The Process of Elimination

Part Three

Purchases from our

chess shop help keep
ChessCafe.com freely
accessible:

In conclusion, I suggest that you solve a few more exercises that will train you
in the use of the process of elimination. The majority of these are more
difficult than the ones we looked at previously.
13. Daly Smith
British Championship, Plymouth 1992

The
Instructor
Mark Dvoretsky

Secrets of Opening Preparation

by Mark Dvoretsky
& Artur Yusupov

Translate this page

[FEN "r4N2/4Pp2/n1kP2p1/6P1/3R4/
PPp1Bb2/K1P5/7r w - - 0 37"]

1.?
14. Polugaevsky Hartston
Las Palmas, 1974
Secrets of Endgame Technique
by Mark Dvoretsky
& Artur Yusupov

Play through and download

the games from
ChessCafe.com in the
ChessBase Game Viewer.

[FEN "8/1P6/3k2p1/8/6p1/2n1R3/
5P2/1r3BK1 b - - 0 43"]

1...?
15. Stahlberg Alekhine
Olympiad, Hamburg, 1930

Secrets of Positional Play

by Mark Dvoretsky
& Artur Yusupov

[FEN "5rk1/1b4p1/1p1p3p/4p1q1/1PPn1r2/
4QP2/5RPP/3RNBK1 w - - 0 31]

1.?
16. Chernin Gindin
Kharkov, 1975

[FEN "3r1b1k/p1qr4/1p1p2p1/4p2p/2P1PPb1/
BP1R2P1/P2Q2B1/3R2K1 w - - 0 29"]

1.?
17. A. Cheron, 1957

[FEN "1b6/8/2P5/6KB/8/
4kP2/8/8 b - - 0 1"]

1...?
18. Kempinski Ogaard
Saint-Vincent, 2005

[FEN "r4rk1/p2qb1pp/2N2p2/1p1n3b/
2pPQB2/8/PPB2PPP/R4RK1 b - - 0 18"]

1...?
19. Smyslov Veresov
Moscow, 1940

[FEN "r4rk1/4ppbp/p1np1np1/qp6/3PP3/
1BN1BQ1P/PP3PP1/R3K2R w KQ - 0 14"]

1.?
20. Timoshchenko Yusupov
Soviet Championship, Frunze, 1981

[FEN "8/1k4pN/4n3/PP6/2K5/8/8/8 b - - 0 59"]

1...?
21. R. Rti & A. Mandler, 1921

[FEN "5k2/6p1/7p/8/7P/7K/8/8 w - - 0 1"]

1.?
22. De Firmian Gheorghiu
Lone Pine, 1980

[FEN "r1b1kb1r/1pNn1ppp/3ppn2/8/1q2PP2/
3BB3/P1P3PP/R2Q1RK1 b kq - 0 12"]

1...?
23. R. Rti, 1929

[FEN "K2n4/8/P7/8/2k5/8/2N5/8 w - - 0 1"]

1.?
24. I. Sokolov Rozentalis
Tilburg, 1993

[FEN "8/8/4N1k1/2pP3p/5R2/3K3P/
3n1pr1/8 b - - 0 46"]

1.?
Solutions
13. Daly Smith
The unsophisticated 37.d7? allows Black to give mate in three: 37...Nb4+! 38.
Rxb4 Rxa3+! 39.Kxa3 Ra1#.
The move 37.Ra4? doesn't save you from mate: 37...Nb4+! 38.ab Rxa4+ 39.
ba Bd5+ 40.Ka3 Ra1#.
Played in the game was 37.b4? Bd5+ 38.Rxd5 Kxd5 (38...Nxb4+ 39.Kb3
Nxd5 is also good) 39.d7 (39.Kb3 Kxd6-+) 39...Kc4! (threatening 40...
Nxb4#) 40.Bc5 Nxc5 (and now threatening 41...Rxa3+!). White resigned.
There remains the only, but fully sufficient defense, 37.a4! Preventing 38.d7
+- can only be done by means of 37...Nb4+ 38.Rxb4 Kxd6, then follows 39.
Rc4 Kxe7 40.Bc5+ Ke8 41.Rxc3 with an overwhelming advantage for White.
14. Polugaevsky Hartston
Where should the knight retreat to? In the case of 43...Nd1?? 44.b8Q+ Rxb8
45.Rd3+ or 43...Na2?? 44.Ra3 Nc1 45.Ra1! Ne2+ 46.Kh2 Black loses a
piece. On 43...Nd5?, the reply 44.Rd3! is unpleasant. But no flaws in the
move 43...Na4! are evident, and by playing this way Black easily achieves a
draw.
In the game, though, there followed 43...Nd5? 44.Rd3!

[FEN "8/1P6/3k2p1/3n4/6p1/3R4/5P2/
1r3BK1 b - - 0 44"]

1...?
44...Kc6?! (44Rxb7?? 45.Rxd5+!) 45.Rd1! Rxb7 46.Rxd5! Black resigned.
It was possible to defend considerably more stubbornly by retreating the king
to a different square: 44...Ke5! For example, 45.Kh2?! Nf4 46.Rd7 Rb2 47.
Ba6 (47.Kg3 Nh5+ 48.Kh4 Nf6 49.Rg7 Ne4=) 47...Rxf2+ 48.Kh1 Rb2 49.
Rd8 Rb1+ 50.Kh2 Rb2+ 51.Kg1 (51.Kg3 Kf5! 52.Rf8+ Kg5 53.Rxf4 Rb3+

54.Bd3!=) 51...Rb1+ 52.Kf2 Rb2+ 53.Ke1 (53.Ke3 Nd5+ 54.Kd3 Nc7=) 53...
Rb1+ 54.Kd2 g3!=.
Lev Polugaevsky gives the variation 45.Rd1! Rb2 46.Bg2 Nf4

[FEN "8/1P6/6p1/4k3/5np1/8/1r3PB1/
3R2K1 w - - 0 46"]

1.?
47.Rd8 Nxg2 48.b8Q+ Rxb8 49.Rxb8, but after the knight retreat to f4 or h4
the position that arises is most likely drawn.
Instead of 47.Rd8 the move 47.Re1+! is stronger, and any king retreat has its
drawbacks:
With 47...Kd6 48.Re8 Nxg2 (48...Kc7 49.Rc8+) 49.b8Q Rxb8 50.Rxb8
Black's king is far from the kingside, which sharply diminishes his chances of
building an impenetrable fortress.
And in the variation 47...Kf6 48.Bc6! g3!? (48...Nh3+ 49.Kh2+-) 49.Rf1!+/(49.fg? Nh3+ 50.Kh1 Nf2+ leads to a draw) the position of Black's king on
the f-file has an unfavorable impact on his position.
15. Stahlberg Alekhine
With his last move, 30h6!, Alexander Alekhine defended his queen, thus
preparing a capture on f3. Gideon Stahlberg, not seeing the biggest danger,
played 31.Kh1?, and after 31Rxf3! resigned in connection with the
impossibility after 32.Qxg5 Rxf2 of repelling two threats at once: 33...Rxf1#
and 33...hg.
31.Qd2? doesn't help (the queen retreats to a defended square) because of a
forced exchanging operation: 31...Bxf3 32.Nxf3 Nxf3+ 33.Rxf3 Rxf3 34.
Qxg5 Rxf1+ 35.Rxf1 Rxf1+ 36.Kxf1 hg 37.Ke2 Kf7 38.Kd3 Ke6 (38...b5) 39.
Ke4 b5-+.
The only defense: 31.Bd3!=/+. Taking on f3 is impossible now, Black has to
look for new resources to conduct his attack.
16. Chernin Gindin
White is better, so the positional sacrifice of a pawn or the exchange 29.f5?!,
leading to unclear consequences, simply isn't necessary.
29.fe? de is a mistake Black has too many threats: 30...Rxd3; 30...Bxa3; 30...
Bxd1.
Exchanging the light-squared bishops is favorable to White in principle, but
on 29.Bf3? there follows 29...Bxf3 30.Rxf3 d5!!.
Alexander Chernin restricted himself to the simple move 29.Rf1! White
retained all the advantages of his position, he intends 30.Bf3 or 30.f5.
29...Bh6 30.Bf3! Bxf3 31.Rfxf3 (intending 32.Rd5, 33.Rfd3) 31...b5!? (31...
h4 32.Qh2!) 32.cb d5 (and here it was probably worth trying 32h4!?) 33.

Rxd5 Qb6+ 34.Kg2 Rxd5 35.ed Qxb5 36.d6 ef 37.Qd4+ (37.gf is more
precise) 37Bg7?! (37Kh7 38.gf Qf5) 38.Qxf4 Qa5 39.Qb4!? (39.Qc1!?),
and White won.
17. A. Cheron
Black has to keep control of the f4-square for that reason he should only
move his bishop for now. The opponent will obviously bring his king to d7
then the black king should make it to b6.
With 1...Bd6? 2.Kf5 Kd4 3.Ke6 Kc5 4.Kd7+- the black king doesn't get there.
1...Bh2(g3)? is also a mistake because of 2.Kf5 Kd4 3.f4+-. We're left with
the only defense.
1...Bc7! 2.Kf5 Kd4! 3.Ke6 (3.f4 Ke3=) 3...Kc5 4.Kd7 Kb6 5.Be8

[FEN "4B3/2bK4/1kP5/8/8/5P2/8/8 b - - 0 5"]

1...?
Accuracy is again demanded from Black bearing in mind the possible return
of the enemy king to the kingside. A draw can only be secured with 5...Bb8! 6.
Ke6 Kc5 7.Kf5 Kd4 8.Kg4 Ke3.
18. Kempinski Ogaard
A capture with the queen on h7 isn't White's only threat: 18...Bg6? 19.Qxd5+!
loses immediately.
Leif Ogaard chose 18...Nxf4? 19.Nxe7+ Kf7 20.Qxf4 Qxe7, but after 21.
Rfe1 his position became hopeless. There then followed 21...Qb4 (21...Qd7
22.Bf5) 22.Qf5 g6 23.Qd7+ Black resigned.
But with 18...Kf7! 19.Nxe7 (19.Qxh7? Nxf4-+) 19...Nxe7 Black would be no
worse. For example, 20.Qxh7?! Rh8 21.Qe4 Bg6 (21...Rae8!?) 22.Qe2 Bxc2
23.Qxc2 Qxd4=/+.
19. Smyslov Veresov
The tempting 14.e5?! only leads to a draw: 14...de 15.Qxc6 Rac8 16.Qb7
Rc7! (16...e4?! with the threat of 17...Rc7 is weaker after 17.Bf4 Rxc3 18.00 Rcc8 19.Qxe7 Black experiences difficulties in connection with the
weakness of the f7-square) 17.Qf3.

[FEN "5rk1/2r1ppbp/p4np1/qp2p3/3P4/
1BN1BQ1P/PP3PP1/R3K2R b KQ - 0 17"]

The simplest is to act on Smyslov's recommendation: 17...Rxc3! 18.bc (18.

Bd2? Rxf3 19.Bxa5 Rf4-/+) 18...Qxc3+ 19.Ke2 e4! 20.Qg3 Qd3+ 21.Ke1 Qc3
+ with perpetual check. But 17...ed 18.Bxd4 e5! 19.Bxe5 Re7 (or 19...Re8)
20.0-0 Rxe5= is also possible.
In the initial position White controls more space and is better, so there's no
point in him going into the drawing variation. The simple 14.0-0!+/= is
stronger, as Vasily Smyslov did play. Then came 14...Rac8 15.Rad1 b4 16.
Nd5!

[FEN "2r2rk1/4ppbp/p1np1np1/q2N4/
1p1PP3/1B2BQ1P/PP3PP1/3R1RK1 b - - 0 16"

1...?
16...Nxd5?! (16...Qb5! is preferable, intending 17...a5 or 17...Nxd5 18.ed
Na5) 17.ed! Na7 18.Bg5!, and White has an obvious advantage thanks to the
pressure on the e-file.
20. Timoshchenko Yusupov
The h7-knight is locked in for now, but if White gets it into play (for example,
after 59...g5? 60.Nf6), he'll easily win.
The game ended like this: 59...Kc7? 60.Kd5 Nf4+ 61.Ke5 Ne2 (61...Nd3+ 62.
Kd4 and 63.Ng5) 62.Ng5+- (the knight has joined the battle and Black hasn't
got any counterplay) 62...Nc3 63.Ne6+ Kb8 (63...Kb7 64.Nc5+ Kc7 65.b6+
Kc6 66.b7 Kc7 67.a6 Nb5 68.Nd7) 64.b6 Na4 65.Kd4 g5 66.Nc5 Black
resigned.
The move Kd5 can't be stopped, but before letting the white knight out to
freedom it's important to force b5-b6, to attack the pawns with the king. The
aim is achieved by means of 59...Ka7!! 60.Kd5 Nc7+! 61.Kc6 Ne6.
62.Ng5 Nxg5 63.b6+ Kb8 64.a6 Ne6= isn't dangerous; 62.Kd6 Nd4 63.b6+
Kb7= is harmless. Only 62.b6+ Ka6 63.Ng5 (63.Nf8 Nxf8 64.b7 Nd7=)
maintains the intrigue.

[FEN "8/6p1/kPK1n3/P5N1/8/8/8/8 b - - 0 63"]

1...?

In this situation the knight is invulnerable; 63...Nd4+? 64.Kc7 Nb5+ 65.Kd7!

Nd4 66.Ne6 Nb3 67.Kc7 Nxa5 68.Nc5+ Kb5 69.Nb3 also loses.
63...Nd8+! 64.Kc7 gives a draw, and now either 64...Nb7 65.Ne4 g5! 66.
Nxg5 Nxa5=, or 64...Kxa5 65.Nf7 Ne6+ 66.Kb7 g5! 67.Ka7 Nc5!= or 67...
Nd4!= (but not 67...g4? 68.b7 Nd4 69.Ka8! Nc6 70.Ne5+-).
21. R. Rti & A. Mandler
Many years ago in a game against Oleg Romanishin I had the chance to
transfer into this kind of pawn ending, but I was unable to calculate it
correctly (see School of Chess Excellence 1: Endgame Analysis, the chapter
"Give me an envelope, please"). Of course, I didn't know this study by Rti
and Mandler at the time.
We can immediately reject 1.h5? Kf7 2.Kg4 Ke6 3.Kf4 Kf6-+ and start
calculating the most natural move 1.Kg4. Our opponent replies 1...Kf7 (but
not 1...Ke7? 2.Kf5 Kf7 3.h5=, of course).

[FEN "8/5kp1/7p/8/6KP/8/8/8 w - - 0 2"]

Neither 2.h5 Ke6-+ nor 2.Kh5 Kf6 3.Kg4 Ke5 4.Kh5 Kf4 5.Kg6 Kg4 6.Kxg7
h5-+ works.
In the case of 2.Kf5 g6+ 3.Ke5 Ke7 Black gets the opposition and then
advances his king without hindrance: 4.Kd5 (4.h5 g5 5.Kf5 Kd6 is hopeless
too) 4...Kf6 5.Ke4 Ke6 (again Black has the opposition) 6.Kf4 Kd5 7.Kf3
Ke5 8.Ke3 Kf5 9.Kf3 h5 (the reserve tempo h6-h5 allowed Black to put his
opponent in zugzwang for the last, decisive time).
The key to understanding the following variation, and indeed the whole
ending, is the fact that seizing the opposition has decisive significance for
both players with a black pawn on g6. But with the pawn on g7 it's essential,
on the contrary, to yield the opposition to the opponent. Which means that on
2.Kf4 it's necessary to reply 2...Ke6! 3.Ke4 g6-+. Finally, in the case of 2.Kf3
the move 2...g6! (seizing the distant opposition with the pawn on g6) wins 3.
Ke3 Ke7! 4.Kf3 Kd6! (a king walkabout is a weapon with the help of which
the strongest player exploits the possession of the opposition) 5.Ke4 Ke6 6.
Kf4 Kd5 7.Kf3 Ke5 8.Ke3 Kf5 9.Kf3 h5.
And so, 1.Kg4? doesn't save White. We have to put the king on another
square: 1.Kg3!! For example,
1...Kf7 2.Kg4! Kf6
2...Kg6 3.h5+ is useless. On 2...Ke6 there follows 3.Kf4!= (with the pawn on
g7 White yielded the opposition to his opponent), and on 2...g6 3.Kf3!=
(seizing the distant opposition).

[FEN "8/6p1/5k1p/8/6KP/8/8/8 w - - 0 3"]

1.?
3.Kh5! (but not 3.Kf4? g6-+) 3...Ke5 (3...Kf5 stalemate) 4.Kg6 Kf4 5.Kxg7
h5 6.Kf6! Kg4 7.Ke5 Kxh4 8.Kf4=.
Black doesn't manage to achieve success with maneuvering of the kings
either: 1...Ke7 2.Kf3! (it's impossible to go forwards:2.Kf4? Ke6!-+) 2...Kf6
(2...Ke6 3.Kf4! Kf6 4.h5=; 2...g6 3.Ke3!=) 3.Ke4! (but not 3.Kg4? Ke5-+ and
not 3.Kf4? g6-+) 3...Kf7!? (3...Kg6 4.Kf4=) 4.Ke3! (with the pawn on g7 you
can't take the distant opposition) 4...Ke7 5.Kf3!=.
22. De Firmian Gheorghiu
Black played a sharp and very risky variation of the Sicilian Defense: 1.e4 c5
2.Nf3 e6 3.d4 cd 4.Nxd4 a6 5.Bd3 Nf6 6.0-0 d6 7.Nc3 Nbd7 8.f4 Qb6?! 9.
Be3 Qxb2 10.Ncb5! (10.Ndb5! is equivalent) 10...ab 11.Nxb5.

[FEN "r1b1kb1r/1p1n1ppp/3ppn2/1N6/
4PP2/3BB3/PqP3PP/R2Q1RK1 b kq - 0 11"]

By choosing 11...Qb4, Florin Gheorghiu allowed a knight fork on c7,

counting on catching the white king in the corner of the board. Subsequently,
11...Ra5 12.Rb1 Rxb5! (forced) 13.Rxb2 Rxb2 was also used. In my view,
what happened in the duel Anand Kasparov (Tilburg, 1991), 14.Qa1 Rb6 15.
Bxb6 Nxb6 16.Qc3! Be7 17.Rb1 secured White an advantage. For those who
are interested in this variation I suggest an idea that hasn't been tried once in
practice yet and wasn't mentioned by Anand in his notes to the above game:
11...Ra3!? 12.Rb1 (having to deal with 12...Qxb5 13.Bxb5 Rxe3, and on 12.
Qe2 possible is 12...Rxd3 13.Qxd3 Qb4 unclear) 12...Qxa2 13.Bd4 e5 unclear.
Let's switch to the position from our exercise that arises after 12.Nc7+. The
move 12...Kd8 looks like the natural retreat. But Gheorghiu justifiably
rejected it in connection with 13.Nxa8 Qa5 14.Nb6! Nxb6 15.Qe1!, and in the
case of 15...Qxe1 the knight is taken with check, and with 15...Qa7 16.a4 it's
lost because of a pin.
12...Ke7! 13.Nxa8 Qa5
Here he wins the knight back; then again, Black's position remains alarming.
14.e5! Ne8! (14...Nd5 15.ed+ is much worse) 15.ed+ Nxd6 16.c4! Qxa8 17.c5

[FEN "q1b2b1r/1p1nkppp/3np3/2P5/
5P2/3BB3/P5PP/R2Q1RK1 b - - 0 17"]

17...Ne8 (17...Nf5 18.Bxf5 ef is hardly better) 18.f5! f6? (18...Qa3 left more
chances for a successful defense), and now White achieves a big advantage by
continuing 19.Bc4!? (E. Geller) or 19.fe!? Ne5 20.Be4.
23. R. Rti
Obviously we have to choose between two king moves: to a7 or b8.
In the case of 1.Kb8? the response 1...Nc6+? 2.Kb7 Kb5 3.Nd4+! quickly
loses (and with 2...Kc5 the move 3.Nb4 is good too). But Black gets a draw
by means of 1...Kb5! 2.Nb4 Nc6+ 3.Kb7 Na5+ (3...Nd8+? 4.Ka7! Kxb4 5.Kb6
+- or 4...Ka5 5.Kb8+- are mistakes) 4.Kc7 Nc6.
1.Ka7! is stronger. If 1...Kc5, then 2.Nd4! decides matters (but not 2.Nb4?
Kb5 3.Kb8 Nc6+). 1...Kb5 2.Nb4!

[FEN "3n4/K7/P7/1k6/1N6/8/8/8 b - - 0 2"]

Black is in zugzwang. His king is forced to occupy the important a5-square.

2...Ka5 3.Kb8 Nc6+ (3...Kxb4 4.Kc7 Ne6+ 5.Kb6+-) 4.Kb7 (4.Kc7? Nxb4 5.
a7 Nd5+) 4...Nd8+ (unlike the variation 1.Kb8?, there's no knight check from
a5) 5.Kc7 Ne6+ 6.Kb8+- (or 6.Kc6+-).
24. I. Sokolov Rozentalis
White's position is alarming. True, 46...f1Q+ isn't threatened yet (because of
the knight fork on f4), but it will be threatened on the next move. For
example, 46.d6? Rxh2! 47.d7 f1Q+ 48.Rxf1 Nxf1, and a winning knight
endgame for Black arises.
It makes sense to move the h-pawn out from under the attack. In the case of
46.h4? c4+ the response 47.Kd4 Rg4-+ can't be played, and 47.Kc3 Ne4+ (or
47...Nb1+ 48.Kb4 c3) 48.Kxc4 Rg4-+ doesn't help. It's becoming clear that
the g4-square has to be taken under control.
46.h3!! Rh2 (46...f1Q+ 47.Rxf1 Nxf1 48.Nf4+; 46...c4+ 47.Kd4) 47.d6 f1Q+
48.Rxf1 Nxf1 49.d7 Rd2+ 50.Kc4=

[FEN "8/3P4/4N1k1/2p4p/2K5/7P/3r4/
5n2 b - - 0 50"]

White wins the rook back and with the h3-pawn still on the board easily holds
the knight ending.
50...Ng3 (50...Rxd7 51.Nf8+) 51.d8Q Rxd8 52.Nxd8 Ne4 53.Ne6 h4 54.
Nxc5 Nf2 55.Nd3 Draw.

2011 ChessCafe.com. All Rights Reserved.

A PDF file of this month's column, along with all previous columns, is
available in the ChessCafe.com Archives.

Comment on this month's column via our Contact Page! Pertinent responses
will be posted below daily.

[ChessCafe Home Page] [Book Review] [Columnists]

[Endgame Study] [The Skittles Room] [ChessCafe Archives]
[ChessCafe Links] [Online Bookstore] [About ChessCafe.com]
[Contact ChessCafe.com]
2011 BrainGamz, Inc. All Rights Reserved.
"ChessCafe.com" is a registered trademark of BrainGamz, Inc.