion. A.K.remenetsky intended the
continuation 22 ttJfl CiJg6 23 CiJd6
'Llh8 (on 23 ... tt:Je5 winning is 24
'it>d5) 24 'it>e7 tt:::l g6+ 25 @fl tt:::l e 5+
26 'it>g7 g4 27 ttJc4, so as on
27 ... CiJc6! to offer a draw, while on
the natural-looking 27 ... CiJD? there
is the beautiful win: 28 'it>h6! CiJg5
29 @g6 @f4 30 CiJd2 ! 'it>g4 3 1 lt.Je4
CiJD 32 ti:Jf2+ h4 33 CiJd3 CiJg5 34
CiJe5 (zugzwang !) and then 35 fl.
..

22 'it>e7 Black resigned.

On 22 ...h4 winning is 23 @f8

'it>g5 24 g7 f5 25 CiJd7.

Knight Endings 3 9

The following example shows the

possibilities of the weaker side,
when he has compensation in the
form of actively placed pieces.
Gufeld-Grigorian
Daugavpils, 1 9 79

White has an extra pawn but

Black has sufficient counterplay due
to the active positions of his pieces.
Now White should secure the elas
ticity of his pawn chain by 1 h3,
retaining the possibility of the ex
change gxf5 for an appropriate mo
ment. However, there followed:
1 gxf5 gxf5 2 ctJf3 ctJf6
It is important to maintain the
possibility of play in the centre and
on the king's flank. Passive is
2 . . . ctJd6 3 @d3 h6 4 ctJd4 with a
wm.
3 ctJd4

On 3 @d3 possible is 3 ...ctJd7 4

ctJd4 ctJc5+.
3 ...@e4 4 ctJb5 @d5

Further activity by 4 ... ctJd5?? ends

in a sudden mate by 5 ctJd6!
5 f3 ctJg4!
The activity of the opponent' s
pieces does not permit White to
strengthen his position without ex
changing pawns. But this allows
Black drawing chances linked to the
possibility of sacrificing the knight
at an appropriate moment.

For examEle:
(a) 6 h3 8h6 7 ctJd4 'it>d6 8 @g3
d5 ! 9 lt>h4 e4 IO @g5 ct:Jf7+ 1 1
@f6 ctJd6 1 2 h4 ctJe8+ 1 3 @fl
@xe3 ! 14 ctJe6 ctJc7 !=;
(b) 6 e4+ @c5 7 exf5 ctJxh2+ 8
g3 ctJfl + 9 lt>g4 c6! 1 0 f6 @d7
1 1 ctJc7 ctJe3+ 12 g5 ct:Jc4 and
ctJd6=.
Probably the maximum that White
can extract from the position is to
organise a passed pawn by 6 h3
ctJh6 7 ctJd4 @d6 8 e4.
The attempt to obtain two con
nected passed pawns, undertaken in
the game, is met surprisingly by an
energetic
black
counterattack,
linked to the activity of the king.

6 ctJd4 ct:Jxh2+ 7 @g3 liJfl + 8

@ f2 @e4 ! !

An apotheosis o f the activity of

the king! On 9 ctJxf5 lLixe3 ! the
white pawns are liquidated.
9 'itixfl @xe3 1 0 lLie6 f3 1 1
Wgl @g3 D raw!

The black king confidently ties

the knight down to the f4 pawn and
does not allow its own white col
league out of the cage. On 12 hl
follows 1 2 ... h6! 13 @gl h5 and then
h4-h3-h2, stalemating the king.
Now let us look at an example of
playing knight endings with the
presence of pawns on opposite
flanks. In this case a decisive role is
often played by the organisation and

40 Knight Endings

energetic exploitation of a distant

passed pawn.
Vladimirov-Novopashin
Volgodonsk, 1 981

6...fS? 7 lL\aS tLldS 8 b7 tLlb4+

9 e2! tLla6 10 'it>f3

And so White blockades the pawn

and with a fine knight manoeuvre
wins it.
10 ... tLlbS 1 1 h4 @es 12 tLlb3 tLlc6
13 tLlcS @fS 14 tLld3!
An important finesse ! If 14 tLld7

Black holds on by 14 ... tLld4+ 1 5

2 tLlc6 1 6 b8==\W lLlxb8 1 7 lLlxb8

'itig4.

14 ... tLld4 lS f2 lLlc6 16 lLlb4 !

tLlb8 17 @f3 @es 18 lL\d3+ d6 1 9
'it>xf4 c6 2 0 @gs 'it>xb7 2 1 'itih6
tLlc6 22 'it>xh7

White's advantage lies in the

possibility of quickly creating a
passed pawn while retaining a good
coordination of pieces which allows
him to support the advance of this
pawn and at the same time to con
trol Black's activity on the kingside.
However the peculiarities of the
position are such that great accuracy
is required from both sides for the
fulfilment of their plans. Now White
can quickly organise the manoeuvre
1 d4 tLldS 2 @cs tLlxf4 3 b5, but
he prefers prophylaxis.
1 tLle3 gS? !

A hasty decision, facilitating

White' s task. After l .. .h6 2 h4 g5!
Black's defence is far easier.
2 fxgS f4 3 tLlc4 @fS 4 bS axbS S
axbS xgS 6 b6

Let us pay attention to the differ

ences in the possibilities for the two
sides. The passed b-pawn distracts
the black knight whereas both white
pieces can attack the passed pawn.
Nevertheless Black can put up stub
b m resistance by breaking through
.
w ith his king to the g4 square6 ...'itig4 L However he chooses
another route-march for his king.

This ending has theoretical sig

nificance. Will the black pieces
succeed in neutralising the h-pawn?
22 ... tLle7 23 hS 'itic6 24 'itig7 dS
2S @f6 tLlg8+ 26 @f7 tLlh6+ 27
'itig7 lL\fS+

On 27 ... tLlg4 winning is 28 lL\2 ! .

28 f6

Knight Endings 4 1

An instructive position! The black

knight can control the h-pawn from
the g4 and f5 squares, therefore to
advance it White must deflect the
knight from these points. This task
can be fulfilled by the knight from
the e3 square: 28 ... lbh6 29 lLif2!
\tid6 30 'it>g6 lbg8 3 1 rJig7 lbe7 32
ltJg4 e6 33 lbe3 .
Black chooses the best plan, en
deavouring to break through with
the king to the h-pawn and at the
same time controlling the e3 square,
and this attempt is rewarded.
28 ... 'it>e4! 29 'it> g5 lbd6 30 lbf2+
rJie5 31 Wg6 lt:\f5 32 lbd3+ rJie4 33
lbcl? 'it>f4 34 lbe2+ g4 35 lbg3
lbe7+ 36 Wf6 lLif5 Drawn.
The king attacks the pawn and de
fends the knight, observing from the
h6 square.
Is White's advantage in the last
diagram really insufficient for a
win? No way! After the game
E. Vladimirov demonstrated a clear
way to realise the passed pawn. In
stead of 29 'it>g5 he gave 29 'it>g6! !
as leading immediately to victory.

Now every move by Black loses:

(a) 29 ... lbe7+ 30 'it>g5 lbf5 3 1
lLif2+ rJie5 3 2 lbg4+ rJie6 (32 ... rJie4
3 3 lbe3 ! ! + -) 33 rJig6 lbh4+ (or
3 3 ... lbe7+) 34 rJig7 'llf5+ 3 5 rJif8 !
Zugzwang! On any king move, win
ning is 36 lbe3 ! Leading to the same
outcome is 30 ... lbg8 3 1 lbf2+ rJie5
32 lbg4+ We6 33 Wg6 lbe7+ 34
rJig7 lbf5+ 35 'it>f8 ! .
(b) 29 ...lbh4+ 30 rJig5 lbf3+
(30 ...lbf5 after 3 1 lLif2 leads to
Variation (a) above) 3 1 rJif6 ! , and
the h-pawn is unstoppable.
But to conclude White missed the
win by playing 3 3 lbc 1 ?. After
Black's 32nd move there arises the
position shown in the last diagram,
but with White to move. With the
tried and tested 'triangulation' king
manoeuvre 33 'it>f6 ! 'll h6 34 rJig7
lbf5+ (34 ... lt:\g4 35 lLif2!) 35 'it>g6
he hands the move over to his oppo
nent and wins.
Returning to the position shown
in the above diagram, it is useful to
bring to mind a fragment from the
game Botvinnik-Simagin (Moscow
1 955), where, by exploiting his
opponent's inaccurate play, M.Bot
vinnik was able to save himself in a
similar ending.
Botvinnik-Simagin
Moscow, 1955

A picture of a position, having an

instructive character! White's pieces
are clearly fulfilling their functions:
the king confidently squares up to
the enemy knight, while the white
knight remains triumphant on the d3
square.

42 Knight Endings

Leading to a win here is 1 .. .@f5 2

tiJf3 ltJg5 3 'Llh4+ c,t>g4 4 'Llg6 tt:Jf3
5 'it>b4 @g5. However there
followed:
1 ...'Lld2? 2 'iitb 4 @f5 3 @c3!

With tempo the king draws closer

to the centre of events and now if
3 ... 'it>xe5 4 'it>xd2 he succeeds in
containing the p awn.

3 ... 'Lle4+ 4 d4 'Llg5 5 'Lld3 c,t>g4

6 'Lle5+ 'it>f5 7 'Lld3 @g4 8 'Lle5+
@g3 9 'Llg6!

In contrast to the previous

example White has this defensive
resource, since the pawn finds itself
one move further from the queening
square.

1 'Lle4 @es 2 c,t>d6!

After 2 'Llxg5 'Llf2 White's task is

more complicated.
2 @f8 3 'Llxc5 'Llf2 4 'Lld7+ @f7
..

5 ltJe5+ @f6

And so White creates a distant

passed pawn. He combines its ad
vance with fine manoeuvring of
pieces, creating on the way threats
to the g5 pawn.
6 'it>d5 'Lid 1 7 c5 'Llc3+ 8 @c4
'Lle4 9 'Lld3 ! 'it>e6 10 'it>d4 'Llf6 1 1
'Llf2 @e7 1 2 c 6 'it>e6

On 12 ... 'it>d6 winning is 13 'Lle4+.

13 @c5 @e7 14 @b6 @d6

9 .'Lle6 1 0 @e3 'Llf8 1 1 'Llxf8 h4

12 'Lle6 h3 13 'Llg5 Drawn.
..

Let us look at some more

examples of a struggle with pawns
on different flanks.
Vasiukov-Timoschenko
Volgodonsk, 1 981

In this position White can quickly

convert his positional advantage to a
material one but playing this ending
is complicated due to the fact that
the pawns are situated close to one
another which might allow Black to
regroup ,his forces successfully for
defence. With fine play White pre
vents this possibility.

How can he realise the extra

pawn? Insufficent is the natural 1 5
'it>b7 'Lle8 1 6 'Lle4+ @e5 1 7 'Llxg5
@f4 1 8 'Llh7 'Lld6+! (for the present
the g4 pawn is untouchable because
of 1 8 ...@xg4 1 9 'Llf6+ 'Llxf6 20 c7,
and Black does not succeed in play
ing ... 'Lle8 since the pawn queens
with check) 1 9 'it>c7 'Llf7 (also poss
ible is 19 . . . 'Llb5+ 20 Wb6 'Lld6) 20
g5 @f5 ! =. On the g5 square the
pawn becomes vulnerable.
However E.Vasyukov finds an el
egant manoeuvre and wins the g5
pawn while the black knight oc
cupies the d5 square-which is
rather poor for the struggle against
the c-eawn.
15 t'll h3! 'Lld5+ 16 'it>b7 'it>e5 17
'Llxg5 @f4 1 8 tLlh7

Knight Endings 43

In this lies the main point of the

refined manoeuvre begun with the
move 1 5 ctJh3 ! . In contrast to the
position looked at in the previous
note, the pawn remains on g4 with
Black to move-and he is in zug
zwang. On 1 8 . . . e4 or 1 8 . .. 'it>e5
follows 19 g5 'it>f5 20 ctJf6 tt:le7 2 1
c7 'it>xg5 22 ctJd5 ctJf5 2 3 'it>c6! and
24 c8=iV. This same manoeuvre
wins on 1 8 ... tt:le7-19 c7 'it>xg4 20
tt:lf6+ 'it>g5 21 ctJd5 On 1 8 ... 'it>xg4
decisive is 1 9 ctJf6+ tt:lxf6 20 c7
tt:le8 2 1 c8='ii'+.
In the game followed . ..
1 8 tt:lc3 1 9 g 5 f5 20 c 7 and
...

Black resigned.

A very instructive ending to the

game.
Kochiev-Lerner
Beltsy, 1 981

In contrast to the previous

example Black's positional advan
tage is felt less in view of the more
active position of the white pieces.
With energetic play, 1 tt:le5+ 'it>c3 2
ctJI! or 1 b4 'it>c3 2 'it>e3, White
could create defensive chances.
However if he deprives his pieces of
activity, Black achieves a decisive
advantage.
1 tt:la7? b4 2 tt:lc6 c3 3 tt:la5
ctJd7

The main thing is to organise a

distant passed pawn and support its
advance. The white knight is in no
position to struggle against the
whole of Black's position.
4 e4 tt:lc5+ 5 f5 ctJxb3 6 ctJb7

Alas, on 6 tt:lc6 follows 6 ... ctJd4+.

6 ctJd4+ 7 'it>xg5 tt:le6+!
..

It is important to restrict the white

knight; now it is deprived of the
important c5 square.

8 'it>f6 b3 9 ctJd6 b2 10 'it>xe6

bl='ii' 1 1 \t>f6 'ii'b 6 White
resigned.

Let us look at another interesting

ending, demonstrating the import
ance of active defence.
Nikolaevsky-Gufeld
Kiev, 1 951

After 1. .. tt:la5 2 b4 cxb4 3 axb4

tt:lxc4 White gradually loses. How
ever analysis shows that he is not
exploiting his defensive possibi
lities. After 2 tt:lcl g6 the first im
pression is that White gets into
zugzwang since 3 e5 @xg5 4 'it>d5
loses because of 4 ... 'it>f4 ! 5 b4
'it>e3 ! ! , and the pawn cannot be
stopped. But nevertheless White
finds a saving resource:
3 tt:le2! ! tt:lxb3 4 tt:lg3+ @h4 5
ctJf5+!

44 Knight Endings

And Black is forced to reconcile

himself to a draw, 5 ... h5 6 ttJg3+,
since he would risk defeat after both
5 ... 'it>h3 6 lLie7, and 5 . . .gxf5 6 g6
ttJd4 7 'it>e5 ! ttJc6+ 8 @d6 ttJd8 9
'it>d7.
With passed pawns on opposite
flanks, as in pawn endings, the
strongest side might win even with
out the participation of the king. An
interesting example is from the
game ...
Azmaiparashvili-Novopashin
Volgodonsk, 1 981

A position which, despite its ap

parent simplicity, is quite insidious.
Black's pawn weaknesses require
due attention from him. The cardi
nal decision in the position is the
exchange of knights, which requires
accurate calculation. For example,
after l ...ttJe6+! 2 lLixe6 'it>xe6 3 g5
fxg5 4 hxg5 'it>d6 5 '\t>g3 e5 6 f4+
f5 6 'it>f3 d4 the chances are even.
In the game, however, there
followed ...
1. h6? 2 hS gxhS?
..

The only chance of saving the

game remains with 2 . 'it>f7 3 hxg6+
'it>xg6 4 f3 ttJd7 5 ttJf5 h5 6 ttJe7+
@fJ, and White' s advantage is not
so tangible.
3 ttJfS+ 'it>e6 4 gxhS lLie4 S f3
.

ttJc3 6 e3 ttJe2+ 7 'it>g4 es 8

ttJxh6 d4 9 ttJf7+ 'it>e6 10 ttJd8+
@d7 1 1 exd4! and White won.
Godena-Lalic
Portoroz, 1998

1 as f6 2 f4 \tg7 3 rs 'it>gs 4
ttJd3 ttJa6 S ttJeS ttJb4 6 f6 c4 7 h6,
and Black lost on time.
Van der Sterren-Douven
Netherlands, 1 985

After 1. ll'ld3 ! 2 b3 'it>dS the

black king cuts off the enemy knight
from the centre.
..

3 lll as 'it>e4 4 ttJc6 es s h4

White presents his opponent with

good chances. Correct was 5 b4 !
d5 6 b5 e4 7 ttJa7! c5 8 ttJc8

Knight Endings 45

@xb5 9 tlld6+ ..t>c5 1 0 tllxe4+ Wd4

1 1 tllg5 with a clear advantage.
5 .....t>d5

Barlov-AbramoviC
Yugoslavia, 1 989

A critical position has arisen.

6 tlla 5 e4 7 g3

Again better was 7 tll c4 tll c 1 8

tlld2, exchanging the b- for the
e-pawn.
7...@d4 8 tll c4 tll c l 9 tll a5 tlld3
1 0 @e2 tllc l+ 1 1 @d2 tll d3 12 @e2

And White reconciled himself to a

draw.

Polnareva-Akhsharumova
Moscow, 1 984

1. .. tlle7+ 2 Wf6 tllc8 3 Wg5 h6+!

4 c;t>h4 If 4 Wxh6 @g4. 4 ...c;t>e4 5
tll g7 @es 6 ..t>h5 @f6 7 @xh6 tlld 6
Drawn.

White is not able to break the

blockade surrounding him.
Vyzhmanavin-Chiburdanidze
USSR (ch), 1 984

At first sight it seems that a com

plicated and long struggle is in pros
pect, but Black, exploiting the bad
position of the white king, quickly
achieves victory.

1 tll e5 f5 2 tll d 7 @f'l 3 tll c 5 We7

4 tlla 6 c;t>d6 5 tllb4 e5 6 tll d3 @d5
7 tllb4+ c;t>e4 8 tlla 6 c;t>d3 9 tllb 4+
..t>e2 10 .!lJ c6 tllf3 1 1 .!lJ e 7 .!lJ e 1+ 12
..t>gl tll d3 13 tllx f5 tll xf2 14 @g2
e4 15 tlld4+ c;t>d3! and White
resigned.

In the next game, White has a ma

terial advantage but Black finds an
interesting possibility to save
himself.

White has the advantage, but the

limited amount of material left com
plicates the task. White's plan is
linked to a breakthrough with the
king to the queen's flank, since
there is nothing in 1 tllb6 tll a 5 2
@d4 because of 2...tll c6+ ! .
1 @e3 c;t>g6 2 @d2 @g7 3 Wc2
'iitg6 4 @b2 tlle7

46 Knight Endings

He cannot wait any longer, the c4

pawn cannot be held. To exploit the
remoteness of the white king, Black
strives to simplify the p osition.

5 lbb6 f5 6 exf6 'it>xf6 7 'llxc4

'llg6 8 'itc2 lDh4

In the pawn ending-8 . . . 'lle 5 9

'llxe5 'it>xe5-White gains victory
by 1 0 'it>d3 '\t>f4 1 1 c4 e5 1 2 c5 'it.?3
1 3 c6 e4+ 1 4 'it>d4 e3 1 5 c7 e2 1 6
c8='i e l ='fl' 1 7 'fl'f5+ Wg2 1 8
'iVe4+ 'ixe4 1 9 'it>xe4.
9 'it>d3 ctJf3 10 h3 'llg l 1 1 'll d2

'ite5 12 'it>e3 'llxh3 13 c4

With this move White rejected

Black's offer of a draw. In fact, if
13 . . . 'll f4 White retains the advan
tage by 14 tbf3+ or 14 Clie4. Also
other defences are no help.

27 ... 'ite5 28 'ite3!

The last finesse: if 28 'llx g5, then

28 .. 'llf2 ! 29 'itxf2 'itf4.
.

28...'it>d5 29 'll f6+ 'it>e5 30 'll h5

Black resigned.

He loses the knight, without man

aging to capture the g4 pawn in
return.
An instructive and complicated
knight ending, where the advantage
of one of the sides lies only in a
slightly better structure, is encoun
tered in the game . . .
Timm an-Ree
Netherlands, 1 984

13 'llg l 14 'll e4 'llh3 15 'll c5

'it>d6 16 ctJd3 e5 17 'ite4 'it>e6 1 8
'llxe5 'll t1 + 1 9 c;t>e3!
.

19 'it>d4 'it>d6 20 c5+ 'it>e6 21 c6

'it>d6 leads to a draw.

19 'lld l+ 20 'itd4 ctJt1 21 c5

ctJhl 22 c6 'itd6 23 c;t>e3 'llg3 24
'llc4 'it>xc6 25 tbd2 'itd5 26 'it>f3
ctJhl
..

1 . ..tbc5

The more obvious l ... e5 looks

stronger. For example: 2 'ite2 h6 3
tbe4 b6 4 'itd3 'itf7 5 'it>c4 'it>e6 6
'it>b5 'llb 8, and it is difficult for
White to obtain real chances of a
wm.
2 b4 'lla 6 3 a3 'll c7 4 'it>e2 h6

In this apparently arid desert a

win can still be achieved due to the
poor _eosition of the black knight.
27 tt:'le4

But not 27 'llfl because of

27 ... 'it>d4! 28 ctJh2 'it>d3 29 'it>g2
'ite3 30 'it>xhl @f2 with a draw.

His defensive problems are not

solved by 4 . . . 'llb 5 5 a4 Cll c 3+ 6 'it>d3
lllxa4 7 'llxe6 and Black has
difficulty involving his knight in the
future play.
5 Clle4 'it>f8

More circumspect was 5 ... b6 6

'lld6 a6.
6 ctJd6 b6 7 'itd3

Knight Endings 47

Accurate calculation was required

for the initiative-seeking 7 tl'ic8 ! .
For example: 7. . . tl:\b5 8 a4 tl:\c3 9
'it>d3 tl:\xa4 1 0 tl:\xa7 tl:\b2 1 1 'it>d4
e5+ 1 2 @e4 tl:\d l 1 3 f3 tl:\f2+ 1 4
@d5 tl:\d3 1 5 b 5 tl:\f4+ 1 6 'it>e4
tl:\xg2 1 7 tl:\c8, and White' s position
is close to a win.
7...a6 8 tl:\c4 tl:\d5 9 'it>d4 rJle7 1 0
g3 'it>d7 1 1 f4 'it>c6 12 'ite5

12... tl:\c7

Black chooses passive defence

and wrongly so. In his comments to
this game Timman wrote that
1 2 ...'it>b5 1 3 tl:\b2 tl:\xb4 14 axb4
'it>xb4 1 5 'it>xe6 a5 1 6 @fl a4 1 7
tl:\xa4 was losing. Considerably
stronger, however, is 16 ... 'it>b3 ! ,
after which 1 7 'it>xg7 'it>xb2 1 8 f5 a4
might lead to a queen ending with
an h-pawn, where White's chances
of winning are problematical.
13 tl:\d6 'it>d7 14 f5 exf5 15 tl:\xf5

With simple and convincing

moves White has succeeded in in
creasing his advantage and placing
Black in a zugzwang position.
15 ...tl:\es 16 g4 lt.Jf6 17 h3

17 lt.Jxh6 'it>c6 1 8 g5 tl:\d7 1 9 'it>e6

is more quickly decisive.

17... h5 18 g5 tl:\h7 19 h4 tl:\f8 20

tl:\xg7 tl:\g6 21 @f6

2 1 'it>d5 is also sufficient to win.

21 ... tl:\xh4 22 tl:\xh5 'it>c6 23 tl:\g3

rJld5 24 a4 b5 25 a5 'it>c4 26 tl:\f5
tl:\g2 27 'it>e5 Black resigned.

The difference between pawn and

knight endings is seen when there
are passed pawns on opposite
flanks. For example, the game. . .
Vukovic-Eingorn
Belgrade 1 987

White should continue 1 'it>g2! h4

2 tl:\d4 'it>e4 3 tl:\e6 'it>e5 4 tl:\g5, and
the outcome of the struggle is still
not quite clear, since it is difficult
for Black to advance both pawns to
the third rank. But in the game
followed...
1 tl:\d4 'it>e3! 2 tl:\ f5+ 'it>f2 3 lt.Jg3

After 3 tl:\h4 g3+ 4 'it>h 1 the

pawns are frozen but the poor posi
tion of the white king is decisive---4
... tl:\e3 5 b6 @fl 6 b7 lt.Jg4 7 b8='i
tt.Jf2 mate.
3...h4 4 tl:\e4+ 'it>e3 5 tl:\d6

On 5 tl:\g5 g3+, with the following

interesting variations:
(a) 6 'it>g2 (6 'it>h3 tl:\f4+ 7 'it>xh4
g2) 6 ... 'it>f4 7 tl:\h3+ rJlg4 8 'it>gl
tl'if4+ (8 ...tl:\e3+ 9 'it>hl h3? 1 0
tl:\xh3=) 9 'it>h l h 3 1 0 b6 h 2 1 1 b7
tl:\h3, and mate in two moves.
(b) 6 'it>g 1 'it>f4 7 tl:\h3+ rJlg4 8
'it>g2 tl:\e3+ 9 'it>gl 'it>xh3 1 0 b6 tl:\g4

48 Knight Endings

1 1 b7 llle 5 1 2 b8= Clif3+ 1 3 fl

g2+ 14 e2 gl= 1 5 xf3 'Yi'g3+.
5 ... g3+ 6 'itgl
Forced. If 6 g2 Cll f4 7 g 1 then
7 ...f3 is decisive.
6... h3 7 CZlfS+ f4 8 CZld4 lt>g4 9
lll e2
The threat was 9 . . . h2 1 0 h 1
'it>h3. Now, however, on this could
follow 1 1 Cllxg3 'iii>xg3 12 b6 with a
draw. In order to achieve victory
Black must broaden the range of ac
tivity for his knight.
9 ... Cllb 6 10 fl lll c4 11 gl
Cll d2
White resigned, without waiting
for Black to promote to a queen in
the variation 1 2 b6 Clif3+ 1 3 lt>fl
g2+ 14 lt>f2 g l='Yi'+ 1 5 Cllxg l h2.
The comparison between knight
and pawn endings is borne out with
the exploitation of zugzwang in the
two following endgames.
Sajtar-Benko
Budapest, 1 954

If in this position we exchange

knights for bishops or rooks then the
game would be drawn. It is interest
ing that Reshevsky managed to win
a queen ending with such an align
ment of forces against Geller, Inter
zonal tournament, Sousse 1 967,
and, with c-pawns, Mikhalchishin
Kasparov, USSR (ch) 1 978.

Of course, according to Botvin

nik, playing a knight endgame, is
like playing a pawn ending. White's
plan consists of advancing the first
pawn with help of the king, after
which Black gradually lands in
zugzwang.
1 Cll f4 CZld6 2 g4+ 'it> gS 3 Cllh3+
g6
If 3 . .. lt>h4, then 4 'it>f4 with the
irresistible threat of 5 Cll f2 and 6 g3
mate.
4 g3 lll c4
Attempting to create counterplay
by means of an attack on the pawn.
5 lt>h4 lll e3 6 Cll f4+ lt>h6 7 gS+
'it> g7 8 'it>hS CZlfS 9 tiJdS!
Inferior was 9 g6 in view of
9 ... Cllg 3+! 10 lt>g5 Cll e4+ 1 1 \t>f5
Cll g3+, preventing White attacking
the coordinated forces.
9 Cll g3+
On 9 . . .lll d6 or 9 ... Clld4 would
have followed 10 Cll e3 ! , and Black
is forced into a worse position.
10 'it> g4 Cll fl
If 1 O . . .Cll e4 1 1 f4 Cll d6 decisive
is 12 Cll e7! .
1 1 'it>f4!
The main principle of the end
game is not to rush! White restricts
the activity of the black knight,
which, in order to get into play, has
to go to d2 or h2, and then the white
king finds itself in a very favourable
position-safe from the black
knight along the diagonal.
11 ...Cll d2 12 Clle3!
Still more restriction of the black
knight.
12 ..@g6 13 CZlfS Cllfl 14 Cll h4+
'it> g7 15 g4! 'it>h7
15 ... Cll e3+ 16 'it>h5.
1 6 g6+ 'ith 6
1 6 .. .'it>g7 1 7 g5.
17 'iitf4 !
Zugzwang in action.
11 ... lll d2 1s @rs lt>g7
..

Knight Endings 49

1 8 ... 'Lic4 1 9 'it>f6.

19 g4 !

Zugzwang in tum. It should be

mentioned that White does not
hurry with the advance of his re
serve pawn, since this cuts off
squares from the king.
19...'it>g8

On a move of the knight follows

20 'it>g5, while on 1 9 ...'it>h6-20
'Lig2 'it>g7 2 1 'Lif4.
20 'it>g5 'Lie4+ 21 'it>f4 !

2 1 'it>h6 'Lif6.

2 1 ...'Lic5 22 'Lif5

Step by step White creates a very

strong position, and Black must
continually watch the threat of a
king infiltration to f6 or h6
22 ...'Lid3+ 23 'it>g5 'Lie5

Black continually attacks the

pawn and accuracy is required from
White. Thus there is nothing in 24
g7 'it>h7, after which the coordin
ation of the white pieces is
destroyed.
24 'Lig3!

The knight crosses to h5, after

which the pawn pushes on to g7.
24 ... 'Lic4 25 ctJh5 'Lie3

25 ... 'Lid6 26 'it>h6 'Lie4 27 g5;

25 . . . 'Lie5 26 'it>f5 and 27 g7.

29... 'Lic4

In this case White wins in the fol

lowin g instructive way.
30 @f6 @g8

30 ... 'Lid6 3 1 We7 and then 32

@f8.
31 'it>g6 'Lie5+ 32 @f5 'Lic4 33
@f6

The familiar 'triangulation'.

33 ...'it>h7 34 'Lig3!

A new reconstruction. Now bad is

34 ...'it>g8 because of 35 'Lif5 and 36
'Lie7, and also 34 . . . 'Lie3 because of
35 'it>f7
34... 'Lid6 35 ctJf5 'Lie4+

35 ... 'Lie8+ 36 Wf7.

36 @f7 'Lig5 37 'it>e7! 'Lie4

37 ... 'it>g8 38 'it>f6 'Lie4+ 39 <tig6 .

38 Wf8 'Lif6 39 'it>f7 'Lig8

Black is on the final frontier.

40 g5 'Lih6! 41 Wf8 'Lig8 42 g6+

And White wins.

It is interesting that precisely the

same endgame was met in the fol
lowing game.
Matulovic-Uitumen
Palma de Mal/area, 1970

26 g7 'it>h7

26 ... 'Lid5 27 'it>h6; 26 ... @7 27

Wh6 'Lixg4 28 'it>h7.

27 'it>f4 'Lidl 28 'it>f5 'Lie3+ 29

'it>g5

White hands over the move to his

opponent and at once places him in
zugzwang. Black, apparently tired
from a difficult defence, here made
a mistake by playing 29 ... 'it>g8?, and
after 30 'it>g6 resigned, since on
30 . . . 'Lixg4 follows 3 1 'Lif6 'Lixf6 32
'it>xf6 'it>h7 33 @7.
There was a chance for him to put
up more stubborn resistance by...

This ending also ended in victory

for the stronger side and we present
it without commentary since here

50 Knight Endings

the same idea was utilised-only

White probably defended in weaker
fashion.

1 ...ClJb6 2 CZJc6 ClJd7 3 0,e7 ClJf6+

4 c;f;>h4 gs+ s '\i>g3 g6 6 CZJc6 WhS 7
llJd4 llle4+ 8 Wf3 tLld6 9 c;f;>g2 llJfS
1 0 CZJe6 g4 1 1 ClJf4+ WgS 12 CZJe6+
c;f;>h4 13 tl:lf4 CZJe7 14 CZJe2 llJdS lS
Wh2 gS 16 CZJg3 ClJe3 17 CZJe4 ClJfS
1 8 c;f;>g2 g3 1 9 0,f6 CZJe3 20 c;f;>gl
CZJg4 21 llJh7 CZJes 22 'ii>g2 QJd3 23
@gl CZJf4 24 Whl g2+ 2S c;f;>h2 g4
26 llJf6 g3+ 27 c;f;>gl c;f;>h3 White
resigned.

What else can he do? He must

give up a pawn-though it was still
possible to attempt a breakthrough
with the king to the d8 square, but
then, with the king on e7, the knight
from e5 goes to d3, and White is
again in a blind alley.

6...@xe7 7 @cs @f6 8 c;f;>dS llJd3

9 CZJc6 ClJel 10 'itie4 ClJg2 1 1 ClJd4
CZJe3 12 ClJfS CZJc4 13 c;f;>d3 CZJeS+ 14
'itie2 c;f;>g6 lS 'ii>f2 CZJc4 16 Wg2 c;f;>f6
17 h3 tl:leS 1 8 ClJd6 '\i>g6 19 CZJe4
c;f;>h6 20 'ii>f2 'ii>g6 21 'ii>e2 Wh6 22
CZJcs 'ii>g6?

It is interesting that in both

examples the stronger side had
doubled pawns on the g-file. But
what will be the case if the pawns
are situated on another file? To us it
seems that the weaker side will
make a draw only with pawns on
the edge file, since then the king of
the stronger side has no exit.
Van Wely-Adams
Groningen, 1 997

Black must have the possibility,

on CZJd3, to reply.... ClJg6, defending
the f4 pawn.
23 h4?

Correct was 23 CZJd3 ! ClJc4 24 h4,

obtaining a winning position. The
transposition of moves gives Black
the possibility of saving himself.

23 gxh4 24 CZJe6 h3! 2S CZJxf4+

@gs 26 CZJxh3+ 'it>h4 27 llJf2 c;f;>g3!
.

1 ClJd4?

Correct was the transfer to a pawn

ending by 1 llJd6 ! '\i>e6 2 ClJe4 tl:lxe4
3 fxe4 xe5 4 @f3 Wf6 5 h4 ! gxh4
6 'it>xf4, and then the king travels to
.
h3.

1 ...ClJd7 2 e6 ClJeS 3 'ii>d2 'it>d6 4

'ii>c3 'ii>e7 S 'it>b4 Wd6 6 e7

The Black king breaks through to

the white pawns and a draw is
inevitable.
28 gS CZJxf3 29 g6 ClJd4+ 30 c;f;>d3
CZJe6 31 We3 Drawn.

An interesting ending arose in the

following game where despite, the
approximate equality, the struggle
was still not over.

Knight Endings 51
lvanchuk-Eingorn
USSR (ch), 1 988

,_._.,, 0
.
-

%..n
m
!
o
,,, .v,

.
.
,,.,,v, .

.0 ,,,,,v, .0
.

1 f3 h4 2 b5 a5 3 'llg l d6

More active is 3 ...@f4 provoking

_elay such as 4 @d5 'llb6+ 5 @c6
tLla4 6 b6 ctJxb6 7 c,t>xb6 @g3 8
@xa5 c,t>g2 9 'll e2 @xf3 1 0 tllg 1 +
g2 1 1 'lle2 with a draw.

4 d4 tll c 5 5 ctJh3 ctJe6+ 6 @c4

@c7 7 @dS @b6 8 @d6 @xb5 9 f4
a4 1 0 c,t>e7 xa3 1 1 @xf7 'll d4

Foreseeing 1 2 @xg6 a4 1 3 f5
ctJxf5 14 @xf5 @b2 1 5 g6 a3 1 6 g7
a2 1 7 g8='fl' al ='iii', the rivals agreed
to a draw.
Sveshnikov-Sokolov
Moscow, 1 991

exploitation of this advantage

requires filigree technique.
1 f3 'll d7

Neither now, nor later is there any

_eossibility of l ...c5 since then 2
tbb5 a6 3 ctJd6 @e7 4 ctJb7 'lld 7,
and the white king goes to the
centre followed by a3 and b4.
2 f2 f5 3 @e3 e7 4 b4 e5 5 a4
@d6 6 @d3 tll f6 7 c5+ We6

Better is 7 . . . c7 8 Wc4 a6.

8 b5! d7 9 c4 Wc7

Bad is 9 .. e4 1 0 @d4.
10 a5 a6!

The threat was a6, bxc6 and ctJb5 .

1 1 b6+

After 1 1 bxa6 follows 1 l ...@b8

12 ctJa2 ct:Jd5 and 1 3 . . . Wa 7.
1 1 ...@b7 12 g3 h5?

He should not freeze his structure.

Correct is 1 2 ... 'lld 7.
13 h4! ctJd7 14 f4!

Closing the way for the king.

1 4...exf4 15 gxf4 tllf8 16 'lle 2

'llg6 17 'll d4 tllxh4
After 17 ... 'llxf4 18 tllxf5 g6 19

'lle7 Black is in zugzwang.

18 'll e 6! c,t>cs

To defend against 'lld 8.

19 tllxg7 ct:Jg6 20 tllx h5 Black

resigned.
Torre-Portisch
Toluca, 1 982

White has the advantage on ac

count of his majority on the queen
side and greater space, but the

52 Knight Endings

The endgame appears very diffi

cult for White but, with a pawn
sacrifice, he activates his king.

Ilincic-Abramovic
Tivat, 1 995

1 b6! axb6 2 ttJc4 bS 3 ttJe3+

@cs 4 fS !

Weaker is 4 ttJg4 ttJe6 5 e3

ttJc7 ! 6 ttJf6 ttJd5, and the pawn
ending is hopeless for White. With
the sacrifice of yet another pawn
White creates his own passed pawn,
which is a principal factor in this
endgame.
4 ...gxrs s h4 f4 6 lLig4 rs 7 ttJh2

lLic6 8 hS ttJeS+ 9 e2 b4 10 h6 b3
11 h7 lLig6 12 ttJf3 b2 13 ttJd2 @d4
14 @f3 c3 lS ttJbl+

Now Black cannot win the knight

on b l because after h7-h8, ttJxh8,
the king captures both black pawns.
He has to try and go with the king to
the h7 pawn.

1S ...'it>d3 16 'it>t1 'it>e4 17 ttJd2+

'it>dS 18 'it>e2 'it>e6 19 d3 f6 20
'it>c2 'it>g7 21 ttJf3 xh7 22 'it>xb2
'it>h6

It looks like White has some prob

lems converting his passed d-pawn
to a win, but really it is very simple.
1. ..g6 2 @d4 'it>f7 3 ttJcS!

Speculating on a transfer to a win

ning pawn endgame, White clears
the way for his pawn and king.

3 ttJb6 4 d6 '\t>f6 S d7 e7 6
eS hS
..

Zugzwang-after 6 . . .ttJc4+ 7 d5
ttJa5 (preventing 'it>c6) 8 g4 g5 9
'\ties the white king penetrates one
flank or another.
7 'it>f4 ttJc4 8 'it>gS ttJeS 9 h4!

And because of zugzwang White

wins yet another pawn. Black
resigned.

Maric-Zaitseva
Tivat, 1995

It is interesting to compare this

ending with the two following ones,
where doubled pawns were success
fully realised.

23 'it>c2 'it>hS 23 'it>d2 'it>g4 24

ttJd4 lLieS 26 e2 ttJf7 27 'it>fl ttJgS
26 t1 ttJe4+ 29 'it>g2 ttJd2

And Black did not manage to real

ise his material advantage.

Knight Endings 53

In many cases 4:3 on one flank is

winning for the stronger side, but
here it isn't so clear.

Usually such endgames with a

distant passed pawn give great win
ning chances.

More or less necessary-White's

plan was 'Lld4, g2-g4 and f4-f5
creating a weakness on e6.

Black has a compact pawn struc

ture and White tries to create some
weaknesses so he can penetrate with
his king, taking advantage of the
fact that Black must spend time to
win the a-pawn.

1 'it>f3 f6

2 'it>e3

Playable was 2 ctJd4.

2 ...'LlbS 3 g3 'it>e8 4 'it>d3 'it>d7 S

'Llb4 fxeS?

Clearly better would be 5 ... 'Lla7 6

'it>e3 ctJc6 7 ctJd3 'it>e7 with the idea
8 ... g5 and transfer of the king via
f7-g6 to f5.

1 ctJeS

1...f6 2 'Llc4 'it>b8 3 @e2 'it>a7 4

@e3 'it>xa6 S 'it>e4 tf:ic7 6 'Lle3 'it>bS
7 ctJdS

6 fxeS 'it>c7 7 <;;>e3 'Lla7 8 ctJd3

'Llc6 9 h4 g6 10 'LlcS ctJd8

Otherwise after h4-h5 Black's

structure would be completely
blocked.

11 g4 h6 12 gS hxgS 13 hxgS
'it>c8 14 @d4 <;;>c7 lS 'it>d3 b6 16
'Lld7+ c6 17 'Llf8 cs 18 'Lld7+!

Rather dangerous was 1 8 'Llxg6

ctJf7 when the e5 pawn is quite
weak.
18 ... <;;>bs 19 'Llf8 'it>b4 20 'Llxg6!

Now is the time.

20...'Llf7 21 'Llf4 'LlxgS 22 'LlxdS!

Draw.

Alexandria-Marie
Tivat, 1995

7...'Lle6

The pawn endgame after 7 . . .c6

8 'Llxc7 'it>xc7 9 @d5 d7 10 f4
<;;>e7 1 1 f5 would be too dangerous
for Black.
8 ctJxf6!

White has no other way of playing

for the win.
8 ... gxf6 9 @rs 'Llcs 1 o f4

The direct ap12roach was 1 0 xf6

'Lle4+ 1 1 <;;>g7 lll xf2 12 xh7 <;;>c5
13 h4 <;;>d5 14 h5 e5 15 g6 ( 1 5
h6 'Llg4 1 6...'Llxh6=) 1 5 ...'Llg4 1 6
'it>g5 lll f6 1 7 h6 @e6 1 8 g6 <;;>e7
1 9 h7 ! gave White winning chances.
But better would be 14 ...'it>e4 1 5
'it>g6 'Llg4 1 6 'it>g5 'Lle5 17 h 6 ct:Jf7+
and l 8 ... ctJxh6=.
10 'Lld3 1 1 g3 'it>c6 12 'it>xf6
'it>d6 13 rs
..

Or 1 3 'it>g7 'it>e6 1 4 'it>xh7 'it>f5 1 5

h4 ! f6! 1 6 h5 ctJf2 ! 1 7 h6 ct:Jg4=.

54 Knight Endings

13 ...lLieS 14 h3

After 1 4 @g7 h5 1 5 'it>h6 'it>e7 1 6

'it>xh5 @f6 1 7 g4 'Llf3 1 8 h4 lLie5
the draw is obvious.
14 lLid7+ 15 'it>gs @es 16 g4
lLif6 17 h4
..

Or 17 'it>h6 Wf4 1 8 @g7 @e5 1 9

@ fl h 5 2 0 g 5 lLie4 2 1 g 6 @xf5 22
g7 lLif6=.
17 h6+! 18 @xh6 lLixg4+ 19
'it>g5 lLif6 20 @g6 lLid5 Drawn.
..

Sermek-Hulak
Slovenia, I 995

9 ...@c6?

The only chance was 9 ... g5 ! 10 g4

( 10 'it>g6 g4 1 1 Wg5 'it>c6 1 2 xg4
'it>d6=) 1 0 . . . 'it>c6 1 1 @g6 'it>d6 1 2
@xg5 'it>e7 1 3 'it>g6 'Llc3 1 4 @h7
(14 f4 lLid5 15 f5 @f8=) 14 ... @f6 1 5
f4 lLie4 1 6 'iith 6 'it>fl with a probable
draw. After the move in the game
White's task is easy.
10 @xg6 'it>d7 11 g4 'it>e6 12 g5
lLid6 13 h7 lLie4 14 g6 lLif6+ 1 5
@h8 rl;e7 16 f4 @f8 17 g7+ @ f7 18
g4 lLig8 19 g5 Black resigned.

This is a very similar endgame to

the previous one but here Black's
pawn structure is weaker.
If now 1 ... lLi d5 2 lLid l with the
idea 2 . . 'it>xa4? 3 lLic3 + -.

Beliavsky-Tratar
Bled, 1 996

1 ...@b3 2 a5 !

And here White is forced to sacri

fice a piece-if 2 lLid l then
2 .. 'Llxa4 3 @e3 lLic5 4 'it>d4 'it>b4.
.

2 ... @xb2 3 a6 lLib5 4 @e3 Wc3 5

'it>xe4 @b4

If the black king tries to go for the

white pawns by 5 ... @d2 then White
creates second passed pawn by 6
g4 ! We2 7 f4 @f2 8 f5 + -.
6 We5 Wa5 7 @f6 @xa6 8 @g7
.
@b6

Or 8 . . lLic3 9 'it>xh7 lLie4 1 0 f4

lLixg3 ,1 1 Wxg6 + -.
.

9 @xh7

White's king is much more active

than Black's and this gives him
chances to make progress.

1 @c6 We7 2 'iitc7 lLia6+ 3 @b6

lLic5 4 'it>c6! lLib3 5 lLixa4 lLid4+ 6
'it>b7! lLib5 7 lLib6 lLic3?

Knight Endings 55

The way to the draw wasn't easy:

7 .. .ltJxa3 8 Wc6 ltJbl 9 ltJc8+ 'iiff6
IO ltJxd6 ltJc3 ( 1 0... ltJd2 1 1 'it>d7
'itig5 12 'it'e6 'it'f4 1 3 'it'f6 ltJb3 1 4
ltJb7! 'it'xe4 1 5 d 6 ltJd4 1 6 d 7 ltJc6
17 ltJa5 !+-) 1 1 'it'd7 'it>g5 12 'it>e6
\t>f4 13 @f6 ltJa4 14 ltJb5 'it'xe4 1 5
d 6 lt)b6 1 6 ltJc3+ 'it'd4 1 7 ltJa4
ltJd7+ 1 8 'it>e7 ltJb8=.
8 a4 ltJxe4 9 a5 ltJc5+ 10 'it'c7

Wrong was 1 0 'it>c6? e4 1 1 ltJc4

lit'd8! 12 'it>d6 e 3 ! 1 3 ltJxe3 ltJb7+=.
IO ltJa6+
..

After 1 0... e4 1 1 ltJc4 @f6 1 2

'it>xd6 e 3 ( 1 2 ... ltJb7+ 1 3 'it>c7 ltJxa5
14 d6! ) 1 3 ltJxe3 ltJb7+ 14 'it>c7
ltJxa5 1 5 d6 'it'e6 1 6 d7 ltJb 7 1 7
ltJd5.
11 @c8!

Once again 1 1 'it>c6? was wrong,

because of 1 1 .. .. e4 12 ltJc4 ltJb4+=;
1 1 Wb7 ltJc5+=.
l l ...e4

Or 1 L.ltJb4 12 'it'b7.
12 ltJc4 ltJb4

Or 1 2 ...Wf6 1 3 'it>d7.

13 'it>b7 ltJxd5 14 a6 'it>d7 15 a7

ltJc7 16 ltJb6+ 'it'd8 17 ltJd5 ltJa8
18 'it'b8! Wd7 19 ltJf6+ Wd8 20
lbxe4 d5 21 t'llc 5 d4 22 'it>b7 d3 23
lbxd3 'it'd7 24 lbc5+ Wd6 25 lba4 !
t'llc 7 26 ltJb6 Black resigned
Zotkin-Kudrin
Moscow, 1 965

Black's task is to create yet

another passed pawn. And so there
followed...
1 ...fxg4 2 hxg4 h5

Now White has no time to win the

d3 pawn.
3 g5+ @g7!

The king must retreat, otherwise

the white knight becomes highly active: 3 ....'it>e6 4 ltJd4+ 'it>d5 5 ltJf3 .
On 5 ... lt'c4 6 f5 gxf5 7 g6 'it'c3
p ossible is 8 'it>f4!. If 3 ...@f5, then 4
tlJe5.
4 ltJd4 h4
Necessary in view of the threat of
5 ltJf3.
s rs h3

After this White himelf obtains a

protected passed pawn.
6 f6+ 'it>f7

If 6 . . .@f8, then 7 lbf3 lbe4 8

'it'xd3 lbxg5 9 tlJh2 'it'f7 1 0 'it>e3
@xf6 1 1 Wf4, and Black is obliged
to let go of the h3 pawn.
7 ltJtJ 'it>e6 8 'it>d2 @rs

The king hurries to support the

h-pawn, while the f-pawn will be
watched by the knight.
9 f7 ltJd7

Of course not 9 ...t'lle 6 in view of

1 0 lbd4+.
1 0 Wxd3 Wf4 11 'it'e2 'it>g3

It seems it's all over. The knight

f3 is doomed, whereas the white
king is too far from the g6 pawn.
But White finds a study-like idea.
12 ltJe5 ltJf8 13 tlJd3 h2 14 lbf2
'it>g2 1 5 ltJbl

In this lies White's idea. After a

few moves the players agreed a
draw. The question arises whether
Black could have won. Instead of
5 ... h3, played in the game, he had at
his disposal the more effective move
5 . .d2 ! , pointed out by I.Zaitsev. Af
ter 6 f6+ 'it'f7 7 'it'e2 (7 'it>xd2
ltJb3+!) 7 ... h3 8 lbf3 ltJ e4 9 'it>d l
decisive is 9 . lbf2+ 1 0 Wxd2 t'llg 4.
.

..

3 Rook Endings
Rook and pawn against rook

The most classical and primitive

(though not for everyone) endgame
which should not, it seems, present
any particular problem for players
of grandmaster rank. But John Nunn
wrote an interesting book about
these endings on the basis of com
puter analysis where he gave quite a
few complicated positions. However
in everyday practice it is much more
simple and tragic (or more confus
ing). Knowledge of precise posi
tions and methods of defence here
have exceptional significance.
The edge pawns

The most frequent case-rook and

pawn against rook.
People have already programmed
computers for this type of ending,
and so for the right method of play
we should now tum to the Endgame
CD !
Here are a couple of characteristic
examples.

Emms-Riemersma
Gausdal, 1 993

1.. ..l:I al?

Correct is 1 . ..a3 ! 2 g2 :a2+ 3
fl a4 4 J::rh 3+ 'it>c2 5 1::rh2+ b l 6
lih3 .l:Ic2 7 c,t>e 1 lib2 8 .l:Ia3 11b4 9
'it>d 1 'it>b2 1 0 lid3 a3 1 1 lid2+ c,t>b 1 ,
and the a3 pawn cannot be stopped.
2 c,t>g2?
But now White misses a draw by
2 g5 ! a4 3 J::rg3 c2 4 .llg2 'it>d3 5
.l:Ig3 'it>e4 6 .l:Ig4 f5 7 l:tb4 e5 8
Wg2 'it>d5 9 lih4 a3 1 0 l:!.h3 ! when
White reaches a well-known
position.
2 a4 3 @f2 a3 4 .l:Ib5+ 'it>a2
...

Rook l:!,'ndings

There was a quicker win by

4 ...c;t>a4!
5 @e2 .l:Ibl ! 6 .l:Id5 @b2 7 l:td2+

57

Tosic-Gyimesi
Yugoslavia, 1 998

c;t>b3 8 .!id3+ c,t>a4 9 .!id4+ l:tb4 1 0

lld8 a2 7 'it>d3 @b3 White
resigned.
Bagirov-Kraidman
Grieskirchen, 1 998

l .l:Ia5?

Simply driving back the king first

by 1 l:tc5+ @b2 2 l:tc8 l:!.xh4 3 'it>d2
draws.
1. ..J:ixh4 2 .l:i.a8

2 l:!.a7 is also enough for a draw.

2 .l:Ib4 3 'it>dl?
..

1 ...:ic5?

Correct is 1 . l:!.g5, holding the

draw as in the previous example.
..

2 'it>d4 J:icl 3 .l:If6+?

Simpler is 3 a5 winning easily.

3 ...'it>g5 4 .l:Ib6 .li(al 4 .!la6

4 .!lb4 led to a draw after 4 . . . 'it>f5,

and the king rushes to c8.

The decisive mistake. 3 .!lc8+!

@b2 4 @d2! .l:Id4+ (4 ... a3 5 l:tc2+! )
5 'iite3 .!ld5 6 .l:i.b8+! saves him.
3 ... 'it>b2 4 d8 a3 5 .!id2+ @bl

and White had to resign.

Herrera-Vasquez
Cuba, 1 998

4...@f5?

Again he should go for the draw

ing mechanism 4 ... l:tfl ! ? and check
on the f-file.
5 a5 .!idl + 6 'it>c5 l:ral 7 'it>b5

@e5

On 7 ....l:i.b l + there is 8 @c6 l:tal 9

'it>b6 .!lb l+ 1 0 'it>c7 l:ta 1 1 1 .!la8
@e5 1 2 a6 'it>d5 1 3 .!ld8+ @c5 ! 1 4
.!ld6! , winning.

8 .!ih6! @d5 9 'it>b6 .!lbl+ 1 0 @c7

.!lcl + 1 1 @b7 l:tbl + 12 .!lb6 .l:Ihl
13 .l:i.c6 .!ih7+ 14 .!lc7 I:l.hl 15 a6
l::tb l+ 16 @c8 'it>d6 17 a7 Black
resigned.

1 ..Jlal?

He can achieve a draw by any

move except this, including even
1 . . Jle7, but best of all is 1 .. ..!le8 ! .
2 a4 'it>f6 3 @d6, Black resigned.

58 Rook Endings

Vyzhmanavin-Lerner
USSR (ch), 1 984

Simferopol, 1 988, where the same

mistake was repeated! 2 .l:Ic6 ! 'it>g5
3 l:!'.c8 ! .l:Ih7 4 .l:Ig8+ leads to a draw.
It is nearly always necessary to at
tack the king from behind, and not
in front, because in the end the king
will get the better of the rook.
2 l:l'.h7! 3 .l:Iel
..

In the above-mentioned game fol

lowed 3 l:l'.h3 'it>g5 4 e2 'it>g4 5
.!lhl h3, winning because of the
poor position of the rook.
3 h3 4 .!lhl 'it>g5 5 'it>e3 'it>g4 6
..

1 'it>d2?

There was a simple draw by 1

b2 .l:Ig3 2 c2 c,t>d5 3 'it>d2 .!la3 4
'it>e2 etc.
1 ...h2! 2 c,t>e2 l:.al ! and White
resigned.

@f2
If 6 1:1'.g l + 'it>h4 7 l:1h l , then
7 ... l:1a7 followed by 8 . . . .!la2, 9 . . h2
and then 10 . . 'it>h3.
.

6 l:1f7+
resigned.
..

7 c,t>gl

l:1a7

White

Since after 8 .!lh2 there follows

not 8 .. c,t>g3? 9 .l:Ig2+ ! !=, but
8 ..lial+ 9 @f2 l:tb l ! with a decisive
zugzwang.
.

Vladimirov-Rashkovsky
Chelyabinsk, 1 975

Novikov-Lalic
Manila, 1992

1 l:1c3?

Correct was 1 .llg 8+ f5 2 l:tf8+

Wg4 3 l:!.g8+ 'it>h3 4 .l:Ig5 and after
the advance of the h-pawn a draw is
It seems that it is not easy for
achieved since the white king is cut Black to defend himself, but he
off only by three files.
finds his only saving resource:
1 ...h4 2 .!le3??

1.. .l:Ih8! 2 'it>b7 dS 3 l:1g4

Interestingly, this mistake is typi

After 3 a6 c5 4 l:tbl l:1h7 the
cal for grandmasters. A mirror im draw is inevitable.
age of this position was encountered
3 'it>cs 4 l:1g7 .l:Ih6! 5 a6 l:!'.b6+ 6
in the game Dvoiris-Kovalev, c,t>a7 .l:Ibl 7 llb7 :!al Drawn.
..

Rook Endings 59
Kamsky-Karpov
Linares, 1 994

champion does not 'go for' for the

side.
1 Wf4?

Correct was I @f6, threatening a

check from g4. If now l . ..a3, then 2
.l:Ig4+ Wd5 3 .li(g5 ! with a drawn
pawn ending.
1. .. l:!e8! 2 'it>f3 'it>d3 3 .l:Ib2 .!if8+
Cutting off the king looks a
deadly blow for White.
4 'it>g3 Wc3 5 .l:!b7 .l:Ifl 6 .l:Ib8

After 6 'it>g2 .l:I a 1 the position is

lost.
6 ... .l:Ial 7 @f3 .!lxa2 8 We3 .l:!h2 9

l:1c8+ 'iii>b2 10 .!lb8+ 'it>cl !

What system of defence should he

Clearly not 1 0 ..'it>al ? because of
choose: wait for the approach of the 1 1 a8 .l:Ih3+ 12 'it>d2 a3 13 Wc l
black king to the a4 pawn or adopt with a draw.
the more active method. It is clear
1 1 .l:Ic8+ 'it>bl 12 .l:Ib8+ .l:Ib2 13
that the second solution is more ef .!la8 .l:Ib3+ 14 'it>d4 a3 1 5 'it>c4 Wb2
fective, but it requires accurate play.
And in this theoretical position
.

1 .l:If6+! 'it>e4 2 .!lg6! 'it>e3 3 .l:Ig4!

The main thing is to maintain con

tact with the opponent's pawns.

3 ... a3 4 .l:Ig3+ 'it>e4 5 .l:Ib3 l:!a2+ 6

Wg3

White resigned.

Rook endings
Several pawns on one flank

This type of ending is met quite

Obviously it is better not to go to
often in practice. With various posi
the first rank.
6 ... 'it>d4 7 f3 l:1al 8 'iii>g2
tions of the pawns on both sides in
And the opponents agreed to a the majority of cases the weaker
draw.
side should hold a draw even in
those
instances where his pawn
Euwe-Alekhine
structure is weakened.
World Championship (m), 1 93 7
Beliavsky-Spraggett
Elista (ol), 1 998

White is i n a dilemma, whether to

go forward or back, and the world

60 Rook Endings

1 h4?

Correct was 1 g4 hxg4 2 hxg4 and

then 3 g5, shutting in the black king,
whereas 1 f3 was also a possibility
with the future transfer to another
type of classical position, looked at
in the next section.

1 6...'it;>g2 17 e2 h3 Drawn.
Bagirov-Berzinsh
Riga, 1 998

1. ..'it;>g7 2 1Ig5 'it>h6 3 g4

Now the plan with f3 promises

less since it is not clear how Black's
position can be breached if White,
after an exchange on f3, puts his
own pawn on e5.

3 ...hxg4 4 .l:!xg4 l:!e5 5 @fl 'it>h5

6 1If4 1Ie6 7 e2 .l:te5 8 'it>el !
White gains a tempo-on 8 'it> d2
there is 8 ...l:rf5 .
8... @h6! !

Black has calculated the pawn

ending clearly.
9 @d2 .l:If5 ! 10 'it;>c3

Black demonstrates a clear and

instructive way to realise his
advantage.
1 . .. J:id5 ! 2 f2 .l:Ig5 3 f3 f5

Rather better was 1 0 .l:Ixe4 .l:Ixf2+

An important link in the plan:
1 1 'it>c3, but even here there are no now White's pawn fortifications are
quickly smashed.
real chances of a win.

10 ...1Ixf4 1 1 exf4 'it;>h5 12 'it>d4

'it;>xh4 13 'it;>xe4

4 gxf5 exf5 5 'it>f4 J:ixh5 6 1lb7

1Ig5 7 .l:Ia7 'it>h7 8 J:ib7 'it>g6 9
llb6+ 'it>f'7 10 1Ia6 g6 1 1 'it;>e5 J:ig3
12 .:i.a7+ e8 13 'it;>f6 h5 14 l:!.h7

Also there is no salvation in 14 e4

fxe4 1 5 1Ie7+ 'it>f8 1 6 lhe4 g5 .

14 d8 15 :Ia7 h4 16 1Ih7 h3
17 1Ia7 h2 18 1Ih7 1Ig2 1 9 'it>e5 g5!
20 d6 'it>c8 White resigned.
..

Madsen-Hansen
corr, 1 974

On 1 3 'it>e5 correct is 1 3 ... 'it;>h3 ! .

1 3 g4 14 'it>e3
..

Clearly not 14 'it;>e5 f3 .

14 h3 ! !
..

Only this leads t o a draw.

14 ... 'it;>h4? is no use due to 1 5 f5 ! .
15 ,'it>f3 'it;>h2! 1 6 'it>e3

After 1 6 'it>g4 'it>g2 1 7 g5 there

is the only move 1 7 . . @f3 ! ! .
.

Rook Endings 61

The position to all intents and pur

1 h4!
poses concurs with the previous
He cannot allow Black the possi
one, and here 1 .!le4 l:tc 1 2 'it>g4 bility of playing . . . g6-g5.
1. .l:!b8
should be played. But White
After l .. .e5 follows 2 'iS?g2 l:rc2 3
cho oses another system of defence.
1 l:l.'.b4 J:icl 2 l:ra4 @ g7 3 l:tb4 g4 J:ic l 4 g5+ 'it>f5 5 l:If7+ Wg4 6
l:!f6 @xh4 7 l:Ixg6 winning .
nn+ 4 'it>e3
Here already no help is 4 c;t>g4 h5
2 'it>g2 l:tg8 3 g4 g5 4 h5
This is more accurate than 4
5 @h3 el 6 l:!b5 e4 - +.
hxg5+ l:!xg5 5 'it>g3 e5 6 l%h6+ with
4... J::!f5 5 .!le4
After 5 'it>e4 g5 6 hxg5 hxg5 7 some chances of victory.
4 ... l:ta8 5 b7 .lih8 6 @fl!
J:tb l Wg6 8 nhl f2 9 c;t>e3 a2 1 0
Now the king can proceed to the
g4 l:Ia5 1 1 @e4 l:!a4+ 1 2 c;t>[3 l:!f4+
other side.
1 3 'it>g3 l:!e4 B lack wins a pawn.
6...l:!g8 7 'it>e2 .!Ia8 8 h6 .!Ial 9 h7
5 ... h5! 6 We2 g5! 7 hxg5 @g6 8
@e3 J::!xg5 9 'it>f2 c;t>f5 10 l:!e3
l:!hl 10 Wd2 @es 1 1 c;t>c3 nb4 12
After 1 0 .!If4 the pawn ending is @c4 l:!h2 13 'it>c5 hl 14 'it>c6 'it>f6
15 l:!d7!
won.
Making a future path for the king.
10 .. ig7 1 1 .Ub3
15... .!Ih2 16 @c7 i::!. hl 17 'it>d8
Or 1 1 @f3 llg4 ! 1 2 'it>f2 h4 1 3
gxh4 I:!'.xh4 14 @g3 l:l.'.g4+ 1 5 @f3 Black resigned .
.l:If4+ 1 6 @g3 @g5 1 7 nel f5 1 8
Beliavsky-Hodgson
l:te4 'it>g6 !, and the king works his
Cacak, 1 996
way round to d5. In a previous ex
ample the king also had the possibil
ity of attacking the h-pawn in this
way.
.

1 1 .. ..l:Ih7 12 l:!b5 h4 ! 13 gxh4

xh4
White resigned, since the pawn

ending after 1 4 l:ta7 'it>xe5 1 5 l:!xf7

J::i.f4+ is lost.
H uzman-Mikhalchishin
Lvov, 1 985

1 .l:If6

Stronger would have been 1 h4!

'it>g7 2 c6 f5 3 f3 !, creating weak
nesses in Black's pawn structure.

1. . @g7 2 llf4 f5 3 h3 'it>g6 4 g4

hxg4 5 hxg4 fxg4 6 Wg3 e3!
.

Black's only chance lies in the

creation of doubled pawns in
White's position.
7 .lixg4+ @f5 8 kif4+ @g6 9 fxe3
nh3 10 @f2 'it>g5 1 1 'it>f3 ki.a3

62 Rook Endings

After l 1 .. ..!lb8 1 2 @e4 .!le8+ 1 3

'it>d4 .!ld8+, adopting a frontal at
tack, secures Black a draw.
12 1If8 .!lal 13 @e4 J:Ia4+ 14

One extreme but typical case

occurred in a practical position.
Szab6-Keres
Moscow, 1 956

'it>d5 l:1a5+ 15 'it>d4 .l:Ia4+ 16 @c5

.l:.a5+ 17 'it>b4 J:ia2 ? !

Again correct was 1 7 ...e5 ! 1 8

.l:I f3 .!le8 1 9 @c4 .!lc8+ 2 0 'it>d3

.!id8+ 2 1 @c3 l:tc8+ 22 'it>d2 .!ld8+

23 @el .l:.a8 24 @f2 .!la4! with a
draw.
18 .!lf2 J:ic2 19 e4 @g4?!

Rather better, though already in

sufficient, was 19 ... @g6.
20 e5 'it>g5

If 20 ...@g3, then 2 1 @b3 ! l:rc8 22

.!lf5 @g4 23 e4 with a win.
21 'it>b5?

Correct is 2 1 J:ig2+! 'it>h6 22 @b5

.!lc8 23 e6 .!le8 24 @cs .!lxe6 25
@d5 ! with a theoretically winning
position.
21. ..@g6 22 e6 J:Ic8 23 J:tf3

B lack has great problems and it is

quite complicated to transpose into a
drawn position with f and h-pawns.
1. ..h6

If 1 . ..h5, then 2 g5+ @f5 3 .!lc6+,

and there is no defence against 4
.l:If6.
2 J:ih7! .l:Id6 3 h5!

Possible is 3 1Ixh6 @g7 4 g5 fol

lowed by 5 f4 and h4-h5.
3 ...@g5 4 .!lg7 'it>h4 5 .!lxg6 .l:.d2+
6 @fl 'it>g3 7 .:i.f6! .l:Ia2 8 @el l:th2!
9 @ d l ! l:tg2 10 'it>cl .!lh2 11 @bl !

The king moves further away

from its opponent.

23 @g7?
.

11 ... .!lg2! 12 J:ixh6 @xf3 13 l:1g6

@e4 14 h6 .l:Ih2 15 g5 @d3 16 .l:i.g7
@c3 17 h7 @b3 18 l:rb7+ Black
resigned.

Missing the typical drawing

mechanism 23 ...@gs 24 e7 .!le8 25
l:!.e3 @f4 ! 26 .!le6 'it'f5.
And here we see how difficult it is
24 e7 .!le8 25 .l:Ie3 'it>f7 26 Wc6 even for a class grandmaster to
.!lxe7 27 l:txe7 Wxe7 28 @d5 Black defend a classic position.
resigned.

Rook Endings 63

Nikolic-Ftacnik
Pu/a, 1 997

Better than 1 4 .!lb6, played by

Capablanca. White's plan, with the
rook on the 8th, is to to play f5-f6,
and Black's defence, with the white
king on f2, is to play . . .h6-h5.
14 .. J!b3?

Correct is 14 ... .l:!e3 !, hindering

f4-f5 .
15 'it>f2?

Correct is 1 5 .!id8 ! with the threat

of f4-f5-f6, e.g. 1 5 . . . .l:Ib2+ 1 6 'it>g3
.!lb3+ 1 7 'it>h4 .l:Ie3 1 8 .l::.e 8 ! and
f4-f5 .
15 ... .l:Ia3 16 .l::.d 7!

1 6 .l:Id8 is bad because of l 6 . . .h5

17 g5 h4 1 8 .!ld7 'it>g6 19 !td6+
1 I:l.c5!
Correctly preparing for h4-h5 in 'it>g7 20 .!lh6 h3 2 1 f5 .ll a5 with a
draw.
order to take on h5 with the rook.

1 ... .!lal 2 h5 .l::.g l + 3 @f3 'it>g7 4

.f:!.a5 gxh5 5 1Ixh5 .l::.a l 6 l:!.d5 .l:!gl

Not yet a mistake but more

technical is 6 ....l::.a4, cutting off the
white king.
7 .l::.d 6 'lt'h7??

A serious mistake-correct was to

attack the e3 pawn by 7 .. Jlal 8 'ite4
:a3 9 .l:!d3 .!la5 1 0 d5 l:i.a3 1 1
'it>d4 .!lb3 or 7 . . . .l:Iel 8 e4 .llf l 9 'it>g3
I:l.g l + 1 0 'it>h3 .l:r.fl , and the rook
does not have the f6 square.
8 e4! .l:Ifl + 9 'it>g3 .!lgl + 10 Wh3

.l::.fl 1 1 I:!'.f6! Wg7 12 e5

16 ...lia2+ 17 'it>f3 .l::.a3+ 18 'ite4

.!la4+ 19 d4!

This is the point of deploying the

rook on the d-file. After 19 @f5
'it>f8 ! according to an analysis by
Kopaev, the game is drawn.
,

19 ... lial 20 f5?

Correct now was to move aside

with the rook 20 .!lb4! .l:Ig l 2 1 'it>f3
.Ila 1 22 .l::.b 8 fl + (after 22 ... lia3+
23 'it>g2 the king goes via g3 to h4)
23 'it>e3 .l::.e l + (23 . . ..l:Ig l 24 f5 !
.l::.xg4 25 f6+ 'it>h7 26 e6!) 24 Wf2
.l::.e4 25 'it>f3 :!.e l 26 I:!'.b7 ! @f8 27
.l::.b4 'it>g7 28 'it>f2 .l::.a l 29 I:!'.b8 h5 30
g5 h4 3 1 'itf3 ! h3 32 1::rb 2 winning.
20 ....l::.e l + 21 'itd5 I:!'.gl ?

This position i s from the game

Duras-Capablanca, New York 1 9 1 3
(with colours reversed), and here
2 1 . . .'it>f8! 22 'itd6 .l::.a l gives a draw.
22 'it>d6 lial 23 .l::.c4 l:!.a8 24 .lac7
lla6+ 25 'it>e7 l:!.a4 26 e6! fxe6

Or 26. . .l:!.xg4 27 Wd6 winning.

White obtains a winning position

which is in every text book.
12 ...J::rf3 + 13 'it>g2 .l::.a3 14 .l::.d 6!

27 f6+ 'itg6 28 f7 J::rf4 29 f8='ii'

.l::.x f8 30 'it>xf8 e5 3 1 .!lc4 Black
resigned.

Even with a very lucky author and

a passed e-pawn it's drawn!

64 Rook Endings

Novikov-Beliavsky
Graz, 1 997

and the plan of the stronger side

consists of creating a passed pawn
on the d-file which promises him
great chances of success.
Bagirov-Veingold
Tallinn, 1 981

1 ...l:!c7 2 .l:Ia5 We6

He had to try the plan with the ap

proach of the king via h6.

3 .!le5+ Wf6 4 J::ra5 Wg7 5 h3

'it>h6 6 I:!'.e5 I:!'.c2+ 7 'it>g3 'it>h5 8 .!le7
h6 9 .lle5 J:ic3+ 1 0 Wh2 l:tcl l l g3!
After 1 1 'it>g3? J::rfl 1 2 l:ta5 there

is 1 2 . . . g5 ! 1 3 .!lxf5 e3 with a win.

Now, however, despite the 'cut-off
king from the first rank, Black does
not succeed in improving his
position.
1 1 ...J:ic2+ 12 Wgl J::rc 8 13 'it>g2
l:if8

How else to promote ... g6-g5?

14 'it>t1 g5 15 We3 g4 16 h4 J::rf6

17 'it>t1 Wg6 1 8 J:ie8 l:!a6 1 9 l:te5 h5
20 @fl 'it>f6 2 1 'it>t1 .l:Ia3 22 lies
.:i.a5 23 'it>e2 @f7 24 l:rh8 'it>e6!?

1 al

With the idea of defending the

pawn and freeing the king.
1...'it>g6 2 .!ifl .!la2
After 2 . . .f5 3 l:!a 1 fxg4+ 4 hxg4
the e6 pawn would be too weak.
3 <tii e4 l:te2 4 We5 Wg7 5 e4!

White's plan is clear-after

preparation to play d4-d5.
5 ...<tiig 6 6 f'3! I:l.e3
If 6 ...Wg7, then 7 'it>d6 and d4-d5,

when 6 ... l:l.'.h2 is too late because of

7 d5 exd5 8 exd5 .!lxh3 9 d6 J:ih2 I 0
But even the sacrifice of two l:td 1 ! .
pawns does not help.
7 d5 ! exd5 8 'it>d4! .!la3 9 exd5

25 I:!'.xh5 .l:Ia2+ 26 'it>e3 l:!a3 27

'it>t1 <tii d5 28 J:ixf5+ 'it>d4 29 .!le5
e3+ 30 <tiifl .!lal+ 31 <tiig 2! l:ta2+ 32
'it>gl 'it>d3 33 f5! .:i.al+ 34 'it>g2 e2
35 f6! and it's a draw!

Rook endings of five pawns

against four are a rarer formation,

'\t>f6 10 :!el !

Cutting off the king is more im

portant than a pawn!

10 ....!lxO 1 1 d6 l:tt1 12 'it>c5

.!lc2+ 13 'it>b6 l;Ib2+ 14 'it>c7 J:ic2+
15 <tii d 8 l:rc3 16 d7 J:ixh3 17 We8
l:td3 18 d8='*l l:txd8 19 Wxd8 h3
20 J:ie2 Black resigned.

Rook Endings 65

Ehlvest-Polugaevsky
Hanninge, 1 990

Complicated Rook Endings

It is interesting to see why grand

masters cannot hold slightly inferior
positions in which the method of
playing for a draw is well known.
Dautov-Alterman
Germany, 1 998

1 h3!

Preparing l:tc5-c4 and g3-g4.

1 ...f5 2 h4!

It is necessary to fix the weak

nesses and not allow . . . g6-g5 and . . .
h5-h4.
2 @f6 3 l:ta6
..

In such positions 'standing still' is

not to be recommended: there are
3 ... f4! ? 4 @g2 .l:Ib2 5 'it>f3 fxg3 6 two active plans:
(a) Petrosian's plan l ...f6 ! ? and . . .
fxg3 J:ib3 7 lia5 lib4
g6-g5, creating a weakness on h4;
Not allowing d3-d4.
(b) leaving the pawns alone by
8 lies e5 9 .l:Ic6+ Wg7 10 l:!c4
J:ib6 1 1 l:ta4 Wf6 12 'it>e4 'it>e6 13 e3 1 . . .l:tg2 2 'it>f4 lib2 3 a4 .l:Ib4+ 4
'it>e5 l:!b3, and White does not suc
There is no other plan for White.
13 ...f6 14 J:Ial lib4+ 15 'it>d5 ceed in advancing the a-pawn very
l:!b5+ 16 @c4 lib2 17 J:ia6+ 'it>f7 1 8 far before the D and g3 pawns fal l.
.li(a5 J:ic2+
In the game followed:
On 1 8 ... 'it>f6 White's plan would
1 ...lic2? 2 a4 l:ta2
After 2 ...J:ic3+ 3 'it>e4 Itc4+ 4
be 19 d4 exd4 20 exd4 l:tg2 2 1 J:ig5,
and then the advance of the d-pawn. d3 J:ib4 5 a5 .l:Ia4 6 f4 lia3+ 7
19 'it>d5 lig2 20 lia7+ 'it>f6 2 1 'it>c4 lixg3 8 lib6 lig4 9 a6 :Ixf4+
J:Ia6+ @f7 2 2 'it>xe5 l:!.xg3 23 a7+ 1 0 Wb5 .l:If5+ 1 1 'it>c6 l:!a5 1 2 'it>b7
g8 24 d4 l:txe3+ 25 f6 :b3 26 White wins.
Unfavourable for White is 3 e3
e5 ! , and the d3 pawn is weak.

d5 lif3+ 27 @xg6 @rs 28 h7! l::tf4

29 d6! 'it>e8 30 'it>xh5 lid4 3 1 d7+
@d8 32 .l:If7 :ds+ 33 @g6 lid6+
34 .l:If6! lidl 35 h5 Black resigned.

3 c;t>e4 lia3 4 a5 :Ia4+ 5 @d5

a3 6 @c6 lic3+

If Black takes the pawn 6 ... J:ixD,

then after 7 J:ib6 .l:Ixg3 8 'it>b7! .i:la3

66 Rook Endings

Material is nothing-activity (in

(8 ... g5 9 a6 .!:!'.a3 1 0 hxg5 + -) 9 a6
f6 10 a7 l::Ix a7 1 1 <tlxa7 g5 12 l:Ib l ! view of the passed pawn)
gxh4 1 3 Wb6 @g6 14 Wc5 <t>f5 1 5 everything.
6 ... hxg3 7 hxg3 gxg3 8 aS gl 9
c;i;>d4 <t>f4 1 6 :n + @g3 1 7 @e3,
White wins, but also the move in the l:!.a2
game does not save him.
7 @b7 l:!a3 8 f4 .l:tal 9 l:!.a8 @f6
1 0 a6 @rs 1 1 f8 .l:tbl+ 1 2 @a7 f6
13 b8 el 14 :b3!

An important reconstruction. The

rook has two defensive functions: to
defend the g3 pawn and control the
b-file in order to free the king.

14 ... @g4 lS <t>b6 .i:te6+ 16 'it>aS

![e7 17 .kla3!

One more function: to support the

a6 pawn.
17 ... gS 18 fxgS fxgS 19 hxgS
es+ 20 <t>b6 xgS 21 a7 g8 22
c3! gg6+ 23 b7 :g7+ 24 @b8
Black resigned.
Mikhalchishin-Eslon
Kecskemet, 1 983

9 ...1:1'.bl+?

Better at once 9 . . Wd6!, not allow

ing in the white king, while on 1 0
a6 :tb l .
.

1 0 <t>cs l:!.b8 1 1 a6 gS 1 2 c6
f6?

In such positions it is useful to

prepare to push away the white
king. 12 ...<t>e6 1 3 a7 :as 14 <t>b7
l:!.xa7+ 1 5. <t>xa7 'it>d5 ! with clear
chances of a draw.

13 a7 l:1a8 14 @b7 nxa7 lS !Ixa7

g4 16 'it>c6 <t>es 17 @cs g3 18 'it>c4
@f4 19 @d3 Black resigned.
Keller-Mikenas
corr, 1 992
1 ....:.a3+?

In analogous situations where the

pawn has not advanced to the fourth
rank, it is correct to activate the
king, but the rook gives a frontal
check.
2 <t>e4 cj;e7 3 'it>dS rs 4 <t>cs h4? !
Again correct is 4 ... l:l'.a8, prepar
ing' to meet 5 'it>b4 with 5 ...b8+.
S <t>b4 l:!.e3 6 a4!

Rook Endings 67

Black must lose because of the

weakness of the e4 pawn.
1 a 6! g6

S 'it>dS 1Ia3 6 a6 .!lxf3?

This is a decisive mistake-after

6 ...'it>f6 7 'bc6 .l:Ixf3 8 1Ib8 .!la3 9
b6+ 'it>f5 1 0 @b7 @g4 1 1 a7 .!lxa7
1 2 'it>xa7 'it>xg3 1 3 .!lb4 f6! there is
an easy draw by ... g6-g5.

After l ...h4 2 gxh4 :!.a l+ 3 'it>e2

a2+ 4 'it>e3 .l:Ia3+ 5 'it>xe4 lhh3 6
f4! xh4 7 .l:Ia8 .!lh l 8 l:tf8 and 9 a7
wins. In the event of passive de
7 b8 a3 8 .l:Ib6 .!lxg3 9 'it>c6
fence by Black, White pushes his .l:Ia3 10 Wb7 gS 1 1 hxgS h4 12 a7
pawn up to a7, then the king goes to h3 13 a8='i' .l:!xa8 14 'it>xa8 h2 lS
e5, the Black rook must keep watch .l:Ih6 f6 16 .l:!xh2 fxgS 17 .!lt1! !
from the a4 square. Then White
We advise paying attention to this
plays 'it>f4 and, on ... 'it>h7, '\t>f5 with move.
a win.
2 a8 'bf7 3 a7 'it>g7 4 'it>el 'it>h7
S 'it>dl 'it>g7 6 @cl Wh7 7 'it>bl I:!'.a6
8 g4!

Determining the pawn formation

on the flank.

8 ... hxg4 9 hxg4 @g7 1 0 gs 'it>h7

11 'bb2 'it>g7 12 'bb3 'it>h7 13 'it>b4
r:j;g7 14 'it>bS a2 lS Wb6 .l:Ib2 1 6
'bc6 .!lc2+ 1 7 'bd6 .!la2 1 8 .!lc8!
lha7 19 .!lc7 and the pawn ending
is easily won. Black resigned.
Lerner-Dorfman
Tashkent, 1 980

Now the white king avoids being

shouldered away and hurries over to
the opponent's pawn, therefore
Black resigned.

Konopka-Shcherbakov
Pardubice, 1 996

1 ...'it>g7?

Passive play-correct is l ...'it>e6

-or l ....:i.a2 detaining the king on
f6.
2 as a4+ 3 'bes .!la3 4 'it>e4
lfa4+?

Simpler is 4 ... @f6 ! 5 a6 'be6 giv

ing a draw.

This position resembles the previ

ous one, but there are a number of
important differences. Black has a

68 Rook Endings

doubled pawn, and the greater num

ber of pawns creates more prerequi
sites for a White victory.

1 g4! g5 2 a4 .l:Ial 3 a5 'it>g6 4 a6

'ith6

f4! exf3+ 1 2 Wxf3 Wg7 1 3 b7

10 'it>h2 'it>h7 11 lif8 Ji(xa7 12

lixf6 Wg7 13 .l:Ie6 .l:Ia4 14 'it>g3 lib4
15 h4 gxh4+ 16 'it>xh4 l:ta4 17 Wg3
.l:Ib4 18 .l:Ie5 'it>g6 19 llf5

White also wins after 4 . . . h5 5

gxh5+ 'it>xh5 6 l:!a8 'itg6 7 a7 'itf5 8
Worthy of attention is 1 ... e4 ! ?, for 'itg3 with the idea 9 f3 .
5 J::ra8 'it>g6 6 a 7 'it>g7 7 'it>g3
the present maintaining the pawn on
fl so as to defend the king against .l:Igl+ 8 'it>h2 l:!.al 9 'itg2! h6
After 9 .. . 'it>h6 White wins by 1 0
checks from the flank.
2 .l:Ib8 e4 3 h3 h5 4 h4 J::rb l+ 5 J::rf8 lixa7 1 1 .l:Ixf6+ 'it>g7 1 2 lie6
c;tie2 lib2+ 6 Wfl J::rbl + 7 <;tig2 lib2 lia4 1 3 cJi>g3 'iitfl 1 4 lih6 'it>g7 1 5
8 b4 :!bl 9 b5 'itg6 10 b6 Wh7 1 1 .l:Id6 lia7 1 6 lle6 .l:Ia4 1 7 h4.
1...f5

On 13 Wf4 there is 13 . . .Wf6 ! with

the idea . . llfl .
.

13 ... Ji(b4 14 We2 llbl 15 Wd2

:gb3 16 Wc2 J::rb 5 17 Wc3 l:!.b6 1 8
Wc4 :!bl 19 Wc5 l:tb3 2 0 'itc6
l:tc3+ 21 'itd7?

Stronger is at once 2 1 c;tib5 l:tb3+

22 'it>c5 ! .

With the idea of llf5-f4, f2-f3 .

19 ... .l:IbS 20 'it>f4 .l:Ie8 21 lie5 !

.l:Ixe5 22 'it>xe5 <;tig5 23 Wxe4 c;tixg4
24 'it>d3 ! Black resigned.
Kozul-Nikolic
Sarajevo, 1 993

2 1...lid3+ 2 2 <;ties!

After 22 c;tixe6 l:txe3+ 23 Wxf5

l:tf3+=.

22 ... lib3 23 'it>e7 l:tb6 24 'it>d7

l:!b3 25 'itc6 llc3+ 26 'it>b5 J::rb3+
27 'it>c5! :gb l

If 27 ...c;tih7 winning is 28 c;tic6

llc3+ 29 Wd6 .S.d3+ 30 'it>e7 J::rb 3 3 1
'it>f6 ! .
2 8 'iit c 6! J::rc l + 2 9 c;tid6 :!bl 30
'it>xe6 .:.b5 31 .l:Id8 Black resigned.
Andersson-Hubner
Ter Apel, 1997

1 a7

White sees no chance of going

with his king to the queenside to
help his a6 pawn. For this he has to
sacrifice the f2 pawn.
1 . .. c;tie5

Also not bad is 1 ... <;tig7 with the

idea of 2 ...f5 or 1 .. .l:ta3 .

2 f3 l:!.a3 3 'it>t1 1ial 4 'it>e2 'it>f6 5

f4 Ji(a3?!

The simplest way to achieve a

draw is 5 ... c;tie7, and there is no way
for the king to approach.

Rook Endings 69

In the game after 7 . . . g6? 8 l:th8

6 @d2 .!ia2+
A move such as 6 ... We7 can al Black resigned.
8 @d2 .!la3 9 i;t>e3 .!IaS 10 @xd3
ways be made.
7 @c3 .!la3 8 Wc4 llaS 9 'it>b4
.U,a2 1 0 @cs J::ra l 1 1 Wc6

:Ia4 1 1 Wc3 .!lal 12 'it>c4 .!Ia2 1 3

'it>cS J:i a l 14 'it>c6 .!I c l+ lS 'it>d6

Preparing to meet @es with the

No use is 1 1 l:th8 .iha7 12 l:txh6+
move .!la5, but now he threatens to
@g7 13 fxg5 .lla 5
1 1.. ..!la2 12 @cs J:Ia6 13 'it>d4 tranpose to a pawn ending.
.

lla4+ 14 @d3 .!la3+ lS We2 l:taS?

lS ... J:ial 16 .!lc8! .!Ia6+ 17 .llc6

Completely losing his sense of .!lxa7 18 J:ic7+ J:Ixc7 19 Wxc7

danger-why not 1 5 ... 'it>g7?

16 h8! l:txa7 17 l:!.xh6+ 'it>g7 1 8

fxgS k.i.a3 1 9 'it>f3 .llaS 20 h 4 I:i:bS
21 ltf6 Black resigned.
Yermolinsky-Seirawan
USA, 1 997

Note an important fine point-the

tempo g2-g3 is decisive. With the
pawn on g3 it would be a draw.
19 ... @e7 20 i;t>c6 'it>e6 2 1 @cS!
Destroying the opposition.

1.. ..!lxa2?

21...'it>f6 22 @dS g6 23 hxg6

@xg6 24 'it>e6, winning. But with
the pawn on g3, 24 .. 'it>hS 25 'it>xf5
.

B lack does not suspect any danger is stalemate.

associated with the the white pawn
Vujala-Smith
on h5. Therefore correct is l .. .f4 ! .
2 f4! d4
corr, 1 993
The other chance is to hide on h7
and wait with 2 ... 'it>h7 3 @fl . But
then the white king enters the game
via b l , attacks the d5 pawn and as a
result of zugzwang captures the d5
and f5 pawns.
3 @fl d3 4 'it>el ! @rs
Losing is 4 ...l:txg2 5 l:l'.d7 .!la2 6
a7 @h7 7 'it>dl, and the white king
captures the f5 pawn, and then goes
to b8.
S .!la8+! 'it>e7 6 a7 @f7 7 'it>dl

lla4

70 Rook Endings

1 g4!

Fixing
nesses.

the

opponent's

weak

not have enough tempi to save

himself.
2 J:ib4 'it>a5

At first sight 2 . . . gxf3+ 3 'it>xf3

1::rg 8 seems sufficient for a draw.
Underlining White's advantage. Actually, in the variation 4 .l:Id4
'it>xb5 (4 ... @c5 is useless because of
4 ...'it>e6 5 g5 ! .
5 hxg6 i;t>xg6
5 l:td7) 5 ltd5+ 'it>c6 6 l:1xf5 i;t>d6 7
If 5 ...fxg6, then 6 1::rc7 with the g4 'it>e6 8 @f4 there arises a theo
retical position where it is Black's
idea Ji(a7,
move-and consequently a draw
6 f3 f6
Better is 6 ... @g7 7 i;t>f2 'it>g6 8 by 8 . . . l:1a8. But White plays the
<tte2 Wg7 9 l:!.f5 <ttg6 1 0 'it>d2 h5 !=. more refined 4 .l:Ib3 ! (threatening 5
'it>f4) 4 ...J:Ig4 5 d3 i;t>xb5 6 l:td5
7 'itg3 .l:Ia4 8 .l:Id5! :g,b4
Now the continuation 8 ... <ttg7 9 i;t>c6 7 l:1xf5 :g,gg 8 g4 <ttd6 9 \t>f4
f4 i;t>g6 1 0 l:!.d7! f5 1 1 l:td6+ i;t>g7 'it>e6. The same position, but with
12 a6 !ta3+ 1 3 <tth4 :g,a4 14 g5 White to move, 1 0 g5, and this is
already a win.
leads to victory for White.
1. .. h6 2 'it>g2 .l:Ia3 3 h4 i;t>f6 4 h5 !
<ttg7

9 J:id8! !ta4

3 l:!.b3 'it>b6 4 'it>f2 as

Or 4 ... gxf3+ 5 <ttx f3 :g,gg 6 <tt f4It is too late for 9 ... h5 10 gxh5+
i;t>xh5 because of 1 1 l:th8+ <ttg 6 1 2 Black does not stand his ground.
5 fxg4 fxg4 6 l:tb4 J:if8+ 7 i;t>e2
l:tg8+ followed by J::rg 4.

10 l:tg8+! <ttf7 1 1 .l:Ia8 al 12 f4

@g7 13 a6 lla2 14 f5! !!'.al 1 5 <ttf3
l:!.a3+ 16 <tte4 J:Ia4+ 17 <ttd5 h5

J:ie8+

18 gxh5 !las+ 19 <tte4 Black

resigned.

13 .l:Ixg4 i;t>cs 14 z:!'.g6 <ttd5 15 <ttf4

f8 16 @g4 e4 1 7 l:1e6+ @dS 1 8
lth6! with an advantage that is quite

In the variation 7 ... @c5 8 .l:Ixg4

<ttx b5 there is 9 J:if4 ! J:ie8+ 1 0 <ttf3
There is also no salvation in 'it>c5 1 1 g4 J:ig8 1 2 l:tf6.
8 <ttf2 l:tf8+ 9 <tte 2 J::l.e8+ 1 0 <ttd3
1 7 ...J:ixg4 1 8 l:ta7+ <ttg8 1 9 .l:Ib7
<ttc 5 1 1 e4 J:id8+ 12 <tte3 <ttxb5
and then a6-a7 + -.
Salov-Malaniuk
USSR (ch), I 988

easy to realise.

Guseinov-Beliavsky
Pula, 1 997

1. .. <ttb6

The sealed move. It is easy to cal

culate that on l ...l:!.xb5 Black does

Rook Endings 71

White has a weakness on c4, but

this is small and Black needs to
create another.
t. .. hS!

2 Wf4

No help is 3 gxf4 'itf5.

3 ...J:ib2+ 4 'itgl fxg3 5 I:!'.f6+

It is necessary at least to force

back the black king.

5 ... @g7 6 I:!'.xe6 l:l.'.h2 !

After 2 gxh5 .!lxh5 3 @g3 .!lc5 4
It is precisely this pawn which is
l!c3 follows 4 ... 'itd7! with the threat
important.
... @c7 and ... d5.
7 J::rd6 J::rxh4 8 .llxdS
2... gS+! 3 'itt3 h4
Now there is a weakness on h3.
After 8 ctig2 l:l.'.g4! 9 .!lxd5 h4 fol
lows h4-h3 and g3-g2.
4 l:rdl
8 ....!le4! 9 e6 h4 10 .!lhS .!lxd4 1 1
Preventing .!le6-el and h i .

4... l:IcS 5 :!:!'.el+ ctid7 6 e4 I:!'.eS!

7 bi.d4 'itc6

Now the king is included in the

attack on the weak c4 pawn.

8 .l:Id3 WcS 9 J::rd l f6! 10 J:id2

bi.e6 1 1 J::rc2 ctid4 12 :!:!'.cl I:!'.e3+ 13
\tg2 .l:r.g3+ 14 Wh2 .llc3+ 15 J::r d l +
d3 16 .l:Icl :n 17 l:l.'.dl+ @cs 1 8
l:rcl .!lf2+ 19 'itgl I:!'.f4 White
resigned.

e7 @f7 1 2 .!lh7+ @es 1 3 Wg2 I:!'.g4!

Zugzwang. 14 'itgl g2. White
resigned.
Mednis-Gurevich
New York, 1 982

Mikhalchishin-Khmelnitsky
Sibenik, 1 990

In this rather untypical position

there is only one possibility for
Black to save the game.
1 ...cS ! ! 2 bS

White can do nothing against the

break ... f5-f4.

After 2 bxc5 'itd5 3 ctid3 l:!.c4 4

@e3 'itie5! Black achieves a draw.

2 ... .!lb4 3 l:l.'.b2! .!lxh4 4 b6 .!ih8 5

@c4 ctid6 6 @bS @d7!

1...l:!.bl ! 2 @f2
It was still not too late to lose:
Bad 2 @f4 .!lei 3 lle3 (3 'itf3 l:te4 6 ... c8 7 .l:Id2+ \ties 8 b7 .l:Ib8 9
with the idea ... f4) 3 ... .l:Id l , and .!id7 'ite6 1 0 'itc6 with a win for
there is no defence.
White.
2 ... f4 ! 3 13
7 @a6 J:Ia8+ 8 ctib7 .l:Ia4 !

72 Rook Endings

b7, and White wins; or 13 ... l:!'.c8 14

b7 l:!b8 1 5 'it>c6 with the same re
sult) 1 3 b7 .l:Ib3 1 4 lte2+! (only
now, when the rook stands on the b3
square!) 14 ... f5 1 5 'it>c7, and
White wins.
Interesting methods of defence
in rook endings

Now if White plays 9 J::rd2+, then

Black has the possibility to trans
pose into a drawn pawn ending by
9 . . l;Id4.
.

9 c2 ! ! c4

If 9 .. 'it>d6 1 0 .l:Id2+ :!:i.d4? 1 1

l;Ixd4+ cxd4 12 c8, and White
queens with check. 9 . . .l:Ia5 1 0 l:!.d2+
is also losing.
.

10 lld2+ e6 11 'it>c6 c3

In defending this typical position,

exceptionally important is the pres
ence of the two white pawns-the g
or h pawns do not win, the others do
quite simply, by advancing them to
the seventh rank. Knowledge of
these types of position is extremely
important in the different variations
when transposing to a rook ending.
12 l:!e2+ f5 13 b7

If 13 'it>b5, then Black achieves a

draw by the manoeuvre 1 3 ... l:Ia3 14
@c4 a6!
13...l;Ib4 14 .l:Ih2

Nothing is changed by 14 c7=.

14 ...J::rx b7! 15 'it>xb7 'it>e4 16 'it>c6

'it>d3 Drawn.

However at the end of the game

nalysis
showed that in the position

m the last diagram, 12 c2! de

served consideration. Now if
12 ... Ita3 ( 1 2 ... l::tc4+ 13 'it>b5 '>t>d5 14

Chaunin-Friedman
Moscow, 1 951

Rook Endings 73

Here White can win easily by 1

Here, an analogous winning
'it>xg3 ! 2 h4 ! , after which the f3 method is demonstrated by Shirov.
pawn goes to f7. However White
1 b6! cxb6 2 .l:Ih8 Black
decided that simpler was 1 hxg3?, resigned.
keeping his pawns connected, but it
Mokry-Pribyl
turned out that after 1. .. g4+! 2 fxg4
Olomouc, 1977
he cannot win even with an extra
pawn.
Polugaevsky-Parma
Sochi, 1 965

After 1...Wh7? 2 Wc6 .l:Ia2 3

.l:Id8! .:i.xa7 4 l:l'.d7+ xd7 5 Wxd7
g5 6 'it>e6! Black resigned.

The method which helped White

to win is typical of such positions. It
thought there was no defence. How is possible that Black did not even
ever after l ...Wg6! and 2 ...Wh7 ! think about the fact that it was still
possible to give up even a centre
Black calmly makes a draw.
pawn.
On
the
correct
defenceShirov-Kramnik
1 . . .Ir.a 1 !-the operation to transpose
Belgrade, 1 999
into a pawn ending is not achieved
since after 2 'itc6 Black checks
along the file until the king is forced
away from the a-pawn. For
example: 2 .. .:i.c l + 3 'it>d6 .:i.d l + 4
'ite6 .l:.a l 5 l:td8 (5 :ge8? l:l'.a6+!)
5 ...na6+ 6 l:td6 xa7 7 l:Id7+ nxd7
8 Wxd7 'it>f6 ! 9 'it>d6 g5 ! 1 0 hxg5+
'it>xg5 1 1 Wd5 e3 ! 1 2 fxe3 Wg4 with
a draw.
And here in a classic game Black
resigned since he saw 1 h6, and

The right method of defence was

not found in the following game.

74 Rook Endings

White was convinced that he

would make a draw but, despite
thinking about this position for
more than an hour, he did not con
struct that 'fortress '. This perfectly
appropriate example is of interest to
the theory of rook endings.
The fact that the pawn stands on
f6, and not on fl, is of no signifi
cance. Now simplest for White was
1 g3! <tle7 2 'lt>g2 lt>d7 3 l:!.a8. In the
end Black reaches the following
position.

Here a mechanism like 2 ... lle l

does not work, since there follows
first 3 l:.ta3+! and only then 4 l:txa2.
There is also no win for Black after
2 ...l:!.c l 3 a3+ l:tc3 4 l:!.xa2 l:.tc2+ 5
:rxc2 <tlxc2 6 'it>f3 @d3.
By comparison with the game
Mokry-Pribyl this pawn ending is
favourable for the stronger side. But
in this case the weaker side achieves
a draw: 7 @f4 lt>d4 8 f3 g5 ! 9
hxg5 fxg5 1 0 e5 ! .
Also here the sacrifice of a centre
pawn saves him!
1 0 ...xe5 1 1 e3 f5 12 lt>f3
with a draw.
Black could play 12 . . . g4-but
then arises a theoretically drawn po
sition--or else he must try to hand
over the move to White which he
will possibly succeed in doing.
But even in this case White
achieves a draw after 1 g4 ! hxg4 2
Wg3 . Therefore Black must take the
pawn at once.
2 @xe4
Reaching the basic position of this
ending.

With White to move, Black wins a

pawn: 1 @h2 @f3 etc. With Black
to move, after 1...<tld3 2 'it>h2 'it>c3
Blck wins in the same way as in
the game Mokry-Pribyl. Then sim
plest is to give up the pawn at
once-2 J:ta8!

It is of no significance at all what

the move is here. It is necessary
only to bear in mind that Black
should not play ... g6-g5, since after
the exchange of pawns, a second
passed g or h-pawn will be created,
which does not win.

Holmov-Timoschenko
Pavlodar, 1 982

..

Rook Endings 75

The only path to victory is to cre with a draw since Black cannot go
ate a passed f-pawn, but how can to the c-file with his king: 1 8 ... 'it>c7
this be done? If Black manoeuvres 1 9 hxg5.
with his king to the d4, c4 squares,
etc, then White gives check along
We return to the game Holmov
the file. But even here White needs Timoschenko.
to take care. For example, after 3
.l:!c8+ @d4 4 .l:Id8+ @es mistaken is
5 .l:r.e8? 'it>f5 6 lia8 g5 7 J::ra 5+ 'it>g6
8 hxg5 f5 ! .
After capturing o n g5 Black man
ages to create a passed f-pawn,
which wins. Instead of 5 lie8? it is
necessary to continue S a8!
Now the above-mentioned ma
noeuvre does not work: 5 ...g5 6
hxg5 f5 7 .l:Ia6! h4 8 g6, and already
Black has to think how to save
himself.
On 5 .l:Ia8 he will try to send the
king to g7, so as then to play
1 0 .l:Ia3+ xe4 1 1 .l:Ia4+?
Also here still possible was 1 1
... g6-g5 : S ... 'it>e6 6 lia6+ 'it>f7 7
.l:Ia7+ 'it>g8 8 J::ra8+ 'it>g7
h2 and then g2-g3.
On this follows a check on the
l l ...We3 12 l:ta3+ 'itid4 13 a4+?
rank-9 .l:Ia7+, and after 9 'it>h6 'it>c3 14 .l:ta8 fS ! lS .l:Ia7 f4+! 16
the manoeuvre 10 .l:Ia6 prevents the 'it>h2 'it>d4 17 .l:Ia4+ 'it>eS 1 8 .l:!a3
'iitfS 19 .l:Ia6 g4 20 lixg6+ Wxh4
advance ...g6-g5.
There is nothing else for Black, 21 a6 'it>gS 22 l;Ia8 h4 23 l:!'.g8+
besides 10 ... fS. Possible then is 1 1 'it>f6 24 .li(a8 h3! 2S gxh3 f3 26 .l:Ia3
..

.l:Ia7 gS 1 2 l::t a6+ 'it>g7

Now Black can choose two paths,

each of which leads to a favourable
result for him. Let us look first at
the direct 1 3 hxg5 h4 14 gxh4 f4 1 5
h5 f3 + 1 6 @ f2 lih l 1 7 h6+! 'it>h7 1 8
l:ta7+ 'it>g6 1 9 h7 .l:Ixh7 20 lia6+!
xg5 2 1 @xf3 with a draw.
The second path also leads to his
objective: 13 lia7+ 'it>f6 14 laa6+.
Black is at the crossroads.
After 14 ... e5 1 5 hxg5 h4 1 6 g6
he might even lose.
But 14 ... 'it>f7 15 lia7+ e8
( 1 5 ... We6 16 hxg5! h4 17 g6) 1 6
l:ta8+ c;t>d7 1 7 lia7+ 'it>d8 1 8 .l:Ia8!

es White resigned.

Activity is more important

than material

It has Jong been known that in

rook endings activity is more impor
tant than material. This means maxi
mum possible active deployment of
the king and rook in coordination
with one's own passed pawns and in
the struggle against the opponent's
pawns. And even the very idea of
the priority of material will be fatal.
Here are a few striking examples.

76 Rook Endings

Larsen-Browne
Las Pa/mas, 1 982

Arbakov-Gurevich
Moscow, 1 9 78

The black rook is hopelessly pass

1 . .Ucl ! 2 @f3 :gc4 3 <;t>e3 e5! 4
ive and White's plan is to go with <;t>d3 .l:!a4!
Clearly not 4 ... e4+? because of 5
his king to h5, place the rook on b6
and break up the black pawns with xe4.
the pawn march f4-f5-f6.
5 <;t>e3 e4 6 .l:Ig3 l:!'.a2 7 _:g4
1 '\t>g4! '\t>e6
Or 7 h4 g4 8 h5 .l:Ia3+ 9 @f2
After 1 .. g6 winning is 2 l:tb6+ xg3 10 '&t>xg3 e3 with a win.
7 . .Ua3+ 8 <;tin l:.d3 White
'\t>g7 3 f5 @h7 4 Wf4 and 5 '\t>e5.
.

2 fS+ e5 3 .:M ! g6

On 3 . .. Wd6 follows 4 <;t>h5.

resigned.

"Active positions of the pieces in

rook endings are worth a pawn"
At a necessary moment the b7 -Smyslov. Here is a classic
pawn is given up in exchange for example.
the win of a pawn on the opposite
flank. This results in a theoretically
Capablanca-Tartakower
winning position.
New York, 1 924
7 ...ki.g7+ 8 <;t>f4 t7+ 9 'it>g5 'it>es
10 g4 .t!'.f8 1 1 @h5 f7 12 g5 'it>fS
13 .l:Ih8 and Black resigned.
4 fxg6 fxg6 5 b6! '\t>d4 6 .l:Ixg6
xb7 7 !rxh6

In the following position the

white rook is finnly posted, but it is
immobile and passive. Black finds a
plan with a transposition to a pawn
ending, exploiting the position of
the rook on g4.

Rook Endings 77
Lilienthal-Smyslov
l @g3 ! l:hc3+ 2 'ith4 l:f3 ?
Moscow, 1 941
Materialism i n such positions i s
simply ruinous. This position
sparked a great debate in 1 998 in
the magazine 64. Master Goldin
maintained that 2 ...a6, intending to
create a passed pawn with maxi
mum speed, gave Black drawing
chances. However after 3 g6 b5 4
axb5 axb5 5 @g5 b4 6 :n+ @g8 7
Ii'.xf5 b3-master Barsky pointed
out 7 . . . l:lg3+! 8 'ith5 ! (8 @f6 .l:rg4!
leads to a blind alley-8 ...b3 9
l:lxd5 b2 1 0 .l:rb5 @g7 1 1 :l:xb2
!:.xg6 1 2 :!c2 l:ld6 1 3 l:.c4 ! @f6 1 4
@g4 c6 1 5 @ f3 'it>f5 1 6 l:Ic5+ 'ite6
There followed:
l . .@e4! 2 1:1'.xcS f4 ! !
1 7 @e4, gradually driving back the
black pieces) 8 lixd5 'itg7 (8 ... li'.c6
B y sacrificing a third pawn, Black
9 @h6 J;;td 6 IO l:.a5 ! l:ld8 1 1 li'.b5 ! creates cover against checks to his
own king.
with a win)
.

3 exf4

Or 3 @fl .l:ral + 4 e2 f3+ 5 @f2

l:a2+ with perpetual check.
3 ... @f3 4 h3 i:lal+ with a draw
four pawns down.
Kozlov-Mikhalchishin
Vladikavkaz, 1 978

9 l:ld7+ @f8 1 0 d5 ! (but not IO f5

b2 1 1 g7+ @g8 1 2 @g6 l:lc6+ 1 3 f6
b l = with check! ) 1 0 . ..@e8 (after
1 0. . .b2 1 1 @f6 @e8 1 2 l:le7+ @d8
1 3 g7 l:tg3 14 ltel White wins eas
ily) 1 1 l:Ih7 b2 1 2 l:lh8+ @e7 1 3
l:lb8 :c2 1 4 l:l'.b7 .l:rg2+ 1 5 @f5 l:lc2
16 d6! xd6 17 @f6, and White
wins.

e4!

A passed pawn in conjunction

The f5 pawn protects the white with a centralised king is worth two
pawns. But where does the capture I
king, but not the black one.
5 ... g8 6 l:lg7+ @h.8 7 :txc7 lie8 .l:rxb3 lead? After l ... J:l.xg3 2 e4
7 'itxf5 li'.e4 8 @f6 l:If4+ 9 e5 l:I g4 h5 3 f4 l:lg4+ with a further ...g5
White is in a deplorable state.
1 0 g7+, etc. White won.
3 g6! l:l'.xf4+ 4 g5 e4 5 @f6!

78 Rook Endings

1. .lhg3 2 e5 'it>g6 3 e6 l:id3+

After 3 ... lie3 4 'it>d6 h5 5 lib8 h4

6 e7 the white pawn queens.
4 'it>e5 h5 5 b8!

There is a very clear path to the

draw.

1. .. a5! ! 2 bxa5 l:id4 3 I:!'.xb5 lia4

4 1::rb 3

Or 4 l:i.b6+ 'it>e5 activating the

5 e7 does not work because of
king.
5 . .@f7, but now this is a threat.
4 ... l1xa5 5 'it>f4 .l:Ia4+ 6 J::rb4
5 .. Jle3+ 6 'it>d6 'it>f6 7 J::rf8+ 'it>g5
8 e7 b2 9 b8 'it>f6 10 .l:If8+! 'it>g5 lixa3 7 J::rb6+ 'it>g7 8 .l:Ie6 lib3 9
.

1 1 J::rb 8 Drawn.

Taimanov-Chekhov
Kishinev, 1 9 76

.l:Ie3! l:!xe3! 1 0 'it>xe3 'it>t7 1 1 'it>d3

'it>e7 1 2 'it>d4 'it>d6 Drawn.

A classic activating of the king

was seen already in this game:
Owen-Morphy
London, 1 858

The threat is 1 e4+, and then 2 a4,

3 'it>g4. Therefore correct is
1. ..I:!'.d3! 2 xb6 l:!a3 3 .l:!b2 .l:!a4

... followed by ... g5-g4 with a

draw.

1 ...'it>t7!

After 1 .. . .l:Id3 2 lic8+ 'it>h7 3 .l:Ie8

A frequently met theme is the
White has great chances of a draw.
sacrifice of a pawn to activate the
2 lic7+ Wf6 3 .l:Ixb7 d3 4 'it>f2
rook.
Barlov-Schiissler
Hanninge, 1 988

l:!xd4 5 We3 e5 6 b6 l:i.b4 7 l:!b8

'it>e7!

Black chooses a plan to liquidate

White's passed p awn. Inferior is
7 ...J::rb 3+ 8 'it>d2 We6 9 'iitc 2, and it
is not clear how to win.
8 b7 'it>d7! 9 g8 xb7 10 xg7+
Wc6 1 1 .:i.g6+ 'it>c5 12 .l:!xh6 l:l'.b3+
13 'it>e2 e4

After the pawn sacrifice all

Black's p ieces have become more
active and his pawn further
advanced.
14 lih8 'it>d4 15 l:!g8 1::rb 2+ 16
'it>dl 'it>d3 White resigned.

Rook Endings 79

Schlechter-Lasker
Berlin, 1 9 1 0

succeed because of 3 ... h3 4 a5 .!lg2+

5 Wfl f4 6 a6 f3 7 a7 h2, and it is
time to resign.
3 ...h3 4 .llb 8 J:ig2+ 5 wn J:id2

Quite possible is 5 . . . l:ta2, but

Black wants to use the rook to cover
against checks from the side, while
he simply pays no attention to the
a-pawn.
6 'it>gl f4 7 .:i.g8+ f3 8 h8
dl +! 9 'it>h2 'it>f2

Black wants simply to promote

the f-pawn to a queen. If now 1 0
.l:!xh3, then 1 o.. f3 1 1 l::!.h 8 l!d3 ! 1 3
a 4 'it>e2 1 4 f8 l:!.e3 ! 1 5 a 5 e7! 1 6
a6 f2 with a win.
.

The best defence is ...

1. ...l:!e4! 2 .l:Ic5 'it>f6 3 .l:Ixa5 l:tc4!

4 l:ta6+ 'it>e5 5 l:!a5+ 'it>f6 6 .l:!a2
e5 7 J:ib2 J:ic3+ 8 'it>g2 'it>f6 9 h3
.!lc6! with a draw.
Kramnik-Beliavsky
Groningen, 1 993

10 a4 f3 11 a5 Wfl 12 a6 .!lal 13
.!las

After 1 3 c;t>xh3 f2 14 .!lg8 .l:Ixa6

15 'it>g3 J:if6 ! the win is straight
forward.
13 ... f2 14 a7 lia6!

The right idea. Now on 15 xh3

'it>g l 1 6 l:!g8+ 'it>hl 1 7 f8 l:l.'.a3+ 1 8
@h4 Wg2! i s reached a postion from
a classic study by Lasker, where
Black wins by shouldering the white
king to the seventh rank, while on
15 .:i.b8 follows 15 ... l:!xa7 16 .!lb l+
'it>e2 1 7 l!b2+ We3 1 8 I:!'.b8 c;t>e4 19
J::!b4+ c;t>e5 20 .l:Ib5+ 'it>e6 2 1 .l:Ib6+
'it>e7 22 l:!b l l:!a3 ! 23 :!:!'.fl l:l.'.f3 fol
1 :!.al ? .l:!g3+ 2 Wf2 Wg4 3 l:r.bl
Here also he had to go back; lowed by the approach of the king.
15 'it>hl h2! 16 J:ib8
counterplay with 3 a4 does not

Correct was the natural 1 l!b8, ac

tivating the rook and not fearing
1 . . ..llg 3+ 2 'it>f2 l::!.x a3, since the
rook ending with the f-pawn is
drawn. But White decides to defend
the pawn and at a suitable moment
to obtain counterplay with the help
of the a-pawn. But this proves to be
a decisive mistake.

80

Rook Endings

If 16 'iitxh2, then 16 ... .!lh6+ 1 7

Wg3 'it>gl 1 8 z:rf8 l:Ig6+ 1 9 'it>h4
.!la6, again reaching Lasker's study.

Bojkovic-Kakhiani
Erevan, 1 9 9 6

16 ... bi.xa7 17 .l:Ibl+ 'it>e2 18 .l:r.b2+

We3 19 .:b3+ 'it>e4 20 .!lb4+ Wd3
21 .i:tbl .!lf7

Also good is 21 . . .l:i.e7 with the

idea of 22 ... Ite l+.

22 bi.fl We2 23 1Ixf2+ Wxf2

White resigned.

Activating pieces also means acti

vating the king to coordinate with a
passed pawn.
Smagin-Naumkin
Moscow, 1983

Let's try to evaluate the position

-White has an extra pawn, but
Black has a strong passed pawn on
g3 plus an active rook plus the pos
sibility of activating his king-it is
Black who is playing for the win.
1 llg2 l:'th3!

Usually the rook is placed behind,

but here Black does not allow 2
'it>c2 to be played because of 2 .. . l:l:h2
winning.

2 l:i.gl Wd6 3 'it>c2 i;t>e5 4 'it>d3

g2+!
If 4 . . . Wf4, then 5 e5 ! 'it>xe5 6 'it>e3

with equality.
5 i;t>c4?

Correct was 5 'it>e2 .U.g3 6 'it>f2

The endgame with f and h-pawns
is drawn, but it is necessary to trans .l:lb3 7 'it>xg2 l:rxb2+ 8 Wf3 .!lb3+ 9
pose to it at once!
'it>e2 brxa3 1 0 .!lb 1 with a draw.
1 'it>g4 Wf7 2 'it>g5 !!b4 3 h6 .l:tbl

Black reluctantly decides to give

up a pawn-which he should have
done earlier. But now, psychologi
cally, he is not ready for defence.

4 xa4 Wg8 5 f5 :tgl +

Correct is 5 ... Wh 7 .
6 .!lg4! 1Ial 7 'it> g6 .:a2 8 f6 .!lal
9 f7+ Wf8 10 h7 l:i.hl 11 f6 Black
resigned.

5 ... z:tg3 6 'it>c5 'it>xe4 7 b4 'it>d3? !

Stronger is 7 . !!g6 with a win.
8 'it>b6
Or 8 b5 bi.g6 9 a4 b6 1 0 d5 We3
.

wmnmg.

8 . .llg7 9 a4 'it>c4 10 b5 Wb4 1 1

a5 .l:tg6+ 1 2 'it>xb7 'it>xa5 1 3 b6
l!xb6 14 i;t>c7 g6 15 'it>d7 'it>b4 16
i;t>e7 '>t>c4 17 'it>f7 l:r.g3 White
resigned.
.

Rook Endings 81

Eliskases-Levenfish
Moscow, 1936

Zugzwang-bad is 19 . . . f4 20
f6=.

20 f7 f4 2 1 g7 'itg5 White
resigned.
Azmaiparashvili-Kupreichik
Kuibyshev, 1 986

White has a pawn more but Black

has an active king and a far ad
vanced passed pawn. White must
play 1 l:!e7 ! c3 2 a4 lia 1 3 lic7 l:!a3
4 a5 e5 5 a6 d4 6 a7 d3 7 f4
Readers will ask-what has all
c2 8 .l:td7 with a draw according to
an analysis by Smyslov. However in this got to do with activating in a
rook ending. This is what it has to
the game he quickly played:
do with it. White has the exchange
1 lia5+? 'ite6!
for a pawn but Black threatens after
Sacrificing yet another one.
2 l:ta6+ d5 3 l:Ixh6 c3 4 lih8 . ..lid7 to seize the initiative. So
White decides to transpose to a rook
.l:Ial 5 l:!c8 l:txa3 6 h3
After 6 f4 g4 7 f5 e5 8 lic5+ ending a pawn down, but in the
d4 9 l:tc8 l:!.a 1 1 0 f6 l:tfl 1 1 llc6 process activating all his remaining
d3 12 :gd6+ c4 13 l:!c6+ 'itb4 ! , pieces.
and h e has to give up the f6 pawn.
1 .l:txc5 ! ! 'l'Wxc5 2 'fkxc5 bxc5 3
6 ...@d4 7 g4

Bad is 7 f4 gxf4 8 gxf4 c2 9 g4

.l:Ic3 +-.
7 ... lia5!

@fl @f6 4 e2 e6 5 d3 d5 6
'itc3 !

Now 7 lid l +! is threatened fol

lowed by c4.

6 ... c4 7 b4! .l:tc7 8 lid4+ e4 9

'Building a bridge'-the main
b6 .l:tc6 10 b5 :gc8 l l l:td7!
manoeuvre in rook endings.
With the threat of lic7.
8 f4 llc5 9 l:!d8+ 'ite3 10 l:1dl
No help is 10 l:1e8+ @f2 1 1 .l:Ia8
11 ...lib8 12 l:1c7! d3 13 l:rc6!
Here is the key move---defending
c2 1 2 l:1a l gxf4 1 3 'itxf4 c l ='ii' 14
llxcl l:txc l 14 g4 lic4+ 1 5 f5 the b6 pawn and keeping in his
g3 ! 1 6 g5 h4 1 7 g6 h5 1 8 g7 sights the c4 pawn, White wants to
take on a4 and obtain connected
lig4 + -.
1 0 ...c2 11 l:!.cl gxf4 12 gxf4 @d2 passed pawns.

13 l:tal cl= 14 lixcl lixcl ! 15

g5 e3 16 f5 e4 17 g4 'ite5 1 8
g6 l:!c6+ 1 9 @g7 :ga6!

13 ... f5 14 @xa4 c3 15 b5 g5 1 6
a 4 f4 1 7 gxf4 gxf4 18 a 5 e4 1 9
'itb4!

82 Rook Endings

The last finesse-now on 1 9 . c2

follows 20 'it>b3.
.

19 ... 'it>e2 20 a6 e3 2 1 a7 li:tf8 21

fxe3 Black resigned.
Van der Doel-Klovan
Gelsenkirchen, 1 998

White has a distant passed pawn,

but his cut-off king, indeed also the
centralised black king, gives Black a
decisive advantage.
1 .l:Ie7+ 'it>d4 2 l:!e6?

Correct is 2 .l:If7! .

2 d5 3 l:!xh6 'it>e3 4 lif6

..

Also bad is 4 lie6+ 'it>xf3 5 l:te 1

d4.

4 ...d4 5 l:!xf5 d3 6 l:re5+ 'it>xf3 7

f5+ 'it>e3 8 .:i.e5+ 'it>d4 9 l:!e8 d2
10 d8+ 'it>e3 1 1 .l:Ie8+ @f3 1 2
I:!'.f8+ 'it>g4 White resigned.

There will be cases when it is nec

essary to sacrifice all one's pawns
for maximum coordination of all the
pieces.
Each of White's remaining pieces
are clearly more active than his op
ponent's. Therefore any delay by
Black would be equivalent to death.

Pelletier-Rozentalis
Erevan, 1 996

1 ...l:!c8 ! !

Forcing the capture o f the pawn,

which allows Black, through the
opening of the d-file, to invade the
opponent's position.
2 .l:Ixa6 l:id8 3 I:!'.b6 l:td2 4 lixb5
xf2 5 a4 lixg2 6 a5 f5+! 7 xe5
f3 8 .l:Ib4 Drawn.
Smyslov-Epishin
Rostov, 1 992

1 ...g4 ! !

Inferior i s l ...f4 2 gxf4 gxf4 3

llxb4 e3 4 fxe3 fxe3 5 l:ta4 .:i.b2 6
b4! 'it>e4 7 l:!a8 'it>d3 8 l:ra3+ 'it>d2 9
li al !=.
2 llxb4 f4! 3 l::ta4 b2 4 gxf4 g3!
5 fxg3 e3 6 f5 'it>e4 7 f6 'it>f3 8 J:fal
!Ig2! 9 f7 e2+ 10 'it>el 'it>e3 White
resigned.

The rule of the two weaknesses

the possibility of transferring an

Rook Endings 83

attack from one weakness to another

until the time comes that the
opponent cannot defend all his
weaknesses-is an exceptionally
universal one.

Fercec-Mikhalchishin
Nova Gorica, 1 99 7

Nikolic-Movsesian
Polanica Zdroj, 1 996

Black clearly has the more active

king and rook, and the method of
realising the advantage lies in the
creation of two weaknesses and at
tack on them.
I . ..liti>e5

1 h5! gxh5

Less logical is l ...g4, and White

can avoid the drawn pawn ending. 2
.l:tf4 .l:xf4 3 exf4 @f5 4 '>te3 gxh3 5
gxh3 c5 6 a4 b6 7 b3 a5 8 'it>f3 d4 9
cxd4 ! ? cxd4 1 0 h4 g6 1 1 'ittg 3 'iti>e4
1 2 'ittg4=. It is equally a draw after 9
c4.

2 !:lxh5 .l:.i.g8 3 .l:.i.h4!

More principled is 2 g4 g6 3 b3 f5
4 .Ug3 f4 5 l:.f3 c5 +.

Even here the two weaknesses on

a6 and c6 would not mean anything
were White not to have chances of
organising another weakness on the
other flank.
If L.g5, then after 2 'itie4 the king
breaks through to the black pawns.

2 b3

2 ... g4 ! 3 ll:fl
Weak is 3 xh6? :xg4 4 ltxf6
3 ltg3? loses at once because of
:lg3+ and the endgame is drawn,
while the passivity of the rook is 3 ... f5. The attempt 3 hxg4 !? de
provisional upon the arrival of the serves attention, but even then
3 ... .l:i'..xg4 4 g3 b5 5 b4 ll:g6 6 a3 lih6
white king to the defence.
3 ... '>td7 4 'itt e4 '.te6 5 liti>f'3 :l'..h 8 6 with the idea 7 ....Uh3 leads to a win
.Uh5 Ith7 7 'it> g3 .l:i'..d 7 8 !fa5! :a7 9 ning position for Black.
'itth4 ..ttn 1 0 'itth s 'iti>g7 1 1 rs 'iti>h7
12 ltcS .l:i'..c7 13 a4! 'iti>g7

3 gxh3 4 gxh3 lth4 5 .:tf'3 b5 6

'>te2?
.

A decisive mistake. Better is 6 b4

g5 7 ltg3 @f5 8 ll:f3+ 'iti>g6 9 l:.g3 f5
14 b5 axb5 15 axb5 .l:.i.b7 16 bxc6 intending 1 0... g4 - + .

After 1 3 . . ..l:i'..b 7 1 4 .Uxc6 .l:i'..xb4 1 5

:lxf6 White wins easily.

l.c7 17 ll:cl l:.c8 18 c7 'it>t7 19 ltc6

'iti>g7 20 lifi>h4 'iti>ti 21 '.tg3 Black
resigned.

6 g5 7 '.tf2 a5 8 .l:.i.g3
.

If 8 'iti>g3 .Ue4 9 '.tf2 b4 - + .

8 ... c5 9 'ifi>g2

84 Rook Endings

If 9 a3 , then 9 . a4 1 0 'it>e2 axb3

1 1 cxb3 h8 with the idea 1 2 . . ..l:!a8
..

- +.
9 ... b4 IO cxb4 axb4 1 1 .!if3 .l:Ih8
12 .llfl .!la8 13 J:Ial 'it>e4- + 14 'it>f2
f5 15 'it>e2 J:ih8 16 .l:!hl .!lh4!

It was still not too late to let the

win slip. 1 6 ... g4? 1 7 h4.

17 .!lgl .!lxh3 18 :xg5 1Ih2 19

Wdl c,t>xe3 20 .llg3+ d4 White
resigned.
Typical mistakes in rook endings

In rook endings there are a great

many typical methods of play, but
also, naturally, also a great many
typical mistakes. We acquaint you
with the most typical of these.
Making Passive
Ilivitsky-Taimanov
USSR, 1 955

1 f3?

He should play 1 h4 ! , boxing in

the black king.
1. ..gS! 2 c,t>g3 'iitg6 3 l:!c2?

Again White sticks to waiting tac

tics. The correct path was 3 :i.d6+!
f6 4 h4! gxh4 5 'itxh4 .l:Ixa2
(5 ....l:.xf3 6 .l:r.a6=) 6 f4 I:!'.a4 7 f5+
'iit g? 8 Itd7, maintaining equality,
Levenfish.
3 f6 4 l:th2?
..

This was the last time White had

the possibility of activating his rook:
4 J::! c 6! I:!'.xa2 5 h4! gxh4 6 'it>xh4
with equality.
4 h5 ! 5 J::!c2
..

Now there is already no saving

himself :
(a) 5 h4 hxg4 6 hxg5 f5-+;
(b) 5 gxh5+ 'it>xh5 6 h4 g4 7 1If2
f5 8 'iitf4 .!la4 9 'it>xf5 g3 - + ;
(c) 5 gxh5 Wxh5 6 :t:Ic2 f5 7 .l:!d2
f4+ 8 @g2 'iio> h4 intending ... l:!c3,
a7-a5-a4-a3 .l:Ic3-cl -b l -b2 - +
(d) 5 h4 hxg4 6 'it>xg4 f5+ 1 'iitg3
g4 8 .!lf2 'it>h5 9 J:ifl J:Ia4 ! 1 0 fxg4
J:Ixg4+ 1 1 @f3 'it>xh4! 1 2 .l:.c l 'iitg5
1 3 a3 l:!a4 14 .l:!c3 'it>f6 1 5 .!lb3 a5
- + , Levenfish.
5 h4+ 6 f2 a6 7 .l:!b2 .l:Ic3 8
'iitg2 a5 9 J:if2 .!la3 IO 'iitfl Wf'l 1 1
f4 gxf4 1 2 .!lxf4 'it>g6+ White
resigned.
..

Vaganian-Schlosser
Germany, 1 994

1 . ...l:!b7?

An analogous mistake. He should

activate his forces at once by 1 ... h5 !

2 g4! 'it>g7 3 'it>e2 J::!e7+ 4 Wf3

:.c7 5 h4 h6? ! 6 'iitg3 .l:!c3+ 7 f3
:tc7 8 'iitf4 :b7 9 h5 .l:Ib4+ IO Wg3
gxh5

If 1 0 . ..l:tb7 White gains the ad

vantage by 1 1 hxg6 fxg6 1 2 f4 fol
lowed by Ita5-a6, f4-f5 .

Rook Endings 85

1 1 gxh5 l:!'.b7 1 2 'i!tg4 1.tbl 13 f4

1 . ..ki.e5 2 h4 h5 3 .a'.c4 l:l'.f5 4 l:i.e4

After 1 3 llxa7 Black activates 'i!ta6 5 .I!e7 g5 6 hxg5 fxg5 7 ti.es

himself by 1 3 ...l:i.g l + followed by h4 8 a'.a8+ 'i!tb7 9 ti.gs 'i!tc6 10
l!g6+?
J;.g5.
A mistake. He should activate the
13 ... 1.tb7 14 >f5 lte7 15 a4 'fri.c7
1 6 'Ot>g4 l:.d7 1 7 lta6 Xlb7 1 8 a5 king by I O c4 ! .

lic7 19 f5 l:.c4+ 20 >g3 ltc5 21

@f4 :Ic4+ 22 'Ot>e5 l:i.c5+

Worthy of attention is 22 ... Jlh4 ! ?

(Schlosser) 2 3 ltxa7 lixh5 24 a6
l:lhl 25 llb7 l:tel + 26 \td6 llal 27
a7 'it>f6.

23 e4 ltc4 24 'i!td5 Jlh4 25 f6+

h7 26 !Ixa7 %Ixh5 27 'i!tc6 \tg6 ?

Better is 27 ...ltf5 !=, Schlosser.

2S a6 Il:a5

White wins also in the event of

28 ... l:th l 29 .:td7 Jlal 30 @b6.

29 'iifb 6 Ital 30 ltd7 Xlbl+ 3 1

c7 @xf6 32 ltd5 ! ! ltcl+ 3 3 @b6
'ii?g6 3 4 a7 l:cS 35 l:ta5 h5 36
aS=W a'.xaS 37 .l;; xaS @f5 3S 'Ot> c5
e4 39 'it>d6 @f4

1 0...@d5 1 1 llxb6 'lt>e5?

Now Black meets mistake with

mistake. Activity of the king is the
main thing in rook endings. Black
maintains the advantage by shoul
dering the opponent's king with
l l . . .d4! . Now however it' s a
draw.

12 'lt>c3 'iif f4 13 Zib4+ 'lt>g3 1 4

l:tg4+ 'lt> f2 1 5 'it>d2 Drawn.
Abramovie-Nikolic
lgalo, 1 994

On 39 ...f5 winning is 40 l!a4 with

the idea of@e5 +-.
40 !:thS! f5 4 1 @d5!

41 ltxh5? would be a mistake be

cause of 4 1 .. .e4 with a draw.

41 ...@e3 42 :es+ @t'3 43 'it>d4 f4

44 @d3 h4 45 lth8 'Ot>g3 46 'i.t;e2
g2 47 ltgS+ 'iif hl 4S @n Black
resigned.
Grunberg-Brunner
Germany, 1 992

1 . ..llf6?

Allowing White to create counter

play on the king's flank. He should
restrict this by l ...f5 ! 2 g4 hxg4 3
fxg4 fxg4 4 @g3 'i!td8 5 xg4 'lt>c8
6 .!Ig7 b5 with advantage to Black.

2 g4 'it>d8 3 g3 cs 4 '11e7 b5 5
lle5 a'.b6 6 gxh5 gxh5 7 :!xh5 b4 8
ltd5

8 l:l'.c5 is rather worse.

8 ... b3 9 ltdl b2 10 llbl 'it>d7 1 1

@ f4 e6 12 g5?

White does not exploit the oppor

tunity presented by his opponent.

86 Rook Endings

The only chance for him was 1 2

ct;>e4! f6 1 3 @d4 Wg6 1 4 @c3
ct;>h5 1 5 I:!'.xb2 .l:Ixb2 16 @xb2 xh4
17 @c3 ct;>g3 18 @d4 xf3 1 9 ct;>e5 !
with a probable draw. Now, how
ever, Black has no difficulty realis
ing his advantage.

Shirov-Morozevich
Amsterdam, 1 995

12 ... l:1bS+ 13 @h6 @f6 14 hS

.l:Ib3 lS @h7 l:1b8 16 f4 1::rb3 17
@gs J::rb s 1 s h6 ct;>g6 1 9 rs+ @r6 20
Wh8 llb3 21 'it>g8 b7 22 @h8
@gs White resigned.
Marie-McNab
Hastings, 1 995

1 l:tg4?

Better is 1 .l:Igl with the idea of

connecting his pawns in the rook
ending by f2-f4-f5.
1 . @e7 2 .l:Ia4? l:ta8 3 f4 Wxe6 4
Wg3 @dS S @g4 @xcS 6 WgS bS
.

This tempo was made possible as

a consequence of 1 lig6-g4.

1 ... aS 2 h2?

Correct was 2 f3 ! a4 3 @g3 l:ta 1 4

f5 ! ct;>xf5 (if 4 ...g5 5 l:l'.a6+ @xf5 6
lia5+ with a draw) 5 l:txf7+ We5 6
lie7+ @d4 7 lia7 a3 8 Wf4 a2 9
a8 with equality.
2 .. f3! 3 l:!a6+ ct;>rs 4 l:!.xaS+

7 1:1'.al llg8+ 8 @f6 llf8+ 9 eS

l:te8+ 10 WdS l:td8+ 1 1 'it>eS l:l'.e8+
12 @dS .l:i'.d8+ 1 3 @e4 .l:Ie8+ 1 4 Wf3
lth8 1s llhl @cs 16 rs @d6 17
@f4 a4 18 @gs @e7 19 f6+ @f'l 20
llbl l:!xh3 21 lib7+ ct;>rs Drawn.
Vaulin-Voikhovsky
Russia, 1 99 7

@g4! S Wg2 l:l'.xf4 6 lla7 Wxh4 - + .

Analysis.

Technically weak play

in rook endings

In the following position White,

of course, has a winning position
but , it is not so easy as it seems at
1 e4? .l:Ia3+ 2 @g2 fxe4 3 llf4
first sight.
'ii>eS 4 l:l'.xg4 @d4 S llg8 @e3 6 .l:Id8
ct;>e2 7 b8

Rook Endings 87

The a-file is inaccessible.

7... l:ra2 8 .:i.bl e3 9 @g3 l:1a8 1 0

b2+ Wd3 1 1 l::tb3+ 'it>d2 12 J::rb2+
i;tic3 13 l:Ib7 .l:!e8 White resigned.
Krasenkov-Iskusnik
Russia, 1 996

with an easy draw and therefore

played
1....l:Ia2?,

But he did not reckon on

2 h4! ,

after which White has a winning

endgame, whereas he could have
made a draw by I ....l:Ib l ! 2 h4 l:1gl +
3 'it>h3 J::rh l + or I . ..1::rb 8! , preparing
against h3-h4.
Materialism instead of activation
Schmitdiel-Mikhalchishin
Berne, 1 994

.
.i
.
.
.

1...'it>f5? After the simple l ... .l:If4

there is an easy draw. 2 c6 @e6 3
c2+ -.

A narrow spectrum of noticing

the opponent's threats
(one threat he sees, the other not)
Kozul-Mikhalchishin,
Bled, 1 996

. , ,

.
f
. .,, , v,. .
. !. . -

.;!,,

A
,,; " fj, 2
D.
,,,---.,,, ,v,.

"

, r,

1 J::rd l?

He should think about the liquida

tion of his opponent's activity (re
stricing the mobility of the pawns)
by I .l:Ia5 ! 'it>f8 2 c,t>g3 @e7 3 c,t>D
@d7 4 'it>e3 'it>c6 5 @d4 with excel
lent chances of a draw.
1...'it>f8 2 J::rx d6 as 3 'it>g3?

White's only chance was to strug

gle against the a-pawn by activating
his c-pawn: 3 c5! @e7 4 l::td 5! a4 5
c6 a3 6 c7 a2 7 J::i.d l , with a
draw-clearly better is 4 ...@e6!
with chances of victory. Now,
however...

3 a4 4 'it>f3 a3 5 dl 'it>e7 6 @e3

a2 7 al 'it>d6 8 @d4 a4 9 f4 h5
10 g4 h4 1 1 g5 g6 and White
resigned.
.

Black saw only one threat I .l:tf5+

and prepared to repulse it by
1 .. .l:rb3+ 2 @g2 .l:!b2+ 3 @fl .l:Ib3

88 Rook Endings

Mikhalchishin-Stangl
Dortmund, 1 992

A well known drawing mecha

nism without the b2 pawn-the
rook goes to the sixth rank (f6) and
upon the approach of the king to b5
begins an endless checking se
quence. But in the game there
followed ...
1 ....l:Ib4? 2 .!lc7 .l:Ia4 3 lic8+?

For what reason? Why not 3 .!lc6

xh7 4 @xb2, winning easily.
3 ...@xh7 4 .!lc6 .l:Ib4?

After 4 ... @g7 5 xb2 @fl 6 b3

.l:Ial 7 @b4 @e7 8 @b5 @d7draw.

1...l:td3?

A technically incorrect attack

necessaty was 1 ...l:td 1 + 2 @g2 lia 1
3 .l:Ia7 J:Ixa3 4 b5 a5=.
2 a4 aS 3 bxaS J:id4 4 a6 .l:Ixa4 S
a7 h6 6 @n gS

Also no help is 6 ... lia2 7 e l ,

and the king goes to b 1 .

7 @e2 .l:ta3 8 @d2 @g6 9 @c2

gxh4 10 gxh4 @rs 1 1 @b2 .l:Ia6 12
@b3 lial 13 .!lc7!

S a7 .!la4 6 "IJ,.c7 @g6 7 xb2 @f6

8 @b3 J::ra l 9 @b4 'it>e6 1 0 @bS
@d6 1 1 .!lc6+ @dS 12 .!la6 :!:!.bl+
13 'it>as @cs 14 .l:Ic6+! and Black
could resign.
Inaccurate technical execution
Ribli-Mikhalchishin
Germany, 1 993

Preventing 13 ... g4 because of

14 .!lc4+ and 15 J::ra4.
13 l:ra6 14 @b4 and Black
..

resigned.

Ignorance of typical drawing

mechanisms
Ivanchuk-Lautier
Horgen, 1 996
1 ...l:!.b2?

Occupying the b2 square, which is

needed for the king-correct is
1 ...J:ic2 ! , and Black has no difficul
ties. Now, however, he lacks a
tempo.

2 hS a2 3 @g2 @c3 4 'itig3 'it>b3 S

f4 J:ib l 6 fS al='iV 7 l:txal J::rx al 8
g6 @c4
8 ... hxg6 9 fxg6 c4 1 0 @2 ! + 9 f6 hxg6 10 f7 ! Black resigned.

Rook Endings 89
Typical Rook Endings

In the endings it is difficult to find

new-everything has
anything
already been played. But among
various positions resembling one
another it is still possible to find
great differences and great simila
rities at one and the same time.
Our attention was attracted to the
ending from the game
Hiibner-J.Polgar
Dortmund, 1 99 6

1 1 .!lf8 ! ! lhf8 12 exf8='YW+ Wxf8

1 3 Wd6 with a won pawn ending.
2 .l:.a6 Wf7 3 .U.f6+ 'it>g7 4 e6 l:id l
5 .!if7+ 'it>gs 6 We4 .!l gl 7 li:tf3

.lle l +

After 7 ... 'it>g7 8 e7 lie 1 9 .!le3 an

easily winning pawn endgame is
again reached: 9 ... l:txe3+ 1 0 <;fo>xe3
@fl 1 1 We4 We8 ! 12 'it>d5 'it>d7 1 3
e8='ii' ! , etc.
8 'it>d5 <;fo>g7 9 .!lf7+ <;fo>g8 10 <;fo>d6,
and Black wins.
This reminds me of a very similar
endgame that I had a year earlier.
Here Black resigned.
1...<;fo>f8

After 1 . . .l:!.b4 Hubner gave the

following variation: 2 .l:!a6 (also
winning is 2 .l:.g7 b6 3 'it>e4 .!la6 4
.!lc7 'it>d8 5 .!lc5 .l:!a4+ 6 'it>d5 .l:!g4 7
<;fo>d6 .l:!xg5 8 l:ta5 + -) 2 .. .Wfl 3
.!lf6+ 'it>g7 4 e6 .l:!b5 5 'it>f4 J:Ia5
(5 ...b4+ 6 We5 .!lb5+ 7 'it>d6 lhg5
8 .!ifl+ <;fo>g8 9 .l:!a7 .!lg l 1 0 l:l.'.a8+
@g7 1 1 e7 + -) 6 e7 l:l'.a8 7 'it>e5
.l:.a5+ 8 'it>d6 .!la6+ 9 'it>c5 .l:!a5+ 1 0
@c6 .!la8

Barle-Mikhalchishin
Slo venia, 1 995

90 Rook Endings

There followed:

'it>f6 obtaining a draw. Correct is 1

e5 ! .l:Ie7+ 2 f4 l:ta7 3 I:!'.b6 l:tc7 4
After 3 Wf2 .l:Ib2+ 4 @fl arises a .l:tf6+ Wg7 5 e5 reaching an easily
position which is very similar to the winning position, known from the
previous one, with the following previous examples.
1.. . .!ld7!
variations: 4 ... e4 5 J::rc4+ @e3 6
The only chance-the threat was
J::[.c 3+ 'it>d4 7 l:ta3 .l:r.b4 (7 ... l:tc2! ?) 8
@f2 e4 9 J::ra2 e3+ 9 e2 e4 1 0 2 e6, followed by e5, 1::rd 5 and
.!lc2 J:id4, again threatening a .!ld7.
2 'it>e4 J:idl 3 .l:Ib7+ 'it>e6 4 l:tb6+
transfer to a winning pawn
endgame.
We7!
3 ...l:rb2 4 lk8 l:!.g2 5 I:!'.f8+ 'it>e6 6
To achieve a draw it is necessary
'it>e4
to give up yet another pawn.
5 l:txg6 J:iel+ 6 @f4 .!ifl+ 7 @g3
After 6 J:ie8+ @f6 7 I:!'.f8+ e7 8
.l:If5 We6 Black wins easily.
l:!.gl+!
6...Ibg3 7 J:ie8+ '\t>f6 8 J:Ixe5
In the game 7 ... .Ile 1 8 J:if6 l:!e4 9
White reckoned on 8 ... .!le3+? 9 I:!'.f4! was played, with a win.
8 'iitf2 l:tg4!,
@xe3=, but more tenacious was 9
And a'ccording to an analysis by
l:rf8+ 'it>g7 9 .l:Ia8 (9 l:tfl g6 1 0
Wxe5 J:If3 - + ) 9. . ..l:tf3 1 0 .!la6 l:tf6, M.Yudovich-it's a draw.
and then the black king goes to g6
More complicated variations with
and g5, winning.
9 .. JU3 1 0 l::t e8 @gs 1 1 .l:Ih8 l:!.ti the addition of the h-pawns are met
12 @e3 g3 White resigned.
in the game
1...g4+ 2 hxg4 hxg4+ 3 e3

A similar endgame was analysed a

very long time ago.

Akopian-Almasi
Ljubljana, 1 995

Filipov-Kopatsny
USSR, 1 968

There followed:
1 ...'iitf4

Inferior is 1 . . . .!lh 1 , since after

There followed:
.l:Ia4 he cannot play 2 ... .!lxh2?
1 e5?
'it>g l .
If 1 .l:tb6 a5+ 2 Wd6 l:txg5 3 e5
2 J::rc2
l:tg i 4 'it>d7 (4 .l:tb7+ 'itf8, and there
After 2 g3+ hxg3 3 hxg3 'it>g4
is no win) 4 .. J::td l + 5 :d6 e l 6 e6 !te2 'it>f5 5 'it>g2 J:ib3 6 f2 J::!d3

2
3
4
7

Rook Endings 91

@g2 e3 the threat of .l:!d2 is

unstoppable.
2 ...@rs

On 2 ...l:ih I Almasi gave 3 g3+

hxg3+ 4 hxg3+ <tig4 5 Itc5 (5 l:ic4
:!Ih2+ 6 'iiitg l l:ie2 - + ) 5 ... l:ih2+ 6
@g l .tie2 7 l:l'.e5 l1e3 8 <tih2 (8 @:f2
!l:f3+ 9 liite2 .l:tf5 ! - + ), but this po
sition is not winning-for example,
8 ...l:te2+ 9 'itigl e3 IO liitfl .tI:f2+ 1 1
@g l z:!'.f3 1 2 <tih2!
3 .l:tc5+ @g4 4 Itc2 l:tb3!,

Preventing g2-g3.
5 Itel h3 6 .tia2

Weaker is 6 ,gxh3+ @xh3 -+.

6 l'!bl + 7 f2 l:[hJ 8 l:Ie2
Bad is 8 lila4 \tf4 9 g3+ @f5 - + .
..

8...@f4

But not 8 ... 1:.xh2? 9 ltxe4+

because of 1 0 'itg3 ! .
9 g3+ lites 1 O 'ite3 l:icl

There is nothing in 1 0 ... l1gl 1 1

lta2 .tlg2 1 2 .tla5+ li>f6 1 3 ltxe4
.l:txh2 1 4 @'3 .

on g3, which also, however, leads to

a draw.
15 Ita5+ 'itid4 16 l'.Ia4+ l:!c4 1 7
l'.Ixc4+!

The transfer to a pawn ending is

here the simplest way to a draw.

17 ... <tixc4 1 8 <tie3 litd5 1 9 <tie2

Wd4 20 Wd2 e3+ 21 @e2 <tie4 2 1
@et <ti n 2 2 @fl e2+ 2 3 Itel with a
draw.

Both of the young grandmasters

played rather inaccurately, and par
ticularly White who allowed his op
ponent to advance with threats
against the g and h-pawn. The right
way to achieve counterplay had al
ready been shown in old games.
Tiets-Forsberg
1 6th World corr. ch. 1 985-91

1 1 l1a2 I:c3+ 12 'ite2 'itd4 1 3

l:[d2+ @es 1 4 l1a2

J ...@dS 2 Itf6 e4+ 3 @e3 l'.Ib3+ 4

'it>f2 l'lb2+ 5 @fl

1 4...g4?

After the correct 1 4 ... @f5 1 5 liit:f2

<it>g4, it seems that there is an irre
sistible threat of . . J:td3, e3 and l1d2
transferring to a winning pawn end
ing-however the pawn ending is
drawn, and, secondly, with the
white king on e2 there is no other
plan besides the sacrifice of the rook

This passivity is forced, since af

ter 5 @g3 h4+! 6 @h3 e3 the pawn
cannot be stopped.
5 ...@d4 6 l:lg6 litd3

After 6 ...'ite3 ? ! 7 l:[g3+ @d2 8

'it.>:f2! :itb8 9 .tia3 '.f8+ 1 0 'iiitg 3 e3
1 1 11a2+ it is very difficult for the
king to escap e the checks.
7 l'ld6+ 'it>e3 8 ltdl l:!.f2+ 9 <tigl
l'.Id2 10 :Iel+ litd3 1 1 @fl e3 12
:al t:if2+ 13 ltgl Itb2 1 4 @fl
@d2 15 g3

92 Rook Endings

White defended i n the following

way:
1 .!le7 d3 2 g4!

An attempt to create very quickly

a passed pawn on the king's flank.
2 ...J::!d 8

After 2 ...d4 3 f3 d8 4 l:!a7!

the position is analogous to the
game.
3 'it>f3 f8+ 4 'it>g2 .!lf4
After 4 .. . e2 5 l:id7+ 'it>el 6 d6!
.l:If2+ 7 'it>gl f8 8 h3 the threat to
Here Black has several different take on h6 saves White.
plans:
7 J::!a 7!
(a) 15 ....l:Ic2 1 6 h3 'it>d3 17 l:!a3+
Transposing the game to a well
'it>d2 (after 1 7 ... 'it>e4 18 a5 'it>f3 1 9 known position with a rook attack
:i.f5+ xg3 20 l:!xh5 with a draw) from the long side.
5 ...e2 6 .!la2+ 'it>d1 7 .l:Ial+ Draw.
1 8 :!al l:!b2 19 gl 'ite2 20 g2
l:id2 2 1 g4 'it>d3+ 22 fl h2 23
.:i.a3+ and a draw was agreed.
The correct plan of defence for
(b) 1 5 ...'it>d3 , and now
this class of position was demon
(b l ) 1 6 'it>g l ? 'it>e4 1 7 l:!a8 :!b l+ strated in the game
18 g2 e2 - + ;
(b2) 1 6 h3 'it>e4 1 7 a8 'it>f3 1 8
Portisch-Pietzsch
Madrid, 1 960
.l:If8+ 'it>xg3 1 9 .l:Ie8 .:i.f2+ 20 'it>e 1
'it>f3 2 1 l:if8 'it>g2 22 e8 f3 23 h4,
and the win is not easy;
(c) 1 6 ... h4! ?, and, for example, 1 7
gxh4 'it>d3 1 8 'it>g l 'it>e4 1 9 h5 .!lb8
20 h6 l:!g8+ 21 @fl 'it>f3 22 .!la3 ! ,
and White maintains equality.

l :-1!
=

-:,:'

An analogous ending is ...

Chiburdanidze-Gal!iamova
Groningen, 1 99 7

a . . ."----

1...:i:!.b l ! 2 h4 .l:In+ 3 '\ties l:!gl 4

l:!c7+ 'it> g6 5 .!lc6+ f7 6 'it>f4 :i:Ifl +
7 'it>g5 h6+!

Weaker is the more passive

7 ... l:!.e l 8 l:!c7+ @f8 9 '\t>f5 .l:Ifl + 1 0
e6 .l:If6+ 1 1 'ite5.
8 'it>h5 :i:!.f4 9 e5 l:!e4!

The black rook has gained maxi

mum activity, but the position still
requires accuracy.

Rook Endings 93

10 e6+ 'it>f6!

Not 1 0... 'it>e7 1 1 c7+ 'it>f8 12

c8+ 'it>e7 1 3 l:i.g8 + -.
1 1 g5+ i;i>e7!

'it>e4 ! 6 ltc4+ c,t>d3 followed by

...l:i.f6 and the advance of the a5
pawn.
4 'it>g2 !td7 5 .l:Ixe5 :i.d2+ 6 'it>gl

Again not 1 l . . .'it>f5 ? 12 l:!c5+ 'it>h3 7 lig5?

White sees a defensive idea, but
'\t>xe6 13 .l:!c7 g6 14 'it>xh6 + - or
l l ...h5 12 hxg5 'it>e7 1 3 'it>g6 + -. his execution of it is inaccurate. Af
12 .Uc7+ 'it>f8 13 .l:!cS+ @e7 14 ter 7 e6! l:i.g2+ 8 'it>h l l:!xg3 9
l:t.g6 ! Black can give up trying to
lic6
After 14 l:!g8 hxg5 1 5 hxg5 .l:Ixe6 wm.
7 .l:Ig2 s 'it>hl :!:i.f2 ! ! 9 'it>gl l:If6!
1 6 lixg7+ i;(f8 draws.
14 ... @fS 15 l:!b6 'it>e7 16 I:!'.b7+
Now, however, it's zugzwang.
..

i;(f8 17 .llf7 i;i>gS 1 S lie7 @f8 1 9

gxh6 gxh6!

10 .l:Ia5 l:!f3 1 1 g4 l::!.g3+ 12 'it>hl

i;i>xg4 White resigned.

20 l:tf7+ @es 21 l:tf6 'it>e7 22

!txh6 .l:r.e5+! and a draw.

Bogoljubow-Rubinstein
London, 1 922

Clearly not 19 ... 'it>xe7? 20 hxg7

J:e5+ 2 1 i;i>g4 + -.

Trabattoni-Barlov
La Valetta, 1 979

Rubinstein is considered the

' king' of rook endings, but his play
in the following ending leaves us
profoundly bewildered.

1. lib7!
The main task is to drive off the
1 ...g6?
Correct is 1 . ..g5 ! , transposing to
rook from the sixth rank and cover
the king against checks along the the position in Barle-Mikhalchishin.
f-file. Then follow preparations for
2 ltbS lia2?
the advance of the e5 pawn.
Clearly better is 2 ....l:Ia3+, gaining
2 lia6 l::!.f7 3 :!:!'.as
some tempi.
Another possibility would have
3 .l:If8 'it>g7 4 !tes 'it>f7 s l::!.b s lih2
been 3 I:!'.a4 lif6 ! 4 !l:b4 lta6 5 .l:Ic4 What is he doing? He can't take the
e4! 6 l::!.c 5+ 'it>g4 7 l:tg5 + 'it>h3 8 @fl pawn!
.l:!f6+ 9 'it>e2 .l:!f3 ! , winning.
6 :cs .l:!a2 7 h4 l:!a7? S 'it>f4 @f6
..

3 'it>g4?!
..

9 .:.rs+ 'it>g7

An interesting plan, but better

And a draw was agreed. Rubin
looks 3 ...f6 4 .:i.b5 lia6 5 .l:!c5 stein's worst endgame!

94 Rook Endings

The fact that matters are not quite

so simple is shown by the game
Fischer-Geller
Curacao, 1 962

ltal @g6 22 .tlbl .l!te5 2 3 @d4 'itf6

24 .tlel?

Stronger was 24 .tlfl+.

24.. Jta5! 25 :l.xe4?!

Again stronger is 25 :.fl .

2S ... 'i.t>fS ! 26 l:e8 'itg4! 27 @e3

@g3!

And in this theoretical endgame

White resigned.

The last game did not answer sev

eral questions of defending the end
game with the e and gpawns. In
modem chess there is one more very
interesting example.
Balashov-Ulibin
Uzhgorod, 1 988

1 g5+? !

An extraordinarily crucial deci

sion-it would be simpler to wait
with 1 l:.c7 or I llb 5
.

1. .. hxg5 2 hxg5+ @g6 3 !1.e7 !1.e3

4 'itf2?

Not to this side--i t was necessary

to exchange the g5 pawn for the e6
pawn and not stick the king's head
out. Therefore 4 @h2! r!e5 5 @h3
leads to a draw.
4 ... e5 5 @13 llf5+ 6 'ite3
After 6 @e4 :.f7! 7 :.xe6+ 'itxg5
the win is easy.
6 ...eS 7 'ite4 lhgS 8 !1.e8?

1 llf8

The preliminary I @f2 does not

threaten 2 e5 because of 2 ...lla5 3
.tle4 @f5.
1 ... lta2+ 2 @13 lla3+ 3 'i.t>f2
l:i.a2+ 4 We3 Il'.a3+ S @d4 lla4+!
Not 5 ...ltxg3?? 6 e5 +-.

Preferable is 8 :a7! Wh5 9 :al

with the idea after 9 .. :g2 to play
I 0 We5 g5 1 1 'ite4 though 1 1 ... @g4
wins. The general idea to win with
the g and e-pawns consists of giving
6 @dS J:.a5+ 7 'itc6 .a:a6+ 8 'itb5
up one p awn and transposing to a
theoretically winning endgame.
l:.a3 ! 9 .:.rs+
8 ... llgl ! 9 'itt3 llfl+ 1 0 'itg3 :l.f5
On 9 e5 :!'.e3 1 0 .li!e8 'i.t>f5 1 1 g4+
1 1 llb8 @gs 12 lte8 'itf6 13 ltf8+ 'itxg4 1 2 @c5 Wf5 draws.
.

'ite6 14 lte8+ 'i.t>d5 lS :as l:!:f7 16

'iPg4 !1.e7 !

When the rook is placed behind,

then half the business is done.

16 :.as+ We6 17 .tla6+ 'i.t>f7 1 8

@13 lle6 1 9 l:.a8 e 4 20 @e3 g 5 2 1

9 ... 'itg6! 10 g4 lte3 1 1 eS .:e4 1 2

'i.t>c6 :.xg4 13 ltf6+ 'i.t>g7 14 d7
lla4

A drawn ending is reached with

the weaker side's king on the
kingside.

Rook Endings 95

15 li c6 l:!a8 16 l:!.c8 a7+ 1 7 .l:Ic7

a8+ Drawn.
How many roads lead to Rome?

We have already repeatedly said

that in the endgame there are usu
ally two paths (moves)-right and
wrong. But at times the number of
paths is surprisingly greater and it is
exceptionally difficult to find the
right one.

easily wmmng and described in

every book on rook endings;
(d) l ... 'it>d3 ! (this was the only
possible continuation) 2 .:i.fl (after 2
l:Ie5 'it>d4 3 :b5 @c4 ! the rook is
continually hounded) 2 ... @e2 3 :f4
'it>e3 4 l:ta4 (there is nothing else)
4 ...lixf5 5 a6 l:if8 6 a7 .l:Ia8 7 'it>xh3
'it>d3 (similar to variation (b), but
the cut-off king has proceeded one
rank further which is of decisive
significance) 8 @g4 @c3 9 \t>f5
@b3 1 0 al @b4 1 1 'it>e6 'it>c5

Beliavsky-Azmaiparashvili
Portoroz, 1 99 7

The black king has six(!) possible

moves, but only one(!) leads to a
draw.
(a) l ...@f3 (f4) loses because of 2
:al l:!a6 (he cannot draw by taking
on f5) 3 f6 winning easily;
(b) 1 . .. 'it>f2 2 l:!a 1 .:i.xf5 3 a6 :f8 4
a7 .l:Ia8 5 .ll a3 ! (a very important
moment to cut off the king) 5 ... 'it>e2
6 'it>h3 'it>d2 7 Wg4 c2 8 'it>f5 'it>b2
9 :a6 @b3 1 0 'it>e6 @b4 1 1 'it>d6!
(shouldering away the king and
preparing to set up a mating net)
l 1 . . .'it>b5 12 :gal @b6 1 3 :i:Ibl + 'it>a6
14 'it>c7 ! 1ixa7 15 'it>c6, and White
wins;
(c) l ...'it>d2? 2 l:!.e5 @d3 3 a6!
l:txa6 4 'it>xh3 'it>d4 5 .l:Ie6, and
Black resigned, the rook ending is

(After l l . . . 'it>b5? 12 'it>d6 a posi

tion is reached from variation (b),
while exerting control over the d6
square is the key to evaluating the
position) 12 'it>d7 b6 1 3 l:!.b l +
'it>c5 ! (the point) 1 4 :b7 :h8 !, and
a theoretical draw has arisen.
Ward-Baburin
Isle ofMan, 1 99 7

96 Rook Endings

White has an extra pawn but

White's cut-off king plus the strong
passed e4 pawn and centralised
Black king means that it is only
Black who can play for the win.
Baburin assessed the position as
winning for Black. Let's have a
look.
(a) 1 .l:!b2 l:!.c7 ! 2 'it>b3 'it>e5 3 l:rc2

Branicki-Sefc
Prague, 1 955

lid7! 4 'it>c3 e3 5 f6 We4 6 l:!.g2

.l:!c7+ 7 'it>b2 J:if7 8 'it>c2 lixf6 9
l:1g3 (On 9 Wd l , 9 ... 'it>d3 is unpleas

ant. Instead, without the h2 pawn,

White saves himself by means of
the stalemate 1 0 l:rd2+!) 9 !!c6+ 10
...

'it>dl 'it>d3 1 1 h4 .lla 6 12 Wcl :Ial+

Here there are several ideas for
13 'it>b2 J:ihl 14 h5 .l:!xh5 1 5 'it>cl + White to realise his enormous ma
l:thl+ 16 'it>b2 'it>d2 White terial advantage, but only one of
them leads to its objective.
resigned.
(b) 1 f6 'it>e5 2 l:1f2 l:tf7 3 'it>b3 e3
(a) 1 .l:!a6 .U.d l 2 g4+ 'it>f6 3 d7
4 llfl 'it>e4 5 'it>c2 e2 6 l:1gl 'it>e3, 'it>g7 4 .l:!a7 'it>f6 5 'itig3 'it>e5 !

and there is no apparent defence

against l:rxf7 or 'it>f2;
(c) 1 l:!b2 .U.c7 2 J:ib5 e3 3 l:1b3
(on 3 f6 there is 3 . . . e2 4 l:rb l l:ra7+,
and an exchange of rooks) 3 ...lie7 4
J:id3 (if 4 f6, then 4 ... e2! ) 4 . . . 'itic5 5
lid l 'it>c4, and again there is the
threat of 6 ... a2;
(d) 1 h4! (logical, it is necessary
to urge the passed pawn on) 1 ...'itie5
2 h5 c;t>xf5 (after 2 ... '\t>f4 3 h6 e3 4
1Ih2 l:!h7 5 f6 'it>g3 6 f7 Black can
not win) 3 h6 l:rh7 4 l:1h2 e3 (On
4 ... 'it>g6 there is 5 l:!.h4 e3 6 J::!e4) 5
Wb2 '\t>f4 6 Wc2 'itig3 7 J:ih5 e2 8
Wd2 @f2 9 l:!'.h2+ with a clear and
uncomplicated draw.
Nevertheless material is material.

(shouldering away the white king,

now after 6 d8='i!V lixd8 7 .l:!f7 l:!g8
a well known draw is reached) 6
@f3 lid3+ 7 'it>e2 l:rd4 8 c;t>e3 .lld l 9
l!b7 '\t>f6 ! (on 9 . ..lld6 there is 1 0 g5
'\t>f5 1 1 g6!) 1 0 'it>e4 .lle l +? and
now after 1 1 'it>d5 l:!d l + 12 'itic6
'it>e7 there is no defence against
1 3 ....l:!d6+ ! , capturing the d7 pawn.
This is how the game went.
(b) 1 g4+ 'itig6 2 d7 'it>g7 3 .l:!b7
'it>g6 4 .l:!a7 'it>h6 5 g5+ 'it>g6 6 'it>g4
lid4+ 7 f3 'it>g5, and a draw;
(c) 1 d7 ! lixd7 2 g4+ 'itie5 3 g5
and after 4 l:!f6 a well known theor
etically
winning
position
is
obtained.
.

And now a few examples on the

theme of choosing the best continu
There was a wide choice for ation. The solutions are given at the
White in the following game.
end of the book.

Rook Endings 97

How the "one-legged" Viktor

Lvovich tested the youngsters in
the endgame

What is correct:

I Wb5; I 'it>c5 or 1 <:ild5?

How does Black make a draw?

3

Brunner-Korchnoi
Berne, 1996

What is correct: l . f3, l ... '13.. e l or

l '.cJ ?
..

...

Viktor Korchnoi's play has al

ways been characterised by the
highest class and technique. Before
the start of the Berne tournamnent
and his match with Lucas Brunner,
the veteran broke his foot and the
organisers offered to postpone the
match but to their greatest surprise
the 'patient' had not even thought
about refusing to play! From chess
history it is a well-known paradoxi
cal fact that grandmasters with bro
ken limbs play very strongly! We
mention just two examples-Jan
Timman, with a broken foot, won
brilliantly at the super-tournament
in London 1 983, while Alexander
Beliavsky, with a broken hand, won
the board one prize at the Thessalo
niki Olympiad in 1 984! Inciden
tally, both breakages were sustained
playing football. It was rather un
usual to see the active Korchnoi sit
ting motionless for all his games
and only at the end with difficulty
moving away on crutches. But he
played splendidly, gaining particular
success in a couple of rook endings.

98 Rook Endings

There followed 1 .l:Id2?

Correct was 1 .l:Id7 f6 2 Wg 1 l:!.e2
3 h3, and it is difficult for Black to
carry out his plan.
1. ..lt>h7 2 Wg2?

Here it was still not too late to re

turn to 2 I:!'.d7 ! .
2 ... g6 3 fxg6+?

The last chance was 3 f6 g5 4 @3

a4 5 l:te2 g4+ 6 lt>f2 e4 7 112+
and 8 l:tf4 with some chances of
holding the game. Now a 3 :2 end
game is reached, which it seems is
practically winning for Black.
3 'it>xg6!
..

This is stronger than taking with

the pawn, since the passed e5 pawn
must be supported by the f-pawn.
4 lla2 h5!
Threatening to create a weakness
on g3 for White after h5-h4.
5 h4 l:tb4 6 a8 !l:b2+ 7 'it>f3
After 7 Wh3 ...

e3 1 6 :Ie8+ e2 17 J::l.e3 follows

1 7 ... f4 ! 1 8 gxf4 l:!.d3 !) 15 ... 'it>e5 1 6
l:te8+ lt>d4 1 7 I:!'.d8+ lt>e3 1 8 l:tf8
@2 ! 19 l:txf5+ @gl , and White
cannot defend against mate.
Well, now we return to the game.
7 ...Wf5 8 J:ih8

If 8 .l:.a3, then 8 ... d2! (with the

threat of 9 . . .J:id3+! 1 0 :xd3 e4+ 1 1
We3 exd3 1 2 'it>xd3 Wg4) 9 lt>e3
l:td4 1 o lia6 .l:Ib4 1 1 Wf3 l:tb3+ 12
Wf2 e 4 with the unpleasant threat of
... 'it>f5-g4
8 ... e4+ 9 We3 J:ib3+ 1 0 'it>f2 Wg4
1 1 .l:Ig8+ lt>h3!

Usually such an approach of the

king leads to a decisive outcome.
1 2 g5

is obtained practically an identical

copy of the famous game, Smyslov
Gligoric, Warsaw 1 947, (only with
White to move and the black pawn
on e4). There the very instructive
continuation was 8 lie8 .l:Ie2 9 :e7
f5 1 0 e6+ 'i;g7 1 1 a6 (If 1 1
.l:Ie7+, then l l ...'it>f6 12 h7 Wg6
13 .l:Ia7 .l:If2 - + ) l 1 . ..llf2 12 .l:Ie6
Wfl ' 1 3 l:txe5 f6 14 lle8 lld2!
(zugzwang) 1 5 J:If8+ (after 1 5 .l:Ih8

There were rather more chances

remaining with the preliminary 12
J:ig7, when Black can choose be
tween 12 . . . f5 and 12 ... :3+ 1 3 We2
xg3 14 lixfl Wxh4.
12 . .lif3+ 13 'it>e2 f6! 14 g6
On 14 l:!.xh5 Korchnoi intended to
play 14 ... lt>xg3 1 5 .:.h6 lt>g4 1 6 h5
f5 17 :h8 lt>g5 18 h6 @g6 1 9 h7
lih3, obtaining two connected
pawns.
.

14 fS 15 lig5 lt>g2 16 'it>el l:!.f2!

17 1ig8
..

If 1 7 Wd l , then 1 7 ... 'it>fl ! and

e4-e3-e2.
17 ... e3 18 J:ig7

Rook Endings 99

Or 1 8 g5 f4 1 9 gxf4+ 'it>f3 20
7 Wc2 c4?!
The question is where is the pawn
:g 1 lte2+ 2 1 Wd 1 lla2 22 fl We4
best placed-on c4 or on c5.
23 :g1 'it'd3 - + .
18 ... f4! 19 gxf4 Wf3 20 c7
We throw in the variation 7 .. @fl
For 20 .l:i:g l see above.
8 Wb3 li.a5 9 :t::!'.e 3 f5 10 Wc4 'it>f6
20 ... Ira2 White resigned.
1 1 Wd5 f4 12 gxf4 gxf4 1 3 l:!.c3
Wf5 14 f3 'it>g5 1 5 'it>c6 Wh4 1 6
Korchnoi-Kengis
Wb6 lta8 1 7 Wxc5 Wg3, and White
Berne Cup, 1 99 6
should not win.
.

8 Wd2

A t first sight the endgame looks

completely drawn-only the passed
a-pawn is potentially stronger than
the passed c-pawn. Now Black can
calmly transfer his king to d6, but
he decides to display activity on the
king's flank, since White obviously
intends to move his king over to the
a3 pawn.
1...g5 2 hxg5 hxg5 3 Wf3 a4
It is always useful to cut off the
enemy king.

s ...Wg6?!

After 8 ..Wfl 9 We3 'it'e6 10 Wd4

Wd6 1 1 g4 ( 1 1 f3 c3 1=) 1 l .. .We6
12 ..tics We5 1 3 Wb5 l:l.'.a8 1 4 a4
..t>d5 Black has his own counter
chances.
.

9 'it'e3 Wf5 1 0 Wd4 'it'g4 1 1 c5!

'it>h3 ?!

It is not quite clear where to go

with the black king. Better looks
l 1 ...lia8 1 2 Wxc4 f5 1 3 Wb3 l:i:b8+
4 We2 'it>g7
Again it is worth trying to go to 14 a2 lle8 1 5 llc2 l:i'.a8, and it is
d6, though here White can prevent very difficult for White to find a
this by 4 ...f8 5 l:te3 while on winning plan.
5 . . .lld4 6 l:i'.d3 l:.c4 7 Wd2 We7 8
12 Wb5 .l:ta8 13 a4! Wg2 ?
:ic3 lla4.
After 1 3 ... f5 ! ? 1 4 a5 l:!.b8+ 1 5
5 .l:.c3 f6(?)
Wxc4 l:.b2 1 6 a 6 l:.xf2 1 7 lla3 l1c2+
The authors would prefer 5 ... c5!? and 1 8 ...li.c8 Black holds the draw.
6 '>t>d2 'it'f6 7 'it'c2 c4 8 'it'b2 'it>e5
If 1 5 'it>c6, then 1 5 ....i!b2 1 6 a6 i:Ia2!
1 7 Wb7 ltxf2 18 lla3 J::[.b2+ 19 Wc6
but Kengis probably feared 6 l:i'.xc5.
6 Wd2 c5? !
c3 20 li.al c2 2 1 a7 l::l'.b l .
Again preferable is 6 ... Wf7 7 Wc2
1 4 f4! Wh3 1 5 fxg5 fxg5 16 a5 g4
(7 11e3 f5 8 Wc2 f4 =) 7 ... We7 8 1 7 a6 h2 18 !Ia3 Wh3 1 9 'it'xc4
blc8+ 20 'it>b5 l:Ib8+ 2 1 Wc6 .l:tc8+
'>t>b3 bla6 9 a4 'it>d6.
=,

J 00 Rook Endings

22 b7 l:th8 23 a7 l:th7+ 24 'it>b6

.l'.Ih8

It looks like Black has defended

himself-White cannot queen since
then a drawn ending arises because
of the distant white king. However
Korchnoi finds a path to victory.
25 'it>c6 .l'.If8 26 .l'.Ib3! 1If6+ 27
'it>b5 l:tf8 2 8 'lta5 l:ta8 29 'it>a6 .l'.If8
30 l:Ib8 .l'.Ifl 3 1 a8=tli l:lal + 32 'ltb5
lha8 33 lba8 'it>xg3 34 Wc4 Black
resigned.

The young grandmasters made a

surprising number of mistakes as
Korchnoi demonstrated by clear-cut
play.

Rook Endings I OJ

Exercises:
Rook Endings
3

Demonstrate the correct

plan of defence for Black.

What is correct:

I . . . @f4

or I a4 ?
...

What is correct: I 'f!d6 or 1 .l:.e8+ ?

Evaluate the position and fin d the

right plan of defence for Black.

102 Rook Endings

Find the right method of defence

for White.

Find the right plan for White to

realise his advantage.

What is correct: 1 Wc6 or

7

c6 ?

Find an accurate order ofmoves

for White.
10

Find the right continuation for Black. Find the right continuation for White.

Rook Endings 1 03

11

14

What plan should White choose?

12

Choose the correct continuation

for Black.
15

How should White conduct

the defence?

Find th e right continuation

for White.

13

16

What continuation should

White choose?

What is correct: 1 'fJ.d5 or 1 'fJ.dJ ?

1 04 Rook Endings

17

20

What is the best wayfor White

to achieve a draw?
18

White to play and win.

How does White win ?

What is White 's winning plan with

an exactly calculated variation.

19

22

What is correct: l .. J:!:x:h4 or L .'1:..g l+ ?

How does Black defend?

21

Rook Endings 1 05

23

26

White to play and win.

What is correct: J . a8=Vlll or 1 @b7?

24

27

..

What is correct: 1 .. @eJ or l @xj3 ?

White to move. How does he win ?

25

28

What is correct: J h3 or J. ..@h l ?

What is the winning move: L.i:Jilg2

or J...11.a l ?

. . .

...

J 06 Rook Endings

29

32

How does Black win ?

Can White make a draw?

30

33

How does White win: 1 h6 or 1 @g4?

What must Black do to draw?

31

34

Can Black save the game?

What gives chances of a win:

a) J l:lxa4 or b) 1 d7 ?

Rook Endings 1 07

35

38

Threatened by l 'IJ.g3, how can

White make a draw?
Calculate the variations.

Where is the clear drawfor Black?

36

39

How can Black make a draw?

How does Black make a draw?

37

40

White to play and win.

How did Black play?

...

I 08 Rook Endings

41

42

How does White win the game?

Can Black make a draw?

43

Point out a winning plan for Black

Shouldering:
the struggle of the kings
4

A reciprocal attack by a piece is

not usually restricted by either time
or, more so, by space. However,
when we find ourselves talking
about kings then the situation is
cardinally changed-the king can
neither attack, nor, more so, capture
the enemy king. Every king creates
around itself some strong squares in
which can enter any enemy piece,
except the king, of course-this
feature makes its biggest impression
in pawn endings.
Manukovsky-Pustovalov
Rosija, 1 992

1 'it>e6! 'it>c3 2 'it>d5! 'it>b4 3 'it>c6

'it>a5 4 cJi>b7 wins.

possible
Also
shouldering.

double

is

Zinar
1 984

1 ...'it>f4 !

Only so. I ...'it>xf3 2 \t>f5 leads to a

draw.
A fter the text Black resigned.
The mixture of shouldering and
zugzwang is indeed deadly.
Or yet another classic study by
Maizelis.

.,
.
.......
,_., .

1 'it>e2 ! g2 2 a4 'it>g3 3 'it>e3

'it>g2 4 a5 'it>n 5 'it>d4 wins.

J l 0 Shouldering

However, it is possible to define

shouldering, particularly in pawn
endings, as a form of zugzwang.

L.Mitrofanov

Mikhalchishin-Azmajparashvili
Tbilisi, 1 980

1 rid7!

Bad is 1 l:Xt7+ @e4 ! 2 Ae7+ 'it>d5

3 .Ud7+ 'it>e6 4 nds .Uc5+ and
5... .Ud5.
1. . 'it>e4 2 @g4!
The only move-if 2 'i.tg6, then
2 . .l:Ic6+ 3 @g7 ltc7 with a win.
.

It is obvious that White is strug

gling for the draw, and the question
is whether Black's king will succeed
in getting back. There followed
1 'il;xt7 al ='ii' 2 Ibal :xal 3 f5!

is dangerous to advance the

other pawn. 3 b5? 'it>c4 4 b6 'it>d5 5
b7 ltbl , and White has problems.
It

3 ...@c4!

If 3 ..@xb4, then 4 f6 'il;c5 5 'itg7,

and there is no defence against
f6-t7.
.

4 'it>e6!

Here is the shouldering-king

against king, shoulder to shoulder.
There is nothing else-if 4 f6, then
4 ... @d5 5 @g7 'it>e6 6 t7 lta7 etc.
4 l:Iel+
...

If 4 ....Ua6+, then 5 'it>e5 !-the

king goes to the other side, but the
main thing is not to allow in the
enemy king.
5 'it>d6! Afl 6 @e6 ltel + and he
had to agree a draw.
This idea was expressed simply
brilliantly in one study.

2 ... :c4! 2 Axd2 @e3+ 4 @g5

@xd2 5 h4 'ite3 6 h5 Ac5+ 7 'it>g4! !

This move was made before, but

nearer to the opponent's king it se
cures a draw, while the 'more ac
tive' 7 'it>g6? loses because of
7 ... 'it>f4 8 h6 l:Ic6+ 9 'itg7 'i.tg5 IO
h7 l:Xc7+ 1 1 'i;;g 8 'it>g6 12 h8=(:ZH
@f6.
Sometimes even the greats make
mistakes in the endgame.
Alekhine-Bogoljubow
World Championship (m) 1 92 9

Shouldering I I I

1 . 'it>g4??
Forgetting about the shouldering.
t 'it>e4 2 b7 f5 3 b8='iV! .l:Ixb8 4
xb8 f4 5 .l:Ie8+ 'it>d3 6 J::!f8 c,t>e3 7
@d5 f3 would draw.
.

...

2 b7 f5 3 b8=1i' .l:Ixb8 4 .l:Ixb8 f4

5 'it>d5 f3 6 'it>d4 f2 7 f8 'it>g3 8
'it>e3 and Black resigned.

Complex Endings

Playing to exploit a pawn

majority on the flank

One of the most important factors

in the endgame is the possibility of
creating a passed pawn-and in the
majority of cases this is formed
from a majority, i.e. from a pawn
advantage of two against one or
three pawns against two. From this
follows that the majority is the pre
cursor of forming a passed pawn
and in itself is an important posi
tional factor in the endgame. The
fewer the pieces, the more important
this factor, and the greater the role it
plays in the position.

i.xf6! i.xg3+ 8 xg3 gxf6 9 c6

WVes+ 10 @f3 rs 1 1 iYcs+ @f7 12
iVb7+ Black resigned.

A classical example.
Botvinnik-Rabinovich
Leningrad, 1 934

Playing to exploit the passed

pawn in the endgame is the most ex
treme case of the maj ority.
Portisch-Ribli
Skelleftea, 1 989

The presence of open lines and

White's pawn advantage on the
queen' s flank determines his su
periority. Among his tasks now is to
create a passed pawn.
1 a3 .l:.b7 2 l:!b2 i.g6 3 b4 'it>f8 4
@f2 rJ;e7 5 rJ;e3 d7 6 i.e2 c7 7
b5 lib8
After 7 . a6 8 a4 axb5 9 axb5
.

White controls the a-file and

vades the opponent's camp.

8 a4 f6

8 . .Wb6 follows 9 lid2.

.

9 a5

More accurate is 9 f4 ! .

9 ... e5 10 lid2 i.f7 11 f4! exf4+

12 xf4 l:!e8 1 3 i.f3 i.e6

Complex Endings 1 1 3

After 1 3 ...ii..xc4 wmmng is 1 4

lic2 ii..d 3 1 5 l:txc5+ @b8 1 6 ii.. c 6
lld8 17 l:tc I ! with the idea of e I .

1 4 ii.. c6 g5+ 1 5 @f3 lif8 1 6 ii..d 5

lldS 17 We3 ii.cs

The pawn ending is hopeless.

lS a2 ii..b7 19 lid2 lies+

On 19 ...ii..c 8 follows 20 l::td 3 with

the sequel We3-d2-c3, h3-h4xg5,
a5-a6, l:id3-h3.
20 Wf2 :ids 2 1 g4 ii.. cs 22 lid3

lif8 23 lie3 f5 24 gxf5 xf5+ 25

'it>g3

White's position is winning due to

the threats b5-b6, a5-a6.
Botvinnik-Kan
USSR, 1 955

12 @d3 .l:Ibl 13 el l:!b6 1 4

Wc2 !

Now White needs to defend the

bl square.
14 ... d6 15 .l:Ie4 b6 16

a3 !

I t i s better to give up the pawn on

the third rank-prophylaxis !
16 ... g6 17 a5! ligS

There is no saving himself by

1 7 ...bxa5 1 8 ii..x a5 l:!a6 19 ii..d2
ii..xd2 20 @xd2 l:Ixa3(!) 21 lle5
with a decisive advantage for White.
The rest is forced.

18 a4! I:l.f8 19 l:!e2 bxa5 20 'it>d3

.lld S 21 l:!e4 'it>c6+ 22 @e2 @b6 23
i.el! i.cl 24 g3 hxg3 25 fxg3 l:thS
26 h4 gxh4 27 gxh4 i.b2 2S ii.. d 2
i.d4 29 Wf3 i.b2 30 ii.gs lif8+ 31
.l:If4 ligS 32 llf5 ii..d 4 33 h5 @c6 34
h6 @d7 35 'it>g4 @e6 36 l:!f4 bS
37 lie4+! @f7 38 lie7+ WgS 39
Wf5 with a win.
Vukic-Pietzsch
Sarajevo, 1 96 7

From what we have said before it

follows that White needs to ex
change as many pieces as possible.
And he sets about realising this
plan.
1 l:id4 ! f5 2 e2 ! g5 3 l:!.xe4
lbxe4 4 liJd2 liJxd2 5 Wxd2 !

The king must head towards the

majority.

5 ... llh6 6 e4 ii..f4+ 7 Wel fxe4 8

:rxe4 c5 9 ii.. c3 a4? !

Stronger i s 9. . ..l:Ia6.

10 bxa4 l:!b6 1 1 We2 @d7

On 1 1 .. . .:.b I would have followed

12 .i e5 .

a
.
,,
,,
,
f,
. . 0. v.

,. ;-
..!.,
. .

,
!ian9n
r---- -
1 . .. lbe8

Intending to transfer the king to

the queen' s flank to neutralise the
opponent's pawn superiority.

2 liJd7! e6 3 Wg2 Wg7 4 d3

l'tJf6 5 liedl liJxd7 6 l:.xd7 h5 7
@n 'it>f8 8 'it>e2 @es 9 l:!xc7 l:l.'.xc7

By placing his pieces in good

positions, White forces his opponent

1 1 4 Complex Endings

to exchange off, thus increasing the

significance of the majority factor.

Georgiev-Khalifman
Pardubice, 1 994

10 d4 f6 1 1 f4 @e7 12 b4 b6 13
c;t>d3 J::[.c8 14 @c3 e5 1 5 fxe5 fxe5
16 J::!d5 @e6 17 a4 a6

1 7 .. .!lf8 ! ? would
problems for White.
.

create more

18 a5 b5 19 l:!c5 ! ! .:i.xc5 20 bxc5

c;t>d7?

And here it was still not too late to

escape with a draw by 20 . . . b4+! 2 1
xb4 @d7=.

21 cxb5 axb5 22 h4! Black

resigned.
Martinovic-Yudasin
Krynica, 1 998

l f5 !

The d4 pawn is well blockaded

and White has a serious advantage
on the king's flank.
1 . f8?
Correct is 1 .. ..l:!e8 2 h4 .l:!xe2 3
@xe2 i.d6 4 h5 i.e5 ! , defending
the black king.
..

2 h4 e8 3 h5 .l:!xe2 4 c,t>xe2 i.d5

Better is 4 ... i.d6 5 f6 gxf6 6 gxf6

i.e5 7 i.h6+ c,t>g8 8 .llg 7! with the
idea h5-h6-h7.

5 f6 gxf6 6 gxf6 @g8 7 i.h6 b4 8

b3! i.e6 9 @f3 i.d5 10 @f2 axb3
1 1 axb3 i.xb3 12 i.g7 i.f8 13 h6

Whose majority is better

White's or Black's? After I c;t>e2
l2'ld6 2 @d3 e4+ problems with the
b3 pawn appear for White. Best was
I f4 l2'ld6 2 @f3 with equality. In
the game, however, there followed
1 g4?

He should only play on his weak

flank when it is possible to create
weaknesses for his opponent.
1...l2'ld6 2 gxf5+ gxf5 3 c5 bxc5 4
bxc5 l2'lb5 5 c6 f4 6 l2'lc2 c;t>d6 7
@e2 'it>xc6 8 'it>d3 c;t>dS

And Black
endgame.

has

winning

with a winning position for White.

Kovacevic-Tosic
Belgrade, 1 998

Complex Endings 1 15

Against doubled pawns, the natu

ral strategy is to exploit the majority
on the other flank.

1 h3! a4 2 'it>f3 l:!.d7 3 l:Ixd7

'it>xd7 4 g4 hxg4 5 hxg4 b5 6 b3!

Black cannot create a passed

pawn.
6 axb3 7 axb3 'it>e8 8 t'llh4
..

Preparing f4-f5.
8 t'Llf7 9 'it>e4 g5 10 Cll f3 gxf4 1 1
.

wxf4 'it>f8 1 2 t'lld 2 t'll h 8 1 3 'it>g5

Wf7 14 'it>h6! t'llg6 15 t'll f3 t'll f4 1 6
'it>h7 c6 17 t'll g5+ 'it>e7 1 8 Clle4 c4
19 bxc4 Clld 3 20 g5 Black
resigned.
Psakhis-Bonsch
Trnava, 1 988

On 1 6 ... t'llxe5 decisive is 1 7 t'llxe5

i..xe5 1 8 'it>a5, and if l 6 ... i.. x e5,
then 17 b5 axb5 1 8 cxb5 i..d4 19
bxc6.

17 i..c5+ We8 18 i..d 6 f6 19 exf6

i..xf6 20 t'llc5 e5 2 1 t'llx a6 e4 22 b5
i..e5 23 i..c5 Clld4 24 i..xd4 ! i.. xd4
25 'it>b3 'it>d7 26 c5 h5 27 g4 hxg4
28 hxg4 g5 29 'it>c4 e3 30 @d3 and
Black resigned.

If 30 ... 'it>e6 winning is 3 1 b6 ! ;

while on 3 0. . . Wc8-3 1 c6 i..b 6 32
t'llb4 and 3 3 Clld5 .

The following game serves as an

example of a bad majority (i.e. a
case when the majority can become
an object of undermining and
attack).
Kan-Keres
Moscow, 1 952

1 'it>fl !

The main trump in White's posi

tion is the pawn majority on the
queen's flank. Psakhis exploits this
advantage very instructively.

1 ...a5 !

Trying to create a weakness.

1 ...'Wc8 2 'it>e2 Cll d 7 3 'it>dl ! i.f8

4 'it>c2 g6 5 i.c3 .1i.g7 6 a3! .1i.f8 7
b4 cxb4 8 axb4 t'llb 6 9 'it>b3!

2 :i:!.xd8 .l:Ixd8 3 l:idl t'lld5 4

Cllx d5+ i..xd5 5 c4?

9 ...e8 10 Cllb 2 'iVd7 1 1 i.d4

'bes 12 i.e3 t'll e7 13 'iVa8! 'iVc6

5 ... i.e4 6 :xd8 'it>xd8 7 i..d l

'it>c7 8 @fl Wd6 9 f4 e5 1 0 g3 i..b l
1 1 a 4 exf4 1 2 gxf4 We6 13 @f2
Wf5 14 'it>e3 h5! 15 'it>f3

14 'iVxc6 tllxc6 15 t'lld3 i..g 7 1 6

a4! Wf8

If 1 7 @f3 i..f5 1 8 'it>g2 f6 1 9

'it>h2 i.e4 ! and Black has a decisive
advantage.

White has in effect an extra piece:

his king actively supports the pawns
on the queen' s flank.

Black is forced to exchange

queens and the remainder becomes
a matter of technique.

Creating a bad majority.

If 1 5 h4 i..e4 ! , zugzwang.

15 ... h4! 16 We3 'it>e6 17 .ig4+

IJ6

Complex Endings

17 ...f5 1 8 'it>f3 'it>f6 19 xf5

'it>xf5 20 'it>e3 f6 21 'it>f3 g5 22 fxg5
fxg5 23 'it>g2 g4 24 hxg4+ 'it>xg4 25
<;tin 'it>f4 White resigned.
Ljubojevic-Ivanchuk
Buenos Aires, 1 995

1 'it>t1 'it>t7 2 g4! ctJe6 3 CLixetS

'it>xe6 4 ctJe4 d6 5 g5! fxg5 6
'2Jxg5+ 'it>f5 7 '2Jxh7 'it>xf4 8 ctJf6
cs 9 '2Je8 f5 10 c3 bl 1 1 a3
a2 1 2 b4 c5 13 '2Jxc7 c4 14 h4!
cxb4 15 axb4 i;t>e4 1 6 ctJe8 <;ties 17
'it>e3 f7 18 '2Jc7 g6 19 ctJb5
'itd5 20 CLixa7 'it>c4 21 ctJc6 Black
resigned.
Toothill-Heemsoth
corres, 1 986

White has a pawn majority on the

queen's flank, therefore he can
boldly improve the positions of his
pieces which only increases his
positional advantage.

1 c4! xg2 2 1Ixd7 .l:!xd7 3 .l:Ixd7

'i!Vxd7 4 'iitxg2 'i/Nc6+ 5 'it>gl f5 6
b6! ! e4 7 b4 e5 8 b5 f4 9 gxf4
exf4 10 bxa6 bxa6 1 1 'i!Vxa6 'ilYel+
12 'it>g2 and White has a decisive

advantage.

Hiibner-Spassky
Candidates (m), 1 985

l ...lif8!
With the unequivocal intention of
driving back the opponent's king
from the centre.
2 :i.d2 f5+ 3 gxf5 gxf5+ 4 'it>f3
c4! 5 a4 a6 6 f4 .l:Id8! 7 .l:Ixd8

Taking into account the previous

comment, more chances for White
rested with 7 lic2, retaining as many
pieces as possible.
7...xd8 8 e3 b5 9 axb5 axb5
10 cl c7 1 1 h3 'it>d5 12 i;t>e3 b4
13 d2 b3! 14 f3 f4+ 15 'it>t1 ..)te5
1 6 .)tel d4+ 17 'it>el ..)te3! White
resigned.
Mutual Majorities

White' s extra pawn on the king's

flank proves a decisive factor.

Each side has his majority, and

the main problem is how to limit
mobility from the opponent's side.
With level pawns, if one of the
rivals has a majority on one flank,

Complex Endings 1 1 7

then for the other it is the opposite

-this is a fact. But often it will
happen that one majority is better
than the other, then we speak of a
qualitative majority.

Yates-Alekhine
The Hague, 192 1

Karpov-Yusupov
Dortmund, 1997

A classical position-White loses

it without a struggle.
1 g3?

Incorrect prophylaxis!

1 ...@f7 2 c5 '\i>f6 3 c4 .tc8 ! !

1 'Llc4!

It is necessary to keep in his sights

the
opponent's
weaknesses
White's plan consists of an ex
change of queens and advance of
pawns on the king 's flank.
i...'Lld7 2 b3 rs 3 c,t>n
Also good is 3 g4 g6 (3 .. .fxg4 4
hxg4 'i/Nf3 5 'ii'f5) 4 gxf5 gxf5 5
'if'c3.
3 ...@f7 4 f3 'ii>e7 5 'iite 2 iie6 6

'ii' c3 'Llf6 7 'it>f2 ! ?

The king will be needed on the
other side.
7 ...'ii>d7 8 g4 'ii>c6?!

Better really is 8 ... fxg4 9 hxg4.

9 'ii'e 5!

Now the
hopeless.

knight

ending

is

9 ...'iYxeS 10 'Llxe5 'ii>d5 1 1 'Llc4

fxg4 12 'Llxb6+! i;;>c6 13 'Llc4 gxf3
14 'Llxa5 'ii>d5 15 i;;>xf3 g5 16 tt::l c4
h5 17 'Lld2! i;;>e5 1 8 e4 'Lle8 1 9 i;;>e3
'Llc7 20 'Llc4+ 'ii>f6 21 'ii>f2! 'Lla6 22
'it>g3 'Llb4 23 h4 ! 'Llc6 24 a5 'Llb4
25 'Lld2!.

A terrible move. For Black the

bishop is more important-so it will
stand firm on the c8 square.
4 a4?

Here it is worth playing 4 f4 ! ?

exf3 5 c,t>f2 l:!d2+ 6 i;;>xf3 .!lxh2 7 b5
with counterplay.
4 ...gs 5 b5 f4 6 c,t>n .!id2 7 gxf4
gxf4 8 @el l:!b2 9 .te2

After 9 l:id l g4 10 .!ld6+ 'ii>e7

1 1 1Id4 f3 12 d5 .l:Ib l+ 13 'it>d2
e3+! 14 fxe3 .lld l + 15 'ii>c3 .!lxd4
1 6 'ii>xd4 d5 -+.

9...'\i>e5 10 c6 bxc6 11 .l:Ixc6 e6

12 d l .!lbl 13 .Iles+ 'ii>d4 14 .llc 2
e3 15 fxe3 fxe3 16 .l:Ic6 .tg4 17
.lld6+ 'it>e5 1 8 h3 h5 ! White
resigned.

But correct prophylaxis was

possible:
(a) 1 f4! (intending c,t>g l -f2-e3)
l:id2 2 c5 i;;>f7 3 .tc4 .txc4 4 .llxc4
i;;>e6 5 b5 i;t>d5 6 c6 bxc6 7 bxc6=;
(b) 1 f3 ! (to create a weakness on
e4) l . ..e3 2 f4 ! .lld4 3 l:!e 1 .l:!xf4 4
l:!xe3 .tc4 5 .txc4 .:i.xc4 6 .:.e7 with
a drawn ending.

J J 8 Complex Endings

Bronstein-Rantanen
Tallinn, 1 9 75

Portisch-Kramnik
Biel, 1 993

Advancing on the other side-but

here it is important to force back
and restrict the black pieces.

If 1 'l'Wa3 tt:Jc4 2 4 b5 3 a4
tiJb6 ! 4 axb5 tiJd5 ! Black maintains
a small advantage due to his control
over the central squares.

1 g4!

1 ...a6 2 g5 tt:Je8 3 a4 l:!'.a7 4 h4

b7

Better is 4 ... @f8 ! ?.

5 l:1d3

Now it is important to control the

d-file.
5 ....l:tc5 6 llel li:td7 7 .l:Iedl !

Tactics. If 7... e5 there is 8 ctJf5.

7 ... g6 8 tt:Je2 !

Typical.

8 .l:r.xd3 9 l:1xd3 b5 10 cxb5

axb5 1 1 .!id7 @f8 12 a5 .!lc6 13
.!lb7 b4 14 .l:Ib8 i.c5 15 tt:Jg3 Black
resigned.
.

A classic example of exploitation

of a pawn superiority on the queen's
flank is presented by the following
position from a modem tournament.
Here Black has a pawn superiority
on the queen's flank, typical for the
Griinfeld Defence, and therefore
White should keep as many pieces
as possible so as to maintain the
tension.

1 .l:tc3

1. .. b5 2 :i.fcl

Bad is 2 l:!'.a3 ? because of the

tactical blow 2 ... tt:Jc4 3 .!lxa6 7.
2 tt:Jc4 3 'ie2
..

On 3 tiJd2 there is the effective

3 ... a5 ! ; also in the event of 3 e4
there is 3 . . . f5 ! .
3... tt:Jb6!

With the idea of capturing on c3.

4 .l:Ic7

4 i.e5 ! ? deserves attention.

4 ... 'ii'e6 5 i.g5

And now the only chance was 5

i.e5 ! ? with a slight advantage for
Black.

5 ...liJd5 6 .l:I7c5 h6 7 i.h4 b4 + 8

WVh2

There is no saving himself by 8

'l'Wc4 l:1xc5 9 'ii'xc5 tt:Jc3 +.
8 tt:Jc3 9 .l:txc8 l:!.xc8 10 'it>hl
'it>h7 ! !
..

Prophylaxis. 1 0 .. .a 5 1 1 a 3 tt:Ja4 1 2
.l:Ixc8 'i!Vxc8 1 3 ii'b3 'ii'c l+ 14 'ith2
'.Wxa3 1 5 '.Wd5 gave White
counter-chances.

Complex Endings 1 1 9

1 1 l:Ial

Other moves are also no help: 1 1

3 tllxa2 ! or 1 1 a3 tlla4.

1 1 ... aS 12 'iVb3 iVxb3 13 axb3 g5

14 i.g3 a4! + .

In complicated endings are

usually to be found complicated
plans (many stages), various pieces
and
asymmetrical
formations.
Therefore it is necessary to pen
etrate deeply into the position, tak
ing care over the exchange of
pieces.
Balashov-Korchnoi
USSR (ch), 1969

White seems to have a very cen

tralised position and a solid pawn
formation, but Korchnoi begins to
undermine the white structure.
1 ...!If8! 2 tll d2 f6 3 .lift

Whit wants to exchange rooks to

neutralise the pressure.
3 ... i.d8! 4 J::rf2 Ji.c7
Increasing the pressure on the
centre-improving the position of
the bishop.
5 Cllfl i.b6 ! 6 :d2 fxeS 7 i.xeS

J::rfi 8 @e2 d6 9 i.c3 i.c4+! 1 0

Wet ii.xfl !

A surprising exchange of the

strong bishop, but now it is more
important to create a passed pawn.
1 1 'it>xfl es 12 'it>g2 e4 !

Not 1 2 . . . exf4 because o f 1 3 .l:!f2 .

13 J:Ie2 dS 14 i.es i.c7!

An exchange of strong pieces.

15 i.xc7 .l::!.xc7

But now the weakness of the b2

and g3 pawns is felt.

16 fl J:ib7 17 c3 bS! 18 l:!c2

'it>c4 19 @e2 .:i.b3! 20 e3 d4+!

A blow!

21 xe4 dxc3 22 bxc3 :xa3 23

g4 hxg4 24 l:!g2 l:!b3! White
resigned.
Mikhalchishin-Beliavsky
Moscow, 1981

Black has a weak pawn on c4, but

together with this there is strong
counterplay on the d-file. White
should combine the improvement of
the positions of his pieces with
prophylaxis.
1 tllc5 ! i.dS

If 1 . .:i.d2, then 2 i.fl ! .!lxa2 3

tllx e6 fxe6 4 i.xc4 with material
advantage.
.

2 f3 .!lc8 3 tll a4 i.e6 4 e4

White's cunning manoeuvre has

cut off Black's play along the d-file
and restricted Black's setup with an
ideal structure f3-e4. Only now does
his king make an appearance in the
centre.
4 ... llld7 s Wf2 lllb 6 6 tllc3 !
The knight is best placed on a
blockading square.

120 Complex Endings

6 f5 7 'it>e3 fxe4 8 fxe4 @f8 9

.l:Ibl !

It is very difficult for White to im

prove his position, therefore he re
sorts to such unnatural maneouvres.
But 3 h4 deserved serious attention.

9... d7 10 'it>d4 rtie7 11 l:!'.b4

.!:i:d8

With each move Black improves

the positions of his pieces-the
same cannot be said of White.

..

Now he has time to improve the

position of his rook-the threat is
l:tb l -b5-a5.

If 1 1 . ..Wd6, then 12 e5+ 'it>e7 1 3

a4!, taking from the knight its last
good outpost on b6.
12 xc4! i.h3+ 13 Cll d5+ Clixd5
14 exd5 xg2 15 We5!

Pawns are equal but the activity of

each of White's pieces is clearly
greater-this is also a typical case of
domination.

15 ... :gd7 16 lib8! f3 17 .l:Ig8 g6

18 d6+! lixd6 19 l:!'.g7+ @f8 20
l:tf7+ Black resigned.

3 ... e5 4 e2 'llg 5! 5 cl Clie6

6 Cll c3 rs 7 a3 c6! 8 b4 i.d4 9

Cllb 5? !

White did not think that his oppo

nent would so readily part with his
bishop, therefore better was 9 i.d2.
9 ...xb5! 10 cxb5 @f6 11 c4
i.c3 !

Preparing the d4 square for the

knight, while preventing a3-a4 or
i.cl -d2.
12 c,t>n 'lld 4 13 f3 f4!

Fantastic technique in a compli

cated endgame was demonstrated by
lvanchuk in the following game.
Dao Thien Hai-lvanchuk
Moscow, 1 994

At first sight it seems incredible

that Black should be playing for a
win-White has a pawn majority on
the queen's flank. First of all Black
improves his structure on the king's
flank an thereby restricts the oppo
nent's pieces.
1. .. f6! 2 Clie3 'it>f7 3 'll d l

A decision by a grandmaster of
extra class-weakening the white
squares rather too much, but in re
turn restricting to the maximum the
white king and bishop c l .
1 4 @f2 h5 1 5 g3
If White waits, then Black pro
ceeds with ...g7-g5-g4 and then
... @f6-g5-h4 with the threat of
...g4xf3 and . . . @h4-h3 .
15 gS 16 Wg2 'll c2 !
..

Surprisingly, after the knight

move many threats appear-for
example, not possible is 1 7 i.b3
because of 1 7 ... Clle l +-this is a

Complex Endings 12 J

consequence of the restricting ma

noeuvre ... f5-f4
17 @fl i.b2!

One cannot leave out the tactical

element-it has been well known
since the time of Capablanca that to
achieve the maximum one must
transpose positional factors into
tactical ones.
18 i.xb2

If 1 8 i.b3, then 1 8 . . . i.xcl 1 9

i.xc2 fxg3 20 hxg3 i.a3, and with
the presence of the passed pawn on
the h-file Black ought to win easily.

18 ...l'De3+ 19 'it>e2 l'Dxc4 20 i.cl

Wf5 2 1 @d3 l'Dd6 22 a4 g4 23
fxg4+ 'it>xg4!

Precisely this-the main thing is

to break through to the h2 pawn.
24 gxf4 exf4 25 i.b2 f3 26 i.d4
'it>h3 27 a5 'it>xh2

Not everyone would pay attention

to such 'trifles' as the b6 pawn.
28 axb6 axb6 29 i.xb6 'it>g2 and

White resigned.

Here is a very complicated re

alisation with pawns on just one
flank.
Romanishin-Nunn
Debrecen, 1 992

4 .l:Id5 \tg6 5 g4 i.f6 6 Wg2 i.g5

7 i.g3 i.e7 8 h4 lic4 9 lid7 i.f6 1 0
f4

Black already has some trouble

-White activates his phalanx to the
utmost.
10 ... .:i.c5 1 1 .l:Id6 'it>g7 12 i.f2
.l:.a5 13 lid3 b5 14 'it>g3

Now that the rook defends the

third rank, White improves the
position of his king.

14 .. Jla5 15 i.b6 .l:!b5 16 i.e3

i.b2 17 'it>f3 i.al 18 dl i.c3 1 9
.l:Icl .l:Ib3 2 0 'it>e4!

Beginning a very important

stage-activating the white king.
20 ....l:!b4+ 21 Wd3 i.f6 22 g5
hxg5 23 hxg5 i.d8 24 c6!

Preventing 24 ... f6, on which fol

lows 25 .l:Id6 i.e7 26 l:!d7 ! + -.
24 ...I:!'.b3+ 25 'it>e4 f5+?

It is not clear why-the fact is that

White does not threaten the move
26 f5 because after 26 .. . .l:Ixe3+! 27
'it>xe3 i.xg5+ Black has a certain
draw.

26 'it>f3 i.e7 27 :gc7 Wf8 28 'it>e2

Ji(b2+ 29 d3 b3 30 'it>d2 We8 3 1
i.c5! i.d8 32 .l:Ig7 i.a5+ 33 We2
.l:Ib2+ 34 Wd3 J::!d 2+ 35 Wc4 l:!e2
36 g6! .l:!e4+ 37 'it>d3 .l:!xf4 38 I:!'.e7+
Wd8 39 g7 .l:Ig4 40 .:i.tl! Black
resigned. A fine piece of work.
Nezhmetdinov-Romanovsky
Moscow, 1 95 7

1 g3 h6 2 i.f4 'it>h7 3 lid2 .l:tc6

Stronger is 3 ... .l:Ic5 and 4 ... h5.

122 Complex Endings

What plan should White choose?

1 g4! !

With the intention of playing

further f2-f4 and e4-e5, gaining
ground, therefore Black's reply is
forced.
1...hxg3 2 ..txg3 Cllf8 3 'it>h2 lll e 6

4 f4 b5 5 .l:Igl g6

White has threats along the g-file,

therefore Black is forced to weaken
himself.
6 Cllh6+ <#;;g7 7 lllg4 f6

How otherwise to deny the knight

access to e5? If 7 ... ..tf6, then 8 e5
and 9 f5 with a strong attack.
8 f5!

Reminiscent of the play in a

famous Lasker-Capablanca game.
8...lllg5 9 .l:Igfl c5 10 e5! ..txf5

On 10 . . . fxe5 White plays 1 1

..txe5+ Wh7 1 2 lixd8 ! l:!xd8 1 3
fxg6+ 'it>xg6 1 4 :i.f6+ 'it>h7 1 5 .l:Ih6+
with decisive threats.
1 1 exf6+ 'it>f8
On l l ...i.xf6 there is 1 2 Cllxf6
'it>xf6 1 3 i.h4.
12 llle 5! b4 13 Cll d5 i.xc2

More stubborn was 13 ... lllf7 .

14 f7 lie6
resigned.

15

llld7+ Black

Lputian-Tukmakov
Moscow, 1 983

1 lic3!

Clearly not 1 a4 because of

l . .. Cllb 4, and the black knight jumps
to freedom.
1...lie6 2 llle5 g5 3 ..tg3 Wg7 4
Cll f3 h6 5 h4 !

Such moves are always unpleas

ant. White latches on to the weak
ness and wants to 'occupy' it. If
5 . . . f6, then 6 h5 Wf7 7 .l:Id3 with the
idea Clld4, though Black can of
course defend himself for a long
time.
5 ... g4 6 Clle l !

It turns out that the knight is head

ing for the f4 square.

6 .....td6 7 ..txd6 1ixd6 8 .l:Ic4 h5 9

<#;;e 2

The exchange of bishops has not

eased Black's position very much.

9 ...lllb 8 10 lies 'it>h6 1 1 llld3

Clld7 12 lia5 a6 13 g3 .l:!b6 14
lllb 4!

By covering the file, White pre

vents the activation of the black
rook.
14 .. Jie6+ 15 'it>fl 'it>g7

If the passive 1 5 ... lllb 8, then 1 6

Clld 3 l:.b6 1 7 Cllf4 .l:Ib5 1 8 .l:Ixb5
axb5 1 9 'it>e2 llla6 20 llld5 ! with a
winning knight endgame. Now
however any capture is quite good.
16 lixa6 .l:!e8

If l 6 ... lle4, then 17 Clld3 llle 5 1 8

Cll f4 ! Clif3 1 9 lllxh5+ 'it>f8 20 lia8+
We7 2 1 lie8+! ! , exploiting the
tactics.

17 Cll d5 lllc5 18 I:!'.c6 lll e 6 1 9 lic4

lib8 20 a4 lib2 21 lllf4 libl +

The rook ending is hopeless, since

the king simply goes to his passed
pawn.

22 'it>e2 Wf6 23 Cll xh5+ We5 24

J::!xg4
And Black soon resigned.

Complex Endings 123

Salov-Khalifman
Candidates (m) 1 994

Nezhmetdinov-Luik
USSR 1 950

White has a majority plus a space

advantage.

White has the advantage on the

king's flank plus a more active de
ployment of pieces-there followed

1 h4! f6 2 h5

Spoiling Black's majority.

1 b4 !

2 ... a6 3 tl:ic3 tl:ie5 4 tl:id5+ c,t>dS 5

b3 b5!?

With the threat of 2 i.c5, whereas

l . ..i.xb4? is impossible because of
2 .!la8+ i.f8 3 i.c5.

6 cxb5 axb5 7 b4 ! i.g4 S a3!

@cs

Completely cramping the oppo

nent-not possible is 3 ... hxg6 4
fxg6+ 'it>xg6 5 tl:ixe5+.

Without this move Black would

not have any real counterplay.
After 8 ... i.xh5 9 tl:ie6+ attacking
the g7 pawn.
9 tl:ie7+! @c7 10 tl:i g6 i.xh5
Or 1 0 ...tl:ixg6 1 1 'it>xg4 tl:ie5 1 2
@f5 + -.
11 tl:ixe5! fxe5 12 c;t>xe5 gs

After 12 ... i.g4 1 3 tl:ie6+ i.xe6 1 4

c;t>xe6 the pawn ending i s simply
lost.

13 tl:ie6 @b6 14 'it>f6 i.e2 15 @g6

J.. fl 16 g3 'it>c6 17 'it>xh6 'it>d5 l S
tl:ixg5 i.e2 19 @g6 'it>c4 20 tl:ie6
cJi>c3 21 @rs 'it>b2 22 tl:ic5 'it>xa3 24
tl:ia6! and there is no defence

against the march of the g3 pawn.

Exploiting a space advantage ...

1...J.. eS 2 i.c5 @f7 3 g6+!

3 ... hxg6 4 fxg6+ @rs 5 h5 i.xc5

Already he has to allow White to

create a passed pawn.
6 bxc5 J::rc 7 7 .!las l:!xc5 S h6!

Here also an advantage on the

king's flank expresses itself in a
tactical form.
S ... gxh6 9 tl:ih4 l:!'.c6 10 tl:ifS .lle6
1 1 l:!bS

White plans to go with his king to

h5, therefore he places his rook to
hold up the b-pawn.
1 1 ... h5 12 'it>g3 b4 13 c,t>h4 b3 14
@xh5 b2 15 'it>h6 c;t>gs 1 6 J::rxb2
And as Black is in zugzwang, he
resigned.

124 Complex Endings

An entertaining endgame motive

was produced in the game

Vaganian-Smirin
USSR (ch), 1 988

It seems that there is no hint of

danger in White's position, but...

1 ..id2 ltJeS 2 'it>g3 ..ic6 3 f4

liJd3! 4 i.e3 eS S fxeS fxeS 6 'it>h4
rJlf6 7 'it>g3 'it>e6 8 c;t>M 'it>f6 9 'it>g3
aS 1 0 ..igl a6 1 1 ..ie3

Black's two doubled pawns

successfully counter White's three
thanks to the excellent position of
the knight on d3, while the passed
e5 pawn has become dangerous,
since the white king cannot get over
to the e-file.

1 c6 is striking with the intention

of winning in the variation l . . ..ixb4
2 c7 lbb6 3 lbc4. But Black plays
I .. .lLlb6! 2 c7 (or 2 b5 ..id6 3 lbg2
'it>f8 with equality) 2 ... ..id6! 3 c8=='i
lbxc8 4 ..ixc8 ..ixb4, and is in no
danger. The game, however, went
.

1 'it>g2

1 1 . ..'it>e6 12 rJlh4 ii.bl 13 rJlg3

'iitd7 14 ltJdl lbb4 lS ltJc3 rJlc6 1 6
'it>h4 ltJc2 17 i. f2 e4 1 8 <;fo>gS e3 1 9
i.g3 'it>xcS 20 ltJe2 i. f3 2 1 ltJgl
..idl 22 h3 hS 23 'it>xg6 hxg4 24
hxg4 ..ixg4 2s 'it>gs ..idl 26 @rs
'it>d4 27 'it>e6 ltJb4 28 ii.el ltJxa2 29
..ixaS ltJcl 30 b4 ..ig4 3 1 rJld6 e2
White resigned.

Space plus control of an open

line.

Surprisingly for White, Black is

saved by a piece sacrifice.

Botvinnik-Balashov
Hastings, 1 967

1...ltJxcS ! ? 2 bxcS ..ixcS 3 ltJg4

..id4 4 ..ic6 c;t>g7 S h4 ii.cs 6 <;fo>g3
..id6 7 'it>f3 i.e7 8 hS gxhS 9 lLlf2
fS 10 liJd3 'it>f6 11 ltJeS i.d6 1 2
lLld7+ rJle7 1 3 i.bS h4
The draw is obvious.
Neverov-Dreev
Lvov, 1 985

1 a4 f6 2 aS

White increases his space advan

tage and thereby drives back the
black pieces.

2 ..id8 3 l:!c8 'it>f7 4 e4 d7 S bS

..ie7 6 ..id2 ..id6 7 l:l.'.a8 b6 8 a6!
..

Now not only is there a weak a7

pawn but, when the opportunity

Complex Endings

presents itself, also the manoeuvre

.l:!a8-b8-b7.
8...i.cs 9 i.e3 @e7 10 h3 d6
1 1 lle8 :c7 12 e5+! 'it>d7 12 .l:!g8!

There is no sense in winning a

pawn at the cost of exchanging
rooks.

12 ... fS 13 :xg7+ Wd8 14 l:!g8+

@d7 15 i.h8 i.e7 16 i.d4 .l:!c2+ 17
'it>d3 .l:!a2 1 8 .l:Ia8+ Black resigned.

125

The invasion of the king is decis

ive-this is more accurate than
i.xe7 and ll:lc8+.

27...ll:lg6+ 28 @hS ll:le7 29 i.xe7

@xe7 30 'it>g6! Black resigned.

Playing to exploit a weakness.

Korchnoi-Pinter
Reggio Emilia, 1 98 7188

Tikhomirova-Morozova
Kiev, 1 966

1 ll:le2 !

White's plan is simple-to push

the pawn to a6 and still further
hamper the opponent, while in the
event of an exchange on a5 the c 5
pawn will be very weak.
1 ll:lcl ! ll:lf5+ 2 'it>d3 e8 3 a4
'it>d7 4 ll:lb3 ll:lh4 5 g3 ll:lf5 6 a5
ll:ld8 7 a6 @c7 8 ll:lbd2

The knight has done its business

and now transfers to the other flank.

8 @b8 9 ll:lf6! ll:lh6 10 ll:lde4

@c7 l l ll:le8+ 'it>d7 12 ll:l4f6+ i.xf6
1 3 ll:lxf6+ rt;c7 14 h3 ctJf5 1 6 g4
.

Now it is time to expand on the

other flank.

16 ... hxg4 1 7 hxg4 ll:le7 18 ll:le4

'it>b8 19 .Itel ll:lg8 20 i.h4 @c7 2 1
@e3 rt;d7 2 2 @ f3 'it>c7 2 3 ll:ld6 f5
24 gxf5 gxf5

Somewhat better is 24 ... exf5.

25 i.g5 @d7 26 @g3 ll:le7 27

@h4!

Hurrying to activate his pieces

would be punished: 1 .l:!b3 ll:lc5; 1
'it>e3 ll:le5.

i...:c8 2 l:!c3 g6 3 I:!'.b4 .l:!c7 4

ll:lf4 :hc8

The threat was 5 ll:ld5 .

5 as @d6 6 llbl ll:le5 7 h3 ctJd7 8

h4

Gradually occupying space.

8 h5 9 'it>e3 ll:le5 10 l:tc2 @e7 1 1

'it>d4! Wd6
..

On l l . .d8+ White had prepared

1 2 ll:ld5+! ( 1 2 'it>xe5 f6 mate)
1 2 ... 'it>e6 1 3 rt;e3 .
.

1 2 :d2 l:td7 13 @c3+ 'it>c7 14

lidb2 rt;d6 15 a6

Securing a bridge-head for an in

vasion along the b-file.

15 ...1:.dc7 16 :b3 ll:ld7 17 l:tdl +

rt;e7 18 ll:ld3 .l:Id8 19 l:tdbl l:tdc8
20 .l:!b7 'it>d6 2 1 f4 ll:lb6

A mistake in a difficult position.

He should not lose control of the e5
point.

126 Complex Endings

22 tl\e5 'it>e6

The threat was 23 tbxfl.

23 t2Jc4 .l:Ixb7 24 axb7 :gb8 25

tba5 !

Leading to complications was 25

tl\xb6? .l:Ixb7 26 f5+ gxf5 27 exf5+
'it>xf5 28 l:tfl + 'it>g4.

25 ... @d6 26 Wb4 'it>c7 27 cl

'it>d7 28 'it>a3 f5 29 e5 tl\d5 30 e6+
Black resigned.
Weaknesses in the endgame
and their exploitation

The presence of weaknesses in the

pawn structure is a great minus, and
the correct technical exploitation of
them is the most important factor in
realising an advantage.
Topalov-Beliavsky
Linares, 1 995

Black has two passed pawns but

at the same time these are serious
weaknesses which need looking
after.
1 c4 .i.g3

Bad is 1 .. .g5 because of 2 .i.f6 ! .

2 'it>c2 i. f4 3 i.f6 g 5 4 l:te7+

Inferior is 4 .lig8 l::tc 5! with the

idea of ... .l:Ic5-f5
4 ...'it>c8 5 .l:Ig7 l::tc5 6 'it>d3 a5 7
i.e7 l:\'.f5 8 c5 !

White activates his pieces to the

maximum and also the passed pawn,
and in the process threatens to enter
with the king on b5.
8... @d7

He must activate at the cost of a

pawn.

9 i.xg5+ 'it>c6 10 i.xf4 :xr4 1 1

:a7 l:.f3+ 1 2 'it>c4 .U.f4+ 1 3 'it>c3
.:i.f3+ 14 'it>b2 'it>xc5 15 l:Ixa5+

Connected pawns in the majority

of cases win against a single pawn.
15 ...'it>d4 16 1:1'.aS e5 1 7 a5 .l:If6

If 1 7 . .. 'it>c5, then 1 8 1:1'.bS ! , cutting

off the king.
18 :es e4 19 b4 e3 20 'it>b3 l:tfl

Or 20 ... l:\'.f5 2 1 a6 l:l'.e5 22 .l:Ixe5

'it>xe5 23 a7, and in the queen end
ing White wins the enemy queen.

2 1 I:!'.d8+ 'it>e4 22 a6 .l:Ial 23 b5

e2 24 .l:Ie8+ 'it>d3 25 Wb4 'it>d2 26
'it>c5 e l=11Y 27 1:1'.xel 'it>xel

28 'it>b6!

A typical means of advancing the

b-pawn to queen-the only way to
victory.
28 ... 'it>d2 29 'it>a7 Black resigned.

As in rook endings a single weak

ness will usually be defensible, to
achieve victory it will often be
necessary to create a second
weakness.

Complex Endings

Kasparov-Andersson
Belgrade, 1 985

!27

Hiibner-Ftacnik
Polanica Zdroj, 1 995

'":iZ

. . ..
.
e-r
nf! - -.
r,ci . ;i; ;i;
?;,, !Y,
f' m.,,,, .e"""!Y,r
+
u
.. .., i",.
,.!Y,r0{'"'!Y,

All Black's pieces are forced to

protect the b7 pawn, while White
must create a weakness in the oppo
nent's camp. To the question
where?-the reply is simple--0n
the g6 square!

1 h4! @n 2 h5 @g7 3 Wg2 l:!e7 4

.:i.b6 lif'l 5 i.d5 .l:Ifd7 6 lilb5 .l:Ie7
7 @g3 lied7 8 @f4 .l:!c2 9 i;i;>g5 !
J::rxf2 10 hxg6 hxg6 1 1 lixg6 @rs
12 i.b3 tt::lf'l 13 Wf6 f4 14 e4 l:!b2
15 e5 f3 16 e6 f2 17 il.c4 ! Black
resigned.

1 g4! i.d5 2 i.c2 g6 3 tt::lg3

@e7?!

The first dubious manoeuvre

better is 3 ... i.d6 4 tt::le4 i.e7
4 tt::le 4 tt::ld 8 5 g5 !

Beginning a weakening of Black's

position.
5 ... fxg5 6 tt::l xg5 h6 7 tt::le4 i;i{f'l

But not 7 . . . Ci:Je6? because of 8

tt::lc3

8 tt::lc5 i.d6 9 tt::l e4 i.e7 10 tt::lc3

i.e6 1 1 @g2 g5?!

Another
1 1 ...tt::lc 6

weakening-better

is

Commentary here is superfluous

since White's play was direct and
logical.

12 i.e4 i.d6 13 tt::le 2 i.c4 14

tt::l g3 We6 1 5 @f3!

Play in equal positions

15 ...i.d5 16 @g4 i.xe4 17 tt::lx e4

tt::lf'l 18 @h5 b6?!

In equal positions first and fore

most the task is to create weak
nesses in the opponent's camp. This
is usually achieved through the
greater activity of one's pieces,
which, with the help of an attack,
force a weakening of the pawn
structure.

With the threat to go to h5

Stronger is 1 8 . . . @f5 19 tt::lxd6

tt::lxd6 20 @xh6 g4, but White right
ly plays 1 9 tt::lg 3 @f6 20 f4.

19 @g6 ii.rs 20 i.c3 tt::lh8+ 2 1

Wh7 tt:::in 2 2 i;i;>g8 i.e7 23 f3 h 5 24
Wg7 g4 25 fxg4 hxg4 26 @g6 a5 27
i;i;>h5 Wd5 28 tt::ld 2 axb4 29 i.xb4
i.xb4 30 axb4 i;i;>c6 31 i;i;>xg4 Wb5
32 f5 @xb4 33 @g6 tt::ld 8 34 @f6
+ -.

128 Complex Endings

Endings with a pawn sacrifice

In principle, a material advantage

in the endgame is the main factor,
but it can happen that other factors,
particularly the presence of a passed
pawn, are decisive.

Barlov-Seira wan
Zagreb, 1 987

Rublevsky-Shariazdinov
Elista, 1 996

Though Black is a pawn down,

the American grandmaster tries to
exploit the active positions of his
pieces and does this, as we shall see,
successfully.

1...f4 2 gxf4 gxf4 3 lZ'ic4 :gd3 4

al h5 5 h4 'itf5 6 .l:Ia8 Wg4

1 a5! i.h4 2 i.d5 i.a6 3 lZ'ic7!

The threat is b4-b5.

3...i.d8 4 lZ'ie8!
If 4 lZ'ixa6? bxa6 5 i.b7 i.g5 !

with the idea 6 ... i.d2=.

4...@h7 5 'itg2

5 lZ'ixd6? i.e7 would be bad

materialism.
5 ... g5

On 5 ... 'itg6 there is 6 h4! .

6 i.f7!

Zugzwang!

6...i.e7 7 lZ'ic7 i.d8 8 lZ'ixa6!

Now that the g5 square

inaccessible to the bishop.

1s

8 ... bxa6 9 i.c4 'it>g7 10 i.a6 d5

11 exd5 i.e7 12 i.c8 i.xb4 13 a6
i.c5 14 @f3 @f6 15 d6! h5 16 We4
i.d4 17 @d5 @f7 18 i.f5 @f6 19
@c6! Black resigned.

Black obviously does not rush to

capture the white pawns, for the
time being improving the position
of his king.

7 g8+ @xh4 8 J:ig6 J:ixd5 9 Wf3

l:td4 1 0 .l:Igl i.c5 1 1 @e2 i.b4 1 2
@ f3 i.c5 13 @e2 l:te4+ 14 @ f3 d5
15 lZ'id2 es 16 J:id l @gs 17 lZ'ibl
d4 18 lZ'id2 wrs 19 :ghl .:i.h8 20
lZ'ie4 i.b4 21 J:igl h4

Slowly the black pawns crawl

along to their queening squares.

22 l:1g5+ @e6 23 _:g6+ Wf7 24

l:tgl d3 25 Wxf4 d2 26 'ite3 l:td8 27
We2 .:i.e8 28 f3 h3 29 .:bl :ds 30
'it>dl

If 30 lZ'if2 Black has 30 ... h2 3 1

l:1xh2 d l=+ 32 lZ'ixd l .:i.d2+.
30 ... aS 31 We2 h2 32 lZJf2

If 32 lixh2?, then 32 ... d l =v&'+ 3 3

@xdl lr.a l + and 3 4...lia2+.

32 ... J:ieS 33 lZ'ie4 .l:td8 34 lZ'if2

i.c5 35 lZ'idl i.gl - + .

Complex Endings 129

Chekhov-Karsa
Lvov, 1 983

1 @f2 e5 2 J:id2 CZJa4 3 ..tn lLib6

4 ClJc2 ..tc4 5 ClJe3 ..tb3 6 g4 h4 7
g5 ..te6 S ClJg2 h3 9 ClJh4 ..tb3 10
..txh3 llJc4 11 .!ide2 .!idS 12 f4
..tdl 13 .li(e4 llJxb2

The strategy is completed and the

game
transposes
to
tactical
channels.

1 b5! cxb5 2 a6 ..tc4 3 ..txc4

bxc4 4 ClJc3 ..tc5

14 @gl .!id3 15 ..tn l:txc3 16 f5 !

..tc2 17 f6 l:teS l S J:Ixe5 .!lxe5 19
.!lxe5 l:txa3 20 .Iles+ 'it>h7 21 .!lf8
..tb3 22 llJf3 l:!a4 23 llJe5 l:!e4 24
lLixf'l .l:Ig4+ 25 @f2 ..txf'l 26 .l:Ixf'l+
'itgS 27 g7+ f8 2S .!lxg6 J:if4+
29 'itel J:Ie4+ 30 @d2 ClJc4+ 3 1
..txc4 l:!xc4 3 2 l:th6 Black
resigned.

5 ..txc5 lLixc5 6 a7 as 7 ClJd5

c3! S llJxc3!

Domination is a superiority in the

functional activity of each piece. In
the endgame this increases many
times over.

Also bad is 4 ... ..tb4 5 llJd5 ..ta5

because of 6 ClJc7! and 7 a7 ! ,
deciding the game.
Black returns the pawn and White
must not fall for 8 tLlxb6? .l:Ixa7 ! 9
l:!xa7 c2 1 0 .!lal ClJb3 with a draw.
S ... ClJb3 9 l:ta6 ClJd4 10 ClJd5 lLic6
1 1 .!lxb6! lLixa7 12 l:ta6! Black
loses a piece, therefore he resigned.
Shirov-Van Wely
A msterdam, 1 995

Morozevich-Balashov
No vgorod, 1 997

. . fa
..
. .,.

-.-)
- (''"" .

-:'fJJ''% % '/ l
. 0;a
.
1
,.,,..

1 h4! lLie6 2 'i1Nf6 h5

There is no other
defence against ClJg4.
3 ClJd5 c2

Black has achieved the ideal

structure for the Sicilian Defence,
but now he begins a principally in
correct advance on the king's flank.

Defending
lLixg6.

apparent

against lLie7+

and

4 ClJe7+ 'it>f8 5 ClJcS! 'fl'e4+ 6 gl

'fl'd4

If 6 .. . 'itg8, then 7 ClJd6.

130 Complex Endings

7 e7+ 'it>g8 8 tt:Jd6 'i!Vdl + 9 Wg2

Or 1 5 ... 'it>h6 1 6 tt:Jf7+ 'it>g7 1 7

tt:Jg5+.

10 'it>h2! ctJe2 1 1 'fl'xti+ 'it>h8 12

e8+ c;i;>h7 13 'fl'd7+ 'it>h8 14 'fl'c8+
'it>h7 15 'iltb7+ 'it>h8

f7+ @h8 19 'ikf6+ c;t>h7 20 @g2!

tllf4+
The last chance.

1 6 'iltbs 'it>h7 17 'Yifxa7 @gs 1 8

Black resigned.

Defence in the Ending

A complicated, important and

exceptionally large theme. We can
touch upon it only partially.

Fercec-Cvitan
Zadar, 1 995

Playing for stalemate

How strange it is that this method,

for example, rather frequent in pawn
endings with an extra pawn, is sel
dom met in its purest form.
Beliavsky-Gelfand
Belgrade, 1 997
1 . .i.g2? 2 11xg3! ! 'ii'xg3 3 'Wh6
'it>g8 4 Wih8+ \t>ti 5 \\Ve8 Drawn.
.

Exploitation of a lucky chance

Gelfand-Lau tier
Belgrade, 1 997

The threat is ... @h3-h2 and then

g2+, so

1 llf.3! ! g2+ 2 \t>xg2 i.xf3+ 3

@xf2 Drawn.

132 Defence in the Ending

Here winning are 1 l:tc 1 or 1 Wd2,

but Gelfand decides to play more
simply.

Difficult practical defence

in the ending
Beliavsky-Mikhalchishin
USSR (ch), 1 984

1 l:tc5??,

Now l .. b4! is winning. However

after...
.

1. .i.c4? 2 'it>d2 Black resigned.

.

Shirov-Lautier
Belgrade, 1 997

Thinking a long way ahead,

Lautier went for the variation
1 l:txe4 Wxe4 2 f6

Reckoning on 2 ... i.e3 3 'it>h4

i.h6 (3 ...Wf5 4 'it>xh5 'it>e6 5 Wg6)
4 'it>xh5 i.f8 5 'it>g6 d5 6 h4 d4 7
i.a5 d3 8 h5 'it>e3 9 h6 with a win.
However there followed
2 ...i.xf6 ! ! 3 i.xf6 d5 4 Wh4 d4 5
'it>xh5 'it>f3 ! 6 'it>g5

Or 6 b3 g3 7 hxg3 Wxg3 8 Wg5

@f3 9 @f5 d3 1 0 i.c3 'it>e2 1 1 'it>e5

d2 12 i.xd2 'it>xd2 13 'it>d5 Wc2,

reaching the pawns.

6 ...d3 7 i.c3 c4 ! 8 @rs 'it>e3 9

Wxg4 d2 1 0 i.xd2+ @xd2

And a drawn ending with a c

pawn is reached.

Black has two pawn weaknesses

on e6 and b7 (For the present Black
is coping with defending them) and
a rather worse deployment of
pieces. The question is-how can he
improve his position? Correct is
1 .. .g5 ! , commencing play on the
king's flank, but Black played
1. . e5? !
.

I t i s not usually worth moving a

weakling.
2 f3 'it>d5 3 lba4 lbe8? !

Too intricate-simpler is 3 ... @d6!

4 lbb6 tbd5 5 We4 lDf6+! with a
draw.

4 lbc3+ 'it>e6 5 'it>e2 lbd6 6 b4

@d7

Black wants to defend the b7

pawn with his king, though also not
bad is 6 . ..c7
7 J:ib6 Wc7?!

Very strong was 7 .. .1H6 ! , and

then 8 .. g5 and 9 . .l:!.h6, obtaining
counterplay.
.

8 lbd5+ 'it>d7 9 b3 h5?!

Defence in the Ending 133

This is only a weakness-better

was the transfer of the rook via f8,
c8 to c6 The text just weakens the
g6 square.
10 lll e3 rbc7 1 1 ll'lc4! l:id7 ! !

After 1 1 ...ll'lxc4 1 2 bxc4 the rook

proceeds to g6 and then to g5.
12 1Ib4

If 1 2 lllx e5, then 1 2 ...lle7 ! 1 3 f4

ll'lf7 with a draw, while the pawn
ending after 12 lllxd6 J:ixd6 1 3
xd6 'it>xd6 1 4 'it>e3 'it>c5 1 5 g4 b 6 !
was also drawn.
12 e4! 13 lll x d6 exf3+ 14 'it>xf3
l:rxd6 15 'it>e4 b6!
..

1 'it>c4 h4!

Cutting down pawn material is al

ways useful for the defence!
2 'it>d3 hxg3 3 hxg3 i.d8 4 'it>e3
..tg5 5 'it>f3 'it>e8 6 ..td6 rbd7 7 i.e5
c,t>es 8 'it>g2 'it>d7 9 'it>f3 'it>e8 1 0
'it>e2 c,t>d7 1 1 'it>d3 ..td8 12 'it>c4
i.b6 13 f3 i.f2 and White cannot

improve his position. Therefore it's

a draw.
Rozentalis-C hristiansen
Groningen, 1 992

Completely equalising by ex
changing the important a5 p awn.
16 .l:!c4+ 'it>d7 17 b4 bxa5 1 8
bxa5 z:!'.c6! and drawn.

A quite rare means of defence is

without pawns-against a central
pawn and bishop he succeeds in
impregnable
an
constructing
fortress.
Gretarson-Magerramov
Groningen, 1 993

l f4

He should try 1 g4 ..tb l 2 g5, pre

paring an attack on the h7 pawn.
1. .. 'it>d6 2 rbf2 l:!e7!

Not allowing the white king into

the centre.
3 i.f3 h5 4 a6!? i.g4 ! ?

A n interesting decision-the ex
change of bishops weakens the d5
pawn.
5 ..te4 ..tf5 !

Seeing the idea through to the

end!

In endings with same-colour

bishops very often an extra passed
pawn 'will not run' .

6 .to ..tg4 7 ..te4 ..trs 8 axb7

llxb7 9 Ji(c6+ 'it>e7 10 i.xf5 gxfS 1 1
'it>e3 l:!xb2 12 'it>d4 l:txh2 1 3 'it>eS
.lle2+ 14 'it>xf5 lie3 15 lic7+ 'it>d6
16 lixa7 lhg3 17 l:!xf7 'it>xd5

I 34 Defence in the Ending

Gurevich-Andersson
Leningrad, I 987

An interesting rook ending

Black must play exceptionally
accurately in order to hold the
position.
19 lih7 Wd6!

The king must come closer at any

price.
20 'it>f6 l:!.h3!

But now he needs to advance his

pawn.

2 1 .l:Ih6! ? 'it>d7 22 .l:Ih8 llhl 23 f5

h4 24 'it>t7 h3 25 f6 h2 26 .a'.h6 c7
27 .l:Ih3!?

White wants to transfer his rook

to the second rank, and then, with
the rook on e2, try to enter with his
king. But Black too activates his
king.

27 ... 'it>d6! 28 1::rd 3+ 'it>e5 29 l:!.d2

\t>f4 Drawn.

Black heads for the h2 pawn and

then plays .l:Ifl .
During defence, mistakes due to
tiredness from protracted defence
can occur.
In the following game, over the
course of the last 50 moves, with the
same correlation of forces, White
had stubbornly tried to improve the
positions of his pieces and pawns,
while Black defended technically
and cooly.

1 g5

Finally White goes over to the

attack.

1. .. hxg5 2 fxg5 lLld7 3 'fl'e4

'fl'b3+ 4 'it>h4 lLlf8 5 tLle7+ 'it>h8 6
'fl'f4 lLlh7

Leading to a quick defeat. It

seems to us that at the very last mo
ment Black could still surprisingly
save himself, in a truly fantastic
way, by playing 6 .. 'it>h7 ! ! . Now 7
'if5+ g6 8 'i'xf8 leads to an im
mediate draw in view of 8 ...'fl'g3+
('i/Nxh3+) 9 'it>xg3(h3) stalemate! .
Meanwhile Black threatens after
7 ... tLlg6 to bring the game down to a
drawn queen ending, therefore it is
necessary to accept the knight sacri
fice: 7 'fl'xf8 'fl'c4+! (but not
7 ...'iib4+? 8 'it>h5 nor 7 ...'fl'a4+? 8
'it>g3 'fl'a3+ 9 'it>g4! 'ili'a4+ 1 0 'fl'f4
'fl'd l+ 1 1 'iVf3, and White wins) 8
g3 'il'd3+! (again avoiding a crafty
pitfall: 8 .. 'fl'c3+ 9 'il'f3 'ie5+ 1 0
g4 ilixe7 D _g6+! mating) 9 '1Wf3
(On 9 'it>g4 'iie4+ 1 0 f4 'fl'xe7
White lands his queen in a poor po
siiton- 1 1 g6+ already does not
lead to mate. On 9 Wg2 follows
9 ...'i/Ne2+, while 9 \t>f4? is imposs
ible because of 9 . .'fl'fl +) 9 . . .d6+
.

Defence in the Ending 1 35

1 0 \t>g4 'ib4+! (this is the point of

Black's idea: he refrains from the
immediate capture of the knight-in
view of 1 1 g6--and with a correct
intermediate check disrupts the
coordination of White's pieces) 1 1
h5 'iVxe7 1 2 'iVf5+ 'it>g8 1 3 'iVc8+
fl with a drawn queen ending.
7 'iVg4 'iVc2 8 'iVf3 Black
resigned.

Gelfand-Shirov
Munich, 1 993

You should always take the op

portunity to spoil the opponent's
pawn structure.
White has two extra pawns and, it
seems, an easily winning position.
But...
1 . .h3! ! 2 gxh3?
.

How could he not take a third

pawn? Correct was 2 g4 tLld2 3
.l:td4! .!Llxe4 4 J:Ixe4 lixc3 5 'it>g3,
and by defending the f3 pawn with
the rook White then wins the h3
pawn.

2 ....!Lid2 3 i.d3 l:1xc3 4 1Id4 c6

5 'it>e3 .!Llb3 6 lid5 .!Lies 7 i.b5
lie6+ s Wf2 .!Llb7 9 h4 .!Lld6 1 0
i.d7 lif6 1 1 i.g4 .!Lle8 1 2 h 5 g5 !

The pawn must hold two white

pawns.
13 lid8 .!Lld6 14 'it>e3

The last chance would have been

14 lih8+ 'it>g7 1 5 I:!'.a8 Wh6 1 6 l:ta5
.!Lle8 17 'it>g3 followed by h2-h4.

14 ....!Llf5+ 15 'it>e4 .!Llh4 16 'it>e5

.l:tf4! 17 :!:i.d3 'it>g7

Black has blockaded everything

excellently-therefore it's a draw.

The isolated pawn

in the Ending

The isolated pawn is a weakness,

but this weakness is felt in different
ways. The strategy everywhere is
similar-blockade
by
utilising
strong blockading pieces or direct
attack on the isolani itself.
We begin with the simplest
example-a pawn ending.

7 'ii>f6 'ii>c4 8 'it>xg6 c,t>d3 9 e4

Black resigned.
Yurtaev-Temirbaev
Elista, 1 998

Ehlvest-Rausis
Riga, 1 995

Here the b6 point is firmly de

fended and White does not succeed
in penetrating via the c5 square.

Sooner or later Black will run out

of useful moves, and he will be
forced to allow the white king into
his camp via the e5 or c5 squares.

1 . . b6 2 a4 c,t>e6 3 f3 @d6 4 a5
bxa5 5 bxa5 'ii>c6 6 'it>e5 'ii>c5
.

After 6 ...'it>b5 7 'ii>xd5 Wxa5 8

'ii>c5 White also wins.

1 h4 h5 ! 2 a4 'ii>e6 3 a5 c,t>d6 4 a6
'it>e6 5 e4

Sooner or later it is necessary to

try and exploit the slightly better
position of the king, but this is
insufficient.

5 ... dxe4 6 fxe4 fxe4 7 @xe4 g6 8

'ii>f4 'it>f6 9 g3 b5 10 g4 hxg4 1 1
'it>xg4 'it>e6 1 2 'ii>f4 'ii>f6 1 3 'ii>e4
c,t>e6 Drawn.

The Isolated Pawn in the Ending 1 3 7

Pelling-Miles
I/ford, 1974

1 i.f5 ! i.a6 2 f4

It is important to exploit his ma

jority on the flank.
2 ... lZJg8 3 h5 ctJf6 4 c;t>f3 i.b7 5
i.g6 @e7

Bad is 5 ... i.c8 6 g5 i.g4 7 Wg3

i.xh5 8 gxf6 i.xg6 9 fxg7 with a
winning position.
6 g5 lZJe4 7 i.xe4!

The knight here is undoubtedly

stronger than the bishop.

1 . .. a5!

Weaker is l .. .h5 2 b4 ! f6 3 h3 g5
4 fxg5 fxg5 5 @e3 e5 6 dxe5 @xe5
7 'itd3 Wd5 8 c;t>e3 @c4 9 c;t>e4 with
a draw.

7 ... dxe4+ 8 @e3 .id5 9 b4 axb4

10 axb4 .tf7 1 1 lZJf5 @rs 12 lZJg3!
.ib3 13 @xe4 Wf7 14 c;t>es hxg5 15
fxg5 .ic2 16 g6+ c;t>g8 17 @f4 .ib l
1s c;t>gs .td3 19 CZJrs c;t>rs 20 lZJd4
.ic4 21 h6 gxh6 22 c;t>xh6 i.a2 23
@gs c;t>g7 24 lZJf5+ and White won.
Panno-Donner
Palma de Mallorca, 1 967

2 b3 a4! 3 bxa4 bxa4 4 h3

After 4 g4 g5 ! White is badly

placed, while on 4 c;t>c3 there is
4 ...e4 5 c;t>b4 @xd4 6 xa4 @c4 !
7 c;t>a5 f6 8 b6 e5, and the queen
ending is easily won since Black
manages to exchange queens.
4 ... h5 5 h4 f6 6 @c3 We4 7 @c4
f5 ! 8 Wc5 We3 9 b4 f3 10 c;t>cs !
Wxg3 1 1 'itd6 @xf4 12 xe6 @g3,
and Black just as easily won the

queen ending.

Rausis-Farag6
Germany, 1 996

How is it possible to win this

position for White?

1 .ib5 lZJe4 2 f3 lZJec5 3 @f2

lbe6 4 i.c3 ctJc7 5 i.a4

The problem is that while for

Black there is no active plan-for
White there is.
5 ...We6 6 i.b4 ctJe8 7 e2 ctJd6 8
.ib3

There appears the first threat of

a3-a4.
8 f5 9 h4!
.

Now already arises a real threat to

exchange the dark-squared bishop,

138 The Isolated Pawn in the Ending

after which all Black's weaknesses,

riveted on white squares, will be
exposed.
9 ... t'Lle5 10 i.c5 t'Lld7

After 10 ... t'Lldc4 1 1 i.a4 t'Lld6 1 2

i.xd6! 'it>xd6 13 i.e8 rt;e7 1 4 f4!
t'Llf7 1 5 i.xf7 White wins in the
pawn ending.
1 1 i.d4 t'Llb5?

Better in fact was 1 l ...t'Llf6 con

ceding the d5 pawn.
12 i.b2 t'Llc5 13 i.c2 rt;d6 14 g4!
hxg4

Two parts of the plan have been

completed, White goes over to the
third.

7....l:Id8 8 c3+ rt;b7 9 g4! :i.d6

10 a3 a5 1 1 h4 i.f7 12 f5 ! l:id8 13
:c2 !

A prophylactic move, eliminating

the threat of l:id6-c6.

13 ....l:Id7 14 g5! hxg5 15 hxg5

fxg5 16 l:l'.g2 :i.d6 17 .l:!xg5 g6 1 8
fxg6 lixg6 19 i.xd5+ Black
resigned.

Also no help is 1 4 ... 'it>e6 1 5 gxh5

gxh5, since he cannot hold the two
weaknesses on f5 and h5.

Nikolic-Portisch
Ter Apel, 1 996

15 fxg4 'it>e6 16 h5! t'Llc7 1 7 hxg6

Black resigned.
Szabo-Penrose
Bath, 1 973

1 .l:!c2 !

White plan consists of three parts:

(a) transfer of the king to d4;
(b) transfer of the bishop to f3 ;
(c ) transfer o f the rook t o the
eighth rank and an advance of the
pawns on king's flank.
1 l:Ia4 a6 2 :b4 l:1d7 3 '>t>c3 rt;e7

Or 3 ...d4+ 4 .i:txd4 .:i.c7+

(4 ... l:Ixd4 5 rt;xd4 i.xa2 6 i.c4+!
with a winning pawn ending) 5 rt;b4
i.xa2 6 b3 lie 1 7 rt;a3 i.b 1 8 rt;b2,
winning.
4 rt;d4 rt;d8 5 i.e2 rt;c7 6 i.f3 b6
7 J:!b3

A very important prophylactic

move-the a2 and :f2 pawns are the
main target for Black's counterplay
and therefore ought to be defended
to the maximum.
1 ...t'Lle4 2 t'Lle5

In these kinds of positions block

ading the isolated pawn is not so im
portant-the main thing is to attack
it, as Bent Larsen said about such
situations!
2...lib7 3 rt;n rt;f8 4 rt;e2 rt;e7 5
g4 t'Llf6 6 h3 g6 7 f3 h5

As the defender is supposed to do,

Black tries to exchange pawns.
8 g5 t'Lle8 9 rt;d3 t'Llc7 10 I:!'.c5 d4

Black has to sacrifice a pawn,

since 1 0 ... rt;d6 1 1 .l:Ic6+ 'it>xe5 1 2
f4+ leads t o mate.

The Isolated Pawn in the Ending 139

1 1 exd4 .!Lle6 12 J::ra5 .!Llxg5 13

.!Lixg6+ Wf6 14 .!Llf4 and Black with
a clear conscience can resign.

16...i.c7 17 .!Llc6 f6 1 8 f4

With the threat of lt>d4, h3, g4

etc.

18 lt>f7 19 lt>d4 lt>e6 20 h3 .l:Ih7

21 .!Llc3 !
..

Magerramov-Kohlmeyer
Bad Worishofen, 1 993

Play over the whole board!

2 1...l:!.d7 22 g4 hxg4 23 hxg4

Ji..d 6?!

More stubborn is 23 ... J::rh 7 24

.!Llxd5 J:id7, in spite of 25 f5+! gxf5
26 gxf5+ lt>xf5 27 .!Lle7+ lt>e6 28
J::rc7 l:!.xc7 29 .!Lixc7+ We7 30
.!Llxd5+ Wd6 3 1 .!Llf6!, Mager
ramov.
24 .!Lla4 i.c7 25 J:ihl ! l:!'.g7 26
.!Llc3 .l:Id7 27 J:ih6 g7

If 27 ... i.xc6 28 bxc6 J:id6 29

l:!.xg6 White also wins.
An instructive example of the
struggle of two knights against two
bishops, with rooks on the board,
this being a rare case of the com"'
plete superiority of the knights.
1 .!Lla4

As Nimzowitsch said, weaknesses

do not tell unless they are attacked!
1...d6 2 acl z:!'.c8 3 J:Ixc8+

Also 3 f3 ! ? deserves attention

-Magerramov.
3 ... i.xc8 4 .!Libs li:td8 5 .!Lld4!

Worse is 5 .!Llxb6?! ii..g4 6 l:!'.xd5

.l:Ib8 with strong counterplay.

s .l:!d6 6 f3 i.d7 7 .!Llc3 h5 8 Wf2

lt>f8 9 lt>e2 Ji..c8

Upon 9 ... @e7 10 Wd3 ii..e 8 1 1

:Ic l nothing good awaits the black
king in the centre.
10 b4! axb4 1 1 axb4 i.a6 12 b5

Fixing a second weakness on b6.

12 ...i.b7 13 l:.al i.f6 14 lt>d3

l:.d7

The only move. If 1 5 ... lt>e7? 1 6

.l:ta7 l:!.d7 1 7 l:!.xb7 ! l:txb7 1 8 .!Lixd5+
White wins.
15 .!Lla4 Ji..d8 16 .l:Icl

Successfully combiniag threats

along the a and c-files.

28 .!Llxd5 d7 29 f5+!
Black resigned in

29 ... gxf5 30 l:!'.xf6 mate.

view

of

A case of an isolated pawn with

bishops was given in the book
Winning Endgame
Technique.
Bishop against knight is also a very
well known endgame, and the side
with the isolated pawn has to main
tain the equilibrium.
Very interesting also is the case of
rook endings with an isolated pawn
for one of the sides .
Bareev-Farag6
Rome, 1 990

140 The Isolated Pawn in the Ending

1. .. h5?

He should immediately take

measures on the queen's flank by
l . . . a5, though after 2 g4 White still
maintains the advantage.
2 b4 'it>e6 3 'it>d4 f6 4 h4 'it>f5 5 t3
'it>g6

If 5 ...g5 6 hxg5 fxg5 7 1If8 'iitg6 8

We5 there is no defence against 9
e6.
6 a4 ! f7 7 a5 e6 8 a6!

The weakness on a7 is now

permanent.

8 ... 'it>d6 9 b5 e6 10 g3 g6 1 1
.l:Ie8+! d6 1 2 .l:If8 'itie6 13 !!c8
d6 14 .l:Ic6+ e7 15 g4 @f7 16
gxh5 gxh5 17 l:tc8

Now it will be very difficult to

hold on at once to the three wea
knesses on a7, d5 and h5.
1 7 e6 18 z:!'.e8+ 'iitd6
..

If 18 Wf7, then 19 l:tb8 and 20

.l:Ib7.
...

19 .l:Ih8 l!i.c7 20 lid8+ e6 21

lixd5 f5 22 e4 and Black can
resign.

Rook and bishop against

rook and knight

In general the line-up of rook plus

bishop is rather stronger than rook
and knight (to a large extent this de
pends on the structure), but a few
grandmasters, for example Evgeny
Sveshnikov, think the reverse. Rook
and knight is stronger when there
are fixed weaknesses or, for
example, when there are doubled
pawns.
We look at a few cases of the ad
vantage of rook and knight, since
with this line-up the methods of
play are more complicated.
Sakaev-Novikov
Moscow, 1 998

White begins a plan associated

with an attack on the a5 pawn or the
forcing of the advance ... a5-a4.
6...c,t>e7 7 l:!.al ! @d6

After 7 . . . d7 8 a4 l!tb4 9 lllxa5

l:!.xa4 1 0 !!xa4 xa4 1 1 b4 and
then 12 c5 White wins the d5
pawn.
8 a4 b7 9 lllxa5 c7+ 10 @d2
d7 1 1 lllb3 c4 12 a5 c6

No help is
winning.

1 2 . .a4
.

13 a6 .llb4 14 lll a 5! a8 15 'itic3

b8 16 lllb3 J:ie8 17 @d3 @c7 1 8
lllc5 @b6 19 !!bl+ a7 20 c,t>d2 !

Preparing an exchange of rooks

-technically the simplest way to
win.
20 ...c6 21 .li(b3 l:re7 22 l:!e3
with an easily winning position.
Mikhailov-Volchok
corr, 1995

1 d2 b4 2 llla 4!

There is no sense in White open

ing the a-fil e-but here it is possible
to open the b-file.
2 ...bxa3 3 bxa3 :h7 4 'it>c3 c,t>f8
5 lllc5 l:!.b5 6 lllb3 !

1 3 a6!

J 42

Rook and bishop against rook and knight

Black has a majority on the king's

flank, but the knight has to carry out
an unpleasant defensive function
and, in addition, the pawns are ham
pered by their white counterparts.

Milov-Pelletier
Biel, 1 997

1 g5! J:id5 2 .!lgl <Jtc6 3 'itc2 l:!.f5

4 J:ih l !

With the interesting threat, after

4 . . . hxg5, of 5 i. xg7 ! cuxg7 6 h6+ -.
4 ...c.t>d7 5 <Jtd3 .!lb5 6 gxh6 gxh6
7 i.xh6 cuf6 8 i.g5 cuh7 9 .!ih2 !

Again prophylaxis-utilising the

rook to the utmost.

9 ... e5 10 <Jtc4 'it>e6 1 1 a4 .!lb8 1 2

1::re2 f6 13 i.h6 l:tg8 14 'it>c5 cu f8 15
'it>b6 Black resigned.
Alterman-Chemin
Beersheva, 1 992

If we remove a pair of pieces

from the board, then White is better.
With the interchanging of threats the
advantage lies with Black.
1 'itigl .!lc3 2 !:!.el cug4 3 e4 @f6
4 h3 CUe5 5 J::rf 1+ 'itie7 6 .!lel hS!

Improving his position!

7 .!le2 h4 8 l:rel CUg6 9 @f2 @f6

10 l:te3 .!lcl 1 1 .!if3+ cuf4 1 2 g3
hxg3 13 .!lxg3 I:l.hl 1 4 a4 b l ! 15
.llc3 <JteS 16 a5 :Ial 1 7 .!lc5+ 'itid6
18 l:l'.b5 l:ta3 19 <Jtg3 'it>c6 20 .!lb6+
'it>c7 21 @g4 .!lxa5 22 .!lb4 CUd3 23
l:l'.d4 cue5 24 @h5 g4! - + .

1 . ..g5! 2 a4 :Ia3 3 1::ra 6?

Better is 3 i.b5 J:i a l + 4 Wh2

cug6+.

3 ... 1:1'.al 4 <Jtg2 cug6 5 i.b5 cuh4+

6 'itih2 cuxf3+ 7 'itig2 tt:::lh4+ 8 <Jth2
e5! 9 i.c6 f5 10 f3 l:!.a3 1 1 l:ta7
cuxf3+ 12 'it>g2 cud4 13 i.d5+ 'itih8
14 a5 e4 15 a6 e3 16 i.c4 f4! 17 h4
f3+ 18 <Jth3 e2! White resigned.

It is very important to control the

moment of exchanging rooks or
minor pieces, which can often alter
the assessment of the position. At
times, with pawns on different
flanks, the knight can even prove to
be the stronger piece, though
usually this is the more long-range
bishop.
Gausel-Agdestein
Reykjavik, 1996

Rook and bishop against rook and knight I 43

Not only are White's chances not

better, but Black simply has a clear
advantage.
1 ....l:Ie4!

A surprising exchange of rooks.

2 .l:Ixe4 lllxe4 3 i.c2 l'llc3 4 g4

d6

It was also possible at once to fix

the opponent's queenside pawn
structure by 4 ... l'llb 5 ! ? 5 a4 l'lld 6.

5 el @c5 6 @d2 l'llb5 7 a4 l'llc 7

s @c3 l'll d5+ 9 Wb2 @b4 10 g5 b5
1 1 axb5 cxb5 12 i.e4 l'll c 3 13 i.c6

White wants to obtain counter

chances by doing away with the
black f7-pawn.

13 ...l'll d l+ 14 @a2 t'llx f2 15 i.e8

l'll d3 16 i.xf7 llle 5!

Now the knight dominates the

bishop, while the king transfers to
the other flank.

17 i.e8 @c5 18 @b2 b4! 19 Wc2

Wd5 20 h5 gxh5 21 i.xhS e4 22
@d2 g6 23 i.dl @f5 24 @e2 'lt>xg5
25 @f2 'it>h4 26 g2 l'll d 3 27 i.e2
lllc5 28 i.dl g5! 29 g3 'it>f6 30
@g4 e5 31 i.c2 l'll x b3! 32 i.xb3
@e4 White resigned.

2 l'llh 3! ..tcs 3 lllg5 Wf6 4 Clio

We6 5 I:!'.b5 !

For the time being activating the

rook.

5 ... i.d7 6 I:!'.b2 Wf6 7 l'll e5 ..ta4 8

lib8 @g7 9 @d2 @f6 10 c3 @g7
1 1 @b2

Also possible was 1 1 d4.

11 ... 'it>h7 12 @a3 i.c6 13 @b3

g7 14 a4 i.b7 15 a5 ..tcs 16 @c3
f6 17 .l:Ib2 g5 18 lih2 ! gxf4 1 9
gxf4 e6 20 d 4 cxd4+ 21 xd4 @e7
22 c5! i.a6 23 .l:Ih8 i.fl 24 lih7+
d8 25 llhl i.g2 26 .llg l i.d5 27
1Ig8+ e7 28 a6 @f6 29 1Ig6+ @e7
30 .l:Ih6!

Zugzwang!

30 ...i.g2 31 lllg6 @es 32 l'll h 4!

i.d5 33 l'll xf5 Black resigned.
Genba-Irzhanov
Russia, I 998

Beliavsky-Neverov
Koszalin, 1 998

Sometimes chessplayers them

selves do not know how to exploit
the possibilities of their own pieces.
1 i.xa6?

Driving the bishop to a passive

position. He should play 1 a4! 1Ia8
2 b5, creating a _passed pawn.
1 ... f5?

Better was l . . .i.f5 2 g4 i.d7 3 g5

i.f5 4 lll e4 i.xe4 5 xe4 f5+ with
chances of holding the rook ending.

1 ...li:ta8 2 b5 t'lld 7 3 1Ia4?

Better really was to leave the

pawn and activate the bishop by 3
i.b7.
3 ...lllc5 4 .l:Ia3 @f6

1 44 Rook and bishop against rook and knight

Now the black king enters the

game with decisive effect.

S 'it>g2 ri;;e7 6 'it>f.3 d4! 7 'it>e2

ri;;d 6 8 Wd2 'it>c7 9 h4 'it>b6 10 ri;; d l
Iif8 11 ri;;e2 .l'.td8! 1 2 lilt3 d3+ 13
'it>dl ttla4!

Kamsky-Cvitan
Palma de Mallorca, 1 989

Nobody wants the bishop on a6!

14 ri;;d 2 ttlb2 lS a4 ttlc4 1 6 'it>dl

d2 17 h!b3 hS 18 aS+ 'it>xaS White
resigned.

With a struggle on one flank the

advantage of knight over bishop in
conjunction with some other pieces
can be very great.
Lastin-Alferenko
Moscow, 1 998

1 ll:lfS+ 'i;;g 6 2 .l.Ig8+ f6 3 ll:le3 !

j_cs 4 ll:ldS+ 'it>e6 s liXe8+ 'it>d6 6
.l'.td8+ 'it>c6 7 l:tc8+ 'it>bS?!

Obviously better was 7 ... Wd6,

though even here after 8 h!c7 Black.
has serious problems.
8 .l.Ic7!

After this
becomes clear.

move

everything

8 ... 'i;;c4 9 l!xti l:ia2 10 .l'.tf.3! ri;;d 4

11 ttlf6 h4 12 gxh4 j_e7 13 hS l:t'.a6
14 ll:lgS! j_gs lS .i:tg3 j_f4 16 l:t'.g6
Black resigned.
Gulko-Sveshnikov
Volgodonsk , 1 983
1 Wf.3 llc7 2 l:bS -*.f6 3 ttlcs
j_e7 4 ttld3 j_f6 S l:b4!

Guaranteeing the centralisation of

his king.
S ....l.Id7 6 We4 .UdS 7 l:b7+ 'it>f8?

Slightly better is 7 ... 'i;;g 8.

8 f4! exf4 9 ttlxf4 Iles+ 1 0 Wf.3

:rs 11 e4 l!aS 12 ttlxg6+ 'it>g8 1 3
ll:lf4! Black resigned.

Rook and bishop against rook and knight 145

Activity of the pieces always

comes before material advantage.

1 .l:Ib7 .l:!xa2 2 .l:Ixb6 'it>g7 3 :g_b7+

'it>h6 4 i::i.e7

White tries to limit the activity of

the black knight.
4 .l:!b2 S l1e3 libl 6 'it>g2 .l:Ib2 7
'it>h3 'Lids 8 :d3 'Lif6!

..

Again trying to control the e4

square.

9 .l:Ie3 ctJdS 1 0 .l:If3 'Lif6 1 1 g4

hxg4+ 12 i.xg4 gS !

The transfer to the rook ending is

lost: 1 2 ... 'Lixg4? 13 'it>xg4 xh2 14
J::[.h3+.
13 fxgS 'it>xgS 14 i.e6 'Lie4!

Finally the knight achieves its ob

jective-the pawn is not enough but
the activity of the scattered pieces is
sufficient to achieve a draw.
IS lifS+ 'it>g6 16 l:!.bS 'Lid2 17 b4
'Lin 18 i.rs+ c,t>f6 19 i.d3 lixh2+
20 lt>g4 .l:If2 21 .l:IcS 'Lie3+ 22 'it>g3
:d2 23 i.e4 .l:Id4 Drawn.

However you cannot always man

age to achieve such coordination
with a rook and knight. Very often a
rook and bishop compliment each
other splendidly.

Preventing both . . f7-f5, and also

... 'Lie8 d6 f5
.

2 ... 'it>d8 3 b4 .l:Ie7+ 4 'it>d3 axb4 S

:g_xb4 lt>c7 6 l:!.bl 'Lif6 7 aS ctJd7 8
al !

With the threat of 9 a 6 bxa6 10

:g_xa6 attacking the c6 pawn.
8...'Lib8 9 h4!

The knight has gone over to the

opposite flank. There he can attack
on the other side.

9 ...'Lia6 10 .l:Ibl 'Lib8 11 i.f3

'Lid7 12 .l:Ial l:te8 13 llfl lie7 14
i.g2 'Lib8 IS i.e4 'Lia6 16 .l:Ibl
'Lib8 17 i.rs 'Lid7 18 l1al g6 1 9
i.xd7! .l:!xd7 20 fl 'it>b8 2 1 .l:If6!
'it>a7 22 hS !

Black's weaknesses
crystallise.

start

to

22 ...'it>a6 23 gS!

The passed pawn comes first!

23... hxgS 24 h6 'it>xaS 2S h7 .l:Id8

26 .l:Ixf7 bS 27 cxbS 'it>xbS 28 .l:!b7+
'it>a6 29 g7 .l:Ih8 30 'it>e4 and

White has a winning ending.

Balashov-Vaganian
USSR (ch), 1972

Karpov-Hort
Tilburg, 1 979

1 'Lic2!

1 .l:Ibl! 'it>e7 2 g4 !

White plans to reconstruct with

his king on d4 and knight on e3.
Now Black should go over to
counterplay by I ...lib5 2 'Lib4 i.e6
3 c,t>d4 f6 with a tenable p osition.

1 46 Rook and bishop against rook and knight

1 ... i.e6? 2 'it>e4! i.fS+ 3 'it'd4

i.xc2?

Vaganian makes the mistaken

decision to play a rook ending,
though even after 3 ... i.e6 4 lLie3
c5 5 f5 i.xf5 6 lLixc4 l:!.c6 7 lLie3
i.e6 8 c4 things are not easy for
him.

18 .. J:tcs 19 .l:1b5 :es 20 'it>xc4

rl.e4+ 21 @d5 l:txh4 21 .tixaS 'it>g7
23 b4 .l:1hl 24 bS h4 25 c4 h3 26
.l:1a2 h2 27 'it>cS @f7 28 rl.e2 Black
resigned.
Tal-Balashov
USSR, 1974

4 rl.xc2 cs S e2+ 'it>f6 6 l:Ie8!

Now it would be a mistake to

transfer to a pawn ending: 6 !f.e5?
:i:lxe5 7 fxe5+ 'it>f5 8 'it>xc4 'it>g4! 9
'it>b5 'it>xg3 1 0 'it>xb6 @xh4 1 1
'it>xb7 'it>g3 1 2 c4 h4, and Black is
quicker.
6 ... bS 7 :!::i.e2 :cs 8 a4 g6 9 Zle8

c7

1 a4!

Depriving Black of the possibility

of . . . a5-a4.
1. .. b6 2 lLid2 !

Now White needs to find a plan to

play for a win, and Balashov dis
covers a splendid transfer of the
rook to b5, where the rook attacks
the b6 pawn and does not allow the
king to penetrate to the fifth rank.
10 lieS ltc6 1 1 :bs ! 'it>e6 12 'it>e4
'it>f6 13 fS gS 14 J:idS gxh4 lS gxh4
'it>e7 1 6 WeS!

Provoking ... f7-f6, so taking this

square away from the opponent's
king.
16 ...f6+ 17 Wd4 cs 1 8 :bs l!c6
16 a5! bxaS 17 J:Ixb7+ c,t>f8 1 8
'it>dS!

All White's pieces are active, and

Black's passive.

Now White improves the position

of the knight, whereas Black's ol
lowing activation assumes an unreal
character.
2 ...i.eS 3 f3 l:.f8
Or 3 ...l!d4 4 'it>e2 and 5 lLie4, and
if Black does not exchange rooks,
then 6 !Ihl and 7 h4.
4 lLie4 'it>c7 S llhl !

It is necessary to open the h-file,

whereas the d-file has no
significance.
5 ... J..f4+ 6 'it>e2 'it>d7 7 h4 gxh4 8
rl.xh4 rtle7 9 l2Jf2!

Now the best place for the knight

will be d3.
9 .. l'l.d8 1 0 lLid3 i.g5 11 .l:lhl
e5! ?
.

A vain attempt t o obtain activity

with a pawn sacrifice.
12 :.dl 'it>e6 13 l2Jf2!

Rook and bishop against rook and knight 147

White again wants to exchange

rooks, but Black is forced to avoid
this exchange.
13 ... .l:Id4 14 CZJe4 i.f4 15 ClJc3!

White's main objective is the b6

pawn.
15 ...i.g5 16 CZJd5 i.d8 17 I:l.d3

c,t>f7 18 lbc3!
It is necessary to drive the black
rook away from the important d4
square.
1 8 ... i.e7?

Better is 1 8 ...i.g5 ! , which was the

only chance.
19 ClJb5 e4 20 l:te3! l:td7 21 .l:Ixe4

i.f6 22 f4 i.g7 23 f5 'itf6 24 I:l.e6+

Black resigned.

The following ending serves as

yet another example where the
knight is stronger than a bishop
when there are rooks on the board.
Pelletier-Arencibia
Ubeda, 1 998

pair of rooks, which is in White's

favour.
3 b3 c,t>g7 4 l::ta 6 l:tbc8?!

He should defend by 4 ... l:1b7 5

CZJc5 .:c7, though after 6 b4 (but not
6 .l:Ixe6? in view of 6 ... i.b2) and
Black has no counterplay.
5 .l:Ic5!

The best way to exchange after

which the white knight appears on
c5.
5 ... l:!xc5 6 CZJxc5 i.b8 7 b4

Also here an exchange on e6 was

not possible because of 7 ... .l:Ic8.
7...@f6 8 l:l'.a5!

The rook not only cuts off the

black king along the fifth rank, but
also improves its own position: its
objective-the b7 square.
8 ...l:l'.d5

Hoping for 9 .l:Ib5 i.d6 1 0 lbe4

c,t>e5 1 1 ClJxd6 'it>xd6 1 2 l:rb 7 a5
with probable equality.
9 a4

Now already the above-mentioned

variation is threatened, because at
the end White has b4-b5 .
9 ...@e7

On the more tenacious 9 . . @g7

White intended to play 1 0 lbb3 !
.lld 7 (the exchange o f heavy pieces
leads to a quick win for White:
1 O. . llxa5 1 1 CZJxa5 a6 12 CZJc6 ii.c7
1 3 b5 axb5 14 a5 ! etc) 1 1 lZJd4 i.d6
1 2 lbc6 'JJ.c7 1 3 l:l'.a6 with the future
march of the king to a4 and b4-b5
and the win of the a7 pawn. On the
other side of the board Black does
not get anything.
.

In order to maintain the balance

Black must activate his rooks along
the d-file and then subsequently at
tack White' s king's flank.
1 h3 i.e5 2 'ite2 llab8

Already Black does not succeed

in organising counterplay along the
d-file because of the exchange of a

10 1Ib5 i.d6 1 1 l:tb7+ 'it>e8 12

CZJe4?!

Probably, sufficient for the win is

12 lba6 l:tg5 1 3 g4 h5 14 gxh5
l:rxh5 1 5 l:!xa7, but simplest was 1 2
ClJd3 a 5 1 3 b5.
12 ...ii.e7 13 l:!xa7 e5

Or 1 3 ...f5 14 lbc3 .

148 Rook and bishop against rook and knight

14 J::!a8+ Wd7 15 l:l:a7+ @es

If the Black king stepped to d8,

then 16 tLlf2 .Jtxb4 17 l!xf7 would
d ecide.
1 6 @d3 i.xb4?

It would be better to play the rook

ending after I 6 . . . f5 1 7 tLlc5 i.xc5
1 8 bxc5 :xc5 19 lixh7 g5 ( 1 9 ... Ila5
20 lih4), though 20 .l:Ia7 was suffi
cient for victory because of the very
bad black king.

17 tll f6+ Wd8 18 Itd7+ Wc8 19

.l:!xti l:[fS

19 ... h5 20 tLld7! ? lid5 2 1 Wc4

il..a5 22 tllc 5 also loses.

20 g4! lif3 2 1 We2 xh3 22

l:.xh7 .l:!:xh7

After 22 ...Itg3 23 @f2 the black

rook is in a 'hole' .
23 tt:Jxh7 @d7 2 4 Wd3 We7 2 5
tLlgS d6 2 6 tllf3 and Black
resigned.
Adams-Almasi
Dortmund, 1 998

For the present beginning an at

tack on the black pawns.

9 . aS 10 tLlc6 Wg7 1 1 tlixa5 lia8

12 l:l:bS i.f6 13 :t;.i.a2 i.c3 14 tll c6
l:!.xc4 1 5 aS
..

The a-pawn, supported by the

knight c6, can only be stopped at the
cost of material loss.

15 .l:Ixa6 16 l:i.c2 licxc6 17 dxc6

l:!.xa5 18 :b7 nal+ 19 Wg2 .Jta5
20 lla7!
And Black resigned, since there
..

is no defence against the manoeuvre

.l:tc2-b2-b7.

The knight is particularly strong

in attacking various weaknesses,
whereas a bishop is incapable of de
fending light and dark squares.
Van Laatum-Mikhalchishin
Eeklo, 1 991

...!.
. .

.. .-0.
1.rz...
...lf... ..-0.

. - . . .

i...tt:Jrs

1 dS! 'tiixdS 2 exdS

White is obliged to transfer to an

ending where for him the knight c6
controls all the important squares
around itself, while even the dark
squared bishop, controlling the long
diagonal, will feel redundant.

2 ... g6 3 l:l.'.abl i.g7 4 c4 i.c3 5

.l:tecl .Jtf6 6 @fl e4 7 g3 i.gs 8
.l::!. c2 ae8 9 tllb 8!

There was another, and perhaps

more unpleasant plan for White:
1 ...f5! ? followed by 2 ...@f6, 3 ... f4
and 4 ... @f5.
2 l:.g2

After 2 i.d5 l!xg3 3 .i.e4 g4 4

.Jtxf5 gxf5 it is not easy to defend
the endgame.
2 tlle3
..

Interesting was 2 . . . tlld4 with the

same plan ...f7-f5-f4 .
3 na2 f6! ?

Rook and bishop against rook and knight 149

A new idea:
. . . 'lle 3-fl.

... g6-g5-g4 and

4 i.e6 f5

After 4 ... g5 White brings the

bishop into the defence by 5 hxg5
hxg5 6 i.h3.
5 d4 f6 6 d5?

This pawn should be 'discarded'

by playing 6 i.c8 and 7 i.b7, trans
ferring the bishop to the long
diagonal.
6... 'llc 4!

Now the knight shows itself in all

its splendour.

7 i.c8 'lld 6 8 i.d7 l:td3 9 i.e6

'lle4 10 J:ig2 g5

There was also another plan:

1 0 ... D and l l ...'ll f2.

1 1 hxg5+ 'itxg5 12 i.f7 g4 13

i.e6 l:te3

Avoiding the trap 1 3 . . . 'llx g3? 1 4

i.xf5+! .
14 i.c8 c3 1 5 i.e6 d3 1 6
i.f7?

It was necessary to give up the d5

pawn, if only by the move 16 i.c8
followed by 1 7 i.a6.
16...'ll g5 ! White resigned.

Rook and two pawns against

rook and knight (many pawns)

It would seem that having more

pieces should confer an automatic
advantage, but the knight has such
limited mobility in a battle with
pawns that quite frequently nothing
can be done with it.
Mikhalchishin-Sveshnikov
USSR (ch), 1 978

3 bxa3 l:rxd4 4 xd4 bxa3 5 'it>c3

a2 ! 6 b2 'itt c 4

After sacrificing a pawn, the black

king rushes over to the opponent's
pawns, it seems without success ...

7 lt:Jd6+ 'it>d5 8 lt:Jf5 'itte4 9 lt:Jxg7

'iilf4 10 h3 'lt>g3 1 1 lt:Jf5+ 'iilxh3 1 2
lt:Jxh6 'it>h4!

But this is the point-and typical

for the struggle of the knight against
pawns-there is no defence against
Cllg 5, therefore a draw.
Smagin-Mikhalchishin
Moscow, 1 989

1 :d2

The decision to exchange rooks

was not easy. Other possibilities
were I lt:Jd2 or the roundabout way
I l:.f2! with the idea of2 l:rf8.
l . . .b5 2 l:rd4?

After 2 l:rxd5 cxd5 3 lt:Jd2 a3 4

d4 l.ta4 5 bxa3 bxa3 6 lt:Jb I ! a2 7
".lJc3+ b3 8 lt:Jxa2 the pawn ending
is won.
2 ... a3!

Bad is 2 ... I:.xd4? 3 'i!ixd4 a3 4

lt:Jd6+! 'lt>a4 5 bxa3 bxa3 6 c3
with a win.

1...lt:Jb4?!

Clearly Black wants to attack the

pawns, but stronger seems L.h6,
though it is not usually recom
mended to play on the side where
the opponent is stronger.
2 a3 lt:Jd3 3 g5 I:.e6 4 .l:.d2 lt:Jc5

After 4 ... .l:.e3+ 5 'lt>g4 lt:Je5+ 6

'ittf4 .t'td3 7 .Uxd3 lt:Jxd3+ 8 'it>e4

Rook and two pawns against rook and knight 151

ctJxb2 9 'it>d4 the white king

threatens to break through to the
pawns on the queen's flank.
5 'it>g4 ctJxb3 6 d7+ 'it>g8

Or 6 'Lid4 gxh3+ 7 'it>xh3 'it>f4.

6 J:ixb3 7 l:txd5 'it>xg4 8 ctJxh6+

'it>f4 Drawn.
..

After 6 ... l:l'.e7 7 :i.d6 there is

equality on the board.
7 h5, and a draw was agreed,
since the threats l:td8, h5-h6 give
White sufficient play.

Korchnoi-Lj ubojevic
Bath, 1973

Mikhalchishin-Savchenko
Postojna, 1 991

1 .:i.bl ?

After the correct l ...Wf8 2 'it>f3

:!.b l Black is in no way worse,
whereas now problems start to ap
pear with the fl pawn.
..

Black has two plans of play-the

best one was the plan to advance the
passed pawn. 1 ...l:1d l ! ? 2 ctJc5 (2
l:ta2 d4 ! 3 .:i.xa6 .:i.d2+ =) 2 ...d4 3
ctJxa6?! d3 4 J:If6 h5 5 :i.d6 l:l'.d2+
with distinct counterplay. But Black
decides to exchange pawns on the
queen's flank, which is clearly
weaker.
1. .. a? ! 2 'Lig7?!

But White decides to play for

mating threats. whereas after the
correct 2 d2! Black would en
counter serious difficulties.
2 b l 3 h3+
If now 3 d2, then after 3 ... 'it>f4 4
ctJe6+ 'it>e3 5 l:l'.xd5 b2+ 6 'it>g3
l:.xb3 7 xa5 We4+ 8 g4 l:l'.b2 9
h3 g2+ 1 0 'it>h5 g3 Black elimin
ates White's last pawn.
..

3 .'it>h4 4 I:i:d2 g4 ! 5 ctJf5+ g5 6

hxg4
..

2 'Lid3 'it>g8 3 .l:!b7 'it>g7 4 ctJe5

'it>f6 5 ctJxf7 l:rcl 6 ctJxh6 c2+ 7
'it>O b2 8 'it>e4! l:l'.e2+ 9 'it>d4 l:l'.h2
10 g5+ We6 1 1 J:rb6+ e7 12 'it>c3
f8 13 l:1xb2 l:i'.xh4 14 .l:!b7 J:if4 15
d3 Black resigned.
Hertneck-Narciso
Berlin, 1 998

J 52 Rook and two pawns against rook and knight

Here Black is three clear pawns

ahead, but after I lLie5 ! with the
idea of f4-e4 and then .l:Ic2 seri
ous problems arise for Black, since
he cannot create a passed pawn. But
in the game there followed. . .

1 'it>f4? @ f7 ! 2 'it>es \tg6 3

.!lxd4?!

14 ltJxb7 g3 15 'it>e3 g2 16 f2
'it>e4 17 'it>xg2 'it>d3 Drawn.
Romanishin-Markowski
Polanica Zdroj, 1 993

A primitive decision, more inter

esting is 3 'it>e6 ! ? or 3 lLih2 ! ? .

3 ... .l:Ixd4 4 'it>xd4 cJi>g5

'it>h4 6 lLif2 g3 7 ltJe4+
ltJg5+ @g4 9 t'll xh7 'it>f.3
'it>f4 11 tll f8 g5 12 lLie6+
tiJd8

5 'it>e5
Wxh3 8
10 c;t>d4
@rs 13

1 e5 ! dxe5?

Also not winning is 1 ..!ld4 2

exd6 l::!.xd5 3 h4 ! gxh4 4 .l:Ie4 l:i.d4 5
d7 ! with a draw. But necessary was
l ....!le4! 2 exd6 .l:Ixe6 3 dxe6 tbxd6
4 Wg2 ltJf5 ! 5 f3 tiJh6! (against
h2-h4, so as to have g5-g4) 6 'it>e4
f6 7 d5 lLif5 8 h3 ltJg7 with an
easy win.
..

13 ...g4 !

By sacrificing yet another pawn,

Black diverts the opponent's king.

2 d6 tiJd4 5 i::i.xe5 'it>f6 4 d7 ltJc6

4 e8 .l:Id4 5 .l:Ic8!

This is the idea-to take out the

knight-now it's a draw.

10

Two minor pieces against a rook

Endings with an uneven align

ment of forces are some of the most
complicated in chess-and also
comparatively little investigated.
Among them can be included end
ings where a rook struggles against
two minor pieces.
In the middlegame, particularly in
positions bearing a closed character,
two minor pieces will be stronger
than a rook even where there is a
deficit in pawns. However, in end
ings where the rook has room for
action, the situation often depends
on the arrangement of the pawns.
Thus, when there is a distant passed
pawn on the board, the rook can
prove to be stronger than two
pieces. However with a balanced
pawn formation the situation may
be reversed.
When playing such endings, it is
very important for the side having
the minor pieces to put right their
coordination, and then the field of
activity of the minor pieces may be
sufficient to struggle successfully
against a rook.
According to the generally held
opinion, in standard situations with
two pieces one can count on having
an advantage. But rooks are not
without their own chances too, par
ticularly when there is a passed
pawn on the board. However, first
and foremost, as we go over to a
study of positions of this type, we
will tum our attention to the possibi
lities of the defending side. In terms

of defensive resources the defender

should strive for an exchange of
pawns, bearing in mind the possibil
ity of transferring to a theoretically
drawn ending, exploiting the stan
dard methods of: (a) from the rook's
side-sacrificing the exchange; (b)
from the minor pieces ' side-sa
crificing a piece.
Let's look at some possibilities in
positions where the rook is playing
against passed pawns. The rook's
specific characteristics depend on
lengthening lines of communication.
The broader they are, the more
chances the long-range rook has in
its struggle against the less mobile
minor pieces, and it is even possible
to think about playing for a win.
You see, whereas the minor pieces
are riveted to the blockade of a
passed pawn, the rook can be uti
lised also on another flank to organ
ise new objects of attack.
To counteract the rook, one of the
postulates of play for the side hav
ing the minor pieces is to put right
their interaction while controlling
the opponent's passed pawn. But the
various p lans of activity of the
forces on the board produce play of
a complicated character, and the
exploitation of one's resources is
not always successful even where
highly rated chessplayers are
concerned.
Naturally, the outcome of the
struggle in great part depends on the
arrangement of the pawns: the more

154 Two minor pieces against a rook

compact they are, the shorter the

communications of the rook, and
the
narrower
their
naturally
possibilities.
Beliavsky-Miles
Thessaloniki (ol), 1 984

Wd3 (after 8 'it>f3 Black forces a

draw: 8 ... l:txc2! 9 lllxc2 Wb2)
8 ... c;tib2, and Black prevents White
from putting right the coordination
of his pieces.
2 tlld l + 'it>a3 3 'itie2 l:1hl
There is no saving the pawn:
3 ...gl 4 lll e3 c3 5 @f2 !i:g5 6 f4.

4 ..txg6 .l:Ih2+ 5 c;tie3 'itib4 6 Wf4

c3 7 llle3

White has won a pawn without

losing control of the opponent's
passed pawn.
7 'it>b3 8 ..td3 'it>b2 9 c;tie4
It is useful to pay attention to the
interaction of the white pieces
now the time for White 's passed
pawn ha,s come.
..

9 'it>cl 10 f4 l:!.h8 11 .tbs .l:Ib8

...

Despite the far advanced passed

pawn, supported by all Black's
forces, the chances of White's
pieces are higher. They firmly
blockade the pawn, while the quite
narrow range of activity between
the flanks allows White to fight for
the creation of a passed pawn of his
own.
1 ..tc2 fl ? !
I f 1 .. .g5? 2 tll d l+ 'itia3 3 'it>c3 the
passive rook is unable to prevent the
win of the c4 pawn: 3 ...lic l 4 llle 3
and 5 lllxc4.
Therefore Black activates his
rook, attacking a pawn. But this is
inaccurate: on the f-file the rook is
in the sphere of activity of the king.
Therefore stronger is 1 .. .g l , de
fending the g6 pawn and riveting
the white pieces to the passed c
pawn. Here, after 2 ll:ld 1 + c;tia3 no
good is 3 'it>c3 :!fl or 3 'it>e3 l:!'.g2 4
..te4 Wb3 . But also on the active 3
f4 l:!'.g3 ! the rook succeeds in stand
ing up to the pieces. For example: 4
tlle 3 lif3 5 llld5 c3+ 6 c;tie2 .l:Ih3 7
etJe3 (7 ..txg6? l:th2+) 7 . .l:!h2+ 8
.

After 1 l .. .c2 12 ..ta4 Black loses

a pawn and cannot derive any bene
fit from the temporarily uncoordinated white pieces: 12 . . . c;tid2 1 3
lllxc2 .l:Ia8 1 4 ..tc6 ! or 12 ... l:ta8 1 3
i.b3 ! l:ta3 1 4 ..txc2 Wd2? 1 5 lll c4+.
12 ..td3 'it>d2

Or 12 ... l:te8+ 13 'it>d4 'it>d2 1 4

lll c4+.
13 lllc4+ c;tidl 14 @f3 lih8 15

..te4 Wel 16 llla 3 Wdl 17 f5

And White, maintaining the

blockade of the opponent's passed
pawn, realised his own one.
Morovic-Yusupov
Tunis, 1 985

Two minor pieces against a rook 155

In this example the sphere of ac

tivity of the rook is quite broad, and
this means the more possibilities it
has. In fact, despite material equal
ity, the advantage is on the side of
the rook. It is necessary only to
create a distant passed pawn, while
not forgetting to prevent it being
blockaded by the minor pieces. This
is achieved by the advance of the
pawn pair after the preparatory
l ...l:!f5 .
But i n the game there followed ....
1 ...g4?,

Breaking the coordination of the

pawn pair, which allows White to
blockade it with the move 2 i.el ! ,
without upsetting the coordination
of his own pieces. However he is
deflected from a direct attack on the
d4 pawn.
2 i.h6?! 'it>e6!

No expense spared for creating

passed pawn.

3 ct:Jc5+?

It was still not too late to return to

the blockadin g plan-3 i.f4 h5 4
i.g3 ! 'it>f5 5 i.h4. The bishop con
trols the e l -h4 diagonal, while the
king has nowhere to go along the
light squares, since on any . . 'it>f3
there is ctJel+. Moreover, also poss
ible is the energetic 3 i.f2!?, taking
aim at the d4 pawn.
.

the king penetrates to the weakened

queen' s flank.
10 ctJf2+ 'it>d5 1 1 b4 g3 12 hxg3
hxg3 13 ctJd3

Also after 3 ctJg4 'it>e4 1 4 t'Lie3

axb4 1 5 i.xb4 l:!h6 the passed pawn
must decide the game.
13 axb4

It was also p ossible to win the ex

change: 1 3 ... 'it>e4 14 bxa5 g2 1 5 a6
g l ='ilY 1 6 i.xgl .l:Ic2+ 17 'it'd !
'it>xd3 1 8 a7 .l:Ixa2 19 'it>cl b4 20
'it>b 1 b3 21 i.f2 Wc4 22 i.e3 'it>c3
(A.Yusupov), but Black prefers to
retain the passed g-pawn.
14 i.e3 l:!.c2+ 15 'it>dl l:tc3 16
'it>e2 'it>e4 17 t'Lic5+ 'it>f5 1 8 ctJd3
'it>e4 19 t'Lic5+ 'it>d5 20 ctJd3 lic2+
21 'it>dl .l:Ih2 !

Now the
stopped.

g-pawn

cannot

be

22 i.gl l:thl 23 ctJf4+ 'it>e4 24

ctJe2 Wf3 White resigned.

In the context of defensive possi

bilities and technical realisation of a
distant passed pawn, we present a
classical example from the praxis of
Yuri Razuvaev.
Razuvaev-Kirov
Sofia, 1 981

3...Wf5 4 ctJd3 lic7 5 'it>d2 .li(c6 6

i.g7?

Now already the creation of the

passed pawn cannot be prevented,
whereas 6 i.f4 ! 'it>e4 7 i.g3 'it>f3 8
lt'iel+ would have set up a solid
blockade.
6... h5! 7 i.xd4 'it>e4 8 i.c5

8 ... l:td6 was threatened.

8 ... h4 9 'it>e2 a5

More energetic was 9 . . g3 1 0

hxg3 hxg3 1 1 b 4 l:i.c8, imposing an
original zugzwang on White: 1 2 a3
1:tc6 1 3 ctJel f6 14 Wd2 Wd5, and
.

With such a specific arrangement

of pawns on the king's flank, the
bishop itself can resist the rook

J 56

Two minor pieces against a rook

when there are no pawns on the

queen's flank. It goes without say
ing that here the king of the weaker
side should be close to the pawns.
Therefore to achieve a draw it is
enough for Black to sacrifice the
knight for the passed a-pawn. He
should also avoid any unnecessary
weakening of the position. White
should take this into account and or
ganise offensive action. Black has
succeeded in constructing a defence:
his minor pieces are well coordi
nated, preventing the advance of the
passed pawn and preparing a break
on the king's flank. In addition the
bishop does not allow in the rook
since on .l:ib7 he can cover up by
i.e7. Here, however, Black need
lessly weakens his king's flank with
the move
1. .. h 6?
... and creates objects for attack,
thus increasing the attacking possi
bilities of the rook. Now, as well as
the h6 pawn, it is possible to add the
invasion point f5, which can be
created by undermining the g6 pawn
with h2-h4-h5. After L.'it?d6 2 h3
(more accurate than 2 'itg2, on
which Black quickly organises an
attack on the passed pawn by
2 ...lt::Je 5 3 h3 lt::Jd? and tl'ic5) Black
has two possible plans of defence:
(A) Active-2. .. tl'ia5 (preparing
the attack a4 after transferring the
knight to c5) 3 'it?g2 'it?c6 4 e5
lt::lb7 5 e8 'it>c5 6 h8 'it>b4 7 xh7
lt::lc 5 8 h4 r;.txa4 9 h5 gxh5 l 0 gxh5
i.g5, and Black, after giving up his
bishop for the h-pawn, reaches a
theoretically drawn ending 'rook
against knight' .
(B) Passive-2 ...'it?e6 3 <;t>g2 @d6
4 @g3 (White presently controls the
5th rank in order to prevent the
move ... h7-h5 exchanging a pair of
pawns) 4 . . .'it>e6 5 h4 'it'd6 6 h5 'it>e6

7 ltb7 ii..e7 8 @f4 r;.td6 9 'lt>e4 @e6.

Now the king cannot break through
to the c4 square without loss: l 0
<;t>d3 lt::Je 5+ 1 1 'itc3 lt::lxg4 1 2 hxg4
hxg4 1 3 'it>c4 (or 1 3 a5 tl'ie5 1 4 a6
lt::lc 6) 1 3 ...lt::Je 5+ 1 4 'it>b5 i.d8.
Upon the attempt to proceed to c4
via the b3 square it is necessary to
reckon with the threat of ... lt::Ja 5.
The rook's intended penetration to
the 8th rank by l 0 lic7 r;.td6 1 1
lic8? i.d8 ends in its capture.
Also nothing is offered by l 0 l.t.b5
i.d8.
2 'it?g2 'it>f6?.

As before, Black senses no danger

and breaks the coordination of his
pieces since now the position of the
bishop is limited by the blockading
function of the knight. Better was
2 ... r;.td6 3 lib7 i.c7 with the idea
... lt::Jc 6-a5.
3 .l:c5 tl'ia5 4 @f'3 @e6 5 'it>e4
i.b6 6 lic8 @f6 7 h4 'lt>e6

8 h5

It is important to fix the weak

pawn. Too hasty is 8 l:Ih8 lt::lb 7 9
xh6 tl'ic5 1 0 @f3 @f7 1 1 h5 gxh5
1 2 l:txb6 hxg4+ 13 @xg4 tl'ixa4
with a drawn ending. Now, how
ever, in the event of 8 ...gxh5, 9 %lh8
is unpleasant.
8 .'it>t7 9 'it>d5 gxh 5 10 gxh5 @f6
..

11 b8 i.e3 12 'lte4! i.d2 13 b2

i.el 1 4 l::te2 i.b4 15 g2 @f7

Two minor pieces against a rook 1 5 7

The weakness of the h6 pawn is

self-evident. Combining an attack
on it and the knight, White breaks
the coordination of the opponent's
pieces. On 1 5 ...i.fS follows 16 g6
@fl 17 a6, and the a-pawn starts
to move.
16 .li(g6 i.d2 17 lld6! i.gS 1 8
J::[.d S!

An ideal square for the rook.

18 tt:Jc6 19 as @e6 20 a6 i.e3

..

Trying to blockade the a-pawn on

the final frontier, but the mechanism
of breaking the coordination of the
pieces again comes into operation.
21 .l:IbS i.gl 22 .li(bl i.f2 23 b2

i.cS 24 .l:Ib7 tt:Ja7 25 h7

And White wins the h-pawn and

the game. A very fine ending!
And here is an example where
pieces blockade the passed pawn
from more distant 'remote-control'
squares.
Karpov-Kasparov
World Championship (m), 1 985

Here the bishop will not allow the

rook to get behind his lines via the
queen's flank (due to the covering
b4 and i.c3), whereas the squares
of invasion along the central files
are covered by the king. There
followed:
l...!tb8 2 i.b4 d8 3 @e2 a3? !

While no additional objects of at

tack have been created on the king's
flank, the advance of the pawns to
vulnerable squares is premature.
More logical was the exchange of
the f-pawn after 3 ...g6, which forces
the white king to remain far away
from his pieces and increases the
scope of the rook.
4 i.c3 f6

Also here, stronger is 4 . .. g6.

5 i.b4

The pawn structure on the king's

flank is stabilised, and Black intends
to attack White's pawn weaknesses.
We look at Black's possibilities on
the basis of comments made by
G.Kasparov in his 1 987 book Two
Matches.
5 ...f7? !

Here the king does nothing, since

there are no prospects for it in the
centre. Much more accurate was
5 .. . h7. In this case the method of
defence applied by Karpov in the
game would place White under
threat of defeat. For example: 6 i.c3
k!b8 7 i.b4? b5 ! 8 g4 l:1b8. A zug
zwang position has been reached
and White is forced to allow the
rook into his camp: 9 @d3 l:i.d8+ 1 0
c2 hxg4 1 1 hxg4 d4 1 2 i.xa3
a4 1 3 b3 lixg4 14 i.cl l:i.g3+ 1 5
lLlc3 li f3 1 6 c2 xf5 17 @d3. A
simple evaluation of such a position
is not possible, but in practice Black
would have excellent chances. But
after 7 lLlb4 ! White holds on:
7 . . ..l:Ib5 8 g4 .l:Ib8 9 @d3 ! lia8 1 0
lLla2 l::ta4 1 1 i.b4 h6 12 i.d2, and
the fortress cannot be breached.
6 tt:Jc3 .li(b8 7 lLla2 bS 8 g4 b8

At first sight there is no basic dif

ference from the above-mentioned
zugzwang position, but. ..
9 d3 l:i.d8+ 1 0 @c4 dl 1 1
i.xa3 :i:Ial 1 2 'it>b3 bl

Not 12 ... h4? 13 lLlc l .

1 58 Two minor pieces against a rook

13 gxh5 lhh3 + 14 'Ll c3

This is where the position of the
black king tells: after 1 4... J:Ixh5
'Lle4 the f5 pawn is inviolable.
14...l:!.B

But now he cannot, without loss,

cope with the h-pawn.
15 .tel ! .l:!xf5 16 h6 g6

Right up to the draw there re

mains the sacrifice of a piece for
two pawns. On 1 6 . . .g5 this is
achieved by 1 7 'Lle4 g6 1 8 h7 .l:!f3
( 1 8 ...xh7 19 'Llxf6+) 1 9 @c4 J:ih3
20 'Llxf6.
1 7 'Lle4 l:!h5 18 ..tb2

And the f6 pawn is lost, since on

l 8 ...f5 follows 19 h7. Drawn.
If, together with the rook, there
are no passed pawns then roles are
reversed. It remains only for the
rook to counteract the opponent's
attempts to create his own passed
pawn. But none the less the defen
der's resources are quite substantial.
Short-Kasparov
Belgrade, 1 989

In this position Black's chances

can only be linked to the creation of
a passed g-pawn. Therefore he en
deavours to win the h2 pawn. How
ever to do this it is necessary to
break through the line of demarca
tion-the 3rd rank-which the rook

allows only 'in exchange' for a

counterattack behind enemy lines.
All this is not easy. Let us look at
the possibilities of the two sides.
based on G.Kasparov's analysis.
1 l:!g3 ctJd5+ 2 @f2 'iitf5
The king moves closer to the h2
pawn.
3 li:tf3+ @g4 4 .l:!g3+ @h4 5 @f3
White strives to activate his king,
since he cannot hold on by the pass
ive 5 l:!b3 CLlf4 with the threat
... 'Lld3+. Now 5 ... 'Llxb4 is not poss
ible because of 6 g4+.
5 'Llf6?!
Stronger is 5 .....td7 ! , keeping
under threat b6, while on 6 We4 the
knight manages to attack the h2
pawn: 6 ... 'Llf6+! 7 'iite 5 'Llg4+ 8
d6 ..tb5 9 llg2 'Lle3, and the black
king draws closer to the h2 pawn.
..

6 'itig2 'Llh5 7 l:te3 ctJf4+ 8 gl

'itig4

Now, with the white king driven

back, Black puts right the interac
tion of his king and pieces.
9 l:1g3+ 'iitf5 10 J:if3 g4
Not letting out the king, since in
the event of 10 ... @e4 1 1 @2 'Lld3+
1 2 @g3 'Llxb4 1 3 .l:!f8 White again
activates his forces.
11 l:!.e3 'Lld5 !
Again insufficient is l l ... 'Lld3 be
cause of the activating of the
rook-12 e7! ( 1 2 @g2? @f4+ 1 3
lle7 ..tc6 in Black's favour)
1 2 ...'Llxb2 1 3 nm @g6 1 4 .l:!f8
'Lld3 1 5 @g2 'Llxb4 1 6 l:tf4.
12 .U.b3 @f4 13 @n
Only the b3 and g3 squares re
main accessible to the rook, but now
he should encircle the king. How
ever, on 1 3 @g2? 'Lle3 14 @2
'Lldl+ 1 5 'iite l .Jla4 1 6 a3 'Llxb2
17 lla2 'Lld3+ 1 8 @d2 .tbs White
loses a yawn without compensation.
13 tiJf6 14 l:!a3 'Lle4+ 15 g2
..

.Jle2

Two minor pieces against a rook 159

The manoeuvre 1 5 ... 'Lid2 1 6 .!lc3

'Lic4 1 7 .li(c 1 ! 'Lixb2 1 8 .:i.c8 allows
the rook to escape to freedom.
16 .l:!'.b3 ..tfl + 17 c,t>gl i.c4 1 8

'it>c2 'Lixh2 33 .!lxh2 g3 and Black

won.

The a6 pawn is indirectly de

fended: 21 l:txa6?? ctJh3+ 22 @fl
i.d3+, but this move, losing time,
also allows the rook to escape to
freedom. Stronger was 20 ... ltJD+ 2 1
'it>g2 'Lies 2 2 @ 2 ..tc4.
21 @m ..tb7

Reti-Bogoljubow
Bad Kissingen, 1 928

l:ta3 'Lig5 1 9 .llc3 i.d5 20 .l:!'.a3

i.e4?!

One more example from the

classics.

1 . b6
..

For the side having the rook it is

very important to create a distant
passed pawn quickly.
22 l:td3?!

White misses his chance to acti

vate the rook by 22 l:ta5 ! , threaten
ing to exchange the b-pawn with
tempo-22...'Lie4+ 23 @g l ctJd6 24
l:l'.c5, and, it seems, maintaining the
equilibrium.
22 ...'Lih3+ 23 'it>el?

After moving the king further

away f,rom the h2 pawn the game
already cannot be saved since now
Black cuts it off from this pawn.
Necessary was 23 @fl ..tc6 24 .!lc3.
23 ... i.c6! 24 .!id4+ 'iitf3 25 .:i.d3+
We4 26 .!lg3 'iitf4 27 l:ta3

More tenacious is 27 l:.c3 ..tb5 28

l:l'.c8, though after 28 ...'it>D 29 l:tf8+
'itg2 30 .!lg8 'Lif2 ! White also loses
the h2 pawn.

27 ..tb5 28 l:l'.c3 'Lig5 29 .!lc8

\t>e3! 30 .!lh8 ctJf3 31 'it>dl i.e2+ 32
..

2 axb6 l:Ixb6 3 c,t>e2

A loss of tempo. He should quick

ly put right the coordination of his
pieces, combining an attack on the
e5 pawn with detention of the a
pawn. For example: 3 ctJc4 l!i.b5
(3 . ...!lb4 4 'Lixe5 .!lxe4 5 'Lic6 in
tending 'Lia5 and ..td2) 4 i.d2 .l:Ic5
5 'Lia5 'it>f? 6 'it>e2 'it>e6 7 'itd3 and 8
'Lic4.
3 ... l:l'.b4 4 'it>f3 'it>f7 5 .lth4 .llb l 6
'Lic4 'it>e6 7 ..tg3

This threat is a sham, since it has

a tactical refutation.
It is useful to stop the pawn as
soon as possible. For this purpose
the blockading move 7 J..d 8 is
appropriate.
7 ... .!lcl ! 8 'Lias

8 'Lixe5 is no good in view of

8 ...1Ic3+ 9 c,t>f4 g5+ 1 0 'it>xg5
lhg3+.

J 60 Two minor pieces against a rook

8 ... l!al 9 lbc4 l:f.a4 1 0 ll'ie3 as 1 1

.tel a3 1 2 'itie2 a4 1 3 ll'ic2 l:f.b3
1 4 i.b4

At last White manages to put right

the coordination of his pieces and
blockade the passed pawn. Trying to
obtain additional obj ects of attack,
Black organises pawn pressure on
the king's flank, supported by the
active rook.

14 ... h5 15 i.f8 gS 16 h3 l!c3 1 7

'it?d2 :b3 1 8 'it?e2 l:tb2 1 9 'itid2 l!bl
20 'it?e2 :.c1 21 'itid2 fl 22 i.cs
l:i.f4 23 'it?d3 l:f.f7 24 i.e3 l:i.d7+ 2S
@e2 l:f.g7

26 g4?

This superficially active move,

pursuing the objective of blockade,
turns out to be the decisive mistake.
White creates in his position a weak
h3 pawn, which the rook can attack,
exploiting the fact that White's
minor pieces are tied down to the
threatened advance of the a-pawn.
After 26 lba3 g4 27 hxg4 :xg4 28
'it?f3 .Ug8 29 i.c5 White defends
himself successfully.
26 ... h4! 27 .Itel ?

The only move was 27 i.d2! s o as

to meet 27 ... l:i.c7 with 28 'itid3, and
27 ... .l:Ib7 with 28 i.b4.

27 ...:.c7 28 'itid3 :b7 29 .txg5

:t.!b3+ 30 @c4 :xh3

And so the main idea of attacking

with the rook is seen in a clear light:

B lack has two distant passed pawns

and the opponent's pieces are in no
state to contain them without ma
terial loss.

31 'it?b4 a3! 32 lbxa3 :xa3 33

i. xh4 :l'.e3 34 'it>c4 :.xe4+ and soon
White resigned.

Let's look at yet another example,

illustrating the strength of the rook
in conjunction with an outside
p assed pawn.
Miles-Van der Sterren,
Wijk-aan-Zee, 1 984

Material advantage is on White's

side. He has a passed pawn, har
monious deployment of pieces and
can reckon on playing for a win.
However the strength of the passed
p awn in conjunction with an active
rook is such that Black's counter
resources prove sufficient to obtain
full equality. For a start he ties the
minor pieces down with the threat to
advance the a-pawn.

l...a3 2 @d3 l:f.a5 3 i.a2 'it>g7 4

c3 @f6 5 ttJf3 l:f.bS

The rook transfers to an active

position. Weaker is 5 ...gS?! -6 e5+
@ f5 7 ll'id4+ 'it?g6 8 e6 @f6 9 @b4
lira 8 1 0 lbc6, with advantage to
White.
6 eS+ 'it>e7 7 lbd4 l:::tb 2 8 i.dS
lln 9 ll'ic2 !te2 1 0 ll'ib4

Two minor pieces against a rook 1 6 1

White hopes to realise his material

advantage and therefore rejects the
draw after 10 l'Llxa3 1:!.xe5.
10 ...'it>d7 11 Wd4 gs

We are acquainted, from the

previous examples, with the method
of exploiting a ' quality advantage'
(in the words of A.Nimzowitsch).
12 h3 hS 13 g4? !

As shown b y the endings previ

ously looked at-this is a risky
weakening of the position. However
White plays for the win and does
not want to reconcile himself to the
'drawing zugzwang': 13 l'Lla2 lld2+
14 c;t>e4 l:l.'.e2+ 1 5 'it>f5 l:i.d2 etc.
13 ... h4 14 i.c4

Analysis shows that i n fact he al

ready has to worry about maintain
ing equality. For example, in
Black's favour is 1 5 Wd5 l:.te3 ! ?
(sufficient for a draw i s 1 5 . . ..i::i.c l ,
but Black is striving for more) 1 6
i.b5+ 'it>e7 1 7 l'Llc6+ 'it>f8 ! 1 8 i.c4
:Ie2 1 9 l'Llb4 l:tb2 20 l'Lla2? [20
'it>c5! t:>. 20 ...l:i.h2? (20 ... 'it>e7 !) 2 1
'it>d6 lixh3 2 2 l'Lld5 Ith 1 23 e6 l:te 1
24 e7 J::rxe7 25 l'Llxe7 h3 26 l'Llf5 + -]
20 ... l:th2 etc or 1 5 e6 'it>e7 ! (but not
1 5 ... 'it>d6?-1 6 l'Llc2 J:ih 1 17 l'Llxa3
.l:Ixh3 1 8 l'Llb5+ 'it>e7 19 l'Lla7 and 20
l'Llc8+, and White has the advan
tage) 1 6 l'Llc2 h 1 1 7 l'Llxa3 l:.xh3
1 8 l'Llb5 l:tg3 etc.
It seems he should reconcile him
self to a draw by playing 1 5 l'Llc2 !
Ith 1 1 6 l'Llxa3 l:txh3 1 7 l'Llc2 ! l:tg3
1 8 l'Lle3=.

lS ... dl 16 'it>e4 J:id2 17 i.c4

llb2 18 l'Lla2 .l:.h2 19 @rs xh3 20
'it>xgS J::rh l

14...:!:!.el !

Black should not let the king go

on to c5, since his own king cannot
hold back the onslaught of the white
pieces. For example: 14 ... l:tb2 1 5
'itc5 l:th2 1 6 e6+ We8 ( 1 6 . . .'it>e7 1 7
l'Llc6+ 'it>e8 1 8 i.d3) 1 7 l'Lld5 l:!.c2 1 8
'it>d6 Ji(xc4 1 9 l'Llf6+, and White
wins; or 14 ... l:.th2 15 Wc5 l:txh3 1 6
e 6 :Ie8 ( 1 6 ...We7 1 7 l'Lld5+) 1 7 l'Lld5
with the threat of 1 8 i.b5+ and a
mating attack.
lS i.d3?!

Continuing to play for the win,

White suddenly breaks the coor
d ination of his pieces, and the initi
ative passes to Black. Could White
have improved his position?

An inaccuracy. Far stronger is

20 ... J:ih2 with the threat of
2 1 ...J::rxa2 and 22 ... h3. White cannot
at the same time hold the two distant
passed pawns (there you are, don't
you recall the move 13 g4?!). For
example: 2 1 'it>f5 J:if2+ 22 'it>g6
l:.txa2 23 i.xa2 'it>e7, and one of the
pawns queens; or 2 1 Wf4 l:tf2+ 22
'it>e3 lic2 23 i.d5 h3 24 l'Llb4 lib2
25 l'Lld3 a2 (or) 25 . .h2), with a de
cisive advantage.
.

21 @f6 h3 22 l'Llc3 !:I.el 23 i.e6

'it>d8 24 i.ds h2 2s gs no+ 2 6
'it>e6 .:!.cl 27 l'Lla2 hl= 28 i.xhl
.l:.xhl 29 g6 .l:.gl?

Having achieved a winning posi

tion, Black makes a mistake---and
victory escapes him. After 29 ...'it>e8!
White's passed pawns are neutra
lised. For example: 30 @f6 l:tfl + 3 1
'it>e6 'it>f8 32 'it>d6 J::rf2 33 e6 J::rd 2+!
etc. or 30 g7 h6+ 3 1 'it>d5 Wfl 32
Wc4 l:th3, with a winning ending.

1 62 Two minor pieces against a rook

30 c;t.n l:tfl + 31 We6 J:igl 32 Wf7
J:tfl + 33 'it>e6 l:tgl Drawn.

The opportunity for the side hav

ing the rook to create a distant
passed pawn quite often appears as
a leit-motif of the struggle.
Novikov-Kaidanov
Vilnius, 1 984

pawn, while 1 0 .id8 loses after

1 0 ... c5) 8 ... fxg5 9 fxg5 (also poss
ible is 9 b3 with the threat of 1 0
ctJb4, but risky i s 9 f5 ? ! because of
9 . .J:ih2 10 ctJa l c;t>fl followed by
the creation of a passed h-pawn)
9 ...xg5 1 0 c;t>f4 and then b2-b3. By
reducing the number of pawns,
White can count on holding the
position.
.

6...Wf7 7 We3

Why not to e4?

7... 'it>e6 8 g4 ri;;dS 9 ri;;d3 J:Ic5 1 0

b3

Allowing a tactical solution to the

position. As seen from the previous
examples, the exchange sacrifice
does no,t achieve its objective after
1 0 b4, but in this case the break
through with the king decides1 0...I:!'.c4 1 1 .id2 J:id4+ etc.
10 ... .l:txc3+! 1 1 Wxc3 'it>e4

1...l:!.c4

Black immediately exploits the

break in coordination of the oppo
nent's pieces. In view of the threat
2 ...l:txc3 White is forced to allow
the creation of a passed pawn.
2 ctJc6 a3 3 ctJd4

The piece sacrifice 3 bxa3 l:txc3 4

a4 does not work because of 4 ... a6!
3...l:!.c5! 4 @f3
The knight cannot move: 4 ctJc2
axb2, or 4 ctJb3 I:!'.xb5.
4...a2 5 ctJc2 l:!.xb5 6 ctJal?

He cannot hope to hold the posi

tion with passive defence. Stronger
is 6 'it>e4 with the idea of transfer
ring the king to c4. For example:
6 ... .l:th5 (on 6 . . . 'i;;fl possible is the
prophylactic 7 g4 followed by 8
'it>d4) 7 h4 g5 8 hxg5 (dangerous is
8 .i.xf6 because of 8 ... gxh4 9 gxh4
Wfl 10 d4 .l:txh4, with the cre
ation of a second distant passed

Black has reached a technically

won endgame, since White's king is
riveted to the a2 pawn and the
knight will not succeed in contain
ing the pawn break on the king's
flank. We present the main vari
ation: 12 g5 @xf4 1 3 gxf6 gxf6 14
@b2 ri;;e4 ! 1 5 ctJc2 f5 1 6 ctJa3 f4 17
ctJc4 'it>d3 18 ctJe5+ 'it>e2 1 9 'it>xa2
f3 20 ctJxf3 'it>xf3 2 1 'ita3 h5, and
White is lacking one tempo to save
himself. With the pawn placed on
b4 it would be drawn.
The game ended in the following
way:
12 f5 'it>f4 13 'itb2 'i;;xg4 14 ctJc2
ri;; h3 15 ctJd4 'it>xh2 16 ctJe6 'it>g3 17
ltJxg7 'it>g4 18 ctJe8 ri;;xf5 1 9 ctJd6
'it>f4 20 ltJe8 'it>g5 2 1 ctJd6 h5 22
ctJe4+ ri;;f4! 23 tt:'ixf6 h4 24 ctJd5+
'itf3 and White resigned.

Let's look at a more complicated

example of the creation and realisa
tion of an outside passed pawn.

Two minor pieces against a rook J 63

Beliavsky-Dolmatov
Minsk, 1 979

l'tJf4 .l:Ic3 5 e2 .l:Ixb3 6 i.c4 lib2 7

d3 b3; or 4 l'tJf2 l:tcl 5 i.e4 .l:Ia l 6
ctJd3 l:!.xa2 7 ctJxb4 (White should
also lose in the event of 7 h4 f2! )
7 . ..l:Ixh2 8 ..id3 f5 9 'ite3 (on 9
.txf5 winning are both 9 ... b2, and
9 ...'it>f6) 9 ... 'it>f6 1 0 @f4 l:l'.f2+ 1 1
@e3 .l:r.b2 1 2 i.c4 'itg5 1 3 i.xf7
'itg4 14 i.e6 'itxg3, and each of the
passed pawns will cost a piece.
.

4 ...J::rc l 5 ctJf4 !!'.al 6 ctJd5 l:l'.xa2 7

l'tJxb4 Itxh2 8 i.e2 .l:Ih3 9 @f2

White has a material advantage,

but the presence of a weak pawn on
a2 and the lack of cooperation be
tween his pieces allows the oppo
nent to count on creating a distant
passed pawn. The position bears a
concrete character and in the event
of a premature fixing of the weak
pawns, similar to the previous
examples, White will manage to put
right the coordination of his pieces
and prevent the invasion of the
rook. For example: l ...a3 2 ctJf4
l:l'.c8 3 ctJe2.
However Black breaks through to
the a2 pawn in a tactical way.
1. .. c4! ! 2 i.xc4

After 2 bxc4 lib8 Black organises

the advance of the passed b-pawn: 3
ctJf2 b3 4 a3 (or 4 axb3 a3 5 .id 1
lixb3 etc) 4 ...b2 5 ii.b l lib3 6 e2
1::rc 3 7 ctJe4 l:l'.xc4 8 ctJd2 lie 1 .

Not allowing the rook to assist

with the advance of the a-pawn,
which would have been the case
after 9 @f3 .l:Ih l 10 i.c4 l:ta 1
(otherwise White can blockade the
a3 pawn by 1 1 ctJa2 and 12 b4) 1 1
ctJc2 J::ra2 1 2 ctJb4 lib2 etc. However
Black finds a new resource.
9 ... f5!

In the event of 9 ...h l White suc

cessfully solves the problem of re
stricting the rook by 10 i.fl .l:Ih2+
1 1 ..ig2 l:l.'.h8 1 2 ctJa2 l:tc8 1 3 i.e4.
Now however the threat of the break
... f5-f4 is irresistible. On 1 0 ii.fl
follows 10 .. .l:th2+ 1 1 i.g2 J::l.h 8, and
the rook breaks through on the
queen's flank.

10 ctJa2 f4 1 1 gxf4 l:.xb3 12 ..ic4

1::rh 3 ! 13 i.fl

Black is preparing a breakthrough

with his king, whereas White cannot
force the rook from the 3rd rank
without material loss. For example:
13 'itg2 l:re3 14 'iitf2 J::l.e4 etc.

2 ...J::l. c8

13 ...lih2+ 14 i.g2 @f6 15 @g3

lih5 16 i.fl lies 17 i.d3 J::rd5 1 8
i. a 6 l:!d2 1 9 i.c4 l:!.d4 White
resigned.

3 ..id3 a3 ! 4 'it>e3

An analogous idea of an ultra

sharp breakthrough was exploited
by Krarnni k against Ivanchuk.

the event of 3 ctJf2? there is the

threat of an exchange sacrifice
together with a pawn break3 ...l:!xc4 ! 4 bxc4 b3 5 axb3 a3 .
In

The most tenacious continuation.

The attempt to place the knight at
once loses quickly. For example: 4

1 64 Two minor pieces against a rook

Ivanchuk-Kramnik
Linares, 1 998

Tavadian-Tseshkovsky
Irkutsk, 1983

8 c;t;>g3 .l:Ial 9 lLixa3 xa3 1 0 @h4

Drawn.

To realise his material advantage

White should adjust the coordina
tion of his pieces in such a way that
the threat to advance the f-pawn
hampers the activity of the rook and
king and wins the pawns on the
queen's flank. Together with this it
is important to finnly blockade
these pawns.
1 ttJc6 .l:Id3+ 2 c;t;>f2 <J;;gs
Black would readily exchange the
b-pawn for the f-pawn, but 2 ...c;t;>f4
does not work because of 3 ttJd4 b4
4 ttJe2+ c;t;>f5 5 i.xb4.

In positions where the side having

the rook cannot actively support the
passed pawn, the minor pieces fre
quently prove to be stronger than
the rook. Therefore an important
task for the defender is to ensure the
activity of the rook and king.
With a small number of p awns a
good defensive resource is served
by the possibility of an exchange
sacrifice with the elimination of all
the pawns or with a transfer to a
position which, although without
pieces, is still a theoretical draw.

White does not find a clear-cut

plan to attack the b-pawn. As a re
sult there still remains an interesting
struggle for tempo, leading to a de
cisive increase in material advan
tage. As R.Tavadian pointed out,
there was a win here with the
brilliant manoeuvre 3 ttJd4 ! @f4 !
(3 . . .b4 does not work because of 4
ttJe6+ c,t>f5 5 ttJc5 xc3 6 bxc3
bxc3 7 ttJe4 c2 8 ttJg3+ and 9
ttJe2+-) 4 <Ji;e2 .:.e3+ (above all
4 ...b4 does not work because of 5
i.d2 etc) 5 <J;;fl ! d3 (he cannot

1. . c4! ! 2 bxc4
.

On 2 i.xc4 follows 2 ....li(d2+ and

3 ...l:!.xc2 and 4 ... .l:Ixa2, and Black
has a passed pawn. However it is
impossible to avoid the loss of a
pawn.

2 ...l!b8 3 c5 .l:!b2 4 c6 @e7 5

lLixg6+ c;t;>d6 6 lLie5 .l:Ixa2 7 ttJc4+
@c7

After 7 ...'3;xc6 8 i.xf5 there is no

threat of 8 . ..l:Ixc2 because of 9
i.xc2 a2 1 0 ttJa5+ and 1 1 ttJb3 .
.

3 ttJb4

Two minor pieces against a rook 1 65

play 5 ....l:Ie7? because of 6 i.d2

lt>g3 7 CZ'lf5+ or 5 ....!le8 6 'it>f2 .!lb8?
7 i.d2 'it>e5 8 CZ'lc6+) 6 'it>f2, and an
elegant 'dance' of the king along the
'triangulation' e2-fl -f2, typical for
pawn endings, culminating in the
win of the b-p awn.

Superficially the impression is

that Black's advantage is of a
charcater.
However
decisive
analysis shows that, even though the
d5 and g7 pawns are doomed, his
defensive resources are in no way
exhausted.

It was still not too late to return to

the above-mentioned plan. However
White's careless advance of the f
pawn removes his control of the e4
square, allowing the rook to develop
great activity with support from the
king. In effect the material advan
tage is now devalued.

In such endings it is very import

ant to activate the rook. This is
achieved by 2 l:Ig5 i.e4 3 f5 CZ'lxd5
(if 3 ... f6 4 g6 CZ'lxd5 5 'it>g4 the
chances are already on White's side)
4 f6 ! CZ'lxf6 5 .l:Ixb5 'it>xg7 6 l:Ib6! ,
with the idea o f 6 . . .d 5 7 lt>g5 . The
position is simplified and White ob
tains definite counterplay, asso
ciated with the possibility of
creating a distant passed pawn on
the queen's flank.

3 ... .!ld7 4 e3 'it>f5 5 f4?

5 ...h7 6 i.e5 lld7 7 'it>f3 lid2! 8

'it>e3 .!ld7 9 CZ'la2 :i:Id3+! 10 lt>e2
lib3

The white king is cut off along the

1 st-3rd ranks, and when the oppor
tunity presents itself an advance of
the b-pawn is threatened. The
chances of the two sides are even.
1 1 'it>d2 lid3 12 'it>c2 f3 13 CZ'lb4
'it>e4! 14 CZ'la6 lt>d5 15 CZ'lc7+ c6
16 CZ'le6 d5 17 CZ'lg5 li.g3 1 8 i.c3
.!lg2+ 19 @cl !!f2 20 i.d2 b4! 21
i.xb4 lixf4 and the game soon
ended in a draw.

Here is a good illustration of the

failure to exploit one's chances.
Miles-Adorjan
Wijk-aan-Zee, 1 984

1 J:ig2 g8 2 g5?

2 ... i.e4 3 .!lg3 CZ'lxd5 4 'it>h6?.

White again misses the opportun

ity to activate the rook, though in a
less favourable light. Correct was 4
f5 'it>xg7 5 g4 tLlf6 ! (5 . . .i.f3 looks
tempting on account of 6 .!id4 f6+ 7
'it>h4 CZ'le7, winning the f5 pawn,
however, by playing 6 l!tg3 !, White
is forced to repeat the position:
6 . .. i.e4 7 .!lg4, since it is dangerous
for the bishop to leave the b l -h7 di
agonal because of the threat f5-f6) 6
l:.g l i.c6 7 h4 ! (it is useful to hold
the check to the king in reserve:
after 7 \t>f4+ h6 it is not easy for
White to hold the position, for
example: 8 .!id l CZ'ld5 9 'it>g3 'it>g5,
or 8 l:tel h5 9 .!lgl CZ'ld5, with
advantage to Black; nevertheless
possibilities of resistance were of
fered by 8 h4 ! 'it>h5 9 g3 ! , but not
9 .!lg7? i.d5- +) 7 ... CZ'le4+ 8 f4
h6 9 .!lg8, and with an active rook
White can successfully defend
himself.
4 ... tt:Je7 5 !!c3 CZ'lf5+ 6 'it>g5 'it>xg7
7 l:tc8 CZ'ld4 8 l:te8 CZ'le6+ 9 'it>g4
i.c6 10 .!lb8 CZ'ld4 1 1 f5?

1 66 Two minor pieces against a rook

Now also the f-pawn is lost. We

mention that on 1 1 l:.td8 Black re
groups his pieces by 1 l ...i.f3+ 1 2
'it>g3 i.e4 followed by 1 3 ... tLlf5, and
the d-pawn quickly advances.
1 1 . ..'it>f6 12 h4 tlixfS 13 hS

Belated activity by the passed

pawn.

13 ... tLle3+ 14 h4 dS lS .l:td8 d4

16 .l:td6+ 'it>eS 17 h6 tiJfS+ 18 'it>h3
tLlxh6 19 lixh6 d3 20 .l:IhS+ rs
White resigned.

Let's look at an example where

the minor pieces have to hold back
an onslaught by an active rook when
there is relative material equality.

1. ..@f8 2 tLld2 e7 3 tLlb3 .!lb4 4

tLl3c5

And so the coordination of the

knights is put right. They defend
one another and therefore are
unassailable .
4 ... fS 5 'it>gl .!ibS 6 'it>f2 'it>f6 7
tiJd7+ 'it>e6

His hopes of driving away the

knight from the a6 pawn are unreal.
Stronger is 7 ... g5, trying to keep
back the king and create a passed
f-pawn.
8 tLlb6!

Keres-Szab6
Candidates (t), Zurich, 1 953

The knights have regrouped and

again are unassailable.
8 ... eS 9 'it>g3 l:tb3+

The initiative lays with Black: his

rook is very active, whereas the
knights are far away from each
other and, what is no less important,
they have no support points. Besides
this White needs to reckon with the
possibility of a transfer to a pawn
ending. White's main task is quickly
to put right the coordination of his
knights. Let us proceed with
P.Keres' logical and beautiful fili
gree manoeuvre which solved the
problem of the defence.

Worth considering is 9 ... f4+, but

even in this case the king is not ob
liged to step back. For example: 1 0
g4 f5+ 1 1 'it>g5 lib3 1 2 tLld7+
(weaker is 12 tLlc4 because of
1 2 ...d5 1 3 tLlbd6 f3 ! 14 gxf3
.!lxf3, and the knights are "hob
bling", while Black is able to create
a passed p awn) 1 2 .. d5 1 3 tLldc5
f3 14 gxf3 l:txf3 1 5 tLlxa6, with a
probable draw.
.

10 h4 l:!.c3 1 1 tLlcS!

A tactical solution to the problems

of defence-typical for such end
ings. White exchanges pieces and
transposes the game to a drawn
pawn ending.

Two minor pieces against a rook 1 6 7

1 1 .. ..l:IxcS 12 tL'ld7+ 'ii>d6 13

liJxcS 'it>xcS 14 'it>gS @bS lS c,t>h6
And the game was drawn after a

few moves.

With the pawns placed on one

flank the task of the defending side
still remains quite complicated.
Miles-Kindermann
Bath, 1 983

Black has the advantage, since he

has the possibility of creating a
passed pawn. But the small number
of pawns gives White a defensive
resource-at an appropriate moment
to sacrifice his bishop and obtain a
theoretically drawn position.
1. . .'it>e6 2 .i.c3
The best practical chance! In the
event of 2 Jl..c 7 l:i.c2+ 3 "it>b6 (after 3
'ii>b5 f4 4 .i.d8 nxh2 White loses a
pawn, since he cannot play 5 .i.xg5
because of 5 ...l:.h5) Black plays not
3 ... f4-4 Jl..d 8 J::rxh2 5 .i.xg5 with
possibilities of defence, but first of
all 3 ... h6, with the idea 4 ... f4 5 h4
gxh4 6 Jlxf4 h3- + .

2 ... I:!'.xh2 3 JldS+ 'it>e7 4 'ii>d4

@f6 S 'it>e3+ '\i>g6 6 .i. el f4+ 7 'it>d4
h5

A speedy advance of the distant

passed pawn should lead to a win
for Black.
8 @es h4 9 .i.e4+ @f7?

Carelessness, after which White

saves the game with a surprising
tactical blow. The natural continu
ation appears to be 9 ...'it>h6, and the
h-pawn swiftly advances. For
example: 10 'it>f6 h3 1 1 .i.b4 .lle2 I 2
.i.g6 .l:Ie6! , or I O c,t>fs .l:!g2 I I .i.b4
h3 I 2 'ii>f6 e2-in both cases with
an easy win.

10 .i.xh4 ! !

Now it all comes down to a theor

etically drawn ending.
1 0 gxh4
.

The situation is not changed by

1 0 ....l:Ixh4 I I '\i>f5 .l:Ih6 I2 'ii>xg5
.:.f6 I 3 Jl..f5 and I4 'it>xf4, with a
draw.
1 1 @xf4 I:!'.g2 12 .i.f5
White would have replied with
this move also after I I ...J::rh l . There
followed:
12 .. ."it>f6 13 Jl.. h3 l:.gl 14 .i.g4
.:.bl lS .i.h3 :i:Ib4+ 16 'it>e3 @es 17
.i.g4 .:.b3+ 1 8 'it>f2 '\i>f4 1 9 'it>g2
.l:!b2+ 20 'ii>h3, and Black soon con

vinced himself of the futility of

playing for a win.

1 68 Two minor pieces against a rook

When there are pawns on one

flank for both sides, the rook can
successfully contend with the minor
pieces.
Dan-Pytel
Austria, 1 978

14 'it>f3 .l:Ib4 15 i.d5 l:ra4 16 i.f7

.l:Ib4 17 'it>e2 1::rb7 18 i.c4 J::rd7

The king breaks through in the

centre, and Black changes the
rook's horizontal resistance to a
vertical one.

19 'it>f3 lk7 20 i.d5 .l:Ie7! 21 i.c5

.l:Ie8 22 i.e4 .l:Ic8 23 i.d4 l:Ie8 24
'it>e3

Black has prepared against the

breakthrough g4-g5. For example:
24 g5 fxg5 25 fxg5 lid8 ! 26 'it>e3
'it>g8, defending against 27 gxh6.
Now in the event of 27 gxh6 gxh6
or 27 g6 Black is ready to sacrifice
the exchange for the dark-squared
bishop, reaching a theoretically
drawn position.
Material advantage is on White's
side, but the inevitability of pawn
exchanges when pawns are ad
vanced, gives the defender the
possibility of exploiting the ex
change sacrifice motive to reach a
theoretically drawn position.
1. ..f6!

The more active l . ..f5 creates too

much space for the bishop's
activities.
2 g4 :g,33

One of the most important el

ements is playing for the restriction
of the king's movements.
3 i.f2 h6 4 h4 .l:Ib3 5 h5
White plans the transfer of the
bishop to d4 and organises the break
g4-g5.
5 ....l:Ib4 6 g3 :gb3 7 i.d4 d3 8
i.b2 .:i.b3 9 i.cl .l:Ic3 10 i.d2 J::rc 2

More principled is 10 . . .l:!d3, hold

ing the third rank and preventing
White' s coming manoeuvre.
1 1 i.e3 lic3 12 Wf2 a3 13 i.e4

.l:!a4

Now defence by the rook moves

on to the fourth rank.

24...l:re7 25 i.c5 .!id7 26 i.f5

.l:Id5 !

In preventing the attempted

breakthrough of the king to the d5
square, Black once agains includes
the rook for defence along the rank.
27 i.d4 lia5 28 e4 lib5 29 i.e6

White can attempt to break

through with the king, by playing 29
i.d7. Here are some sample vari
ations: (a) 29 . . . l:!a5? ! 30 i.b6 J::r a6
3 1 i.c5 lia5 32 Wd5 'it>h7 33 Wd6
Wh8 34 i.d4t; (b) 29 ...l:!b7 30 i.c6
l:!c7 3 1 i.d5 .!ld7 32 i.e6 !Ib7 33
i.c5 l:!bl 34 i.d7 .!lc l 35 i.f8t; (c)
29 . . .l:!b l ! 30 i.c5 l:l.'.c l 3 1 i.f8 c;t>g8
32 i.a3 l:l.'.al 33 i.e6+ 'it>h7 34 i.c5
.!lcl , and Black succeeds in defend
ing himself.
29 ... .l:Ia5 30 i.c4 a3 3 1 i.e6

As before 3 1 g5 fxg5 32 fxg5

does not work in view of 32 ... :gh3
33 gxh6 J::rh4+ 34 'it>e3 l:!xd4! with a
draw.
31 ... I:l.a5 32 i.d7 .l:Ia2 33 g5

Reconciling himself to the inevi

table draw. On 33 'it>d5 the balance
is maintained by 33 ...1::rd2.
33 ... fxg5 34 fxg5 I:!'.a6

Two minor pieces against a rook 1 69

Also possible is 34 .. Jlh2 3 5 gxh6

J::[.h4+ 36 @e3 l:rxd4 37 c,t>xd4 gxh6
with a draw.
35 g6 .!id6
Drawn.

On 36 i.g4
3 6 ... l:!'.xd4+! .

there

follows

Ponomariov-Plaskett
Hastings, 1 999

king to f6 or h6. However it should

not be enough to win.

1 1.. ..!le7 12 ctJb4 J:ie8 13 i.g6

J:ib8 14 ctJd3

It is difficult to reach the h6 pawn,

since Black does not allow the white
king admission to f5 . All his hopes
lay in 'humanitarian aid from the
West'.
14...c,t>d5 15 'it>e2 'it>d4 16 'it>d2
f3?

He can't keep himself waiting !

17 ctJel .l:!b2+ 18 lt.Jc2 'it>e5 1 9

'it>e3 J::rb3+ 20 i.d3 .!lc3 2 1 lt.Jel
:!cl 22 ltJxf3+
And White won shortly.

Changing the material balance

(obtaining a position with rook and
pawns against two minor pieces) is
very often exploited by chessplayers
of the highest rank.
The well-known Scotch ending
with two white minor pieces against
a black rook. The game smoothly
transferred to this drawn position,
there only remains to make a solid
move before the time control...

Petrosian-Tal
Moscow, 1 964

1. ...!la2?

In accordance with his active

style.
2 i.xd5 J::rd 2

2 . . . 'it>xd5 3 lt.Jb4+.
3 i.e4 i;t>e6

Despite the win of the pawn, as

before it is not apparent how White
wins.
4 h4 @d6 5 i.f5 .!le2 6 'it>f3 I:!'.e8
7 i.g6 .!le7 8 lt.Jcl J:iel 9 ctJd3 J::re7
1 0 i.e4 l:!e8 1 1 h5

Not an obvious decision. It was

possible all the same to exchange on
g5, attach himself to the black
pawns, place the bishop on f5, the
knight on e4, and try to run with the

l ... CZJxe3!? 2 i.xb7 CZJxdl 3 l::tx dl

.!lad8 4 CZJc3 l:!'.xdl + 5 CZJxdl J::i. d8 6
i.f3 l:id2

For the two pieces Black has not

quite enough material, but activat
ing the rook gives Black sufficient
play.
7 @fl a5 8 i.e2 e5! 9 a4 i.d4!

1 70 Two minor pieces against a rook

After creating his passed pawn,

Black can rivet White 's forces to the
struggle against it.

Formanek-Mikhalchishin
Hastings, 1 985

1 0 .txd4 exd4 1 1 'it>el lta2 12 g4

'it>t'8 13 f4 'it>e7 14 lllf2 .l:!al+ 1 5
.tdl lia2 1 6 h 4 f6 17 h 5 <;fte6 1 7
tlld3 g 6 1 8 hxg6 hxg6 1 9 .te2 .l:!c2
20 'it>dl :f.c3 21 @d2 g5
And Black quietly held the draw.
Beliavsky-Geller
Erevan, 1 9 75

1 ... tll x b3!

'it'xt'8+

2 'ihe7!

tll x cl !

There is nothing else for White.

3 ... @xt'8 4 :xcl .txa2

Black has a rook and two edge

pawns for two pieces. To win it is
sufficient for him to exchange the
bishop.
5 lll g5 f6! 6 lll e4

1 ...J:.b4!

Black stands to win the a2 pawn,

and White finds himself faced with
,problems beyond his control. Now 2
tlJd2 does not work because of
2 ...l:tb2. Therefore White has to
confine his bishop to passivity.

After 6 lllxh7+ 'it,;Jf:-1 7 tllxf6 Zhc3

8 11xc3 'ifi>xf6 the edge pawn must
decide the outcome of the game.

2 .td3 J:.a4 3 .Jlbl g5!

A pawn advantage on the flank is

only real when its exploitation gets
under way.

4 @e2 @g7 5 llld 2 f5 6 @d3 .l:!f4

7 'it,;le2
After 7 f1 .l:.h4 8 h3 h5 and 9 g4
...

a weakness arises for White on g2.

7....l:!h4 8 h3 J:.a4

And a draw was agreed.

6 ... f5 7 tll f6 :f.a7! 8 lll xh7+ i:r.xh7

9 .tb4+ l:!c5! 1 0 .txc5+ bxc5 1 1

Two minor pieces against a rook 1 71

lixc5 a7 12 i..c 4 i..x c4 13 xc4

a4

Again there has been a transform

ation of the advantage. This time the
game passes on to a rook ending
with positional advantage for Black.

1 4 lic2 a3 15 l:l'.a2 @e7 16 @fl

'it>d6 17 We2 'it>c5 18 'it>d3 'it>b4 19
'it>c2 e5 20 h4 e4 2 1 @bl lid7 22
@c2 lic7+ 23 bl 'it>b3 24 l:id2
l:!h7 White resigned.

11

Rook against knight and pawn

The rook is a little stronger than a

knight, and it would seem that with
pawns on one flank the knight, in
view of its capacity to jump, could
compete with a stronger piece.
However, practice shows that in this
type of position the knight is even
weaker than the bishop in the
struggle against a rook.
Farago-Csom
Budapest, 1 984

@g7 1 3 c,t>xf4 @f6 14 g3+ - or

1 0 ... lt:lxg2 1 1 'it>g5 f3 1 2 It.a2 and
then 1 3 J:if2, winning the f3 pawn,
and then also those on g6 and h5.
11 l:ras lt:Jg4+ 12 @n @h7 13

lla7! 'it>h6 14 'it>f8!

Avoiding 14 'it>g8 ctJf6+ 15 @h8

g5 !

14 ... ctJf6 15 J::!f7 lt:Jg4 16 J::!g 7!

Black resigned.

We don't need to talk about a big

alignment
of
forces,
when
brilliant
provide
miniatures
examples ...
Romanishin-Rodriguez
Moscow, 1 985

1 'it>h3!

Black is in a 'little zugzwang' and

he is forced to let out the white
king.
1 ...ctJd6 2 l:1c7 lt:Je4 3 a7 ctJf6 4
'it>g3 lt:Jg4 5 'it>f4 lt:lf6 6 e5 lt:Jg4+
7 We6 f4

Bad is 7 ... g5 8 'it>xf5 gxh4 9 a6+

'it>g7 10 'it>g5 lt:Je3 1 1 J:ia2, and
White attacks the pawns.
8 @n 'it>h7 9 l:!a5! lt:le3 1 0 'it>f6!
'it>h6

There is no saving himself by

1 0 ... tllf5 1 1 l:!xf5 ! gxf5 1 2 'it>xf5

Is it possible, perhaps, for a

grandmaster to lose this position?
As it turns out, yes-it is possible!
1 ...h5

Simpler is 1 ...ctJg7 and 2 ... ctJf5,

creating a typical drawn setup.

2 'it>e4 lt:Jg5+ 3 \t>f4 lt:le6+ 4 e3

lt:Jg5 ,

Rook against knight and pawn 1 73

Now, however, after 4 ...'Llg7 5

We4 'Llf5 6 'it>f4 and 7 l:ta6+
weakening the g5 square where the
white king will infiltrate.

When defending one should re

duce the amount of remaining ma
terial, therefore he should prefer 6
h5 ! .

Black loses his nerve. After the

correct 5 ...'Llt7 6 Wf4 'Llh6 7 .l:Ia6+
Wg7 8 'it>e5 'Llg4+ 9 We6 g5 ! it is
still a draw. Now, however, a rook
against pawn ending is reached.

White intends to attack the h7

pawn.

5 h4 'Lle6?

6 We4 g5
gxh4 9 'it>xe6
i;t>g5 12 lihl
lig2+ Wf3 15

7 1If5+ Wg6 8 'it>e5

h3 10 We5 h2 1 1 fl
Wg4 13 I:!'.xh2 h4 14
l:ra2 Black resigned.

Hulak-Beliavsky
Bled, 1 998

6...f5 7 h5 .lla3 8 h6

8 ... @c4 9 @e2 :gb3 10 'Llhl

Trying to provide a covering de

tachment for the black king.

1 0 ... l:rb2 1 1 'it>f3 i;t>d3 12 CLlf2+

Wd2 13 e4 l:!b3+ 14 i;t>g2 fxe4 15
'Llxe4+ We3 16 'Llc5

After 16 'Llf6 Wxf4 17 'Llxh7

Black plays 1 7 . . . :gb8 ! followed by
h8, and all the white pawns are
attacked.

16 ... lib4 17 Wg3 l:!xf4 18 'Lld7

.llf3 + 19 'it>g4 l:tfl! 20 'Llf6 l:!gl+ 21
'it>h3 'it>f3 White resigned.

Incidentally, Vidmar lost the same

classical endgame against Alekhine.
Salov-Kamsky
Buenos A ires, 1 994

Defence in such positions is not

easy and requires, in the first in
stance, an understanding of the best
distribution of the functions be
tween his pieces. Therefore White
should place his knight on h3 and
wait for the black king. Instead of
this he decides to post his knight on
g3, which proves to be incorrect.
1 'Lle2 lial 2 'Llg3 'it>e5 3 'it>g2
l::t a4 4 f4+?

Too active, but, you know, this is

not a rook ending. He should prefer

4 Wf3.
4 ... Wd5 5 i;t>o f6! 6 g5?

1 . ..'Llh7!

A correct transfer of the knight to

g6 in order to defend the h4 pawn
and attack the e5 point.
2 @h3

After 2 g5 'Llxg5 3 l:!xh4 the

knight all the same goes to h7!

1 74 Rook against knight and pawn

Beliavsky-Kupreichik
Budva, 1 995

2 f8 3 Ite4 g6 4 gs Wg8 5
Wg4 Wh7 6 @n <tig7 7 We3 @f8 8
@d4 @e7 9 \it>cs \it>d7 10 Wb6 lit>e7
11 lit>c6 lit>f8 12 lit>d6 'it>g7 13 l:t.e2
@f8 14 :n li>g8 lS .i:lf6 Wg7 and
..

White cannot break Black's de

fence, therefore draw.

It is well known that with pawns

on one flank the knight is rather
stronger than a bishop. But against a
rook the knight proves very weak,
since it often cannot escape from the
pursuit of a more mobile rook.
Beliavsky-Short
Linares, 1 995

1 ... li>e6?

Despite all the suffering, he

should allow l ...xe3 2 Itb5 li>f6 3
li>g3 Wg6 4 Ite5 or 3 tlJfl+ 4 li>f2
tlJh2 5 :lf5 ! followed by 6 Wf2-g3.
...

2 .l::Ib3 WdS 3 :i:t'.bS+ lit>c6 4 :S:xgS

tlJxe3 5 Wg3 tlJd5 6 :ifs e3 7 <tin
Wd6 8 l:tf8 @es 9 ll.a8 @e6 1 0
:S:a5! Wd6 1 1 g 5 tlJe7 1 2 :a'.a6+
Black resigned.
Suba-Chiburdanidze
Dortmund, 1 983
1 ltc3!

Inferior is 1 We3 g5 ! with the

threat of . . .f5-f4.
1. e4 2 b6 d2
.

Forced.

3 @e2 ext3+ 4 l:ixt3 'it>c6

If 4 ...e4, then 5 l:.b3!,

5 Itxf5 e4 6 .l:te5! xg3+ 7

'xd2 g5

After 7 . @xb6 8 @e3 g5 9 hxgS

hxgS 1 0 :Xg5 tlJfl + White plays 1 1
lit>d3 followed by t:tg2 and @e2.
8 hxg5 hxg5 9 Wd3 g4
On 9 'it>xb6 again follows 1 0
llxg5 tlJfl 1 1 !Ig2.
..

...

IO l:te6+ Wb7 1 1 <tic4 rs 1 2

.l:Ir6 tlJ3+ 1 3 Wc5 Black resigned.

One of the few positions where,

with equal pawns, the knight can
hold the position against a rook.
1 Wh5 2 :a'.a4
...

Rook against knight and pawn 1 75

On 2 .ll a6 there is 2 ... lt::\f5, while

on 2 'it>h3 tt::l e4 3 a5+ g5 4 g4+
fxg3 5 k!e5 tt::ld 6! 6 hxg5 lt::\f7 with a
draw.

A.Petrosian-Rashkovsky
Minsk, 1 976

2 'it>g4 3 'it>gl lt::\fS 4 i;t>f2 lt::\g3 !

..

Of course not 4 ... tt::lxh4? because

of 5 g3 ! .
S llb4

If 5 a6, then 5 ... g5 ! 6 hxg5 tt::le4

and 7 ... lt::\g5 with a drawn setup.

S ...tt::lfS 6 e4 gS 7 hxgS WxgS 8

1:1'.eS 'it>g4 9 lle8 tt::l h4

Also possible is 9 ... lt::\d 4.

10 J:Ig8+ @rs 1 1 'it>e2 tt::lg 6! 1 2

i;t>d3 tt::leS+ 13 'it>d4 f3 ! and a draw!
Stein-Bobotsov
Moscow, 1 967

Is there a basis for Black to play

for a win? To this question the
grandmaster himself answered in
the affirmative and to start with
went after one of the pawns. There
followed:
1 ...l:l.'.d3 2 e4 l:!a3 3 @f2 .:i.h3 !

Black suggests that his opponent

leaves the g-pawn. His argument ap
pears to be 4 i;t>g2 l:th4 5 tt::l e3
l:txg4+! 6 tt::lxg4 i;t>xg4 with a win.
The Erevan grandmaster agrees with
the evidence and tries to organise a
defence with an equal number of
pawns.
1 ...'it>fS 2 lid4 bS 3 axbS axbS 4
lids 'it>e4!

Activity!

S xbS tt::l d3+ 6 'it>e2 f3+ 7 'it>d2

f2 8 i;t>e2 i;t>d4 9 .:i.b8 'it>c4 10 @fl
'it>d4 l l lib7

In the event of 1 1 b4 'it>c4 1 2 b5

'it>c5 13 b6 tt::lb4 Black holds on.

1 1...i;t>c4 12 'it>e2 i;t>d4 13 J::i.d7+

'it>e4!

The only, but sufficient, move.

14 l:!b7 'it>d4 Drawn.

4 'it>e2 lig3 S @f2 l:rxg4 6 i;t>f3

l:!h4 7 'it>e3 l:th3+ 8 'it>d4 lla3 9
'it>c4 .!Ia6 1 0 'it>d3 'it>g4 1 1 i;t>e2
On l l 'it>g3 there is the reply 12
...

We3, and it is not apparent how he

can improve his position. Rash
kovsky in fact allows the second
argument, the essence of which con
sists of transferring the turn to move
to his opponent.
1 1. .. lla2+! 12 'it>e3 l:ta3+ 13 'it>d4
.!Ia6 14 'it>e3 i;t>g3 lS lt::\f4 l:!a3+ 16
tt::ld 3 l:l'.b3 17 'it>e2 J:tb8! 18 'it>e3
lle8 19 tt::lf2 fS 20 tt::l h l+ 'it>h2 ! 2 1

1 76 Rook against knight and pawn

lLif2 'it>g2 22 'it>f4 'it>xf2 23 exf5

.l:If8!

And here is the last argument,

bearing in mind that the ending has
assumed a study-like character.
24 'it>g5 e3

On 24 @e5 Black would have

proceeded on the other side-24
... g3 .
25 f6 @e4 White resigned.

6 h4! g5 7 hxg5 h4 8 .l:Ixh4

And Black resigned because of

8 ...tt.Jdl + 9 d2 liJf2 10 g6 @f6 1 1
.l:If4.
The zugzwang also plays an im
portant role in the next example.
The extra pawn on the other flank is
usually lost after a few moves.
Adorj an-Morozevich
Alushta, 1 994

Chemin-Geo rgiev
Dortmund, 1 991

The defence, even with play on

one flank and an extra pawn, isn't
easy. Here the most unpleasant
thing is zugzwang.
1 lie8! lLib2 ! ?

O r 1 . ..d6 2 l:!.g8 lt.Je5 3 @e4 g5

4 @f5 winning easily.
2 .:i.d8+ e6 3 lid4

Now White's idea is to hunt for

the runaway knight.
3 ...f6 4 e4 h5

After 4 ... @e6 White wins by 5

:!:i.b4 liJdl 6 1::rb6+ 'ii'f7 7 llb3 ! Ci:Jf2+
8 'it>e5 @g7 9 .:i.f3 lLihl 1 0 'it>f4 g5+
1 1 g4 g6 1 2 lifl h5+ 1 3 'it>f3
and the knight is caught.
5 e3 ! @e5

Or 5 ...g5 6 e2 followed by 7
llb4.

1....lld l ! 2 lLig6+ @h7 3 lLie5

:!fl!
A 'little zugzwang'-White has to
release the black king or start push
ing his pawns somewhere.
4 c4 g6 5 h4 @g7 6 Ci:Jg4 h5 7
lLie3 :!:!'.bl !

The white king is riveted to the h4

pawn and his black counterpart is
ready to start his 'promenade' .
8 tt:'ld5 @f7 9 lLie3 @e6 1 0 'it>f4

White tries his last chance.

10 .l:Ixh4+ 1 1 g4

Or 1 1 'it>g5 I:!'.e4 12 lt.Jd5 l:!xc4 - + .

1 1 . .l:Ihl 12 gxh5 lixh5 13 'it>g4

e5 1 4 c5 e6 White resigned.
.

Now two very similar positions

with just one small difference-the
white pawn.

Rook against knight and pawn 1 77

Sokolov-Khalifman
Minsk, 1 986

llJg6+ 2 'it>e4 'it>g5 3 .l:Ig7 'it>h6 4

.!la7 'it>g5 5 .!la3 CZJe5 6 .l:Ig3+ 'it>h4 7
Wf4 ClJc6 8 !Ic3 ClJe7 9 .!id3 'it>h5 1 0
.!id7.
1. ..@h4 2 h8+ 'it>g5 3 'it>d5?

This king transfer isn't correct.

The king has to go towards his own
pawn when White has serious win
ning chances.
3... QJf3 4 'it>e6

It was still possible to return to his

pawn.

4 ... CZJe5 5 'it>d5 ClJf3 6 Wd6 CZJe5 7

'it>e6 .!i:Jf3 8 'it>e7

Black played
1 ...'it>g5

... trying to activate his king. The

game continued

2 'it>f3 'it>f5 3 l:tal 'it>g5 4 .li(a5+

CZJe5+ 5 'it>e4 Wh4

Black tries a counterattack on

White's pawn, but he hasn't enough
time.

White has definitely decided not

to go back to his pawn.
8 . .CZJe5 9 .l:Ia8 CZJg6+ 10 @f7 ClJf4
1 1 .l:Ia3 f5 12 l::tc3 ClJxh3!
.

Black has calculated this known

endgame very carefully.

13 lhh3 \t>g4 14 .llh6 f4 15 .:i.g6+

'it>h3 16 I:!'.f6 Wg3 1 7 'it>e6 f3 18
'it>e5 f2 19 We4 'it>g2 Drawn.

7 . . ClJf3 doesn't work because of 8

.!la3 ! 'it>g2 9 .l:Ixf3 Wxf3 1 0 h4+ -.

A knight is generally more useful

than a bishop when the pawns are
all on one flank, but against a rook
the knight has a more difficult time.

Dokhoian-Shirov
Klaipeda, 1 988

Horvath-Sherzer
Brno, 1 993

6 'it>f5 'it>h3 7 .!la2 ClJd7

.

8 .l:Ia7 Black resigned.

The game continued

1 'iite4?!

This is not the most exact way to

win-more direct would be 1 \t>f4

1 'it>g3 'iit e6 2 .!la3 g5

Black has to build some sort of
wall against White's king, but it
cannot be done.

1 78 Rook against knight and pawn

3 l:!a5 Wf7 4 'it>f3 'it>g6 5 We3

'Lih5 6 J:Ia6+ Wg7 7 'it>e4 'Lif4 8
@f3

Another good possibility was 8

l:!.a2 'it>f6 9 h4 'Lig6 l O h5 'Lif4 1 1
g4 and after a check the king ob
tains the very important f5 SCI!lare.

7 @e3 h4 8 l:!.f3 'Lixg2 9 @f2 h3

1 0 'it>g3 'Liel 1 1 l:!c3 'Lig2 1 2 'it>xh3
'Lif4+ 13 Wh4 'it>f6 14 J:Ia3 'Lig6+
15 'it>h5 'Lif4 16 @h4

There is nothing in 1 6 @h6 'Lid5,

followed by the advance of the
f-pawn.

8.. 'Lih5 9 'it>g4 lt.Jf4 10 g3 tll e2 1 1

l:!e6

16 ... 'Lig6 17 'it>g3 e5 18 h4 'Lif4

19 .!la6+ Wg7 20 'it>f3 'it>f7 2 1 J:Ia5
'\t>f6 22 .!la6+ Wf7 Drawn.

1 1 'Lid4 1 2 .l:!e4 'Lib5 1 3 'it>h5

'Lid6 14 .!le7+ Black resigned.

Zaichik-Sorin
Erevan, 1 996

Another method would be 1 1

.!id6, 1 2 l:!d2 and 1 3 'it>h5.
..

Two instructive and I would even

say amazing endings played by an
international master from Argentina
against experienced grandmasters.
Sorin-Alterman
Erevan, 1 996

For the exchange White has three

pawns and an easily winning
position-for example:
1 'Lig6 l:!.a6 2 f5 .!lb6 3 f6+! .!lxf6
4 h6+ rt;f? 5 h7 .l:!xg6+ 6 'it>h5 .l:!g8
7 hxg8='it'+ c;t>xg8 8 'it>g6.
But Zaichik did not see this plan
and chose another one.

White
threatened
to
play
J:ig5-g3-c3 and, after capturing the
b2 pawn, winning the well-known
ending of rook against knight with
three pawns on the flank. But Alter
man finds an elegant way to save
the game.
1...'Lie3+! 2 'it>xb2 'Lid5 3 'it>c2

If 3 g3?, then 3 ...'Lif6, and the

rook is in a trap after 4 ... h6.
3 ...'Lixf4 4 Wd2 Wf6 5 l:!g8 'it>t7 6
llg3 1!5!

Activity comes first!

1 c;t>f5 .l:!f6+ 2 'it>e4 .!la6 3 g5 :!.al

4 '\t>f5
Not bad was 4 f5 :a4+ 5 'it>f3
l:!a3 + 6 Wf4 .!la4+ 7 'it>g3 l:!a3+ 8

'it>h4 l:ta4+ 9 'Lig4, and there is no

defence against f5-f6+ and g5-g6.
4 ... a2 5 g6 'it>h6!

The threat was 6 'it>g5 and then 7

h5-h6.
6 'Lid7

6 'Lig4 led to roughly the same

thing.

6....!la5+ 7 '\t>f6 l:ra6+ 8 Wf7 .!la7

9 'it>e7 .!la5 1 0 'Lie5

Rook against knight and pawn 1 79

Better really was 1 0 ll'if6, prepar

ing the return of the king.

10 ... l:!.a7+ 1 1 'it>f6 l:ta6+ 12 @f7

lia7+ 13 @f6 .l:Ia6+ 14 'it>f5 @xh5
15 g7 .l:ta8 1 6 ll'ig4

Martinovic-Mikhalchishin
Jgalo, 1 994

There is nothing in 1 6 'it>f6 'it>h6.

16... .l:ta5+!

Now Black's main task is to keep

the king as far away as possible
from the f4 pawn.

17 @e4 l;Ia4+ 18 'it>d3 a3+ 19

'it>e2 lia2+ 20 @f3

If the white king heads towards

the rook, then at a necessary mo
ment follows lia8, ll'if6 'it>g6, g8='iV
l:txg8, ll'ixg8 'it>f5, and a draw.
20 ... lia3+ 2 1 ll'ie3 .l:Ia8 22 ll'if5
@g6 23 g8='iV+ lixg8 24 ll'ie7+ @f7
25 ll'ixg8 Wxg8 26 'it>e4

1 b5?

Nothing is gained by 1 @fl!

l:l.'.xb4 2 @g2.
1 ...'itid5 2 ll'if5 @e4 3 ll'id6+

There is no saving himself by 3

ll'ie3 'it>f3.
3 'it>f3 4 'it>d 1
..

In the event of 4 ll'ie8 l:l.'.b 1 + 5

'it>d2 l:txb5 6 ll'ixf6 after 6 ... h5 with
the idea of 7 ...f5 Black has a de
cisive advantage.
4... l:txt1

4 ... 'it>xf2? 5 ll'ie4+ 'it>f3 is mistaken because of 6 ll'ixf6 lixb5 7

'Llxh7.
2 6...'it>fS!

A clear-cut move--and a draw!

5 @cl

The result is not changed by 5

ll'ie8 @xg3.

5 .l:ta2! 6 b6
Or 6 ll'ie8 5 7 ll'if6 h6.
6 lia8 7 @c2 lib8 8 b7 'it>xg3 9
'it>d3 'it>xh4 10 'it>e4 'it>g4 1 1 'it>d5 h5
White resigned.
..

In the following position White

has good compensation for the ex
change, but he should immediately
start building a fortress for his king.

12

Queen Endings

In queen endgames the main prin

ciples are slightly different from
other endings. And, to be precise,
though material advantage (for
example, a pawn) naturally has
great significance, much more im
portant is a passed pawn. Thus the
main principles can be placed in the
following order:
(a) Passed pawns
(b) Activity (approach) of the king
(c) Activity of the queen
(d) Combinational exploitation of
various factors
Passed Pawns

2 gxf7 e2+ 3 cl xf7 4 h5!

White must exploit his chance to

the utmost-the passed h-pawn is
also ready to be exchanged for the
b5 pawn.
4 ...c4+ 5 'it>d2!

Inferior is 5 'it>b 1 b4, and after the

advance of the b-pawn the white
king is threatened with mate. White
does not fear the transfer to a pawn
ending.
5 ... 'iVc6 6 'iVh8!

White wants to advance the h

pawn further.

6 'ifd5+ 7 @cl 'ifc4+ 8 b2

'ifd4+ 9 '>t>b3 !
..

Meeting the pawn half-way!

9 'iVa4+ 10 'itb2 'ib4+ 1 1 'it>a2

'iVd2+ 12 b3 'iVd3+ 13 'itb4
'ifd4+ 14 'it>a5 'iVa4+ 15 Wb6
.

Mikhalchishin-Gufeld
Nikolaev, 1 981

The white king is up in the

'clouds' and Black reconciles him
self to a draw.
Ed.Lasker-Marshall
New York, 1 923

There followed:
I g6+! 'it>g7

If l .. .'it>xg6, then 2 'iVg5+ with

perpetual check or l ... fxg6 2 'iV e7+
with the same result.

Queen Endings 181

A difficult struggle with a passed

pawn hoping to queen but with the
king unable to get over to help it.
1. ..'ii'e2 2 'ii'c l g5!?

Striving to create counterplay.

However 2 ...'ii'e 5 deserved the pref
erence, intending to transfer the
king to the queen's flank with an
exchange of the h7 and g6 pawns
for the pawn on a5 .

3 a5 f4 4 'i'fl 'ii'a 2 5 'it'b5! g6 6

'iib6+ 'it>h5 7 a6 g4 8 a7 g3 9 fxg3
fxg3 1 0 'iVc5+ 'it>g6 1 1 iid6+ @f7
12 iif4+ e6 13 'it'e3+ d7 1 4
@fl! 'ii'a6+ 15 el '\tics 16 'iVxg3
'iVxa7 17 'ii'g8+ 'it>b7 18 xh7+ and
19 xa7 + -.

Again not 7 'ixg6? in view of

7 .. .h2+! . It is important to activate
his king via the g5 square.
7 ...'ii'a 2 8 Wg5 'iVf2

No help is 8 . ..'if7 9 'iVf6! .

9 f4 \Wb6 1 0 iif7+ Wc8 1 1

'iVxg6
Black can resign since l l . . .'it'e3 +

1 2 Wxh5 \Wxf3+ is no use because

of 1 3 \Wg4+.
Activity (approach) of the king

One of the most important factors

in queen endings.
Ribli-Hertneck
Dortmund, 1 986

Taimanov-Gligoric
Santa Fe, 1960

1 'ii'd3 !

Defending against . . .'ic2+ and

preparing a4, therefore B lack de
cides on a pawn sacrifice with the
aim of activating his forces.

1 ...'it>f6 2 'ii'xd5 'ic2+ 3 h3

'ii'c l

Chances of saving the game re

mained with 3 ...'ii'f2 ! 3 f4 fifl+ 4
'iVg2 iid 1 .
4 'iVd4+! 'it>e6

Not possible is 4 . . . @f7 because of

5 'it'f4+, nor 4 . . . 'it>g5 because of 5
'Yi'e5+.
5 'ii'e4+ Wd7 6 h4 \Wd2 7 e5 !

1 h3 h5

Otherwise White will play g3-g4

obtaining real winning chances.
2 'id7+ '\t>f6 3 'iVe8 a6?

The only defence was 3 .. .'it>g7 4

'ii'e 7+ @h8 5 iif8 'it>h7 6 Wiffl+ h6
7 'ii'g8 'iWh2 and White must return
to the plan with g3-g4, creating a
passed e-pawn.
4 \Wf8+ 'it>e6 5 'it>f3 !

Suddenly there has appeared the

idea of penetrating to the king's
flank with the white king-very
typical in fact.
5 ... a4 6 'ii'g8 'it>f6 7 'iVd8+ 'it>e6
8 'ii'd4!

182 Queen Endings

And Black resigned because

there is no defence against Wf4-g5.
For example 8 ... 'ifa2 9 'ifg7 or
8...c2 9 e4! .
Mikhalchishin-N esto rovich
Budva, 1 994

1 'it>c3! 'ii'cS+ 2 'ii'c4 iVe3+ 3 Wb4

d2+ 4 'it>c5 'ii'a5+ 5 'it>d4 'ii'd 2+ 6
'it>e5 d8 7 'ifd4 'ifh8+ 8 'it>e4
'iVhl + 9 'it>d3 11Vt'3+ 10 'it>d2 'ii'g2+
1 1 'it>cl 'ii'fl + 12 'it>b2 'ii'b5 13 'ifd8
Black resigned.
Hoi-Mikhalchishin
Budapest, 1 989

1. 'it>e7?
.

In queen endings the king must

either be ultra-safe or ultra-active!
Better here would be l .. .'ii'a3 ! cut
ting off the white king.
2 e5 'ii'b6
If 2 . 'ifa3 then 3 'ii'd 6+ 'it>e8 4
.

e6+ -.

3 'it>t'3 'ifc7 4 'it>f4 'i1Nb6 5 'it>g5

'it>e8 6 'it>h6! Black resigned.
Rittner-Bernstein
corr, 1 968

The king usually approaches its

own pawn, but very often it also
heads towards the opponent's king
to create mating threats.

1 'iVrs+ 'it>g6 2 iVgs+ 'it>rs 3

'ii'h7+ 'it>e5 4 'i'e7+ Wd4 5 1!Vd8+
'it>e3 6 'ii'e7+ Wf2 7 'i'a7+ @fl 8
'ifgl+

As a result of the approach of the

black king White has had to mis
place his queen.

8 ... We2 9 'ii'd4 c2 10 'il'g4 'it>d3

1 1 'ii'g_3+ 'it>d2 12 'il'f4+ 'it>c3 13
'iVcl d6+ 14 'it>h3 'iVd3+ 15 g3
'ii'd7+ 16 g4 'ii'd 4!

The black queen is centralised to

the utmost, there is no threat of per
petual check, and the march of the
a6 pawn is decisive.

17 'it>g3 'ife5+ 18 'it>h3 a5! 19 a4

'it>b3 20 'it>g2 'ii'b2 21 'ife3+ 'it>a2
White resigned.

Queen Endings 183

The active position of the queen

This allows the king to be utilised

in the struggle against enemy
pawns.

Marshall-Alekhine
Bradley Beach, 1 929

Gufeld-Minev
Sofia, 1967

1 ...iVa6+! 2 @g2 Witd6

In the first instance Black block

ades the opponent's pawn.
3 iid4 @gs 4 11Vc4 @f7

The white queen controls the

position and attacks all Black's wea
knesses. It is only necessary to in
clude the king.

1 @fl fS 2 h4 'i'c7 3 @e2 'iib6 4

'ikf7+ @h6 S b3 'i'd6 6 'i'c4 'it>h7 7
@el 'iib6 8 'iib s Wi/c7 9 @d2 'i'eS
1 0 @c2 iial 1 1 'i'xcS 11Va2+ 12
@c3 iixf2 13 11Vc7+ 'it>g8 14 'it>d4
'i'b2+

If 14 ...11Vd2+, then 1 5 e5 'i'xe3+

1 6 @f6 with a decisive invasion.

lS 1!Vc3 'i'f2 16 fixaS 11Vxg3 1 7

iid8+ @h7 1 8 a S Wi/el 19 'iib 6 f4
20 'iVc( 'it>g8 21 a6! 'i'd2+

After 2 1 .. .'Wixe3+ 22 'it>e5 one

advanced passed pawn is quicker
than two.
22 'it>eS 'ii'h 2+ 23 @e6 'ii'xb3+ 24
f6 'ii'h2+ 2S 'iVeS Black resigned.

In the following position, who

stands better? He who has the
further advanced pawn or he who
advances it further? Let us see.

It turns out that the white d5 pawn

can be attacked by the king.

s Wilbs iVd8 6 f4 11Vc7 7 @f3

1!Vc3+ 8 'it>g4

Clearly not 8 @e4 'ir'b4+!-the

pawns in this formation usually win
for Black.

8...'it>e7 9 'iib6 hS+! 10 'it>xhS

'i'h3+ 1 1 @gs 'i'g2+ 12 @rs
'i'xdS+ 13 'it>g6 'i'd6+! 14 'i'xd6
@xd6 lS fS a4 1 6 @xg7 a3 and
Black won.
Keres-Aronin
Moscow, 1 951

1 84 Queen Endings

1 ...@e6?

Better is 1 . . h5 or 1 . ..e4, trying

somehow to exploit his pawn major
ity on the king's flank.
.

2 g3 'iVdl+ 3 @g2 'li'd7 4 h4

'itb7+ 5 @h2 'il'd7 6 'iVf8! h5?

Combinational exploitation of
various factors in queen endings
Ballon-Mikhalchishin
Berne, 1 995

A decisive mistake. Correct is

5 ... f5, endeavouring to advance his
pawn.

7 IVgs+ @rs 8 c5 'ii'd 4 9 'iVc8+!

@e4 10 Wg2 @d3 11 \'Wa6+ @d2 12
'iVd6

Also not bad is 1 2 'il'xf6.

12 ...@e2 13 1.Wxd4 exd4 14 c6 d3

15 c7 d2 1 6 c8='iV d l=\li' 1 7 'ii'e6+
Wd3 1 8 'li'd7+ We2 19 'ii'e 7+ @d3
20 'ii'x a7 and White won easily.
Romanovsky-Sthlberg
Moscow, 1 935

Black has a protected passed

pawn plus a much better p awn
structure.
1 ...'ii'f5! 2 @gt Vi'd3!

By speculating on a transfer to a
winning pawn ending, Black
threatens to advance his passed
pawn.
3 'li'f4 c3 4 'ii'g4 !

White wants to put the black king

in a box, but it boldly ploughs on.

4 ... f5 ! 5 'ii'h5+ @e6 6 ik'e8+ @d5

7 'ii'e5+ 'i!tc4 8 d5 c2 9 ik'f4+ 'it>b3
1 0 Wh2 'iVc4 White resigned.
1 \'Wd5!

After this centralisation White

threatens to transfer his king to e4,
therefore B lack must hurry with his
counterattack.

1 ...\'Wb2+ 2 @e3 Yl'cl+ 3 @e4

'iie l+ 4 @f5 'iix g3 5 iixc5+ @g8 6
'i'c4+ @h8 7 'ii'g 4!

With the threat to exchange White

obtains the g6 square for his king.

7...li'el 8 'itig6! 1.Wct 9 Vi'd7

V!\Ygl+ 10 @f7 @h7 1 1 'iVf5+ g6 12
'it'd7 Black resigned.

Botvinnik-Donner
Noordwijk, 1 965

Queen Endings 185

Black has the better structure, and

therefore White's only chance lies
in the creation of a passed a-pawn,
which will be quicker than the
passed h-pawn.
1 'iVd7+ @gs 2 WVcs+ 'itih7 3
'iVxa6 'ili'g5+ 4 @fl 'tlkxh5 5 'ili'd3+!
'itigs

A trap-after 5 ...'iVg6? 6 'ie4! the

pawn ending is winning for White.

6 a4 'iVhl+ 7 'itie2 h5 8 'iVd5+

'itih8 9 'iWdl !

Typical-by speculating on the

transfer to a pawn ending, White
forces back the opponent's queen.
9 ...'iVh2?

Better is 9 ...h4, though even

then White's position is preferable.
10 b4 h4 1 1 fl !

The queen is trapped!

1 1 ...g5 12 a5 bxa5 13 bxa5 g4 14

a6 Black resigned.
Korolev-Palm
corr, 1 980

2 c5 e5

On 2 ...xa3 would have followed

3 1!fc7+ @f6 4 c6 'licl + 5 'itih2 e5 6
'@'d6, leading to a position occurring
in the game.
3 'iVc7+ 'itie6 4 d6+ @f7 5 c6
The pawn is already on c6 and
there is no apparent perpetual check.

5 ...'i!Vcl+ 6 'itih2 'i!Vf4+ 7 'itigl

'i!Ve3+ S @fl f4+ 9 'itie2 e4+ 10
'itid2

The king will boldly go behind

enemy lines, the pawn already has
no significance.

10 ...'i!Vf4+ 1 1 'itic2 VWe4+ 12 'itib2!

'i!Vxg2+ 13 'itic3 'ia2 14 'i"d7+ 'itif6
15 c7

White has achieved his objective

-a pawn on c7. Now he needs to
shelter his king from the checks.
15 ...'i!Vxa3+ 16 'itic4 'iVcl + 17
'itib4 'i!Vel+ 18 'itic5 'i!Va5+ 19 'itic4

Of course, not possible is 1 9

@c6? because of 1 9 ..'iVh5+ with a
perpetual check.
.

19 ...'ili'a2+ 20 'itib4 'i!Vb2+ 21 'itia5

a3+ 22 'itib6 'YWe3+ 23 'it>b7 'iVb3+
24 @cs ib4 25 'i\Vd5 Black
resigned, since the c7 pawn will in

evitably go on to queen.

Karpov-Timman
Buenos Aires, 1 980

Thanks to his far advanced passed

c-pawn and strong queen position
on d6, White has a slight advantage.
1 . ..'iVb2?!

On l ...g5 there would have fol

lowed 2 1!fd3 'itif6 3 'i!Vd4+ @f7 4
'it>h2 'iVc7 5 'itig l with a slight ad
vantage. Stronger is 1 ...'ili'e4! , cen
tralising the queen with chances of a
draw.

White has problems since the

Black king is in play while White's

186 Queen Endings

is placed far from the scene of

action.
1. ..b6 2 'ifh7

If White had sacrificed a pawn to

create a passed pawn by 2 h4, then
Timman gave the variation 2 ... gxh4
3 g5 h3 4 g6 h2 5 1!fe3+ 3 6
1!fg5+ 'it>t2 7 1!fd2+ l'We2 8 'iVf4+
@g2 9 'ilkg5+ Wh3 ! 1 0 l'Wh6+ \t>g4
1 1 g7 'iVd l+ 12 \t>a2 1lfd5+ 13 b3
1!fg2+ followed by 14 ... h l .

Correct is 3 ..."i/Vb2 4 l'Wf5+ 'it>e7 5

l'Wc5+ @e6 6 c6+ 'it>e7 7 'iVe4+
@d6 8 f4, and all the same White
has a dangerous passed pawn.
4 c8 We7 5 c7 @e8 6 \t>f6!

Fantastic domination by the white

pieces, leading to a winning pawn
ending_.
6...1\fb2 7 e5 Black resigned.

2 ...d4

The exchange of the a7 pawn for

that on h3 favours Black.

Marosi-Yudasin
Budapest, 1 982

3 'iVh6 'iVdl+ 4 @a2 l'Wd5+ 5 'it>al

a5 6 h7 1!fd4 7 l'Wh6 l'Wdl + 8 'it>a2
d5+ 9 @al b5

Threatening a mating attack after

. . .b5-b4-b3.

10 h4 gxh4 1 1 g5 b4 12 axb4
axb4 13 b3

After 1 3 g6 b3 ! 14 ife3+ Wg4 1 5

l'Wgl + \t>h5 ! Black wins.
13 ...e5+ 14 \t>a2 e2+ 15 @bl
ife3 White resigned.
Yusupov-Gerusel
Moscow, 1 981

1 ...1!fd5!

This centralisation is better than

simply playing to eliminate White's
pawns.
2 ife8+ @h7 3 ifb8 d7

He should not give the opponent

counter-chances after 3 ...'Wxf5 4
1!fxa7 't\Vbl + 5 @g2 xb2 6 l'Wxf7.
4 l'Wf4 'iVc6 5 g4

White's only chance to change the

course of the struggle is to open up
the position of the opponent's king.
5 ... a5 6 h4 b5 7 g5 ifc4 8 g3

Despite the material equilibrium,

Black has problems in connection
with the excellent positions of
White' s king and queen.
1 l'We6+ @f8 2 'it>g6 g2+ 3 'ifg4
a2? ,

e4! 9 gxh6 gxh6 10 c7 xf5 1 1

l'Wxa5 'ilfhl+ 1 2 'itig2 1!fxb2

As a result of the exchange of

blows Black's positional advantage
has been transformed into a material
one.

13 'ifc7 b4! 14 @fl 'Wi>l+ 1 5

\t>e2 e4+ 16 Wd2 1!ff5 17 iVh6

Queen Endings 187

On 1 7 'ii'g3 Black
achieves his
\IUr
ob/i ect1ve after 1 7 ... wd5+ 18 = c2
,w
wc4+ 19 ,-.h.2 'i/Nc3+! 20 wxc3
bxc3+ 2 1 xc3 g6 followed by
'""'
... ..,.,h5
and ... xh4 .

1 7 'i/Nf4 18 @d3 'ii'xh4

Further comment'ilN
is unnecessary
Wr
'ii'f6 20 a7 b3 21 "iV S
It> 19 'we3
g6 22 iVf8 "iVxf2 23 'ig8+ @h5
White resigned.
..

Mikhalchishin-CvetkoVIC
"'
i rnava, 1 988
,

.
Sometimes a com phcated method
has to be adopted to create a passed
pawn.
1 h6! 'i/Nc4+

. 1 ...'i/Nxh6 2 'ii'c8+ and 3 'i/Nxc6

winning.

@f3

@
l

;:f

'ii' 6 3 e5! 'ia3+ 4

2
xh6 6 'i'c8+ h
"iVxb4 5 ex
'i1Nd7+ @g6 8 g7+ @h5 9 f7 'il'e4+

-/. ,.
0%rut.@
., ...

-=

. .
.

.
, .
-

.*

w
,
if:
"

8
1
p,
.

lffilj

D
@!' '.t>h2 'ii'f4+ 11 @gl 'ii'c1 + 12

As a result of the 'triangulatio

more check
dance' Black h
'
e game.
and so he resign

1;:

188 Various Endings

Exercises:
Various Endings
1

Indicate White 's winning plan.

Find a plan to realise White 's

extra pawn.

Find the right continuation

for Black.

Can White win?

Various Endings 18.

What is right:

l . . . ilbl

or l ii.e6?
...

How does Black realise his extra

pawn ?

How does Black defend after

I 0.xa5 ?

Can Black make a draw?

10

How does Black continue?

How does Black defend?

1 90 Various Endings

11

14

Find the right line for Black.

Demonstrate a plan to realise

the extra pawn.
15

12

What is White 's winning plan?

13

Point o u t the correct plan of

defence.
16

Whit to move. What should be

the result of the game?

Can White win by I g3 hxg3 2 h4,

creating a passed pawn?

Solutions to Exercises
PAWN ENDINGS
1
1 Wf2 rs 2 @f3 '\ties 3 g4
hxg4+ 4 'it>xg4 'it>e4 S hS fS+ 6
Wh3 ! f4 7 h6 f3 s h7 f2 9 g2 1-0,

d l =iV 7 c8+ and a queen ending

was reached with an f-pawn for
Janvarev-Schcherbakov,
White,
Moscow, 1 994.

Botvinnik.

White has a winning position, but

in the game he lost.
1 'be3?? (1 a4 'iii>d 6 2 aS 'it>xdS 3
a6! +-) l .. .d6 2 We4 c4 3 a4 c3 4
'it>d3 <t>xd5, 0- 1 , Guliev-Tukmakov,
Nikolaev, 1 993.
3

(a) 1 gxf5? gxf5 2 'it>e2 'it>e7 3

d3 h5 4 xd4 h4, th- th, Yermo
Iinsky-Ivanov, USA (ch) 1 996
(b) 1 gS! 'bf7 2 'it>e2 'it>g7
(2 ... e7 3 'it>d3 c,t>e6 4 'itxd4 cJi>d7 5
c4 'it>d8 6 e6 e7 7 Wd3 ! 'it>xe6 8
'it>d4 +") 3 e6! (3 'it>e2 h5 !==) 3 . Wf8

(a) l ... g6? 2 'it>e4 'it>d6 3 h5 g5 4

'it>f5 b5 5 cxb5 c,t>d5 6 b6 'it>c6 7
'it>e4 cJi>xb6 8 'it>d3, 1 -0, Dreev
Anastasian, Tbilisi, 1 989;
(b) 1. ..f6! 2 'it>e4 (2 h5 g6 3 a4
g5=) 2 'it>e6 3 hS 'it>f6 4 'it>f4 g6 S
..

'ite4 gxhS=.

6
1 g4! hxg4 2 hS c,t>e6 3 a3 ! g3 4
e2 ! (4 'it>f3? f5 !) 4 ... f7 S @f3
'it>g7 6 'it>xg3 'bh7 7 'it>g4 'it>h6 8
@f5 'it>g7 (8 . . .'it>xh5 9 xf6+-) 9 eS!
fxeS 10 'it>xgS +-, Evreinov.
7

..

4 'it>e2 @es s c,t>d3 e7 6 c4

'it>xe6 7 'it>xd4 +-.
4

1 . ..hxg4?-Black thought that he

could achieve a draw as he pleased,
but correct was 1 fxg4 2 fxg4 hxg4
3 hS @es 4 g3 But suddenly
there followed 2 f4 ! ! 'it>c4 3 h5 d4 4
h6 d3 4 h7 g3+ 5 'it>xg3 d2 6 h8=ii
..

White resigned (?) because he did

not see 1 fS 'it>xc4 ( 1 ...'it>b4 2 c5)

2 h4! 'itb4 3 'it>g6 xa4 4 'it>xhS

b3 S Wg6 , with a draw. The f6

pawn helps White.

(a) l . . .'it> e5? 2 h3 ! ! (2 h4? f5 3

g3 g6=) 2 . . .f5 3 h4 'ite5 4

'it>g4 'it>e4 5 h5 f5+ 6 'ith3 ! e3 7

I 92 Solutions

h6, 1 -0, Padevsky-Latinov, Sofia,

1973;
(b) 1 ...g5! 2 e4 g4 (2 ... f5+?
3 f3 !+-) 3 'it>e5 f5 4 h3+ g5 5 h4

'it>g4 6 h5 f4=.

h5 g4 6 h6 g3 7 'it>e3 d4+ 9 @f3

d3=) 4 ...d6 5 g5 fxg5 6 fxg5 @e6!
7 g6 @f6 8 @d5 @g7=. 2 ... @d6 3

h5 'it>e6! 4 'it>c5 f5 5 g5 hxg5 5 h6

@fi! 7 fxg5 f4=.

14

(a) l .. .g4? 2 We3 ! f4+ (2 ... gxh3 3

@f2) 3 'ii> e2! gxh3 4 f2 f5 5
@g l @e5 6 @h2 @d4 7 xh3 cJi>c4
8 xh4 @xb4 9 'it>g4, 1 -0,
Rufenacht-Orseth, corr, 1 996;
(b) 1 ...@f6! 2 @e3 'ii> e5 3 f4+

Yes, but not by 1 axb4 a3 2 @c3

cxb4+ 3 b3 d5, and White re
signed, Sofia Polgar-Smyslov, Lon
don, 1 996, but 1 c4! !, creating an
impregnable fortress.

king hurries to f8 .

1 h6! ( 1 cJi>c6 f5 2 @d5 'it>f4 3 a4

'it>g3=, Filipescu-Citron, Romania,
1955) 1...'it>xh6 2 c6 @g5 3 'it>d5
f5 4 a4 f4 5 'it>e4 +-; 1. .. gxh6 2 a4

gxf4 4 @f3 @d5 5 @xf4 @c4 6

'it>xf5 @xb4 7 @g4 @c5, and the
10
1 c4! ! cxd4 ( l .. .dxc4 2 dxc5 bxc5
3 'ii> e2 +-) 2 cxd5 bS 3 'ii>e 2 b4 4
a4! e5 5 h5 e4 6 d6 b3 7 d2,
1-0, Degraeve-Hansen, Germany,

1 998.

11

(a) 1 'it>f4? g6=, Ilyin-Zhenevsky

Botvinnik, Leningrad, 1 93 8 ;
(b) 1 f3! g6 ( 1 ...@e7 2 \t>f4 @e6
3 g3) 2 hxg6+ xg6 3 f4 h5 4 g3
+12

Yes, 1. .. cS! ! 2 'it>g2 cxd4 3 exd4

exd4 4 'it>f2 @e7 5 @e2 @e6=,

Kozakov-Asparuhov, Pernik, 1972.

13

(a) l...'it>e6? 2 'it>c5 f5 3 h3 fxg4 4

hxg4 d4 5 'it>xd4 d6 6 f5, 1 -0,
Nimzowitsch-Chigorin,
Carlsbad,
1 907;
(b) 1 ...'it>c6 2 h4 2 h3 'it>d6 3 h4
@c6 4 h5 (4 g5 fxg5 5 fxg5 hxg5 5

15

f5 3 a5 f4 4 a6 f3 5 a7 f2 6 a8='YW
fl='i!V 7 g8+! +-.
16

(a) 1 @e3 ? d4+ ! 2 cxd4 @d5 3 b5

'it>c4, 0- 1 , Havsky-Yuferov;
(b) 1 @d2 d4 2 c4 d3 3 c5 @d5 4
a5=.

17
1 'it>e6! ( 1 @g6 g3 2 'it>f5 @f3=)
1. ..@h3 2 'it>f5 ! @h4 (2 . . .g6+ 3
g5 ! +-) 3 c4 ! g5 4 c5 g4 5 'it>f4 g3
6 'it>f3 @h3 7 c6 g2 8 c7 +-, Hoch,

1 972.

18

(a) l ...e4 2 fxe4 g4 3 xf7! g3 4

f6 g2 4 @e8 gl='if 5 fl +-;
(b) 1 ...g4 2 fxg4 e4 3 g5 e3 4
xfl e2 5 g6 +-;
c) 1 . ..'it>g8 2 f6 g4! ! 3 fxg4 e4 4

g5 e3 5 g6 e2 6 gxfi 'it>h7 7 f8='if

el=if+ 8 'it>ti e6+!! stalemate.

Solutions 1 93

19
1. ..c,t>e2! ( 1 ...'ii>f2 2 'ii>e5 +-) 2 h6
@d3=.
20

(a) 1 @g5? a4 2 h6 a3 3 g7 a2 4
h7 a l ='fl' 5 g8='1'W 'iYc l +, 0-1 , Wells
Hector, Oxford, 1 999;
(b) 1 'ii>f4 ! !=.

1 -0, Anastasian-Khalifman, Erevan,

1 996.
24

1 c4? led to defeat. White should

give preference to 1 b4! cxb4 2
cxb4 @f'l 3 b5 'ii>e7 4 g3 f'l 5
@d6 @f6 6 'it>c7 'ii>e7 7 c,t>b7 'it>d7 8
@xa7 'ii>c7 9 g4! e5 10 gxh5 gxh5
1 1 'ii>a8 e4 12 a7, and a draw.

(a) 1 ...'ii>h5? 2 b4 'ii>xh4 3 b5 g5 4

a5 g4 5 b6 axb6 6 axb6 g3 7 @f3 !
+-
1. ..c,t>g7!=,
UhlmannCb)
Robatsch, Marienbad, 1965.

The game ended 1 ...'ii>f7 2 @d6

@f6 3 c6 'ii>e7 ! 4 b7 'ii>d7 5
xa7 'ii>c7 6 'ii>a8 e5 7 g4 hxg4 8 h5
gxh5 9 b4 cxb4 1 0 c5 b3 1 1 a7 b2
12 cxb6 xb6 13 b8 b l ='l'W 1 4
a8='1'W 'l'Wh7! , 0- 1 , Mortensen-Plueg,
Hamburg, 1 997.

22

25

l ...axb5? (after the correct 1 ...a5 !

Black holds. On h3-h4 he replies
... h6-h5, while on g2-g4 forced is
... g6-g5 !) 2 axb5 @d7 3 @f3 e7 4
h4 h5 5 'ii>e4 'it>d7 6 d4 'ii>c7 7
'ii>c4 'ii>d7 8 b4 'it>c7 (matters are
not changed by 8 ...c8 9 c,t>a5 (9
c6) 9 ... c7 1 0 g3!) 9 c,t>a5 @c8 1 0
'iitb6 e 5 1 1 fxe5 fxe5 1 2 'ii>a5 ! 1 -0,
Speelman-Saltaev, Las Palmas,
1 998.

Here White resigned, calculating a

variation with a necessary exchange
on f6 However, correct is not to
take the pawn.

21

23

l ...b5? (1...h5 ! 2 f3 [2 h3 'ii>b5 3

g4 hxg4 4 hxg4 'it>c6 5 g5 fxg5 6
fxg5 lt>d6=] 2 ...@b5 3 g4 hxg4! 4
fxg4 lt>c6 5 h4 'it>d6=) 2 g4 'ii>c6 3
g5 e5+ 4 fxe5 fxg5 5 f3 h5 6 'ii>c4,

1 f5 'ii>xc4 2 h4! @b4 3 'ii>g6

'ii>xa4 4 'it>xh5 'it>b3 5 'ii>g6, and the

pawns queen at the same time, while

the f6 pawn 'helps' White,
Klovan-Elistratov, Moscow, 1 963.
26

@fl

(a) l ...'ii>xb4 2 a6 'ii>c3 3 'iite l !

'ii>d3 4 a7 f2+ 5
e2+ 6 'ii>xf2
'ii>d2 7 a8=11V e l =+ 8 'ii>g2 xh4=,
Papendieck-Geiger, Austria, 1998.
(b) 1 ...'ii>c4! 2 a6 'ii>d3 3 @el e2 4
a7 c,t>e3 - +

J 94 Solutions

TRANSPOSITION TO A PAWN ENDING

No,

1 ... llle5??

I ... llld6 2 g5 (2 h4 llle4) 2 ... fxg5

3 lllxg5 c4+ 4 bxc4 lllxc4 5 lllf3
llle3 6 g4 lllg2=
2 lllxeS fxe5

2 ... @xe5 3 h4 \t>f4 4 h5 @g5 5

'it>c2 'it>h6 6 'it>b2 'it>g5 7 g3 @h6 8
a4 bxa3 9 'it>xa3 A 'it>a3-b2-c2-d3
b3-b4 +3 gS e4+ 4 @e2! 'it>e5

4 ... d3 5 'it>dl !+-

s h4 'it>f5 6 g4+ 'it>f4 7 g6 d3+ 8

'it>dl e3 9 g7, 1-0, Ivanov

Sagalchik, Kramatorsk, 1 989.

2

No. 1 .l:.xfi+? @xii 2 f3 @f6 3

@e4 gS! (3 ... @g5 4 xd4 g4 5
'it>e5 !=) 4 bS d3 S 'it>xd3 @es 6 'it>c4

g4 7 'i!i cs 'it>e4 8 'it>c6 'it>d4! !, 0-1,

Renet-Olafsson, Reykjavik, 1 993.

Instead I :c4 .l:Id7 2 @f2 d3 3
@el d2+ 4 'it>d l @f6 5 .l:Ic5 .l:Id3 6
.!la5 draws.
3
1. l:l'.d5 2 .!la2 g6 3 Zlf2 fS 4
J:id2 l:l'.xd2 S 'it>xd2 'it>g5 6 'it>e2 'it>g4
7 @f2 'it>f4! 0-1, Balashov

After 1...b5? 2 J:ixb5 J::!xb5 3

axbS+ 'it>xb5 4 e4 'it>c6 5 e5! fxe5 6
g5 hxg5 7 f6!, Black had to resign,

Averbakh-Bebchuk, USSR, 1 963.

6

No, it is not worth it. I i.g2? (1

i.xg6 'it>xg6 2 b5! @rs 3 'it>d4 +-)

l ...e5 2 i.f3 b5=, Adams-Xie Jun,

France, 1 994.
7

No. 1 llld 3?? ( I 'it>f4=) 1.. .. llle5

2 lllx eS 'it>xe5 3 @gs b5 ! - +,

Bronstein-Timman, Tallinn, 1 973.

8

(a) 1. ..i.c5 !, transposing to a

drawn rook ending;
(b) I ... 'it>f8? 2 J::rxd6 J::rxd6 3 i.c5
'it>e7 4 'it>e3 'it>e6 5 i.xd6 'it>xd6 6
d4 'it>e6 7 'it>c5 h5 8 g5 @f5 9
'it>xd5 'it>g4 I O 'it>e5 'it>xh4 1 1 f5,
1 -0, Ricardi-Sorin, Buenos Aires,
1995.

Tiviakov, St.Petersburg, 1 993.

4

1 cS! .!lxc5 ( I ...bxc5 2 d3 A

:Ixa5+-) 2 J::rxc5 bxcS 3 'it>d3 eS 4

@c4 fS 5 'it>xc5 h5 6 b4 axb4 7

'it>xb4 f4 8 a5 e4 9 @c3 e3 10 'it>d3,
1-0,
Kuzmin-Petrosian, USSR,

1 97 1 .

No, he cannot.
1 ...l:l'.gxd7?

1 ...f4!

2 l:l'.xd7+ .!lxd7 3 .!lxd7+ xd7 4

f4! g4

4 ...gxf4 5 'it>f3 'it>e6 6 'it>xf4 \t>f6

7 g3 +-

5 g3 gxh3 6 gxh4 @e6 7 'it>g3

'it>f6 8 hS! 'it>g7 9 xh3 h7 1 0
'it>h4 @h6 1 1 b3 ! , 1-0, Estrin

Gusev, Moscow, 1 963 .

Solutions 1 95

10

lS

(a) 1 .. .ll:\g5? 2 ..tf5 'it>f6 3 'it'e3

ll:\e6 4 i.xe6 'it'xe6 5 it>e4 +-,
Petrosian-Larsen, San Antonio,
1 972;
(b) 1 ...ctJeS 2 'it'e3 ll:\xg6=.

1 'it>f2? (1 it>g2 ! J:ic2+ 2 .!lf2

.!lxf2+ 3 'it'xf2 i.gS 4 'it'g3=)
l . . .!lc2+ 2 .!le2 i.c5+ 3 i.e3 i.xe3+

11

4 'it>xe3 J:Ixf3+ 5 'it'xf3 J:Ixe2 6

it>xe2 it>g6 7 'it>d3 it>f5 8 'it>d4 a5,
0- 1 , Almasi-Portisch, Budapest,
1 994.

1...gS? ( 1 .....te6 ! ?) 2 ll:\xdS 'it>xdS

3 g4! (a typical breakthrough)
3 ... gxh4 4 gxhS 'it'e6 s it>g2 'it'rs 6
f4 !, Black resigned, Deak-Horvath,

1 .. ...tf4! 2 ..te3 'it'gS 3 l:!.xf4 J:ixf4

4 i.xf4+ 'it>xf4 S b4 b6
- +,

12

17

Hungary, 1 994.

There followed l .....tc4+?? 2

ll:\xc4+ 'ifxc4 3 'iVd3+, 1 -0, Tal
Suetin, Sochi, 1 973.
Correct was 1 ...'iVf4!=.

16

Szelaig-Pinski, Poland, 1 997.

No, it is not possible.

1 ti:Jf6+? ll:\xf6 2 'iVxf8+ 'it>xf8 3

exf6 gS ! 4 f4 'it>g8 S fxgS 'it'h7 - +,

Adorjan-Chemin, Budapest, 1 995.

13

18

1 ...l:1dS? ( 1 .. .'it>f6 2 .l:Ixb6+ it>e5 3

l:!.c6 el 4 b4 lixe3 5 b5 .!ld3=) 2
.l:IxdS 'it>xdS 3 'it>a3! (going round
the flank) 3 ... 'it'cS 4 'it>a4 g6 S h4 hS
(5 ...g5 6 hxg5 hxg5 7 g4 +-) 6 g3
'it>c6 7 b4! (7 it>b4? b5 8 'it>a5? 'it>c5
9 it>a6 'it>b4 -+) 7... it>c7 8 'it>bS

(a) l ....l:i'.h5? 2 'it'b4 .!lxh2 3 c5,

Cruz-Seirawan, Moscow, 1 994;
(b) 1. ...l:IaS+! 2 'it>b3 .l:Ixa2 3

'it>b7 9 'it>c4 'it>a6 10 it>c3 ! it>b7 1 1

'it>d4 'it>c6 1 2 it>xe4 it>bS 1 3 it>d4
it>xb4 14 e4, 1-0, Ryumin -

Ilyin-Zhenevsky, USSR (ch), 1 93 1 .

14

1 ....l:Ixg2+? (1... hxg2 - +) 2 l:1xg2?

hxg2 3 'it'xg2 a5, 0- 1 , Valvo-Levitt,
Chicago, 1992;
Necessary was 2 'it'xh3 ! l:Ixg4 3
'it>xg4 a5 4 'it>f5 ! 'it>c6 5 e5 a4 6 e6
a3 7 'it>g6!=.

'it>xa2 it>d6 4 it>b3 'it'cS S 'it'c3 eS 6

it>d3 rs 7 it>c3 e4 (7 ...h5 8 'it'b3 f4 9
it>c3 f3 1 0 'it>d3 'it>b4 -+) 8 'it>b3 hS
9 it>c3 f4! 10 gxf4 e3! 1 1 fxe3 h4
1 2 rs it>d6! ! ( 1 2 ...g3 ? 1 3 hxg3 hxg3
14 f6 it>d6 1 5 c5+!) 13 'it'b4 it>eS!
14 cS 'it>xfS lS c6 'it'e6 16 'it'cS g3
17 hxg3 h3 ! 1 8 it>b6 h2 1 9 c7
it>d7!, and the pawn queens with

check!

19

(a) l ...ctJxe4? 2 fxe4 it>d6 3 e5+!

it>c5 4 'it'c7=, Mestel-Chekhov,
Tjentiste, 1 975;

(b) 1 ...it>d6 2 it>xa7 it>cS 3 i.c6

ctJfl - + .

J 96 Solutions

20

(a) l ...i.c4? 2 i.xc4 dxc4 3 e4

@b5 4 f4 @c5 5 g4 @b5 6 e5! fxe5
7 f5 ! ! 'it>c5 8 g5 +-, Heinig-Liebert,
DDR, 1979;
(b) 1 ...d4+! 2 exd4 d6=.
21
1 i.xb6! i.xb6 2 axb6 axb6 3
@f2 @e6 4 @e3 'it>dS S c4+ @cs 6

g4 bS 7 rs gxfS 8 gxfS @d6 9

Solozhenkin-Anchesi,
@xe4,

Reggio Emilia, 1 998.

22

l .. .@e5 ? 2 i.xd6+ ! 'it>xd6 3 'itf4

cJi>d5 4 b5 ! axb5 5 a6 'it>c6 6 @xe4

b4 7 f4, 1 -0, Larsen-Uhlmann, Las

Palmas, 197 1 . 1...lZJbS!=.

Solutions 197

ROOK ENDINGS

In the game followed l .. .l:ra l ? 2

J::rd 5+ 'it>e6 (On 2 ... 'it>c6 follows 3
g5+) 3 d4! J:ifl 4 1::r e4+ 'it>d7 5
xa4, 1 -0, Gurevich-Rechlis, Israel,
1 989.
He should play 1. ..J::rf4 ! 2 g7

l:!.g4+ 3 'it>f6 l:1f4+ 4 'it>g6 'it>e6 s

l::ta6+ 'it>e7 6 .l:!a7+ 'it>e6 7 'it>g7
J::rg4+ 8 f8 d6! 9 'it>e8 1:!.e4+ 1 0
d8 J::rf4 1 1 l::ta6+ 'it>cS 1 2 rj;;e7
llxf7+! 13 xf7 'it>b4 with a draw.
2

(a) 1 1:!.d6? e3 2 J:ie6+ '\t>f2 3

J:id6 'it>e2 4 lie6+ 'it>d l 5 g5 d2 6
\t>f4 I:!'.a2, 0- 1 , Topalov-Kasparov,
Geneva, 1995;
(b) 1 l:te8+! f3 2 .l:If8+ We2 3
1:!.e8+ 'it>dl 4 :i::!. a8 d2 S a7=.
3

(a) l . ..'it>f4? 2 h4 l::t a2+ 3 g l

l:1c2 4 h5 1:!.c5 5 h6 h5 6 h7 f5 7
a7, I/i-I/i, Leko-Akopian, Ubeda,
1 997;
(b) 1 ...a4 2 h4 l:1a2+ 3 'it>g3 a3 4
hS J:ial S .:.a7 a2 6 'it>g2 'it>gS, and
the f-pawn advances.
4

(a) 1...e7 2 l:te6+ 'it>f7! 3 :xd6

.l:td4 4 'it>g3 'it>e7 S l::te6+ 'it>f7 6 :!:!.es

Wf6 7 l:thS 'it>g6=;

(b) l .. .'it>e7 2 lie6+ 'it>d7? 3 f6

e7 4 1:!.f5 l!tc3 5 h4 llb3 6 h5 .l:Ia3
7 h6 l:ta8 8 h7 l:th8 9 h5 'it>f6 1 0
'it>g3 'it>g7 1 1 g4 g6 1 2 1:!.hl 'it>f6
1 3 'it>f4 'it>g6 14 'it>e4 'it>f6 1 5 d4
'it>g7 ( 1 5 ... 'it>e7 1 6 'it>c4 d7 1 7
a l ! +-) 1 6 'it>c4 .l:tc8+ 1 7 'it>b5 h8

18 b6 1:!.c5 19 1:!.h5 llc8 20 1:!.h6

+-, Lesiege-Ivanov, Bermuda, 1 995.
s

(a) 1 Sb6?? laa4 2 Wh3 l::txg4 3

1:!.xf6 J:ih4+ ! ! , 0- 1 , Gulko-Gurevich,
Parsipanny, 1 996;
(b) 1 J::rc4 'it>h4 2 1:!.c6=.
6
1 'it>c6! ( 1 c6? e5 2 fxe5 fxe5 3
Wb6 e4 4 .lld7+ f6! 5 l:rd8 e3 6
bi.e8 l:rb 1 + 7 rj;;a 7 .l:Ial + 8 b7 J:ib 1
9 'it>c8 11e l !=, Short-Topalov, Nov
gorod, 1996) I. . eS 2 fxeS fxeS 3
dS f6 (3 ... e4 4 Itd4 +-) 4 c6 e4
S 11f2+ e7 6 Ite2 .:.ct1 + 7 'it>xe4
'it>d6 8 J::rc 2 +-.
.

Correct was 1. ..llc8! 2 c3 f6 +.

In the game followed 1 ...:i::!.b 8?. 2
.llb3 .:.xb3 After 2 ... 1:!.c8 3 c3 d4 4
ltb7+ 'it>f6 Black holds the rook
ending without problems. 3 axb3

f6 4 gl 'it>eS S 'it>f2 e4 6 We2

'it>d4 7 'it>d2 aS (7 ...e4=) 8 e2 eS
9 d2 e4? (9 ... 'it>c5=) 10 c3+ @cs
1 1 'it>e3 'it>c6 12 'it>d4 'it>bS 13 h3 h6
14 h4 c6 lS c4 1-0, Kupreichik
_

Zheliandinov, Gome!, 1 968.

8

1 f6 'it>g8 ( l ...'it>e8 2 e6!+-) 2

1:!.g4+ 'it>f8 3 l::ta4 l:i'.d8? (3 ... 'it>g8 4

l::t a 8+ 'it>h7 5 f8 lld l 6 l!txf7+ 'it>g8

7 :a7 fl+ 8 rj;; e6 l::t e l ! , with a
theoretical draw) 4 l::t a7? (4 bi.h4!
'it>g8 5 e7 bi.d5 6 l:tg4+ h7 7
'it>xf7 +-) 4 ... l:tb8? (4 . .. g8 5 .llxf7
1:!.dl 6 .l:!a7, again draws) S l::txf7+

'it>g8 6 :i::!.g7+ f8 7 e6 l::tb6 8 l:!:a7

J 98 Solutions

l:tb8 9 lih7 'it>g8 10 .l:tg7+ 1-0,

Kirov-Kosic, lnformator 59.

9

1 lib7! ! ( 1 'it>g6 :g3+ D. 2 ... J:!.b3)

i..J:Ixh5 2 'it>g6 'it>rs 3 h6 l:tb 1 4
l:!.b8+
1-0,
Levenfish-Lisitsin,

1!z-1!z,

Karpov-Yudasin, Madrid,
1 992;
(b) 1 gxh5 1ih3 2 .l:!xa5 f5!
(2 ... I:!'.xh4 3 'it>d3 :Ixh5 4 cJi>c4 f5 5
'it>c5 fxe4 6 'it>c6 +-) 3 exf5 1ixh4 4
'it>e3 'it>e7! (, Yudasin) 5 l:ta6!
:xh5 6 'it>e4 +-.

Moscow, 1 936.

10

14

1 .. .l:td2? (1 ...l:!b2!=) 2 b4! l:td4

(2 ... I:!'.b2 3 l:te4 +-) 3 l:tb3 'it>e7 4 b5
'it>d8 5 b6 'it>c8 6 c3+ 'it>b7 7 c7
'it>xb6 8 .l:!xf7 1id3+ 9 'it>f2 l:!d4 1 0
'it> f3 l:td3+ 1 1 'it>e4 l:!g3 1 2 f6
lixg4+ 1 3 'it>f5 .Ugl , 1 -0, Timman
Van Wely, Amsterdam, 1 995.

1 f5! !-a typical breakthrough

-1. .. exf5 2 e6! fxe6 3 'it>xg6 D.
h5=.
The game continued 1 'it>f6? 'it>b5
2 l:!.al a4 3 f5 exf5 4 e6 fxe6 5
'it>xg6 f4 6 h5 f3 7 h6 e5! 8 l:tel (8
@f5 !Ih7 9 l:!.hl f2 -+) 8 ... a3 9
l:txe5+ 'it>c4 10 l:tel a2 1 1 h7 J:!.a8
(1 l .. .J:ixh7? 1 2 'it>xh7 'it>b3 1 3 J:!.fl =
or 1 1 ...2 1 2 :!:!'.fl Wd3 1 3 :!:!'.al !) 12
'it>g7 (12 :!.al f2 13 J:Ixa2 1ixa2 14
h8='fl' .l:ta6! - +) 12...f2 13 Ital
'it>b3 14 :!:!'.fl al='ili' 15 l:!xal J:!.xal
16 h8='ili' J:igl+ 0-1, Lasker-Leven

(a) 1 g4? l:td l ! 2 l:!c8 :!.b l 3 I:!'.f8

l:txb4 4 .l:txf7 l:tc4 5 .l:!xh7 b4 6 .llh l
b3 7 .l:.fl b 2 8 l:tb l .l:!b4 9 f5 cj;c7,
1!z-1h, Kasparov-Short, Novgorod,
1 994;
(b) 1 l:tc8! f6 2 gxf6 :xf6 3 g4

11

16

fish, Moscow, 1 935.

1 'it>b6? ( 1 l:te3! l:!.e8 2 'it>c6 +-)

1 .. .e4 2 b5 J:ie8 3 Wa7 e3 4 l:th l e2
5 .l:tel Wxh6 6 b6 Wg5 7 b7 'it>g4 !,
I/z-1h, Bologan-Kramnik, Germany,
1 994.

15

J:ic6 4 :hs+ 'it>a6 5 f5 h5 9 l:!g8! +-.

(a) 1 l:td5+? f5 ! ! 2 J:Ixf5+ 'it>h6

- +

(h) 1 I:!'.dl a2 2 l:tal :a3+ 3 'it>g2 !

l:ta4 (3 ... 'it>xg4 4 h6=) 4 'it>f3=,

Permiakov-Petkevich, Riga, 1 988.

12

17

1 h7? (1 l:!.e8 l:!. f7 2 'it>c3 'it> f2 3

'it>d3 l:.d7+ 4 'it>c2 e2 5 l:tf8+ 'it>el 6
0- 1 ,
l:tf6=; 1 l:tf8 ! ?) 1 ...1if7 !

(a) 1 'it>d4? b4 2 'it>d3 'it>b5 3 'it>d2

b3 4 :!:!'.c l l:.c4 ! -+, PiampuuLancava, Leningrad, 1 992;
(b) 1 :!:i.a2 ! b4 2 :c2=.

13

18

(a) 1 g5? fxg5 2 hxg5 h4 3 d6

l:ta2+! 4 @f3 h3 5 'it>g3 h2 6 a8+
'it>d7 7 .!:l.h8 'it>xd6 8 l:txh2 a4 ...

(a) 1 d6? 'it>g7 ! 2 b5 @f6 3 d7+

'it>e7 4 I:!'.d6 'it>d8 5 I:!'.f6 'it>xd7 6
J:ixf7 'it>.e6 7 I:!'.xh7 :b3 8 b7 l:tb2 !

'

Lautier-Y ermolinsky,
1 993.

Tilburg,

Solutions 1 99

9 b6 e4 1 0 li:tb8 e3 1 1 @fl f6 1 2
h4 Wg7 1 3 g4 'it>h7 1 4 h 5 gxh5 1 5
gxh5 Wg7 1 6 .l:Ib7+ 'it>h8 1 7 h 6 'iitg 8
18 .l:Ig7+ 'it>h8 19 b7 b l+! 20 'it>e2
.l:Ib2! 2 1 Wxe3 .!lb3+ 22 Wd4
1Ib4+=, Yudasin-Kramnik, Candi
dates (m), 1 994;
(b) 1 .!lc8! 'it>g7 2 b6 .l:!b3 3 d6 +-.
19

(a) 1 ....!lxh4? 2 .l:Ixg5 Wc3 3

.!id5 ! ! Wd3 4 'it>cl .l:Ihl + 5 'it>b2 l:!el
(5 ...We3 6 @c2 .l:Ih2+ 7 Wd l ! d3 8
.lle5=) 6 .l:Id8, Draw, Dreev
Beliavsky, Odessa, 1 989;
(b) 1.. .!lgl+ 2 e2 (2 @d2 l:!g2+
3 We 1 g4 4 fl .l:!h2 ! 5 l:!xg4 'it>c3
6 'it>g 1 l:!a2 D. d4-d3 - +) 2 d3+ 3
.

Wd2 .!lg2+ 4 Wdl g4 s hS Wc3 6

J::!c8+ 'it>d4 7 h6 .l:Ih2 We3 9 l:!e8+
'it>f3 10 lih8 'it>g3 1 1 h7 .l:!h4 ! , and

there is no defence against the

march of the g4 pawn.
20
1 f6! z:!'.c6+ 2 @es z:!'.c8 (2 ... .l:!c5
3 d6 l:!c8 4 l:!el +-) 3 .l:!g6! ! Wh7 4
.l:Ic6! l:Ia8 S Wf6 t:>. lie6, Cvitan.
21
1 g6+ 'it>f6 2 l:! f8 'it>eS 3 f6! ! (3
.!17? .lla l =) 3 ....l:!xf6 4 l:!ti! 'it>e6
(4 ... l:!f5+ 5 'it>g4 l:If6 6 'it>g5 +-) s
l:txg7 :n 6 1Ia7 +-, Bayer-Polasek,

Luxemburg, 1 986.

22
1...z:!'.aS! ( 1 . ..Wg7? 2 a5 'iitf6 3

'it>f4 We6 4 @e4! h5 5 f4! @f6 6

Wg3 @g5 7 .!la4+-, Holmov-Hasin,
Minsk, 1 962) 2 f4 @g7 3 We4 hS
4 'it>d4 h4 S 'it>c4 h3 6 b4 l:!'.hS 7
.:!.al h2 8 .l:Ihl Wf6 9 aS @rs 10 a6
.l:Ih7!=.

23
1 'it>c4 l:ta4+ 2 Wb3 .l:Ia3+ 3 'it>c2
.l:Ic3 4 Wb2! +-, Salwe .
24

(a) l ...'it>e3? 2 Wg3 l:!a l 3 .l:Ie8! (3

a7? l:tg l+ 4 Wh2 'it>f2=) 3 ... 'it>d4 4
'it>f4 l:!xa6 5 l:!.d8+! @c5 6 @g5, 1 -0,
Ulibin-Nevednichi, USSR, 1 986;
(b) 1. . xf3 ! ! 2 a7 .l:!a3 ! !=,
Gelfand.
.

2S

(a) l .. .h3? 2 g3 +-, Bykova

Rubzova, m, 1 958;
(b) 1 . ..'it>hl ! 2 Wg4 h3 3 Wxh3
.l:Ixg7! , stalemate.
26

(a) 1 a8='ii' .llx a8 2 l:l'.xa8 Wf5 3

.l:Ih8 g4 4 c5 h3 5 'it>d4 g3 6
We3 Wg2!=, Dammes-Sosonko,
Leningrad, 1 963;

(b) 1 'it>b7! J::!ti+ 2 @a6! l:!f6+

(2 .. .S.f8 3 .l:.b5 ! h3 4 l:Ia3 ! +-) 3

@bS li:tf8 4 as=iw .l:Ixa8 s .!lxa8
'it>fS 6 c4 h3 7 d3 +-.
.

27

(a) 1 'it>f4? 'it>c4 2 @e5 'it>d3 ! 3

Wf6 'it>e4 4 Wg7 l:!.a8 5 h7 Wf5 6 g6
Wg5=, Gutman-Tseitlin, Riga, 1 976;
(b) 1 Wg4 Wc4 2 'it>hS 'it>ds 3 g6
+-.
28

(a) l ...Wg2? 2 li:tf6 f3 3 g6 .!la5 4

:i.7, r;H/i, Morovic-Agdestein,
Havana, 1998;
(b) 1 ....!lal+ 2 'it>d2 l:!.gl 3 g6
l:!g3 ! ! D. 4 ...g2 - + .

200 Solutions

29

(a) l ...f5? 2 h4 ! a6 3 e3 e4+

(3 ... a5 4 e4 f4 5 gxf4 exf4 6 e5) 4
fxe4 fxe4+ 5 c;t;>xe4 :Ixc4+ 6 l:txc4
c;t;>xc4 7 @e5=, Wirthensohn
Mikhalchishin, Lenk, 1998;
(b) 1. ..h4 ! 2 gxh4 f5 3 e3 a6 4 h5
(4 b3 J:Ixc4 5 xb6 e4+! ) 4 ... gxh5
5 h4 a5 - + .
30

(a) 1 h6? @d3 2 @3 l:th3+ 3 @g4

l:rhl 4 @f5 e2 5 f6 l:th3 ! 6 c;t;>f7
l:if3+ 7 @g6 e3=, Piket-Sokolov,
Dortmund, 1995;
(b) 1 g4! @d3 2 g6 e2 3 @g5
el='Yi' 4 llxel J:Ixel 5 g7 +-.
31

(a) 1 ...b3? 2 J:ig7+! Wh8 3 l:!b7

l:rxb4 4 @g6, 1 -0, Salov-Yudasin,
St.Petersburg, 1 997;
(b) 1...I:!'.f3+! 2 c;t>xe5 I:!'.b3 3 c;t;>d6

1ixb4 4 e5 l:td4+ 5 @e7 l:ta4 ! ! 6

J:ib7 @g7 7 e6 lia8 8 'iitd 6 c;t;>f8 9
@d7 @g7=
.

32

Only 1 a3 ! ! makes a draw.

(a) 1...gxh4 2 l:!c3+ c;t>d5 3 J:ic2

e4 4 @h5 c;t;>f3 5 llc3+ c;t;>e2 6

l:!c2+ d3 7 h2=;
(b) 1...g4 ( L.c;t;>d5 2 lia5+ @e4 3
a4+) 2 J:Ic3+ (2 c;t;>g5? c5 ! )
2 ... c;t;>d5 3 :gc2 c;t;>e4 4 @g5 W f3 5
J:ic3+ @g2 6 c;t;>xg4 l:Igl 7 l:.a3
al='iV 8 xal l:.xal 9 h5=.
33

(a) 1 ....ll a l ? 2 I!c8 xa7 3 l:!c7+!

J:Ixc7 4 xc7 c;t;>h7 5 Wd7, 1 -0,
Benko-Gereben, Budapest, 1 95 1 ;

(b) 1 .. .dl+! 2 e6 :!al 3 l:id8

J:Ixa7 4 l:td7+ l:ixd7 5 @xd7 @h7=.
34

(a) 1 xa4? l:!e3+ 2 c;t;>d8 f5 ! 3

gxf5 'it>f6 4 l:ta6 @xf5 5 @c7 g4 6
d7 l:te7 7 @d6 xd7+ 8 Wxd7 g3=,
Olafsson-Tal, Portoroz, 1 958;
(b) 1 d7 J:Ie3+ 2 @d8 a3 3 I:!'.a8 !
+-.
35

(a) 1 @g l ? J:ig3 2 c;t;>f2 c;t;>e6 3 :ga5

J:Ixg4 4 l:!.a4 @f5 5 .!Ib4 l:!e4 6
:gb5+ c;t;>f4 7 l:rb6 f5 8 l:id6 g4 9
J:id8 g3+ 1 0 @fl h3, 0- 1 , Ioseliani
Onischuk, Lucerne, 1 997;
(b) 1 lif5 ! l:tg3 2 d5 J:Ixg4 3 d6+
'it>e6 4 lid5! @d7 5 .llf5 f4 6 .llxf4
gxf4 7 h3 c;t;>xd6 8 Wxh4 @e5 9
c;t;>h5!=.
36

(a) l ...We8? 2 g4 Wd8 3 @g6 @e8

4 c;t;>f6 Wd8 5 g7 l:!b5 6 l:l'.xg5
c;t;>xg5 a5 8 @h6 a4 9 g5 a3 1 0 g6 a2
1 1 g7 a l =ii 1 2 g8 + '\t>d7 1 3
'iVg7+, 1 -0, Lein-Suetin, Bad Wild
bad, 1 997;
(b) 1...l:!a3 ! ! 2 xg5 (2 g4 l:!xh3
3 xg5 g3 4 @f6 @g8 5 g5 :3+
6 Wg6 f8 !=) 2 ... J:Ig3+ 3 @f6 g8!
=

4 l:!xa6 :gxg2=.

37
1 llc2! e6 2 :c3 ! ! a3 3 f7 l:l'.xf7
4 l:te3+ 'iitd5 5 Wxf7 a2 6 I:(a3 1-0,

Konstantinopolsky-Fridman, Lvov,
1 940.
38

(a) 1 ...1:1'.b l ? 2 h5 ligl 3 g5 fxg5

4 f5 ! c;t>f8 5 f6, 1 -0, Zaitsev
Hiibner, Busum, 1 969;

Solutions 201

(b) i. .b4 ! 2 rs .!lbl 3 'it>hs .!lgl

4 'it>g6 .!lxg4+ S 'it>xr6 .!l gl ,
.

(b) 1 .!lh4+! Wxb3 2 aS +-.

Larsen.

39

(a) l ...'it>h4 2 .l:!c8 .l:!h7+ 3 lt>e6

.l:!xd7 4 'it>xd7 g4 5 'it>e6 g3 6 'it>f5
g2 7 'it>f4 +-;
(b) 1 ...'it>g4 2 'it>r6 :i.f8+ 3 c;t>e6
.!id8! 4 l:rdS (4 .l:.c8 .l:!xd7 5 lt>xd7
\t>f5 ! 6 .!lc5+ lt>f4 6 lt>e6 g4 7 .!lc4+
'it>g5 8 lt>e5 g3=) 4 ...r4! S .!lrs+

'it>g4 6 .!in 'it>h3 7 rs g4 s r4 g3

8 lt>f3 lt>h4, 11z_11z, Keres-Mikenas,

Stockholm, 1937.

40
1...d3 ! ! 2 cxd3 (2 'it>xe4 dxc2)
2 ... c4 ! ! 3 bxc4 c2 4 'it>xf4 cl=i#+
S lt>e4 i'd l ! , 0-1, Munios-Salazar,

Dubai, 1986.

41

(a) 1
'it>a3 ! !=)
Wd2 b4
Draw,
1 973.

lt>f2? b5 2 axb5 (2 a5
2 ... axb5 3 'it>e2 lt>a3 ! ! 4
5 Wc2 .!lc8+! 6 lt>d2 J:ih8,
Levy-Peev,
Cienfuegos,

42

(a) 1 ....!lb7? 2 .!lg4! We8 (2 ... b3

l::t a4 .!lb8 4 .!lf4 We8 5 d7+ +-)
.l:!c4 l:tb8 4 .l:!c7 b3 5 .!lxg7 lt>f8
.!if?+ 'it>g8 7 d7 b2 8 :!.fl Wg7
.!lb I ! l:.b6+ 1 0 c;t>e7 .!lb7 1 1 'it>e!
1 -0, Gelfand-Lerner, Norilsk, 1 986
(b) 1...f2 ! 2 .l:!xg7 .l:!e2+ 3 lt>d

b3 4 I:i:b7 b2 S g4 'it>c8 6 J:ib

lt> d7=.

43

It seems that White has succes'

fully completed a difficult defenc(
and Black must agree a draw afte
l .. .a3 2 Wg6 .!lg l + 3 Wh6 Ji(hl
with a perpetual check or l ...fl
.!lxb7 l:i.f8 3 .l:!a7. But, as it turn
out, there is a third possibilit)
There followed the disheartenin
1 .. . .!lb6! !, and, playing on throug
inertia, 2 lt>g6 cS+ 3 'it>rs cxd4
:!::tx dS dxe3, White acknowledge
defeat, Lopyshnoi-Dreev, Maikor
1 998.

202 Solutions

VARIOUS
1

(a) 1 'itb3? ctJb8 ! ! 2 ctJd6+ (2

Cbxc5 ctJc6 3 ctJd3 ctJd4+ =) 2 ... 'ite6
3 ctJe4 Cbc6 4 gxf4 Cbd4+ 5 'itc3
ctJxf3 6 ctJxc5 f5 7 fxe5 Cbxe5 8
ctJd3 Cbxc4 ! 9 'itxc4 g4=,
Sale-Mikhalchishin, Sibenik, 1 990;
(b) 1 'itd2 ! ctJb6 2 @d3 ctJd7 3

(a) I . ...te6? 2 'it>e4! ;:,,, 3 ctJf5 +-,

Vuki6-Vujosevic, Vmjacka Banja,
1 998;
(b) 1....tbl !=.

'itc3 @e6 4 'itb3 ctJb8 5 gxf4! Cbc6

6 Cbxc5 @rs 7 Cbd3 Cbd4+ 8 'it>c3
Cbxf3 9 fxe5 ctJxh2 10 c5 +-.
2

6
1 Cbxa5 'itb2! 2 'itc4! (2 b4
@a3 ! =) 2 ....txa5 3 b4 .tb6 ! ! 4 a5
.tf2 5 a6 (5 b5 'it>a3=) 5 ... .ta7 6
'itb5 'itc3=, Parma-Gligoric, Bled,
1961.
7

(a) 1 . ...txg2! 2 axb5 'ite6 3 Cbxf7

@xf7 4 'itxg2 'ite6 5 @f3 'it>d5 6 b6

@c6 7 b7 'itxb7 8 'it>e4 g4 - + ;
(b) I . ...td3? 2 axb5 e7 3 Cbc8+

'itd7=, Van der Wiel - Larsen,

Brussels, 1 987.
3

1 g4! ! ( 1 ctJf2 'itf6 2 f4 'itg7 3

g4 hx_g4 4 Cbxg4 Cbxc4 ! ) 1...Cbxg4 2
f4 tt:lf6 (2 ...@f6 3 e5+! ! Cbxe5 4
Cbxe5 dxe5 5 'ite4, and the pawn
ending is won) 3 e5+ dxe5 4 tt:lxe5
d6 5 Cbxg6 Cbxd5+ 6 cxd5 c4 7
e4 cxb3 8 ctJe5, 1-0, Razuvaev

Ostoji6, Berlin, 1988.

4

(a) I .te3? 'it>c7 2 'itc5 g2 3 'it>d4

'itc6 4 'ite4 'it>b5 ! 5 @f3 'itc4 6
'itxg2 'it>d5 7 .tf4 'ite4 ! intending
... 'it>f5, f6=, Kudrin-Cebalo, Berne,
1 988;
(b) 1 .ta5 ! ! g2 2 .tb6 'itb8 3 'itc6
c;i;>c8 4 c;i;>d6 +-.

1 ... g5+! 2 hxg5 h4 3 ctJc6 Cbg6+ 4

'ite4 h3 - + Markowski-Gdanski,

Ksiaz, 1 998.

He should choose 1...f6! 2 'if8+

'ith5 3 'ii'g7 h6 - + .

In the game White played l . ..f5?

2 'l'We7 ! d2+ 3 'itg3 el + 4 'ith3
'iWh I + 5 g3 'i!Vg l + 6 'ith3 f4 7
'ii'f8 + 'it>h5 8 'ii'x f4, tfi- 0.,
Slovenia,
Mikhalchishin-Kavcic,
1 997.
9

(a) l ...c2? 2 'ii'xe6 xf2+ 3

'ith3 'l'Wfl + 4 'ith4 'iVb I + 5 'i1Yh3 ! +-,
Polugaevsky-Bronstein,
Tallinn,
1 964;
(b) 1 ...'i!Vf5 !=.
10
1 .. .'ii'x e5? ( l ...fxe5 2 'l'Wxh5
'ii'xa5=) 2 'i!Vg3+! @f7 (2 ...'ii'x g3+ 3

Solutions 203

'it>xg3 +-) 3 'fl'xe5 fxe5 4 i;t>f3 @e7 5

@e4 'it>e6 6 f3 @f6 7 f4 +-.
11

1 ...'it>c8? 2 'fl'g4 and 3 'fl'd l =,

Ehlvest-Topalov, Novgorod, 1 994
1 ... 'it>c7! 2 'fl'e7+ i;t>c6 3 'i!Ve8+
i;t>cs 4 iif8+ 'it>d4 -+.
12
1 h7! xh7+ 2 Wg6 :i.h4 3 rs+
d6 4 ctJh6 hl 5 ct:Jf7+! 'it>c7 6
CZJg5 @d6 7 f7 l:!.h8 8 @g7 lla8 9
f6! (9 f8=11Y+? l:txf8 1 0 'it>xf8
'it>e5=) 9 ... @e5 10 lLih7 I:!'.a7 1 1
c;t>g6, 1-0, Krumpachnik-Maksimen

ko, Ptuj, 1998.

13
1 l:!.xe5 xe5 2 g3, Black re
signed, But he could have saved
himself by 2 ... i;t>g6! 3 j_xe5 @h5

and after 4 ... f4-draw.

14

The game continued 1 i.xf5+?

exf5 2 'it>xf5 @xh6 3 'it>e6 lLib8 4

'it>d6 ctJc6 5 i;t>c7 i.a8! Black's only

chance, which White underesti

mated. Losing is 5 ... ct:Jxe7 6 xb7
i;t>xh5 7 'it>xa6 lbc6 8 'it>xb5 +-. 6

i.c5 xh5 7 'it>b6 'it>g4 8 'it>xa6

Wf3 9 'it>xb5 'it>xe3 10 'it'b6 ct:Jxb4! ,

11z_11z, Chiburdanidze-Maric, Bel

grade, 1 996.
But better is 1 i.c2 ii.as
( I ...i;t>xh6 2 i.dl j_c6 3 i.g5+ i;t>h7
4 h6 A 5 j_h5 +-) 2 e4 ! ! dxe4
(2 ...fxe4 3 i.dl 'it>xh6 4 i.g4 j_c6 5
i.g5+ 'it>h7 6 i.xe6 +-) 3 i.b3 i.d5
4 i.xd5 exd5 5 i.d6! (5 j_g5?
ctJb8! 6 'it>e5 ctJc6+ 7 'it>xd5 ctJxb4+
8 e6 ct:Jc6 9 d5 ct:Jxa5=) 5 ...Wxh6
(5 ... ctJf6 6 'it>xf5 ctJxh5 7 'it>g5 +-) 6

i;t>xf5 c;t>xh5 7 'it>e6 e3 8 i.f4! e2 9

i.d2 lbf8+ 10 i;t>xd5 i;t>g4 1 1 i;t>c6
Wf3 1 2 d5 'it>rz 13 i;t>b7! +-.
15

(a) l ...e5? 2 l:txc6 exd4 3 exd4

d3+ 4 @f4 d l 5 e4 g6 6 d5
:!:!'.e l + 7 j_e3 'it>g7 8 d6 I!dl 9 i.f4,
and at last he resigned. Osterman
Mikhalchishin, Bled, 1 995;
(b) 1 ...e6! 2 xc6 i.d8! with the
idea i.d8-a5-e l with a sufficient
defence.
16

No, because of 2 ...i.a4 ! ! 3 e2 (3

h5 i.xb3 4 h6 i.xc2 5 h7 b3+ 6
@e2 g2 -+) 3 ... i.xb3 4 i.d3

xe2+

Wxe2

i.dl+! !,

0-1,

Timman-Salov, Amsterdam, 199 1 .

204 Solutions

CHOOSE THE BEST CONTINUATION

IN ROOK ENDINGS
1 Alatortsev-Chekhover, USSR,
1937:
(a) I Wb5? :Ixa7 2 .!lxa7 'it>xf2=;
(b) 1 Wd5? Wfl 2 'it>c6 .!lxa7 ! 3
l:.txa7 Wxf2=;
(c) 1 @c5 ! J:Ic8+ ( 1 . .Jha7 2
l:.txa7 'it>xf2 3 @d4 + -) 2 @b6 .U.e8 3
@c6! (3 'it>b7 J::re7+ =) 3 . . . @fl
(3 . . ..!lh8 4 @b7 J:ih7+ 5 @b6 .!lh8 6
.!lc2 + -) 4 @b7 J:ie7+ 5 @b6 .!le8 6
J:ic2! @g2 7 @b7 .!le7+ 8 @b8!
J:ie8+ 9 .!lc8.
2 Karner-Renter, USSR, 1 955:

(a) 1 . ..l:th2? 2 h4 'it>h5 (2 . . . f2 3

l:.tf3 + -) 3 .l:Ia5+! @g4 4 .!lg5+ @h3
5 h5 a2 6 h6 .!la6 7 h7, 1 -0

(b) ! . ..'it>h5? 2 .llxf3 l:1h2 3 @gl

l:.txh3 4 Wg2 + -.
(c) I . . .@g5 ! ! 2 xf3 .!lh2 3 h4+
'itg4=.
3 Cuartas-Zuidema, Skopje,
1972:
(a) l ...f3 ? 2 c7 f2 3 J:id2 D. 4
l::!.b 2+ -.
(b) 1 . ...l:Ie 1? 2 c7 f3 3 .l:td8+ '3;;e7
4 'it>b7+ -.
(c) 1 .. ..l:Ic l ! 2 c7 l:.tc4 ! ! 3 J:id5
'it>e7 4 Wb7 @e6 5 .l:Ib5 @d7 6
d5+ We6=, Karpov.

Index of Players and Composers

Abramovic-Nikolic
Adams-Almasi
Adams-Lautier
Adams-Lutz
Adams-Xie Jun
Adorjan-Chemin
Adorj an-Morozevich
Akopian-Almasi
Alatortsev-Chekhover
Alekhine-Bogoljubow
Alexandria-Marie
Almasi-Portisch
Alterman-Chemin
Anastasian-Khalifman
Anastasian-Romanishin
Andersson-Hilbner
Arbakov-Gurevich
Averbakh-Bebchuk
Azmaiparashvili-Kupreichik
Azmaiparashvili-Novopashin
B agirov-Berzinsh
Bagirov-Kraidman
Bagirov-Veingold
Balashov-Korchnoi
Balashov-Tiviakov
Balashov-Ulibin
Balashov-Vaganian
Ballon-Mikhalchishin
Bareev-Farag6
Barle-Mikhalchishin
Barlov-Abramovic
Barlov-Schiissler
Barlov-S eirawan
Bayer-Polasek
Beliavsky-Azmaiparashvili
Beliavsky-Dolmatov
Beliavsky-Gelfand
B eliavsky-Geller
Beliavsky-Hodgson
Beliavsky-Kupreichik
B eliavsky-Mikhalchishin
B eliavsky-Miles
Beliavsky-Neverov

85
148
I7
II
3I
33
1 76
90
97
1 IO
53
32
1 42
29
I9
68
76
31
8I
44
60
57
64
I19
30
94
145
I 84
1 39
89
45
78
1 28
I 04
95
1 63
I3I
I 70
6I
I 74
I 32
I 54
1 43

Beliavsky-Short
B eliavsky-Spraggett
Beliavsky-Sveshnikov
Beliavsky-Tratar
Benko-Gereben
Bogoljubow-Rubinstein
Bojkovic-Kakhiani
Bologan-Kramnik
Botvinnik
Botvinnik-Balashov
Botvinnik-Donner
Botvinnik-Kan
Botvinnik-Rab inovich
Botvinnik-Simagin
Branicki- S efc
Bronstein-Rantanen
Bronstein-Timman
Brunner-Korchnoi
Bykova-Rubzova
Capablanca-Tartakower
Chaunin-Friedman
Chekhov-Karsa
Chemin-Georgiev
Chiburdanidze-Galliamova
Chiburdanidze-Maric
Cruz-Seirawan
Cruz-Seirawan
Cuartas-Zuidema
Cvitan
Dammes-Sosonko
Dan-Pytel
Dao Thien Hai-lvanchuk
Dautov-Alterman
D eak-Horvath
D egraeve-Hansen
Dokhoian-Shirov
Drasko-Vratonjic
Dreev-Anastasian
Dreev-Beliavsky
Ehlvest-Polugaevsky
Ehlvest-Rausis
Ehlvest-Shirov
Ehlvest-Topalov

I 74
59
I7
54
1 06
93
80
1 03
25
I 24
I 84
I 13
I I2
4I
96
II8
31
97
l 05
76
72
1 29
1 76
92
1 90
19
33
97
1 04
I 05
168
1 20
65
32
26
I 77
I3
26
I 04
65
I 36
I0
1 90

206 Index

Eliskases-Levenfish
Emma-Riemersma
Estrin-Gusev
Euwe-Alekhine
Evreinov
Farag6-Csom
Fedotov-Arkhipov
Fercec-Cvitan
Fercec-Mikhalchishin
Filipescu-Citron
Filipov-Kopatsny
Finkel-Mikhalchishin
Fischer-Bisguier
Fischer-Geller
Fischer-Letelier
Fischer-Larsen
Flohr-Vidmar
Formanek-Mikhalchishin
Gausel-Agdestein
Gelfand
Gelfand-Lautier
Gelfand-Lerner
Gelfand-Shirov
Genba-lrzhanov
Georgiev-Khalifman
Gligoric-Fischer
Godena-Lalic
Greenfeld-Golod
Gretarson-Magerramov
Grunberg-Brunner
Gufeld-Grigorian
Gufeld-Minev
Guli ev-Tukmakov
Gulko-Gurevich
Gulko-Sveshnikov
Gurevich-Andersson
Gurevich-Rechlis
Guseinov-Beliavsky
Gutrnan-Tseitlin
Havsky-Yuferov
Hector-Speelman
Heinig-Li ebert
Hellers-Eingom
Herrera-Vasquez
Hertneck-Narciso
Hoch
Hoi-Mikhalchishin
Holmov-Hasin
Holmov-Timoschenko
Horvath-Sherzer
Hiibner-Ftienik

81
56
31
59
26
1 72
34
13 I
83
27
90
20
23
94
21
23
18
1 70
142
1 05
131
1 08
135
1 43
1 14
22
44
ll
!33
85
39
1 83
25
1 02
1 44
1 34
101
70
1 05
27
9
33
12
57
151
28
182
1 04
74
1 77
1 27

Hiibner-Polgar
Hiibner-Spassky
Hulak-Beliavsky
Huzman-Mikhalchishin
Ilincic-Abramovic
Illivitsky-Taimanov
Ilyin-Zhenevsky-Botvinnik
Ioseliani-Onischuk
I vanchuk-Eingom
I vanchuk-Kasimdzhanov
lvanchuk-Kramnik
lvanchuk-Lautier
Ivanov-Sagalchik
Janvarev-Schcherbakov
Kamsky-Karpov
Kamsky-Cvitan
Kan-Keres
Kamer-Renter
Karpov-Hort
Karpov-Kasparov
Karpov-Kasparov
Karpov-Timman
Karpov-Yudasin
Karpov-Yusupov
Kasparov-Andersson
Kasparov-Short
Keller-Mikenas
Keres-Aronin
Keres-Mikenas
Keres-Szab6
Kirov-Kasie
Klovan-Elistratov
Kochiev-Lemer
Konopka-Shcherbakov
Konstantinopolsky-Fridman
Korchnoi-Kengis
Korchnoi-Ljubojevic
Korchnoi-Pinter
Korolev-Palm
Kovacevic-Tosic
Kozakov-Asparuhov
Kozlov-Mikhalchishin
Kozul-Mikhalchishin
Kozul-Nikolic
Kramnik-Beliavsky
Kramnik-Lautier
Krasenkov-Iskusnik
Kremenetsky-Razuvaev
Krumpachnik-Maksimenko
Krumpachnik-Polak
Kudrin-Cebalo

89
1 16
1 73
61
52
84
27
1 07

SI
20
1 64
88
30
25
59
1 44
1 15
97
145
12
157
1 85
1 03
1 17
1 27
1 03
66
1 83
1 07
1 66
1 02
29
43
67
1 07
99
15!
1 25
1 85
1 14
27
77
87
68
79
17
87
37
1 90
8
1 88

Index 207

Kupreichik-Mikhalchishin
14
Kupreichik-Zheliandinov
I 02
Kuzmin-Petrosian
19
Kuzmin-Petrosian
30
Larsen
1 07
Larsen-Browne
76
Larsen-Uhlmann
33
1 44
Lastin-Cvitan
Lasker-Levenfish
1 02
Lasker-Marshall
1 80
Lautier-Y errno l insky
1 03
1 07
Lein-Suetin
101
Lek6-Akopian
Lerner-Dorfman
67
Lesiege-Ivanov
101
102
Levenfish-Lisitsin
Levy-Peev
108
77
Li lienthal-Smyslov
Ljuboj evic-Ivanchuk
1 16
23
Lombardy-Fischer
Lopyshnoi-Dreev
1 08
Lputian-Tukmakov
1 22
Maciej a-Grabarczak
8
Madsen-Hansen
60
1 39
Magerramov-Kohlmeyer
Manukovsky-Pustovalov
1 09
86
Marie-McNab
52
Maric-Zaitseva
Markovic-Ivanovic
9
Markowski-Gdanski
1 89
Marosi-Yudasin
1 86
Marshall-Alekhine
1 83
1 79
Martinovic-Mikhalchishin
1 14
Martinovic-Yudasin
Matlak-Tseshkovsky
13
49
Matu!ovic-Uitumen
71
Mednis-Gurevich
33
Mestel-Chekhov
141
Mikhailov-Volchok
Mikhalchishin-Azmaiparashvili 1 1 0
119
Mikhalohishin-Beliavsky
1 87
Mikhalchishin-Cvetkovic
66
Mikhalchishin-Eslon
Mikhalchishin-Gufeld
1 80
Mikhalchishin-Kavcic
189
Mikhalchishin-Khme!nitsky
71
Mikhalchishin-Nestorovich
1 82
Mikha!chishin-Stangel
88
Mikhalchishin-Savchenko
151
Mikhalchiishin-Sveshnikov 1 5 0
1 65
Miles-Adorjan

Miles-Kindermann
Miles-Van der Sterren
Milov-Pelletier
Mitrofanov
Mokry-Pribyl
Morovic-Agdestein
Morovic-Yusupov
Morozevich-Balashov
Mortensen-Plueg
Munios-Salazar
Natapov-Schuravlov
Neverov-Dreev
Nezhmetdinov-Luik
Nezhmetdinov-Romanovsky
Nikolaevsky-Gufeld
Nikolic-Ftacnik
Niko!ic-Movsesian
Nikolic-Portisch
Nimzowitsch-Chigorin
Novikov-Beliavsky
Novikov-Kaidanov
Novikov-Lalic
Olafsson-Tal
Oil-Benj amin
Ostenstad-Kuzmin
Osterman-Mikhalchishin
Owen-Morphy
Padevsky-Latinov
Panno-Donner
Papendieck-Geiger
Parma-Gligoric
Pelletier-Arencibia
Pelletier-Rozentalis
Pelling-Miles
Perrniakov-Petkevich
Petrosian-Larsen
Petrosian-Rashkovsky
Petrosian-Tal
Piarnpuu-Lancava
Piket-Sokolov
Plaskett-Rowson
Polgar-Smyslov
Polnareva-Akhsharumova
Polugaevsky-Bronstein
Polugaevsky-Parma
Ponomariov-Plaskett
Portisch-Kramnik
Portisch-Pietzsch
Portisch-Ribli
Psakhis-Bonsch
Rausis-Farag6

1 67
1 60
1 42
1 10
73
1 05
1 54
1 29
29
1 07
8
1 24
1 23
121
43
63
83
138
27
64
1 62
58
1 06
18
11
1 90
78
26
137
29
189
1 47
82
137
1 03
31
1 75
1 69
1 04
1 06
I0
27
45
1 89
73
1 69
118
92
1 12
1 15
137

208 Index

Razuvaev-Kirov
Razuvaev-Ostojic
Renet- Olafsson
Reti- Bogolj ubow
Ribli-Hertneck
Ribli-Mikh alchishin
Ricardi-Sorin
Rittner -Bernstein
Rornanishin-Markowski
Rornanishin-Nunn
Romani shin-Rodriguez
Rornanovsky- Stahlberg
Rossolirno-Fischer
Rozentalis-Christiansen
Rublevsky- Sh ar iazdinov
Rufen acht-Orseth
Ryurnin-Ilyin-Zhenevsky
Sajtar-Benko
Sakaev-Novikov
Sale-Mikhalchishin
Salov-Karnsky
Salov- Kh a lifrn a n
Salov-Malaniuk
S alov-Yud a sin
Salwe
Sch andor ff- Speelrn an
Schlechter-Lasker
Schrnittdiel-Mikhalchishin
Serrnek-Hulak
Shirov-Krarnnik
Shirov-Lautier
Shirov-Morozevich
Shirov- Tirnrn an
Shirov-Van Wely
Short-Kasparov
Short- Top alov
Srnagin-Mikhalchishin
Srnagin-Naurnkin
Srnyslov-Epishin
Sokolov-Khalifrnan
Solozhenkin-Anchesi
Sorin-Alterman
Speelrnan-Saltaev
Stein-Bobotsov
Suba-Chiburdanidze
Sulipa-Gricak
Sveshnikov-Sokolov
Svidler - Lobron
Szabo-Keres
Szabo-Penrose
Szelaig-Pinski

155
188
30
159
181
88
31
182
152
121
172
184
24
133
128
26
32
48
141
188
173
123
70
106
105
10
79
87
54
73
132
86
7
129
158
102
150
80
82
177
33
178
28
175
174
14
51
8
62
138
32

Tairnanov-Chekhov
Tairnanov-Gligoric
Tal- Bal ashov
Tal-Suetin
Tavadian- Tseshkovsky
Tiets-Forsberg
Tikhornirova-Morozova
Tirnrnan-Ree
Tirnrn an-S alov
Tirnrnan-Van Wely
Toothill-Heernsoth
Topalov- Beli avsky
Topalov- Kasparov
Torre-Portisch
Tosic-Gyirnesi
Trabattoni-Barlov
Uhlrnann-Robatsch
Ulibin-Nevednichi
Vaganian-Portisch
Vagani an-Schlosser
Vaganian-Srnirin
Valvo- Levitt
Van der Doel-Klovan
Van der Sterren-Douven
Van der Wiel-Larsen
Van Laaturn-Mikhalchishin
Van Wely-Ad arn s
Va siukov- Tirnoschenko
Vaulin-Voikhovsky
Vladirnirov-Novop ashin
Vladirnirov-Rashkovsky
Vujala-Srnith
Vukic-Pietzsch
Vukic-Vujosevic
Vukovic-Eingorn
Vyzhrnan avin-Chiburdanidze
Vyzhrnanavin-Lerner
Ward-Baburin
Wells-Hector
Wirthensohn-Mikhalchishin
Yates-Alekhine
Y errnolinsky-Ivanov
Yerrnolinsky- Seir awan
Yudasin-Kr arnnik
Yuda sin-O snos
Yurtaev-Ternirbaev
Yusupov-Gerusel
Zaitsev-Hiibner
Zaichik-Sorin
Zinar
Zotkin-Kudrin

78
181
146
32
164
91
125
46
190
103
116
126
101
51
57
93
28
105
15
84
124
32
82
44
188
148
50
42
86
40
58
69
1 13
189
47
45
58
95
28
106
1 17
25
69
1 04
9
136
186
107
178
109
55