350 SILMAN’S COMPLETE ENDGAME COURSE
.._ Optimizing your position before undertaking a decisive move forward has to be useful. Usually one doesn’t have the luxury
of wasting time in this fashion, but if you have carte blanche to
tidy up every imperfection, why not take advantage of it?
You might find that you’re burning up too much time on
the clock in an effort to find the final, winning maneuver. If another time control can be painlessly reached by treading water (thus giving added time in a safe fashion), grab the gift of free minutes with both hands!
Our first example features two famous endgame aficionados-Grandmasters Pal Benko and Yuri Averbach-going head to head, with Benko relishing his ability to inflict long, slow torture on his esteemed foe.
Diagram 351
8
7
6
‘j
4
3
2
a b c d e f g h
Benko-Averbach, Portoroz 1958
White to move
Benko had this to say about the position: “Black’s dream is to trade the b-pawns and the minor pieces and play a Rook ending with four vs. three on the kingside. Such an endgame is usually drawn. Dreams are one thing, though, and reality is quite another. I would never allow such a situation to come to pass.”
l.Nc5 Ra7
1…Bxc5 2.Rxc5 is winning for White because he has two advantages: his Rook is more active than its black counterpart (due to the need to defend b5), and he’s a solid pawn up on the kingside.
2.Nd3 Ral+ 3.Kg2 Rdl 4.KB
White’s winning plan is very simple: he intends to slowly improve his posi tion on the kingside (e3-e4, gaining space and defending the Rook, is a useful prelude), defend the Knight with his King (the immediate threat is Ke2), and then win the pawn on b5.
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 351
4…Rd2
This stops Ke2.
5.e4
Preparing Ke3.
5 …h5
This move not only prepares …h5-h4, but also intends to meet 6.Ke3 with
6…Bh6+.
6.e5
Gains space and gives white’s King access to e4.
6…h4
Let’s borrow a bit more from Benko: “Another pawn is about to fall so Black tries one last idea: if he can trade as many pawns as possible, he might be able to sacrifice a piece for white’s remaining pawns.”
7.gxh4 Rdl 8.Ke2
The pin is finally broken and White will get to eat the b-pawn, giving him a powerful passer.
8 …Rhl 9.Rxb5 Rxh2 10.Rb8 Rxh4 11.h5 Kg7 12.b6 Rhl 13.Rb7 Rhl
14.f4
Diagram 352
8
7
6
5
4
3
2
a b c d e f g h
Benko is happy to solidify his position and keep black’s counterplay to almost zero. The very tempting 14.e6 Kg8 15.e7 (15.exf7+ Kg7) 15 …Bxe7 16.Rxe7
Rxb6 would have left White with only one pawn. Why deal with such a hard technical task when you can torture your opponent at your leisure.
14…Kg8 15.Kd2 Ba3 16.Kc2 Rb5 17.Kc3 Bf8 18.Kc4 Rh119.Rb8
White’s King is now near its passed pawn. With all his pieces so well placed, the win must be just a few moves away.
19…g5
352 SILMAN’S COMPLETE ENDGAME COURSE
Desperation, but 19 … Kg7 runs afoul of 20.Rxf8! Kxf8 2l.Nb4 Rcl+
22.Kb5.
20.fxg5 Kg7 21.Rxf8! Rxh6
Black wasn’t happy about this, but 21…Kxf8 22.Nb4 Rcl+ 23.Kb5 ReS 24.Nc6 is hopeless.
22.Rd8 Rg6 23.Rd7 Kf8 24.Kd5 Rxg5 25.Ke4
Diagram 353
a h c d e f g h
The game is clearly won, but Benko goes about it with extreme patience. Ob viously he can’t allow the exchange of the last pair of pawns, but other than that his goal is clear: make use of the power of the 7th rank Rook combined with the free-range King and Knight to attack f7 or create potential mating nets.
25 …Rgl 26.Nf4 Rel+ 27.Kf5 Rfl 28.Ra7 Ke8 29.Rb7
Cat and Mouse. In a way, White is asking his opponent, “Are you enjoying this? I certainly am!”
29…Kf8 30.Kg5 Ke8 31.Nd5 Rgl+ 32.Kf5 Rfl+ 33.Ke4 Kf8 34.Ne3 Rf2
35.Nf5 Ke8 36.Nd6+ Kf8 37.Rc7 Rfl 38.Kd5
Benko finally decides to put his opponent out of his misery. The King will march down the board and prepare to create a winning King and pawn vs. King position by sacrificing his Rook and Knight for black’s Rook and pawn.
38…Rdl+ 39.Kc6 Rcl+ 40.Kb7 Rxc7+
40…Rbl+ 4l.Kc8 Rfl 42.Rxf7+ Rxf7 43.Nxf7 Kxf7 44.Kd7 amounts to the same thing.
41.Kxc7 Ke7 42.Nxf7!, 1-0. The pawn promotes after 42 …Kxf7 (or 42 …Ke6
43.Kc6) 43.Kd7.
If you thought that game showed patience, let’s let Benko take us even deeper into the world of Cat and Mouse against another great endgame player-the legendary Victor Korchnoi!
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 353
Diagram 354
a h c d e f g h
Benko- Korchnoi, Curacao 1962
White to move
l.Nd4
Benko: “I was willing to trade my Bishop for the Knight because I felt that, ulti
mately, my Knight would prove itself to be more useful than the enemy Bishop.”
l…Nc4+ 2.Bxc4 Bxc4 3.Rc2 Ba6 4.Rhc1 Rxc2 5.Rxc2 Kd7 6.e5
Diagram 355
a h c d e f g h
Benko: “In such positions, little things make all the difference. Here I’m gaining central space, getting my pawns off the color of black’s bishop, and (by preventing an eventual …e6-e5) making sure the d4-square will be permanently available for use by my Knight or King.”
6…Rc8 7.Rxc8 Kxc8
Benko: “The Rooks are traded and, to the uninitiated, the game appears to be headed for a draw. However, the position is actually very pleasant for White because only he can play for a win while Black will be defending for a long time to come. Why is White better? The flexibility of the Knight is one huge factor (the d6-square in particular is calling to it!), and my central space advantage and superior King also combine to make black’s life unpleasant.”
354 SILMAN’S COMPLETE ENDGAME COURSE
Note how Benko is raving about the superiority of his Knight over black’s Bishop. We’ll delve far deeper in this “Knight worship” in the section The Pleasures of a Knight.
8.Nf3 h6 9.Nd2 Kd7 10.Ne4 b6 ll.f4 Bfl 12.Kd4 Bg2 13.Nd6 f5
Benko: “White’s position has improved, but Black has everything covered. Now starts a long game of Cat and Mouse: I make little jabs at his weak points and torment him endlessly.”
14.Nb5 Bfl 15.Nd6
Benko points out that 15.Nxa7 Kc7 16.a4 Kb7 17.Nb5 Bxb5 18.axb5 would be drawn because his King would not be able to penetrate into the enemy camp.
15…a5 16.b4 axb4 17.axb4 Bh3 18.Nc4 Kc6 19.Nd6 Kd7 20.Nc4 Kc6
21.Ne3 Kh5 22.Kc3 g5 23.Nc4 gxf4 24.gxf4 Kc6 25.Kd4 h5 26.Nd6 h4
27.Nc8 Bg2 28.Ne7+ Kb5 29.Kc3 h3 30.Ng8 Kc6 31.Nf6 Bf3 32.Kd4 Kb5
33.Nd7 Bd5
Avoiding 33 …Kxb4 34.Nxb6 Kb5 35.Nd7 Kc6 36.Nf8 Bd5 37.Nh7 Bg2
38.Ng5 Kd7 39.Ke3 when White wins a pawn by Kf2, Kg3, and Nxh3.
34.Kc3 Ba2 35.Nf8 Kc6 36.Nh7 Kd5 37.Nf6+ Kc6 38.Kd4 Kb5 39.Kc3
Kc6 40.Nh7 Kd5
Diagram 356
8
7
6
5
4
3
2
a b c d e f g h
White has “Cat and Moused” his opponent into a near state of coma and now
(with Black bleary-eyed and exhausted) decides to carry out his final assault.
41.Ng5 b5 42.Kd3
Benko: “Suddenly Black is facing some serious problems. White is willing to enter the race by giving up his b-pawn because the passed e-pawn or h-pawn (whichever one White creates) will prove faster. Note that the immediate
42.Nxh3? would have been an error due to 42 …Ke4.”
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 355
42 …Bc4+ 43.Ke3 Bfl 44.Kf2 Bg2 45.Kg3??
Evidently Korchnoi wasn’t the only one tired from the battle. Benko could have forced a win by 45.Ke2! (this allows the King to stop black’s b-pawn in the key variation) 45 …Kc4 46.Nxe6 Kxb4 47.Nd4 Kc5 48.Nxf5 b4 49.Kd2, etc.
45 …Kc4 46.Nxe6 Kxb4 47.Nd4 Kc4??
Diagram 357
8
7
6
5
4
3
2
1
a b c d e f g h
Benko: “Korchnoi finally cracks. He had to try 47…Kc5! when the game would be drawn: 48.Nxf5 b449.Ne3 b3 50.Ndl Kd4 5l.e6 Bd5 52.e7 B£7 53.Kxh3 Kd3
54.f5 Kd2 55.Nb2 Kc3, and a peaceful conclusion can’t be avoided.”-Benko.
48.Nxf5 b4 49.Ne3+ Kd3 50.Nxg2 hxg2 51.Kxg2 b3 52.e6 b2 53.e7 bl=Q 54.e8=Q
The rest is easy for White.
54…Qa2+ 55.Kg3 Qa7 56.Qh5+ Ke4 57.Qe5+ Kd3 58.f5 Qh7 59.h4
Qg8+ 60.Kf4 Qc4+ 61.Kg5 Qg8+ 62.Kh6 Qf8+ 63.Qg7 Qd8
63 … Qxf5 64.Qg6 Ke4 65.h5 is also game over.
64.f6, 1-0.
Are you getting a feel for the power of Cat and Mouse yet? Remember, the concepts in this chapter are not things to be memorized, they are ideas that need to be embraced.
Since Benko is doing such a great job teaching us about the effectiveness of Cat and Mouse, I’ll let him drive the message home by borrowing one more example from Pal Benko: My Life, Games, and Compositions. This time he Cat and Mouses one of the greatest endgame players of all time.
356 SILMAN’S COMPLETE ENDGAME COURSE
Diagram 358
8
7
6
5
4
3
2
a b c d e f g h
Benko- Smyslov, Zagreb 1959
White to move
White, with an extra pawn, a superior King, and a strong Knight vs. a Bishop should win. However, it won’t be easy. This is what Benko had to say: “The game was adjourned here and after analyzing at leisure, I realized the win would take quite a long time. If victory is to be achieved, the white King will have to find a way to penetrate. Unfortunately, black’s King and Bishop make this difficult. As a result, I decided that it was time for a bit of Cat and Mouse: I would patiently maneuver, hoping to make Smyslov less alert. I knew that such long-term torture from a passive position would be very hard for Black to deal with, so I sat down and resumed the game ready for a lengthy siege.”
l.f5+
Benko: “The first order of business is to push back the black King. It isn’t optimal to place my pawn on a vulnerable white square, but I had no choice in the matter.”
I should add that, after l.f5+, Black is completely passive and can’t do anything but go back and forth and hope that White can’t find a way to finish him off. Such positions virtually scream for Cat and Mouse treatment!
l…Kd7 2.Nc4 Bb5 3.Kc3 Ba6 4.Kb3 Bb5 5.Nd6 Be2 6.Kc3
Continuing his Cat and Mouse torture, but the more straightforward 6.Ka4 seems to win by force: 6… Bd3 7.Ka5 Kc7 8.h4 h6 9.h5 Be2 (9…Kd7 10.Kb6 Bc2 ll.Kb7 Bd3 12.Nf7 Bxf5 13.Nxh6 Be6 14.Nf7 Bxf7 15.h6 Bg6 16.b5) 10.Ne8+ Kd8 ll.Nxf6-analysis by Benko.
6…Ke7 7.Nc4 Kd7 8.Ne3 Ke7 9.Nc2 Kd7 10.Nd4 Bflll.Kd2 Bc4 12.Ke3
B£7 13.Kf4 Bc4 14.Kg4 B£7 15.Kh4 Kc7
Smyslov wrongly assumed that, after 15 …Kc7, White wouldn’t be able to win the h-pawn because the Knight gets trapped. However, Black was dead lost in any case: 15 … Be8 16.b5 cxb5 17.Nxb5 and the threat ofNd6 or even Nc3-e4,
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 357
mixed with ideas of pushing the c-pawn and/or advancing the King to h5 and h6, is overwhelming.
16.Ne6+ Kd7
16…Kc8 17.Nf8 Bg8 18.Kh5 is easy for White.
17.Nf8+! Ke7 18.Nxh7 Bg8 19.Ng5! fxg5+ 20.Kxg5
Diagram 359
a h c d e f g h
White’s pawns are unstoppable since Black must also take a b4-b5 break into account.
20…Kf7 21.h4 Kg7 22.f6+ Kf8 23.h5 Bc4 24.Kf4 Kf7 25.Ke5 Bd3 26.h6
Kg6 27.Ke6 Bc4+ 28.Ke7 Bd5 29.f7 Bxf7 30.h7, 1-0.
Obviously, Cat and Mouse and patience go hand in hand. Our next example shows IMJack Peters (who dominated Southern California chess for almost three decades!) picking his opponent apart in unhurried and impressive fashion.
Diagram 360
a h c d e f g h
R Schutt-J Peters, Los Angeles 1989
White to move
Such bad Bishop positions are almost always lost-the fact that white’s pawns are on light colored squares mean they are targets that will ultimately be picked
358 SILMAN’S COMPLETE ENDGAME COURSE
off by black’s pieces, while an exchange of Bishops usually leads to a lost King and pawn endgame since black’s King is already deep in white’s position.
In the present case the position seems resignable for White. Black has man aged to penetrate with this King to el and an attack against white’s pawn on c4 via …Bdl seems to ensure that material losses are impossible to avoid. However …
l.Be2!
A shock! Taking the Bishop leads to immediate stalemate.
l…Bh3 2.Bf3
Black’s King can’t scamper over the queenside due to the white Bishop’s con trol over the dl-h5 diagonal, and black’s Bishop isn’t able to penetrate either, though he gives it one more shot.
2…Bfl 3.Be2!
Yes, I guess it’s true! It really is stalemate! Has White managed to save the game?
3...Bg2 4.Bg4 Kfl
No, Black is still winning. Having accepted white’s miracle defense (and that’s an important emotional step, since a shock like white’s stalemate defense can easily throw a player for a loop), Black calmly (and very, very patiently!) plans the following:
He will swing his King around to g3 (which is where it came from earlier, by the way!). White’s King would then be stuck keeping it out of the decisive f2- and f4-squares.
He will force his Bishop into white’s camp by leaping into dl.
Once that is done, Black will make use of tempo moves and attacks against the white pawns to decisively penetrate with his King.
If White tries to prevent this (i.e., black’s Bishop reaching dl) by placing his Bishop on the dl-a4 diagonal, black’s King will march to el and reach a position where the Be2 stalemate defense is no longer possible (it needs to be on the dl-h5 diagonal).
5.Bh5 Kgl 6.Be2
Forced. A move like 6.Bg4 would have lost immediately to 6… Bfl 7.Be2
Bxe2 8.Kxe2 Kg2 (This decisively takes the Opposition, though to be fair, the
King and pawn endgame wins for Black in most instances here. White would
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 359
have no defense, no matter who has the Opposition, if the black King reaches g4, h4, f3, g3, f2, or g2.) 9.Ke3 Kg3 IO.Ke2 Kf4 ll.Kd3 Kf3 and White loses all his pawns.
6…Kh2
Still patiently following the plan discussed on move four. Note that 6…Bfl ?? fails to 7.Bxfl Kxf1 8.Kf3 when White has reached one of the few positions where his King can successfully keep black’s away from his pawns.
7.Bd3 Kg3 8.Be2 Bh3
Diagram 361
a h c d e f g h
Black’s King has reached g3. Next on the agenda is getting the Bishop to d1!
9.Bd3
No better is 9.Bd1 Bg4 when 10.Bxg4? Kxg4 is a winning King and pawn endgame for Black.
9…Bg4 10.Bc2 Bf3 ll.Bd3
White allows the Bishop to d1 without a fight.Jack Peter’s analysis (comments by Silman) shows that 11.Bb3 would not have changed the result: 11…Bg4 12.Bc2
Kg2 (Heading back to el. Note that as long as white’s Bishop is stuck on a4/b3/c2,
the Be2 stalemate defense won’t be possible.) 13.Bb3 (13.Bd3 Bd1 14.Kd2 Ba4
15.Ke3 Kg3 16.Bfl Bc2 17.Be2 Kg2 18.Bh5 Kf119.Bg4Ke1 20.Bf3 Bb3 2l.Be2
Ba2 22.Bd3 Kdl 23.Bfl Kc2 and white’s pawns will soon fall from the board like
[I’m quoting from dozens of old chess books] “ripe fruit.”) 13 …Kf114.Bc2 Kel
15.Bb3 Be2! (Again, Black must avoid 15 … Bd1?? since 16.Bxd1 Kxd1 17.Kd3 gives White the Opposition and a draw) 16.Ba2 Bd1 (Now that white’s Bishop no longer controls the d1-square, black’s King will be able to swing over to the queenside via d1 and c2) 17.Bb1 (Also easy for Black is 17.Kd3 Kf2 18.Kd2 Be2
19.Bb3 Kf3 20.Ba2 Bfl and White loses material.) 17 …Bb3 18.Bd3 (18.Kd3 Kf2)
18…Kd1 19.Bfl Kc2 with …Kc3 and …Bxc4 to follow.
360 SILMAN’S COMPLETE ENDGAME COURSE
ll…Bdl 12.Kd2 Ba4 13.Ke3 Bb3 14.Be2
Also losing are 14.Bfl Bc2 15.Be2 Kg2 16.Bh5 Kfl (when the black King will make decisive inroads into the white position) and 14.Ke2 Kf4 15.Kf2 Bd1, idea
…Bf3 and …Bxe4.
14…Bc2 15.Bf3
This allows a pretty finish. However, 15.Bfl wouldn’t have made any differ
ence: 15… Bb1 (forcing white’s Bishop to give up control over g2) 16.Be2 Kg2
17.Bh5 Kfl 18.Bf3 Ke1
Diagram 362
a b c d e f g h
Now this is easily winning
Black wins this position (compared to our starting position) because his Bishop has already penetrated into the enemy camp! 19.Bg4 (Now 19.Be2 is blasted away by 19… Bxe4) 19…Ba2 20.Be2 Bb3 (Zugzwang! Either the Bishop must give up control over c4 or d1, or the King must hand Black access to f2 or d2.)
2l.Bd3 (2l.Kd3 Bxc4+; 2l.Kf3 Kd2) 21…Kd1 22.Bfl Kc2 when Black has finally achieved his dream penetration.
15…Bd3!, 0-1 since 16.Kxd3 Kxf3 leads to the loss of all the white pawns.
Diagram 363
a b c d e f g h
Kuznecov- Silman, Oregon 1986
Black to move
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 361
Black fixes white’s pawns on light squares before stepping into the minor piece endgame. This shows how one can prepare a favorable endgame right in the middlegame.
l…e5!
A strong move that clamps down on d4, prepares …Nc5-e6-d4, and also fixes white’s e-pawn on the light-colored e4-square. White’s position is now extremely uncomfortable.
2.Qdl Rxd2 3.Qxd2 Ne6 4.Ne2 Qe7 S.Ncl Kg7 6.Nb3 c5 7.Ncl Bc6
8.Qdl Qd7!
Simultaneously hitting a4 and dl and forcing White to enter a very poor minor piece endgame.
9.Qxd7 Bxd7 10.b3
Diagram 364
8
7
6
5
4
3
2
a b c d e f g h
A lost minor piece endgame
With so many of white’s pawns on light squares, it’s hard to imagine White saving the game. All Black needs is patience-slow improvements in pawn struc ture, kingside space, and piece position, followed by penetration into white’s camp, should be enough to eventually claim the full point.
10…Nd4
Now white’s Knight is stuck doing guard duty to the b3-pawn.
ll.Kgl f5
Adding to black’s list of advantages by grabbing kingside space.
12.Bd3 Bc6
Putting pressure on e4 and letting White worry whether Black will leave the pawn on fS or push to f4 with even more kingside territory.
13.exf5 gxf5 14.Kf2 Kf6 15.Ke3 Kg5 16.Bbl h5 17.Bd3 h4
This threatens to undermine the B-pawn by …h3.
362 SILMAN’S COMPLETE ENDGAME COURSE
18.h3?
This is too much! Putting another pawn on a light square can’t be healthy. He should have played 18.Be2 with the understanding that he was going to passively suffer for a long time to come.
18…Be8! 19.Be2 Bg6
Preparing …f4+ followed by a Bishop penetration on the bl-h7 diagonal.
20.Bdl
Diagram 365
8
7
6
5
4
3
2
a b c d e f g h
Also possible was the desperate 20.f4+ exf4+ 2l.Kf2, though Black wins com
fortably enough. One clear continuation is 21…Nxe2! 22.Nxe2 (22.Kxe2 Bh5+
23.Kf2 Bdl 24.Kel Bc2 is completely hopeless) 22 …Bh5 23.Ngl (23.Nc3 Kf6
24.Nd5+ Ke5 25.Nxb6 Bdl is resignable) 23…Bdl 24.Nf3+ Kf6 25.Nd2 Ke5
26.Kel Bc2 27.Nf3+ Kf6 28.Nd2 f3! 29.gxf3 Ke5 and it’s time for White to shake hands.
20…f4+ 2l.K.f2 Bbl 22.Ne2
Things end quickly after this, but even the more resilient 22.Kel couldn’t hope to nullify the accumulated Black advantages. Both 22 …e4 and 22 … Bf5
23.Na2 Nc2+ would ultimately drag White down to defeat.
22 …e4 23.fxe4
Of course, 23.Nxd4 e3+ followed by 24…cxd4 is horrific.
23 …Nxe2! 24.Bxe2 Bxe4
The game is over. White can’t stop …Bc2 (winning material) and/or …Kf6- e5-d4.
25.Bg4
25.Bdl Kf6 is also easy for Black.
25 …Bc2 26.KB Bdl+ 27.Ke4 Bxg4 28.hxg4 Kxg4, 0-1.
I made Cat and Mouse the first topic in Part Eight since the idea of patience will overflow into many kinds of endgames.
In the middlegame, it’s quite common to see a player saddling his opponent with as many weaknesses as possible. In general, it’s hard to tear down a defensive position if there is only one weakness to attack. However, two or more weak nesses (these can be weak squares, weak pawns, material disadvantage-anything that causes a problem for the opponent) often spread the defense enough to allow breakthroughs to occur.
Here’s a nice middlegame example that spills over into the endgame.
Diagram 366
Unzicker-Donner, Goteborg 1955
White to move
\Vhite is clearly better thanks to his passed pawn (which can push to d6 at will) and the loosening of black’s King position. However, the passed pawn alone won’t win the game-another weakness will be needed to stretch black’s defense to the breaking point.
l.a4!
Intending aS followed by axb6, weakening the b6-pawn and the bS-square. In some lines the a-pawn might also march on to a6 where, with NbS to follow, it turns a7 into a target.
l…a5?!
Too panicky! Now \Vhite has a strong passed pawn to push, a weak b6-pawn to attack, and a juicy bS -square to live on. More than enough to force a win.
363
364 SILMAN’S COMPLETE ENDGAME COURSE
2.Qb5 Qc7 3.d6! Qc6
In the actual game, Black played 3 …Qd8 and lost with hardly a whimper after 4.Nd5 Bxd5 5.Rxd5 e4 6.Radl Kg7 7.d7 Rf6 8.h3 Re6 9.Qc4 Re7 10.Rd6
Ra7 ll.Qc3+ Kg8 12.Qf6 Ra8 13.Qxe7, 1-0 since 13 …Qxe7 14.d8=Q+ Rxd8
15.Rxd8+ Kf7 16.Rld7 leaves White with an extra Rook.
The more aggressive looking 3 …Qc6 allows us to see how an “innocent” weakening of the b6-pawn can have profound repercussions on a defender’s position.
4.Qxc6 Bxc6 5.d7! Rfd8 6.Rd6 Bxd7 7.Radl
The pin is killing Black.
7…Ra7 8.Nd5
Diagram 367
a b c d e f g h
The weakness of b6 haunts black
The weakness of b6 has become extremely important since White threat ens both Nf6+ (which can be defended by …Kf7) and Nxb6 (which can’t be stopped).
S…K£7 9.Nxb6 and now after 9…Ke7 White can torture his helpless (and quite lost) opponent in several ways (lO.B, 10.b3, and 10.R6d5 all leap to mind). Clearly, the creation of a second weakness/target in the enemy camp is a far from trivial accomplishment!
An untrained eye might consider things (in diagram 368) to be fairly balanced. However, White enjoys certain advantages that decisively tip the scales in his favor. Here’s the short list:
–… Superior King position.
A passed c-pawn after a well timed c3-c4 push.
– – … Black’s a6 pawn is a target that can be attacked if white’s Rook is able to penetrate into the enemy position.
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 365
Diagram 368
8
7
6
5
4
3
2
a h c d e f g h
Rubinstein- Cohn, Carlsbad 1907
White to move
Two problems/questions must be addressed. One is whether black’s e-pawn is a strength or weakness. The other is finding a way for White to infiltrate with his Rook.
l.Rh3!
This maneuver aims to turn black’s kingside pawns into weaknesses. Defending the h- and g-pawns will force Black to make concessions which, when taking white’s other plusses into account, should prove decisive (after all, there’s only so much “bad news” a position can handle!).
l…h6
Mihail Marin, in his fantastic book, Learn From the Legends-Chess Cham pions at Their Best, gave a couple ways White can successfully deal with 1…g6. This is one of them: 2.Rxh7 Rxc3 3.Ra7 Rc6 (3…Rxb3 4.Rxa6+ Kf7 5.Rb6 also loses for Black) 4.Rg7 Kf6 5.Rb7 Rc3 6.Rb6+ Kg5 7.Rxa6 Rxb3 8.Re6 b4
9.Rxe5+ Kg4 10.Re6 g5 ll.a6 Ra3 12.Kd4 Ra2 13.Kc4 Rxg2 14.a7 Ra2 15.Re7
Kh3 16.Kxb4Kxh217.Kb3 Ra618.Re2+Kg3 19.Ra2 Rxa7 20.Rxa7Kf3 21.Rf7+ Ke3 22.Rg7 Kf4 23.Kc2 stopping black’s pawn and winning the game.
2.Rg3
The two new weaknesses on g6 and g7 have placed Black under serious pres
sure to which there is no completely adequate response.
2 …Kf6
2 …g5 is met by the “obvious” 3.Rh3, a move low in subtlety but high in strength.
3.Kd5
White’s vast superiority is now strikingly clear: White’s King is completely dominant, while black’s e-pawn has now been labeled as weak.
366 SILMAN’S COMPLETE ENDGAME COURSE
3 …Re6 4.RB+ Ke7 5.g3!
A common “don’t hurry” Rubinstein move. As discussed in the Cat and Mouse section, a player shouldn’t hesitate to make quiet, “fixing moves” if his opponent is helpless. Placing the pawn on g3 gets a pawn off the 2nd rank in case black’s Rook makes its way to d2 after … Rd6+.
5 …g6 6.Re3 Kf6 7.c4
Diagram 369
a b c d e f g h
The c-pawn finally becomes a passed monster. Black’s position has, as if by magic, become hopeless!
7 …bxc4 8.bxc4 ReS 9.Rf3+ Kg5
Hoping to gain some counterplay by attacking white’s kingside pawns.
10.c5 e4 ll.h4+ Kg4 12.Rf4+ Kxg3 13.Rxe4 Ra8
This looks strange, but 13.. .Rd8+ 14.Kc6 ReS+ 15.Kb6 was no picnic either.
14.c6 gS 15.hxg5 hxgS 16.c7 Kf3 17.Rel, 1-0. White will win black’s Rook for the c-pawn, give up his own Rook for black’s g-pawn, then finally have a meal on a6 and promote the aS-pawn.
Diagram 370
8
7
6
5
4
3
2
1
a b c d e f g h
Smyslov- Simagin, Mosow 1966
White to move
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 367
White has an annoying edge. Aside from his superior kingside pawn structure, both his Rook and Knight are more active than their black counterparts. The main target is the b6-pawn, but before laying siege to it, White fixes black’s kingside pawn weaknesses.
l.g4!
A huge move! Now the f6- and h7-pawns are long-term additions to black’s
“list of problems.”
l…h5
Black can’t wait for White to play Rc6 followed by Ke4-d3-c4-b5 with a total rout.
2.gxh5 Kh6 3.Kg4 f5+ 4.Kh4 f4
Trading off one of his potentially weak pawns makes good sense.
5. ReS fxe3 6.fxe3 Rd5 7.Rh8+ Kg7 8.Ra8 Nc3 9.Ra7
Threatening Nxe6+ but also preparing an attack against black’s King by Nf3 and Ng5.
9…Re5
This stops Nxe6+ but gives White time for his kingside dreams.
10.Nf3 Re4+
And not 10… Rxe3 ll.Ng5 when f7 can’t be defended.
ll.Kg5 Rxe3?
Black gets overwhelmed after this natural but poor move. Far better resistance was offered by ll…Nd5.
12.h6+ Kg8 13.h7+ Kxh7 14.Rxf7+ Kg8
Diagram 371
8
7
6
5
4
3
2
a h c d e f g h
Both sides only have two pawns left, so one might expect Black to have real chances to survive. However, his King is in serious trouble and-by mixing
368 SILMAN’S COMPLETE END GAME COURSE
threats to black’s King and pawns-White can eventually pick up the bit on b6 and walk away with a won Rook endgame.
15.Kg6 Ne4
Black needs to have his Knight participate in the defense, but the jump to e4 doesn’t get the job done. Sadly, neither does 15 …Nd5 16.Ng5 eS (Losing instantly is 16…Nxb4 17.Nh7 when the threat ofNf6 mate might prove a tad annoying) 17.Ne6 (the threat ofRf8 mate forces black’s reply) 17 …Nf4+ 18.Nxf4 exf4 19.Rxf4 Rh3 20.Kf5 Rxh2 21.Ke6 Rd2 22.Rf5 with a decisive advantage since White can eventually pick up black’s b-pawn by Rd5, Kd6, b4-b5, and Kc7xb6.
16.h4 Nd6 17.Rd7 Rxf3 18.Rxd6 Kf8 19.Rxe6 Rg3+
As good as resignation is 19…Rf4 20.Rf6+.
20.Kf5 Rgl
20… Rh3loses to 2l.Kg5 (2l.Kg4 is also good) 2l …Rg3+ 22.Kh5 b5 23.Re5 with White owning two extra pawns.
21.Rxb6, 1-0.
Our final two examples show how “fishing for new weaknesses” is a common strategy even when material ahead. In the first, Alekhine turns what might be a complex technical task into something so smooth that it stuns the eye.
Diagram 372
a h c d e f g h
Alekhine- Samisch, Baden Baden 192 5
White to move
White has an extra passed b-pawn but doesn’t wish to push it right away since that would expose his King. Instead, he patiently plays to create a second weak ness in black’s position.
l.Qd4
Alekhine: “By this and his following move White selects the correct winning plan, which is the advance of his kingside pawns. The passed b-pawn must be
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 369
advanced only later, when with the exchange of Queens the danger of perpetual check will be eliminated.”
l…Qe7 2.Bd3! Qc7 3.g4 Kfi
Of course, 3 …Qxh2?? loses the Knight to 4.Qd7+.
4.h4 Nb6 5 .h5 gxh5 6.gxh5
A second weakness has been created-the h7-pawn.
6…Qc6 7.Be4!
White could have exchanged Queens with 7.Qe4 but decided to only do so when the h7-pawn was fixed, thus forcing the black King to stay on the kingside and baby-sit the weak h-pawn.
Note that White didn’t fall for 7.Bxh7 QxB 8.Qxb6 Qdl+ 9.Ka2 Qa4+ with a draw by perpetual check.
7 …Qb5 8.h6
Now the h7-pawn is stuck on a light square, where it will be forever vulner
able to the attentions of the Bishop.
8…Qb3
Also bad is 8…Qfl+ 9.Kc2 Qe2+ 10.Kc3 Qel+ ll.Kb3 Qa5 12.Bxh7 Qb5+
13.Qb4 Qd5+ 14.Kc2 Qc6+ 15.Qc3 and Black’s done since 15 …Qa4+ is killed by 16.Qb3+.
9.Bc2! Qb5
Diagram 373
8
7
6
5
4
3
2
a h c d e f g h
10.Qd3
Only now is White ready to force the trade of Queens. What’s the difference between this position and the position after 6…Qc6 (when 7.Qe4 would have forced this exchange)? In the earlier position black’s h-pawn could have advanced to the safe square on h6, which would have allowed black’s King to rush over to the queenside and help stop the b-pawn. Now the h7-pawn is a permanent target and black’s King is stuck doing guard duty in Siberia.
370 SILMAN’S COMPLETE ENDGAME COURSE
By the way, 10.Bxh7 also wins, but why allow Black to go bonkers with checks when you can trade Queens and win without any bother at all?
10…Qxd3 ll.Bxd3
White has achieved the exchange of Queens in the most favorable circum stances: the black King is tied to the defense of the h-pawn and the Knight is quite unable to cope with the passed b-pawn supported by King and Bishop.
ll…Nc8
Also hopeless was 11…Kg8 12.b4 when 12 …Nd5 13.Bc4 picks up the pinned Knight, while 12 …Nc8 13.b5 with Kb2-b3-b4-c5 to follow is not worth play ing out.
12.Bxh7, 1-0.
Amazing what a second weakness in the opponent’s position can do, isn’t it. Our last example is a very smooth victory by the 5th World Champion where
we see simplification, a bit of Cat and Mouse, and the creation of more than
one new weakness (the more the merrier!).
Diagram 374
8
7
6
‘)
4
3
2
1
a b c d e f g h
Euwe-van Doesburgh, Dutch Championship 1938
White to move
Before moving on, I can imagine some of you screaming, “Wait a second! White’s a pawn up and should be winning easily. Why bother using such an example?”
This is a fair question, but I can answer it with a question of my own: “How often have you (meaning all of us) failed to convert such a decisive plus?”
The beauty of this game is that White doesn’t rest on his laurels, nor lose his concentration. In fact, he does what we all should learn to do in such a situation: Don’t give the opponent any counterplay whatsoever and constantly strive to make new positional gains in the form of multiple enemy weaknesses.
l.Qc5!
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 371
This more or less forces Black (who would love to retain the Queens and keep things as complicated as possible) to accept the trade of Queens, since both 1…a6 and 1…b6 run into 2 .Qe5! when the Rook is pinned to its King.
l…Qf5 2.Qxf5 Rxf5 3.Kfl
Wisely bringing the King into the center. Absolutely horrible is 3.Rc5?? (What’s the hurry?) 3…Rxc5 4.dxc5 Kf6 5.f4 (Trying to prevent …Ke5/…Kd5/…Kxc5)
5 …Kf5 (and not 5 …Ke6? 6.e4) 6.Kf2 Ke4 when Black’s the one in charge.
3 …Kf8 4.Rc5
Now White is happy to transpose into a King and pawn endgame, but this time Black must avoid it since a pawn on c5 could now be supported.
4…Rf6
4…Rxc5 5.dxc5 Ke7 6.Ke2 Ke6 7.Kd3 Kd5 8.b4 aS 9.e4+ Ke5 10.bxa5 is winning for White.
5.Ra5!
A strong move that forces the creation of a hole on b6 that, after a4-a5, will freeze black’s queenside pawns. Note that White isn’t simply going to try and win by using his extra pawn. Instead, he wants to create as many weaknesses as possible in the enemy camp until the defense simply falls apart. You can never give the opponent too many weaknesses!
5 …a6 6.Re5
Stopping the enemy King from coming towards the center. Note how patient White is, how he is slowly playing for new enemy weaknesses, and how he is going out of his way to prevent enemy counterplay.
6…Rd6
6…b6 7.a5 b5leaves the c6- and a6-pawns very weak, and the c5-square ready
to accept the white King as a squatter.
7.a5
Freezing black’s pawns on a6 and b7. If Black plays passively, white’s King will calmly walk all the way to b6.
7…f6 8.Rc5 Ke7 9.g4
Now that Black is more or less helpless, White begins a new plan: he intends to push this pawn to g5, exchange it for the f6-pawn, and create a passed e-pawn.
9…h6 10.Ke2 Re6 ll.f4 Rd6 12.h4 Kf7 13.h5!
13 .g5 hxg5 14.hxg5 f5 15.b4 is another way to win since his plan of marching the King all the way to b6 is hard to deal with, while a well-timed d4-d5 push is also in the air.
372 SILMAN’S COMPLETE ENDGAME COURSE
13 …gxh5
No better is 13 …g5 14.fxg5 fxg5 (14…hxg5 leaves White with a powerful passed h-pawn) 15.Rf5+ (Trapping the King on one side or the other.) 15…Ke7
16.Kd3 when e3-e4 gives Black few chances to resist.
14.Rxh5
Suddenly Black is not only a pawn down, but also must care for weaknesses on h6, f6, and b7!
14…Kg6 15.Kd3 Rd7 16.e4 Kh7
Things were looking grim, so Black allows a little sequence that leads to the exchange of a few pawns. However, this fails to ease black’s pain.
17.g5 fxg5 18.fxg5 Kg6 19.Rxh6+ Kxg5
Black no doubt was happy that he has managed to trade off his weak pawns on f6 and h6; the defender is usually delighted to exchange pawns. Unfortunately, this position is, in reality, dead lost for Black.
20.Rh2!
Threatening Rf2, trapping the black King away from the action.
20…Kf6
No choice.
21.e5+ Ke6 22.Rh6+ Kd5
Again, there was no choice. 22 …Ke7 23.Rh7+ trades Rooks and enters a won King and pawn endgame, while 22…Kf5 23.Rf6+ forces the black King to no man’s land.
23.b4
Diagram 375
a b c d e f g h
Another “No counterplay allowed!” move. Now that …c6-c5 has been stopped, Black finds himself in zugzwang: any pawn moves loses material, a King move is
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 373
illegal, any move of the Rook away from the d-file allows Rd6 mate, and moving back to d8 (which is all he has) let’s white’s Rook target b7.
23 …Rd8
Forced. The Rook must stay on the d-file since moving away allows Rd6 mate.
24.Rh7 Rh8 25.Re7!
Accurate to the end. Black must either give up his b-pawn or push it.
25 …h6
25…Rh8 is met by 26.Rd7+ (Don’t give Black counterplay after 26.Rxb7 Rh3+)
26…Ke6 27.Rxb7 when Black should resign.
26.Rd7+ Ke6 27.Rd6+, 1-0.
One would think that King and pawn endgames would be easy to master since all the pieces have been traded. Doesn’t the lack of mutual armies make things far less complicated? The answer to this is a resounding “no!” In fact, many King and pawn endgames are so complex that they even boggle the minds of the world’s elite players.
Due to this complexity, masters know that entering into any King and pawn
endgame is an “all or nothing” decision. Leaping into one is perfectly okay (and often the only way to win or draw a game), but make sure you have a firm handle on the resulting position before doing so!
The following examples are meant to do two things: Teach you to appreciate their beauty.
Act as a warning. Consider it a “Danger Will Robinson,
danger!” kind of thing. Don’t ever enter a King and pawn endgame unless you have no choice, or are sure that it leads to the desired result.
Diagram 376
H
7
6
5
4
3
2
a h c d e f g h
Shirov- Grischuk, New Delhi 2000
Black to move
374
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 375
Okay, I can hear the outraged protests. This isn’t even close to being a King and pawn endgame! True. But it turns out that one could occur by force after black’s sacrificial move.
l…Bg4!? 2.hxg4
Mter pondering the upcoming King and pawn endgame, Shirov decided on
2.Qb7 in the actual game and ended up winning, though the Queen move didn’t give him any advantage whatsoever. It turns out, though, that 2 .hxg4 wins by force. But, as Shirov admits, it took him three years and computer assistance to finally prove this to himself!
2…Qxg4 3.Rfl f3 4.Qa2 Qh4+ 5.Kgl Qg3+ 6.Rg2 Qxg2+ 7.Qxg2 fxg2
8.Rxf8+ Kxf8 9.e6 g5 10.Kxg2 h5
Diagram 377
8
7
6
5
4
3
2
1
a b c d e f g h
We’ve reached the position that Shirov couldn’t quite solve during the game. It’s clear that white’s only hope of winning is to create a pawn mate, a concept we’ve already looked at in Strange Races in Part Six. To accomplish this, White needs to reach one of the following two positions:
Diagram 378
8
7
6
5
4
3
2
a b c d e f g h
White wants Black to have the move
376 SILMAN’S COMPLETE ENDGAME COURSE
Diagram 379
a b c d e f g h
To win, White needs the move
The following tempo moves (from diagram 3 77) are all designed to make one of these two diagrammed positions occur.
ll.e7+ Ke8 12.Kh2!
A key Triangulation maneuver. The straightforward 12 .Kf3? h4 13.Kg4 Kf7 leads us to diagram 3 78, where White wins if Black has the move, but only draws if White has it (which is the unfortunate case here). The note to black’s 12th move will demonstrate why this is so.
12 …Kf7
12 …h4 13.Kh3 Kf7 14.Kg4 gives White the position we labeled as winning in diagram 3 78. Black has the move and must step back to the fatal e8-square.
14…Ke8 15.Kxg5 h3 (else White would take the h-pawn) 16.Kf6 h2 (16…Kd7
17.Kf7 and white Queens with check) 17.Ke6 hl=Q 18.d7 mate.
13.Kg3 Ke8 14.Kg2!
Diagram 380
H
7
‘5
4
3
2
a b c d e f g h
Zugzwang!
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 377
Incredibly, this seemingly innocuous move leaves Black in zugzwang! We’ve already seen that 14…h4 fails to 15.Kh3 Kf7 16.Kg4 when Black loses since he’s left with the “fatal” move. However, a King move or pushing the pawn to g4 both allow White to create our second key diagram (3 79).
14…g4 15.Kg3 Kf7 16.Kf4 Ke8 17.Ke5 Kf7
Ke6 can’t be allowed.
18.Kd5 g3 19.Kc6 g2
No choice. 19…Ke8 20.d7+ Kxe7 2l.Kc7 g2 22.d8=Q+ is easy for White.
20.Kd7 gl=Q 2l.e8=Q+ and White wins. Sure enough, after 2l…Kg7
22.Qe5+! Kg6 3.Kc7 Black can’t prevent White from promoting his d-pawn.
Though this game might seem unfathomable to some, at least the basic ideas (pawn mates and queening with check) are easy to grasp. Our next example is even more outrageous.
Diagram 381
H
7
6
5
4
3
2
a h c d e f g h
Martz- Benko, Torremolinos 1973
Black to move
Black is a pawn up and he has the move, so how hard can this be? Judge for yourself as we follow Benko’s comments from his book, Pal Benko: My Life, Games, and Compositions:
“So we find ourselves in an interesting King and pawn endgame. This kind of ending rarely occurs in modern tournaments, so I was delighted to get the chance to play such a position.
“Here the game was adjourned and I had to seal my move. A quick calcula
tion convinced me that 41…Kxc5 would be a blunder: 42 .Kf7 b5 43.Kxg7 h5
44.Kg6 Kb4 45.Kxh5 c5 46.Kg6 c4 47.bxc4 bxc4 48.h5, with a draw. I must admit that, in the little time I had left, my first plan was 41… Ke4, when 42.Kf7
Kf5 43.Kxg7 h5 would be an easy win. Also winning for me is 41 …Ke4 42 .Kd6 a5 (or 42 …h5 43.a3 Kd4) 43.Kc7 Kd5 44.Kb6 g6 45.a3 h6 46.b4 axb4 47.axb4
378 SILMAN’S COMPLETE ENDGAME COURSE
Kc4 48.Kxb7 Kb5!. Then I noticed that 41…Ke4 42.b4! created some difficul
ties for me: 42 … h5! 43.a4! and now:
Diagram 382
8
7
6
5
4
3
2
a b c d e f g h
Black to move
—- 43…Kd4 44.Kd6 (44.Kd7? a5!) 44…Kc4 45.Kc7 Kxb4
46.Kxb7 Kxc5 47.Kxa7 (47.a5? Kd6! 48.Kxa7 Kc7 49.Ka6 g6!
wins) 47 …Kb4 48.Kb6 Kxa4 49.Kxc6 is a draw.
—- 43…Kf4 44.b5! g5 45.hxg5 Kxg5 46.b6! (46.a5!? a6!) 46…axb6
47.cxb6 h4 48.a5 isn’t what Black wants. However, Black
can improve his chances with 46…a5! (instead of 46…axb6)
47.Kd7 h4 48.Kc7 h3 49.Kxb7 h2 50.Ka8 hl=Q 5l.b7 Qd5
52.b8=Q Qxc5, and now White can force a draw with 53.Qb6! Qc3 54.Kb7 c5 55.Kc6 c4 56.Qb5+ Kf4 47.Kc5, =.
—- 43…a6! 44.a5! Kf4 45.Kd7 g5 46.hxg5 Kxg5 47.Kc7 h4
48.Kxb7 h3 49.Kxa6 h2 50.Kb7 hl=Q 5l.a6 and a strange position has come about:
Diagram 383
a b c d e f g h
“White’s pawn only stands on the 6th rank (which usually means that it’s time to resign) but, despite this, Black has to show a bit of care if wants to reel in the
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 379
point: Sl…Qbl? (Correct is 51…Qh7+ 52.Kb8 Qg8+ 53.Kb7 Qf7+ 54.Kb8
Qf4+! and wins) 52.a7 Qxb4+ 53.Kc7 (not 53.Kxc6?? Qe4+ followed by 54…Qa8. Also bad is 53.Kc8?? Qxc5 54.a8=Q Qf8+ 55.Kb7 Qxa8+ 56.Kxa8 c5 and wins)
53 …Qa5+ 54.Kb7 and Black doesn’t have time to take the pawn on c5.
“The position after 43.. .a6 is full of interesting variations, e.g., 44.a5! Kd4 (instead of 44…Kf4) 45.Kf5 Kc4 46.Kg6 Kxb4 47.Kxg7! Kxc5! 48.Kg6 Kb4
49.Kxh5 c5 50.Kg4 c4 51.h5 c3 52.h6 c2 53.h7 cl=Q, and Black should win.
After considering all of this, I came to the conclusion that I must find an easier solution. As it happens, there is one in the position.”
The move Benko found was 41…g6!! when there followed 42.Kf6 Ke4!
43.Kg7 Kf4 44.Kh6 Kg4 45.Kxh7 g5! 46.hxg5 Kxg5 47.Kg7 Kf5 48.Kf7
Ke5 49.Ke7 Kd5 50.Kd7 Kxc5 51.Kc7 b5 52.Kb7 a5 53.a3 b4 54.a4 Kd4
55.Kxc6 Kc3 56.Kb5 Kxb3 57.Kxa5 Ka3, 0-1.
I don’t know about you, but I have two thoughts whenever I look at this end game. One is, “It’s very beautiful.” The other is, “My god! I would never have found any of that over the board!”
The message should be clear: never enter a pawn endgame lightly!
Transposition into a pawn ending often marks the final stage of the opponent’s demise. However, masters have learned to be very wary of all King and pawn endgames since strange, often unforeseen, complications can turn an easy win
into a draw or even a loss!
Diagram 384
8
7
6
5
4
3
2
1
a h c d e f g h
Aronin-Smyslov, Moscow 1951
White to move
Most experienced masters would note the possibility of entering an appar ently winning King and pawn endgame with l.Rg8. However, their spider sense would tingle and they would hold off on such mass exchanges. \Vhy should they trade? Their Knight is a monster, their pawns are safe, and their Rook is active. Compare that with black’s garbage Bishop, his passive Rook, and his vulnerable pawns.
380 SILMAN’S COMPLETE ENDGAME COURSE
One would expect such a strong player as Aronin to only go into a winning King and pawn endgame after tidying up the position. Thus, l.Ke2, stopping black’s Rook from penetrating into white’s camp and also centralizing white’s King, would easily win the game.
l.Rg8? Kh7 2.Rxg7+??
White completely missed black’s brilliant defensive trap. It’s easy to miss! But why go into something that might turn complex when simple moves keep the win safely in hand? For example: 2.Rb8 c5 3.Rc8 Kg6 4.Ke2 (the c-pawn isn’t going anywhere) when Black will lose material without having any counterplay whatsoever.
2…Rxg7 3.Nxg7 Kxg7 4.g4
This was white’s idea. It seems to end all kingside counterplay for Black since leaving the pawn on g4 results in a winning endgame for White, while taking en passant gives White a “winning” outside passed pawn. It is winning, isn’t it?
4…hxg3!
I guess it isn’t! White had evidently hoped for 4…Kf7 5.Ke2 Ke6 6.Kd3 Kd6
7.Kc4 aS 8.c3 bxc3 (8…c5 9.Kb5 is no better) 9.bxc3 Kd7 10.Kc5 Kc7 ll.c4
Kd7 12.Kb6 and wins.
5.fxg3 g4!! 6.h4
Diagram 385
8
7
6
5
4
3
2
1
a b c d e f g h
A positional draw
White has nothing after 6.hxg4 Kh6 7.Kf2 Kg5. However, it does seem that white’s outside passed pawn should bring easy victory. Incredibly, this doesn’t prove to be the case.
6…c5!
An important move that deprives white’s King of the d4-square once the e5- pawn moves forward. Okay, this doesn’t make sense now, but it will soon!
7.Ke2 Kh7! 8.Kd3 Kh6!
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 381
Black avoids the g6-square since, in the tactical lines that might follow, …f6-f5 would allow an unfortunate check after exf5.
9.c3
9.Kc4?? actually loses to 9…f5! 10.exf5 (It’s already too late to defend: 10.Kd3 f4 ll.gxf4 exf4 12.Ke2 Kh5 13.e5 Kg6 wins) 10…e4! ll.c3 a5 12.f6 Kg6 when the e-pawn turns into a Queen.
9…a5 10.cxb4 axb4, 1h- 1h. Mter ll.Ke2 Kh7 White can’t make progress. Are you convinced yet that almost any King and pawn endgame might turn
out to be a minefield that you should think twice about before walking through?
No? Yes? Let’s look at a game won by IMJack Peters. The transition into a King and pawn endgame was the perfect solution here, but the calculation necessary to prove this is far from easy in an over the board situation.
Diagram 386
8
7
5
4
3
2
a h c d e f g h
S Hughes-J Peters, Los Angeles 1994
Black to move
Queen and pawn endgames are notoriously difficult so, in the interest of mak ing his life easier, Black must consider the very forcing 1…Qf3 when it’s clear that, in the ensuing King and pawn endgame, both sides will eat each other’s pawns. After a deep count it might seem that each side will promote with a resulting Queen and pawn vs. Queen endgame. Is that what Black wants, or is something else going on here?
l…Q£3!
It turns out that this is a dead win! But that decision can only be made if Black notices a subtle point at the end of a long, forcing sequence.
2.Qxf3
White doesn’t have much choice. 2.Qh4 Kxb3 is hopeless, 2.Qc7 Qxf2+ is even worse, and 2.Qgl Qh3 mate is the worst of all!
2…gxf3 3.Kg3 Kxb3 4.Kxf3 Kxa4 5.Kg4 Kb5!
382 SILMAN’S COMPLETE ENDGAME COURSE
Getting out of the way of the a-pawn and simultaneously preparing to stop white’s pawn with the King.
6.f4 Kc6 7.Kg5 Kd7 8.Kg6 a4 9.f5 a3 10.f6 a2 11.£7 Ke7!, 0-1. This was seen by Black when he played l …Qf3. White resigned since 12.Kg7 is met by
12…al=Q+.
Such transitions from “messy” positions to the supposed serenity of a King and pawn endgame can be a blessing or a curse. In the following example White had to decide whether or not to leap into unknown King and pawn waters.
Diagram 387
8
7
6
5
4
3
2
a b c d e f g h
Akiko Uto- Tran, Paris 2005
White to move
In the game, White avoided the King and pawn endgame and took a walk down the wrong path with l.Nd7+. The real questions concerning this position are:
…. Is the King and pawn endgame worth considering after l.Nxd5?
If White retains the Knight, who is better and why?
The King and pawn endgame that arises after l.Nxd5 exd5 2.g4 proves to be extremely interesting.
Diagram 388
8
7
6
5
4
3
2
a b c d e f g h
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 383
It seems that white’s plan of creating a passed e- and h-pawn (which will be unstoppable due to their distance from each other) will give her the full point:
2 …d4 3.h4 (and not 3.f5?? g5 when the g-pawn stops two white pawns) 3.. .Ke7
4.f5 followed by h5 when Black must resign.
Is it really that simple? Of course not! It’s a pawn endgame and that means there are almost always surprising resources lurking behind the scenes. So let’s try (after 2.g4) 2 …g5. A bit of analysis makes it clear that this is a dismal failure:
3.fxg5 Ke7 (or 3…Kg7 4.h4 Kg6 5.Kc3 Kg7 6.h5 Kh7 7.g6+ fxg6 8.e6 Kg7 9.h6+ and a pawn will queen) 4.h4 d4 (4…Ke6 5.h5 Kxe5 6.h6 forces the promotion of the h-pawn) 5.h5 Kf8 6.h6 Kg8 7.g6! fxg6 8.e6, 1-0.
Can Black try …f7-f6 at some point? Let’s see: 2.g4 Ke7 3.h4 f6 4.h5! and we have a very pleasing pawn structure that guarantee’s \iVhite success: 4.. .g5 5.h6
Kf7 6.e6+!, winning.
How about 2.g4 f6! (threatening to win by 3…fxe5 4.fxe5 g5!, fixing two pawns with one) when 3.e6 Ke7 4.f5 g5, =.Are we on to something? Suppose
\iVhite chops off the pawn after 2 …f6 3 .exf6 Kf7 4.g5. This paints a nice visual
picture, but it’s only a draw after 4…d4 5.Ke2 c3 6.Kd3 Ke6-black’s c- and d-pawns handcuff white’s King, while white’s four kingside pawns are stopped by black’s King and pawn.
It seems we might have found a defense for Black, but one last calm look will give us the “answer”: 2 …f6 3.h4! fxe5 (3…Ke7 is met by 4.h5) 4.fxe5 Ke7 5.h5 and \iVhite’s back in the win column.
At this point I can see \iVhite snapping off black’s Bishop and rushing into this endgame, confident that victory will be hers. However, we’ve missed something! Take one final look at the position after 2 …f6 3.h4
Diagram 389
8
7
6
5
4
3
2
a b c d e f g h
Black to move and save himself
It turns out that 3 …g5!! saves the game:
4.hxg5 fxe5 5.fxe5 d4 6.a4 Kf7 7.Kc2 Kg6 8.e6 Kg7, =.
384 SILMAN’S COMPLETE ENDGAME COURSE
4.exf6?? gxh4 and Black wins.
.._ 4.h5 gxf4 (4…fxe5?? 5.fxe5 wins for White) 5.e6 B 6.h6 c3+
7.Kc2! (This allows White to follow up with Qc8, an idea that isn’t possible after 7.Kxc3 f2 8.h7 fl=Q 9.h8=Q+ Ke7 10.Qc8?? [10. Qg8 was correct]lO…Qcl+ picking up the Queen) 7 …f2 8.h7 fl=Q 9.h8=Q+ Ke7 10.Qc8! Qe2+ when only Black has chances.
.._ 4.fxg5 fxe5 and Black wins ifhe can reach the following key position with White to move:
Diagram 390
a b c d e f g h
Whoever moves loses
White to move: l.Kc3 e3 2 .Kc2 d4 and black’s pawns are like a tidal wave. Black to move: l…Kg8 2.h6 Kh8 3.g7+ Kh7 4.g6+ Kg8 5.Ke5 e3 6.Kf6 e2
7.h7 mate.
It turns out that Black can indeed achieve the sunny side of the diagram after
5.Kc3 Kf7 6.h5 e4 7.Kd4 a4 8.g6+ Kg7! (and not 8…Kf6?? 9.g5+ Kg7 10.a3 and White wins, as shown in the diagram above) 9.a3 Kf6! (Avoiding 9…Kh6??
10.Ke5! e3 ll.Kf6 e2 12.g7 el=Q 13.g8=Q) 10.g5+ Kg7 and White possesses the fatal move and, as a result, loses.
So it turns out that White can’t do more than draw the King and pawn end
game, once again confirming our adage that “all King and pawn endgames are confusing!” Instead, returning to diagram 387, an experienced player would quickly look over the mess we just explored and then refuse to make the ex change, even if he wasn’t able to plumb the King and pawn endgame’s depths. Why would he make this decision? Because it should be obvious that the minor piece endgame gives White a huge advantage with no risk whatsoever.
From diagram 387:
l.Kc3
White fixes the c-pawn on c4 and ties down black’s Bishop to its defense.
l …Kg7
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 385
Moving towards the middle doesn’t help: l…Ke7 2.g4 Kf8 3.h4 Ke7 4.h5 gxh5 5.gxh5 Kf8 6.h6.
2.h4!
This not only gets white’s kingside majority into motion, but also stops all
…g6-g5 based counterplay.
2…Kh6 3.g4 Kg7 and now White has many ways to win the game. She can take her time and play “fixer” moves like 4.a3, or she can go into a winning King and pawn endgame by 4.Nxd5 exd5 5.f5 (this is far superior to the King and pawn position we entered earlier-White will follow with h5 and promote one of her pawns), or White can go right for the win of material via 4.g5 Kf8
5.Nd7+ Ke7 6.Nb6 when Nxc4 can’t be stopped.
After looking over this endgame, World Championship candidate Yasser Seirawan said: “My thought process would be that if the King and pawn ending was ‘close,’ White should start with l.Kc3 in order to always get the King and pawn ending she wants by handcuffing the d5-Bishop to the c4-pawn. As Black
‘passes,’ White advances her kingside pawns to g4 and h4, getting a position
where black’s King is on g7 before playing f4-f5. Then she can transpose into the King and pawn ending.”
Note that Seirawan made it clear that he would have no interest in entering
a “close” position from a position of strength. In other words, why take risks when you can step into the King and pawn endgame on your own (highly fa vorable) terms?
We’ve seen players of all strengths having trouble correctly assessing some very complicated King and pawn endgames. However, it’s not uncommon for even world-class players to botch positions that should be well within their range to solve. Our next game, between the legendary Miguel Najdorf and World Championship contender Henrique Mecking, is winning for Black, but he doesn’t come close to winning it!
Diagram 391
8
7
6
5
4
3
2
1
a b c d e f g h
Najdorf-Mecking, Wijk aan Zee 1978
Black to move
386 SILMAN’S C 0 M P LE TE END GAME C 0 U RSE
In the actual game, Black tossed certain victory away by playing 1…e3+?? 2.Ke2
Kc4, but after 3.h5 and the game was drawn since 3…Kxc3 loses to 4.g5. Instead, l…exf3! 2.Kx:f3 Ke5 wins for the following reasons:
Diagram 392
a b c d e f g h
White to move
—- At some point White will be forced to advance his kingside pawns, whereupon Black will be able to munch on them with his King.
…. Trying to hold Black off with King moves (such as Kf2) fails since …Ke4 followed by …f3 and …Kf4 still picks off the kingside pawns.
Black has two tempo moves on the queenside to white’s one. This is of critical importance!
J.K£2 Ke4 4.h5
Lines like 4.g5? Kf5 5.Kf3 a6 6.a3 c6 7.Kf2 Kg4 and 4.Ke2? f3+ 5.Kf2 Kf4
6.g5 Kg4 and 4.Kfl? Kf3 5.g5 Kg4 are even easier for Black.
Diagram 393
H
7
(J
‘5
4
5
2
a b c d e f g h
One move wins
After 4.h5 we have a “Black to move and win” situation.
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 387
4…Ke5!
I can only guess that Mecking missed this move. Other tries don’t get the job done:
…. 4…h6?? 5.g5! wins for White!
….. 4…8 5.g5 KfS 6.g6 hxg6 7.hxg6 (and not 7.h6?? Kf6 S.KxB Kf7 9.Kf4 Kg8 10.Kg4 Kh7 ll.Kg5 a6 12.a3 c6 and, thanks to that extra queenside tempo move, White has to drop his
h-pawn and will lose the game) 7 …Kxg6 S.KxB Kf5 9.Ke3
Ke5 10.Kd3 Kd5 ll.Kd2 when ll…Kc4 12.Kc2 and ll…c5
12 .Kd3 both get nowhere for Black.
…. 4…a6 S.Kfl! (S.Kel is the same thing since S …KB 6.g5 Kg2
7.g6 hxg6 8.hxg6 f3 9.g7 f2+ 10.Kd2 fl=Q ll.g8=Q+ would make White happy) 5 …Ke5 6.Kf2 and suddenly black’s win is gone since he has tossed away one of his precious queenside
tempo moves: 6…Kf6 7.KB Kg5 8.h6! Kxh6 9.Kxf4, =.
5.KB
Nothing saves White: 5.g5 (or 5.Ke2 Kf6 6.KB Kg5 7.a3 a6 8.h6 c6 and White loses all his kingside pawns.) 5…Kf5 6.g6 hxg6 7.h6 Kf6 S.KB g5 and it’s all over.
5…h6! 6.a3 a6 7.Kf2 Kf6 8.Ke2 Kg5 9.KB
Now, after black’s quiet push with his c-pawn, we have a “whoever move’s loses” situation. Thanks to that final queenside tempo move, Black gets the prize.
9…c6 and it’s time for White to resign.
Diagram 394
8
7
6
5
4
:3
2
1
a h c d e f g h
Bacrot- Kramnik, Dortmund 2005
White to move
White is two pawns up and is clearly winning, yet claiming a theoretical win and actually proving it are often two different things. In the present situation
388 SILMAN’S COMPLETE ENDGAME COURSE
\Vhite has many good moves (l.Rb5 and l.h4 are good choices), but offering a Rook exchange by l.Re5 is extremely tempting since, if the resulting King and pawn endgame is indeed a win, it would avoid the potential complications of a long, drawn out Rook endgame.
Thus we come face to face with a common question: does entry into the King endgame make life easier, or are we setting ourselves up for some strange King and pawn endgame pitfall?
l.Re5!
Best, but the two sets of doubled isolated pawns will demand some skill from
\Vhite before he gains the full point.
USEFUL ADVICE
If you find yourself facing a decision like this and you have any doubts about the King and pawn endgame that results, avoid it like the plague!
1…Rxe5 2.fxe5 g5
Stopping white’s King from penetrating and also gaining counterplay by set
ting his own pawn majority into motion. Completely hopeless is 2 …Ke7 3.Kf4
Ke6 when both 4.h4 h6 5.c5 Kd5 6.h5 gxh5 7.Kxf5 Kxc5 8.e6 Kd6 9.Kf6 and
4.c5 Kd5 5.c4+ Kxc5 6.Kg5 win quickly.
3.c5
The material in Part Five should have already taught you that side-by-side passed pawns, separated by one file, guard themselves. For example, 3…Ke6 is met by 4.c6 when 4…Kxe5 5.c7 leads to a new Queen for \Vhite.
3…h6 4.c4
4…h5
Diagram 395
8
7
6
5
4
3
2
a b c d e f g h
Trying a waiting game by 4…Kd8 also fails: 5.Ke3 Ke7 (5 …Kc7 6.e6 is easy while 5…Kd7 6.Kd4 intending Kd5 will shortly force Black to lay down his arms: 6…Ke6 7.c6 f4 8.c7 Kd7 9.e6+ Kxc7 10.Ke5 Kc6 ll.e7 Kd7 12.Kf6 Ke8
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 389
13.Ke6 when the advance of the c-pawn leads to mate in six) 6.c6 Ke6 7 .cS Ke7
8.Kd4 g4 9.Kd5 f4 10.e6 (10.Ke4 g3 ll.hxg3 fxg3 12.Kf3 is another way to go about it) 10…f3 ll.exf3 gxf3 12.c7 f2 13.c8=Q fl=Q 14.Qd7+ Kf6 15.Qt7+ Kg5 16.Qxfl, 1-0.
5.h4 g4+
Black gets no hope at all from 5 …gxh4 since 6.Kg2 followed by 7.Kh3 and
Kxh4 is ridiculously easy.
6.Kf4 Ke6
Similar lines to the actual game occur after 6…Kd8 7.c6 Kc7 8.e6 Kxc6 9.c5
Kc7 10.e3 Kc6 (10…Kd8 ll.c6) ll.Ke5.
7.c6 Ke7 8.c5 Ke6
Black has no choice since 8…Kd8 9.e6 forces a new Queen in a clear, simple manner.
9.c7!
Diagram 396
a b c d e f g h
The key to the position. White gives up this pawn so his e-pawn can march down the board with tempo.
9…Kd7 10.e6+ Kxc7 ll.e3!
Most accurate. Since 11…Kd8 12 .c6 loses immediately, Black must move his
King to c6 where it will be in check if the pawn promotes on e8.
ll…Kc6 12.Ke5!
This leads to a winning Queen endgame.
Though 12.e7?? Kd7 would win for Black, White could have played 12.Kxf5. He decided to leave that pawn on the board since, in some lines, it would block checks on the f-file by black’s soon-to-be-born Queen.
12 …g3 13.e7 Kd7 14.Kf6 g2 15.c6+ Kc7 16.e8=Q gl=Q 17.Qd7+ Kb6
18.c7, 1-0. Black would run out of checks after 18…Qa1+ (18…Qxe3 19.c8=Q Qh6+ 20.Kxf5) 19.Kg6 Qg1+ 20.Kh7.
390 SILMAN’S COMPLETE ENDGAME COURSE
Our final example shows me entering an extremely complex King and pawn endgame simply because it was my only chance for survival!
Diagram 397
a b c d e f g h
Milat-Silman, National Open 1998
Black to move
My opponent had been thrashing me for the whole game (White had missed at least one instant win), and the creation of various perpetual check themes was my last line of defense. After offering my opponent several bottles of free beer, which he refused (seemingly intent on winning the game), we reached the position in the diagram.
Black has a choice between two legal moves, 1 …Kd7 and 1…Ke7. Only one
is playable.
l…Ke7!
The only good move, and it shows how selection between seemingly similar choices can have an enormous impact on a game’s result. When faced with this kind of decision, you can’t take it lightly. It’s very important to hunker down and try and discern what the differences are, and how those differences will ultimately affect your chances. In this case, 1…Kd7?? would have led to a lost King and pawn endgame after 2.Qe6+! Qxe6 3.fxe6+. Note that this move comes with check, meaning that Black must deal with the pawn immediately. Unfortunately,
3 …Kxe6 allows 4.g4 when Black would have a very inferior version of the actual
game since black’s kingside pawn majority would be crippled.
2.Qe6+!?
At the time I was positive that this allowed me to save myself, though much of that view was based on ignorance. I was more worried about 2.Qxh7+ Kd6
3.Qb7 though I was pretty sure that 3...Qd1+ would give me sufficient play. I reasoned that my King was active, white’s King was vulnerable, and the white Queen was out of play and unable to stop perpetual check. Whether this re ally drew didn’t matter at the time, the game had been going on for hours and
PART EIGHT – ENDGAMES F 0 R MASTERS ( 2 2 00-2 3 99) 391
my old, exhausted brain was willing to grab hold of any happy result, real or imagined, that it saw.
2 …Qxe6 3.fxe6 f5!
Made possible by 1...Ke7 since his 3.fxe6 was no longer check. Now my kingside majority is mobile and keeps his King at bay.
4.b3 Kxe6 5.c3
Diagram 398
8
7
6
5
4
3
2
a b c d e f g h
I began to think that I might even have winning chances! At this point a well-known grandmaster happened to stroll by. He glanced at the position for a couple of minutes and walked away. Later he confided that he also thought that I was on top (two minutes isn’t time enough for anyone to figure things out, but it does show that even a world class player won’t find these positions easy).
Now I began to think. I thought and thought and thought and became more and more depressed as the minutes ticked by. Finally I convinced myself that I was lost. White can create an annoying outside passed pawn on the queenside. True, black’s King is more active than white’s but, during the game, I couldn’t see how this was going to save me against best play! Cursing my luck, I struggled to find some saving scheme and finally came up with something. Notice that I didn’t say “something good.”
5 …Ke5
At the time I thought this was very important. The idea is to meet 6.Ke3 with
6…f4+. Unfortunately, it turns out that this check is a losing blunder. However, during the game I fully intended to play it! Embarrassing but true! Let’s see why 6.Ke3 f4+ fails:
7.KB (7.Kd3 Kd5 8.b4 c4+ 9.Ke2 g4 creates a drawing blockade. This is one
of the important points of 5 …Ke5) 7 …Kf5 8.b4 g4+ 9.Ke2 Ke4 10.a5 bxa5
11.bxa5 Kd5.
392 SILMAN’S COMPLETE ENDGAME COURSE
Diagram 399
8
7
6
‘j
4
3
2
a b c d e f g h
Position after 11…Kd5
I really believed I was drawing here since 12.Kd3?? is met by 12 …c4+.
I happily rested in my fool’s paradise until I got home. Setting up this posi
tion on a board, I immediately noticed that 12.c4+! wins: 12 …Kc6 13.Kd3 Kb7
14.Ke4 f3 15.gxf3 gxf3 16.Kxf3 Ka6 17.Ke4 Kxa5 18.Ke5! Ka4 19.Kd6 Kb4
20.Kd5 and Black is toast since he’s on the wrong end of a Trebuchet! Fortunately, my opponent agreed with my incorrect assessment and tossed
out:
6.g3
Now the game was eventually agreed drawn due to a line I’d worked out when
I played 5 … Ke5:
6…Kd5 7.Ke3 Kd6 8.Kd2 Kd5 9.Kc2
And not 9.Kd3?? c4+! when Black wins.
9…Kd6 10.Kd3 Kd5 ll.b4 Kc6 12.Kc4 cxb4 13.cxb4 h5+! 14.axh5+ Kb6
15.Kd5 Kxb5 16.Ke5 f4 17.gxf4 gxf4 18.Kxf4 Kxb4 19.Kf5 Kc5 20.Kf6
Kd6 2l.Kg7 Ke7 22.Kxh7 Kf7, ‘h-‘h.
At the time I was impressed with myself. As already pointed out, though, my elation turned to self-loathing when I returned home. It seemed (incorrectly, as it turns out) that I was indeed lost after 6.Ke3. In a way, I was lucky to have been deluded. \Vho know what I would have done ifl’d seen what was “really” going on? In this case, ignorance really was bliss.
Now other voices started to join in on this King and pawn endgame de bate. First, IM Jack Peters offered up some new ideas from the position in diagram 400.
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 393
Diagram400
8
7
6
5
4
3
2
a b c d e f g h
White to move and win
I have already pointed out that 10.a5 (10.bxc5 bxc5 11.c4 also does the trick)
10…bxa5 11.bxa5 Kd5 12.c4+ wins for White, but Jack noted, quite correctly, that Black can put up a better but ultimately hopeless defense after 12 …Kc6
13.Kd3 Kb7 14.Ke4 with, instead of my 14…f3, 14…Ka6 15.Kxf4 Kxa5 16.Kxg4
Kb4 17.Kf3! Kxc4 18.g4 with four possibilities:
18…Kd5 19.g5 Ke6 20.g6, queening.
18…Kb3 19.g5 c4 20.g6 c3 21.g7!, when White queens with check.
18…Kd3 19.g5 c4 20.g6 c3 21.gxh7 c2 22.h8=Q cl=Q
23.Qd8+ Kc2 24.Qc8+ Kb1 25.Qxcl+ Kxcl 26.h7 and White
WlnS.
18…Kb4 19.g5 c4 20.g6 c3 21.Ke2 Ka3 22.gxh7 c2 23.Kd2
Kb2 24.h8=Q+.
I wasn’t too upset by these additions. They basically verified my own conclu sions. Then I found a bombshell waiting in my e-mail. A strong grandmaster (who wishes to remain anonymous) claimed that my analysis was completely flawed and that Black was better after all (that makes two grandmasters who were off their rocker!), though a draw would ultimately be the correct result!
Shocked, I glanced at his e-mail “blindfolded” (as the years roll by, I get lazier and lazier and often don’t bother looking at a chessboard) and took his word for it (yes, I too am influenced by a high rating!). Sending the analysis to 1M John Watson, he lambasted me for listening to anyone (a very wise bit of criticism, by the way!). This prompted me to take a closer look and I immediately saw that the grandmaster was hallucinating.
394 SILMAN’S COMPLETE END GAME COURSE
Diagram 401
a b c d e f g h
Black to move and demonstrate grandmaster dementia
5 …Kd5
The “unknown grandmaster” considered 5 …Ke5 (the actual game continu ation) to be a mistake. As we shall eventually see, it turns out that, in reality, both 5 …Kd5 and 5 …Ke5 lead to a draw.
6.Ke3 Kc6
Black’s idea is to play … b6-b5. I must admit that I’d never considered such a plan.
7.Kd3
The grandmaster claimed a Black win after 7.b4 Kd5 8.Kd3 f4 9.a5 bxa5 (this analysis is horribly flawed. Here, 9…c4+ is the correct way for Black to ice the game) 10.bxa5 (and now 10.c4+ gives White the better game!) 10…c4+ ll.Ke2 g4, but IM John Watson and Patrick Hummel pointed out that he was a bit off since 9.c4+ wins easily for White. IM Ron Burnett came back with a chal lenge: “Just how does White win after 9.c4+?” Fortunately, Mr. Hummel had answered this several days earlier and saved me the trouble of doing any real work: 9.c4+ Kd6 (9…Ke5 10.bxc5 bxc5 ll.a5 Kd6 12.Ke4 is easy for White)
10.b5 Ke5 ll.Ke2! g4 (also hopeless is ll …Kf5 12.a5) 12.Kd3 and white’s King will penetrate to e4 and enjoy a feast.
7…b5
I later discovered that this is a losing blunder but that Black could still draw with a correct seventh move, though I’m leaving that to the Master or near master (don’t forget, this is a master section!) who is reading this analysis to figure it out for himself.
After 7…b5, our mystery grandmaster claims a draw by: 8.axb5+ Kxb5 9.Ke2
Kc6 10.g3 Kd5 ll.Ke3 Ke5 12.Kf3 Kd5. However, IM Watson once again points out that White wins easily with 8.a5 (instead of 8.axb5+??) 8…b4 9.cxb4
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 395
cxb4 10.Kd4 Kb5 ll.Ke5 Kxa5 12.Kxf5 Kb5 13.Kf6 Kc5 14.Kg7 and it’s time for Black to hang up his Rooks.
So what does this mean? Are international masters better endgame players
than grandmasters? Was the grandmaster pulling my leg? Is there any meaning to unearth? I tend to think that the grandmasters in question were too lazy to give this position a serious look. But it does show that King and pawn endgames can be very tricky, and that anyone is capable of completely misjudging such a complicated situation.
We’re done with the adventure. It’s time to see what the truth is:
Diagram 402
8
7
6
c;
4
3
2
a b c d e f g h
Black to move and draw
5 …Ke5
As played in the actual game. At the last second, I also discovered that 5 …Kd5 draws in all lines! A small sample: 6.Ke3 Kc6 7.b4 Kd5 8.Kd3 Kc6 9.Kc4 cxb4
10.cxb4 b5+ ll.axb5+ Kb6 12.Kd5 Kxb5 13.Ke5 f4! and Black draws as in Peter’s subsequent analysis to 5 …Ke5.
6.Ke3 Kd5
Instead of my original intention of 6…f4+??, which analyzed earlier.
7.Ke2 Ke4 8.b4 cxb4!
Worse is 8…Kd5 9.Kd3 f410.c4+! Kc6 (10…Kd6 ll.b5 Ke5 12.Ke2 g413.Kd3 isn’t any better) ll.b5+ Kd6 12.Ke4 Ke6 13.g3 (or the simple 13.Kf3) and black’s defense breaks down. Analysis by Patrick Hummel.
9.cxb4 Kd4 10.a5 bxa5 ll.bxa5 Kc5 12.Kd3
This was given as a White win by Mr. Hummel and John Watson, butJack
Peters points out that the position is a draw.
12…Kb5 13.Kd4 Kxa5 14.Ke5 f4!
396 SILMAN’S COMPLETE ENDGAME COURSE
This very important, but simple, move (found by Peters) was missed by a whole team of titled players. Black’s King looks awfully far away, doesn’t it?
15.Kf5 Kb6 16.Kxg5 Kc6 17.Kf6
A book draw follows 17.Kxf4 Kd7 18.Kf5 Ke7 since black’s King gets to h8.
17…Kd7 18.Kg7 Ke7 19.Kxh7 Kf7 and now the point of 14…f4! is clear: White doesn’t have a tempo move that will enable him to regain the Opposition. Thus, the game is drawn (20.Kh8 Kg6 2l.Kg8! [21.h7?? Kf7, 0-1] 21…Kxh6
22.Kf7 Kg5 23.Ke6 Kg4 24.Ke5, =).
Remember, all this is merely meant to be a warning. The fact is, complex King and pawn endgames are a bit of a rarity and most masters (on up) love the idea of playing them. But they also know that such positions should be avoided if a safe, clear alternative is available.
I constantly hear players arguing over the relative values of Bishops and Knights. Which is really the stronger piece? The majority of voices point to the Bishop as the superior entity. The following quote from Jose Raul Capablanca is particularly pointed: “The weaker the player, the more terrible the Knight is to him, but as a player increases in strength, the value of the Bishop becomes more evident to him and of course there is, or should be, a corresponding decrease in the value of the Knight as compared to the Bishop.”
That’s all well and good, but then how do we explain the following game?
Diagram 403
8
7
6
5
4
a h c d e f g h
Capablanca-Reshevsky, Nottingham 1936
Black to move
Note the holes on b4, c5 and d4, plus the weak pawn on d5. If Black plays
“normally” (…Kf7-e6) White will achieve a sweet “torture position”: l…K£7
2.Nb4 axb3 3.axb3 Bb7 4.Ke2 Ke6 5.Kd3 Ke5 (5 …Kd6 6.Kd4) 6.Nc2 with Nd4 and Kc3 to follow.
Due to this, Black tries to make use of his Bishop’s long-range powers by creating an outside passed pawn on the kingside.
l…g5! 2.hxg5 fxg5 3.Nb4
Capablanca avoided 3.f4 gxf4 4.exf4 due to 4…d4! when Black would even tually lose his d-pawn, but his active Bishop will give him serious chances to save the game. The third World Champion was one of the greatest endgame
397
398 SILMAN’S COMPLETE ENDGAME COURSE
players of all time and always avoided unclear situations in favor of “iron control.”
3 …axb3 4.axb3 Bb7
Black now hopes to play …h5-h4 with an outside passed pawn.
5.g4
White prevents black’s plan and now intends to swing his King over to d4 with total control. Refusing to die passively, Black realizes his only counterplay is based on an …h5 advance. Thus …Kh7-g6 and …h5 will exchange pawns (always a good defensive idea in the endgame) and allow the King a measure of activity.
5 …Kg7 6.Ke2 Kg6 7.Kd3 h5 8.gxh5+ Kxh5 9.Kd4 Kh4 10.Nxd5 Kg3
Diagram 404
8
7
6
5
4
3
2
a b c d e f g h
Black has lost a pawn but his active King and the limited remaining material gives him some practical chances to save himself. Nevertheless, White should win with best play.
ll.f4 g4
Black loses after 1l…Bxd5 12.Kxd5 g413.f5 Kh3 14.f6 g3 15.f7 g2 16.f8=Q
g1=Q 17.Qh8+ Kg2 18.Qg8+.
12.f5 Bc8
Still twitching. Far worse is 12 … Bc6 13.Nc7 (now black’s Bishop can’t stop white’s f-pawn from promoting) 13 …K£2 14.f6 g3 15.f7 g2 16.f8=Q+ BB 17.Qg8 g1=Q 18.Qxg8+ Kxg8 19.e4 and it’s all over.
13.Ke5 Bd7 14.e4
Now 14.Nc7 doesn’t work: 14… Bxf5! 15.Kxf5 Kf3 16.Ne6 g3 17.Nf4
Kxe3, =.
However, it’s not clear to me that 14.e4 is really necessary. More straightfor
ward is 14.f6 Be8 15.Ke6 Kh4 (or 15 …Kf3 16.Ke7 Bh5 17.Nf4 g3 18.Nxh5 g2
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 399
19.£7 gl=Q 20.f8=Q+ Kxe3 [20…Ke2 21.Nf4+ Kxe3 22.Ne6] 21.Qf5 with an easy win) 16.Ke7 Bh5 17.Nf4 g3 18.b4 Kg4 (18…Kg5 19.Nxh5 g2 20.Nf4 g1=Q
21.Nh3+) 19.Nxh5 g2 20.£7 g1=Q 21.Nf6+ Kh3 22.f8=Q Qxe3+ 23.Kd6 when
23…Qd4+ 24.Nd5 and 23…Qb6+ 24.Kd5 are both hopeless.
14…Be8
Diagram 405
a b c d e f g h
15.Kd4??
This blunder throws the win away. Bondarevsky showed that White could still claim victory by 15.f6! Kf3 16.Nf4 g3 17.Kf5 Bd7+ 18.Kg5 Be6 19.Nxe6! g2
20.£7 g1=Q+ 21.Kf6 and White wins. This theme of allowing Black to promote first comes up more than once.
lS …K£3 16.e5 g3 17.Ne3 Kf4?
Also losing is 17 …Bd7 18.e6 Bc8 19.e7 Bd7 20.f6 Be8 21.Nf5-Alekhine. However, Bondarevsky found a draw: 17 … Bf7!! 18.e6 Bg8 19.e7 (or 19.b4
Kf4!) 19… Bf7 20.f6 Kf4! when White can’t improve his position.
18.e6 g2 19.Nxg2+ Kxf5 20.Kd5 Kg4 21.Ne3+ Kf4 22.Kd4, 1-0 since Black has no defense against e7 followed by swinging the Knight to c7.
The “explanation” for Capablanca’s obvious favoritism towards Bishops while he used his Knight to good effect against Reshevsky is simple: Bishops tend to be better in the majority of cases (i.e., open positions, passed pawn races where the Bishop’s long-range abilities crush the slow Knight, etc.), but Knights have their own set of advantages:
….. They can land on light and dark squares. This means that nothing is safe from them!
….. Knights are more flexible than Bishops.
….. Knights are often superior in closed positions.
400 SILMAN’S COMPLETE ENDGAME COURSE
Knight-hops confuse amateurs and masters alike! After I managed to draw a pawn down endgame (multiple pawns and two Knights each) against the strong IM Georgi Orlov, he lamented, “Hop, hop, hop! The Knights are hopping all over the place! How could I possibly figure out what was going on?”
What this means is that the master has an open mind and understands that many positions are highly favorable for Knights. He might prefer the Bishop in a general sense, but won’t hesitate to rush into a Knight vs. Bishop endgame if he feels the horse will reign supreme.
Our next example shows what happens when a “Bishop is better than a Knight” mentality becomes set in stone (or in this case, in metal).
Diagram 406
a b c d e f g h
Mephisto Computer- Silman, Los Angeles 1989
Black to move
At the time, computers hadn’t quite made their mark at the highest levels of tournament chess. This event was historic in that Mephisto did very well against several strong players and Deep Thought dragged down Bent Larsen and my self (Larsen didn’t take the machine seriously, while I was “introduced” to the computer’s incredible defensive skills in the final round-I outplayed it, achieved
a position that would have given me excellent winning chances against any hu man, and then ran into a “non-human” defensive wall while in time pressure). Though my loss to Deep Thought was traumatic (and costly!), I managed to uphold the integrity of the human race earlier in the same event against Me phisto. The game was a long, grueling affair that demanded accurate play and tremendous amounts of energy. I managed to win a pawn and began the long Cat and Mouse process of torturing the cursed bucket of bolts in the resulting
endgame. In the diagrammed position, I played …
l…Ba4
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 401
It was here that I discovered a glitch in the machine. Any experienced human would understand the tremendous flexibility of a Knight in such a position, but the machine was programmed to believe in the superiority of Bishops over Knights. Thus, it allowed me to take on b3 and enter a winning Knight vs. Bishop endgame.
2.Kel?? Bxb3!
I wasn’t going to let it change its mind!
3.Bxb3 Kc5 4.Ba4 N£7
Diagram 407
a b c d e f g h
The only fly in the ointment is black’s vulnerable pawn on g4. Thus, I swing my Knight to e5 where it defends g4, deprives the Bishop of the d7-square, and also threatens to decisively leap into d3 (targeting f2) or c4 (pressuring e3 and making a … d5-d4 push serious business) at the ap propriate time.
5.Bb3 Ne5 6.Bc2 Nc4 7.Ba4 d4 8.Bc2 Kd5 9.Bb3 Kc5 10.Bc2 dxe3 ll.fxe3
Nxe3 12.Bxe4 Kd4 13.Bg6 Nc4 14.Kf2 Nd2 15.Bf5 Ne4+
Diagram 408
a b c d e f g h
Nothing is safe from the mighty Knight! Now it chases white’s King off of f2 and threatens the pawn on g3. As will be seen in the next note, it will also
402 SILMAN’S COMPLETE ENDGAME COURSE
train its sights on fl, breaking any effort White makes to stop the f-pawn from turning into a Queen.
This highlights the Knight’s main advantage over a Bishop: it can control any color and thus potentially place pressure on any square on the board.
16.Bxe4
A sad necessity, but 16.Kgl Ke3 17.Bxg4 f2+ 18.Kg2 Nd2 promotes my remammg pawn.
16… Kx:e4
The rest is basic King and pawn ending fare.
17.Kfl Ke3 18.Kel f2+ 19.Kfl Kd3 20.Kxf2 Kd2 20.Kfl Ke3, 0-1.
The edge in flexibility that a Knight often has over a Bishop is graphically shown in our next game, played between two World Champions.
Diagram 409
a b c d e f g h
Smyslov- Euwe, Groningen 1946
White to move
Black is in serious trouble. The reasons for this:
– – White’s King is more active than black’s.
– – The white Knight will dominate black’s Bishop.
Black has potential pawn weaknesses on c7, f6, and h5 (that weakness will move to g4 if White plays f3 followed by fxg4). White will push his f-pawn to f3, so that will only leave g3 in need of defense. Simply put: Black has far more weak pawns than White does.
– – In general two Knights don’t work well together.
However, in this case black’s Knight is dominated and, if it moves, an exchange will take place that will leave
PART EIGHT – END GAMES F 0 R MASTERS (2 2 00-2 3 9 9) 403
White with a crushing Knight versus a rather impotent
Bishop.
…. The c3-c4-c5 pawn lever allows White to break through on the queenside in many different variations.
…. Correctly handling a Knight takes more skill than correctly handling a Bishop. As a result, lower rated players fear Knights-in a way, mastery of Knights is the test that separates the men from the boys.
l.Kc2 Bel
Smyslov analyzes 1… Bg5 (intending to transpose into a King and pawn end game) and shows that it also loses: 2.Kb3 Bxh4 3.gxh4 Ne7 4.Nxe7 Kxe7 5.Kc4 fS 6.f3 f4 7.Kb5 Kd7 8.c4 Kd8 9.Kc6 Kc8 10.c5 dxc5 11.Kxc5 Kd7 12.Kc4
Ke7 13.Kb4 Kd6 14.Kb5 g3 15.Kb4 Ke7 (15…c6 16.dxc6 Kxc6 17.Kc4 Kd6
18.Kb5 wins) 16.Kc4 Kd6 17.Kb5 Kd7 18.Kc5 Kc8 19.d6, winning all of black’s pawns.
Perhaps this is true, but John Watson feels that 5…f5 is a bad move since it
later deprives the e5-pawn of some much-needed protection. Instead, he rec
ommends 5 …Kd7 6.Kb5 Ke7 7.Kc6 g3! 8.f3 (8.fxg3 f5!, =) 8…Kd8 9.c4 Kc8
10.c5 dxc5 11.Kxc5 Kd7. His logic is shown by the position that results from the following moves: 12.Kb5 Kd6 13.Kb4 c6 14.dxc6 Kxc6 15.Kc4 Kd6 16.Kb5
Kd7 17.Kc5 Kc7 18.Kd5 Kd7, =,since black’s e5-pawn is solid thanks to black’s f-pawn remaining on f6.
This fits nicely into the chapter, King and Pawn Endgames Are Confusing. Whatever the truth might be, our focus here is on the upcoming Knight vs. Bishop battle, which shows the Knight running rings around its “faster” foe.
2.f3
White intends to play c4 and then march his King all the way to c6. Black would like to “greet” the white King by moving his own monarch to d7, but then Ng7+followed by Nxh5 would follow. As a result, Black has no choice but to exchange a pair of Knights.
2…Ne7 3.Nxe7 Kxe7 4.fxg4 hxg4 5.Nf5+ Kf7 6.c4 Kg6 7.Kb3 Kg5
Black desperately seeks counterplay on the kingside. Trying to defend on the queenside by 7 …Kf7 also falls on its face (see diagram 410 on the next page):
404 SILMAN’S COMPLETE ENDGAME COURSE
Diagram 410
8
7
6
‘)
4
3
2
a b c d e f g h
8.Ka4 Ke8 9.KbS Kd7 10.Nh6 Bxg3 1l.Nxg4 Bh4 12.Nh6 BgS 13.NfS Bd2
14.Ka6 Kc8 1S.Ka7 Bf4 16.Ka8 Bd2 17.g4 Bf4 18.Ne7+ Kd7 19.Nc6 Kc8
20.Na7+ Kd7 2l.Kb7 wins according to Smyslov. Let’s verify this (note how the Knight dominated both the black King and black Bishop): 21… Be3 2 2.Nc6
BcS 23.Nb8+ Kd8 24.Kc6 Bd4 2S.Na6 Bb6 26.Nb8 and now:
26…Bd4 27.Nd7.
26… Be3 27.Nd7 BgS 28.Nf8 and Ne6 will prove decisive.
—- 26…Ke7 27.Nd7 BaS 28.cS Bb4 (28…dxcS 29.NxcS Bb6
30.Nb7 Kf7 3l.d6) 29.Kxc7 dxcS 30.Nb8! BaS+ 3l.Kc6
c4 32.KcS c3 33.Nc6+ Kd7 34.NxaS c2 3S.Nb3 and Black must resign. “White’s Knight gives the impression of being everywhere at once, while the Bishop seems ponderous and completely ineffective.
8.Ka4 Bxg3
The only chance, otherwise “White just marches his King to c6 and feasts on black’s queenside.
9.Nxg3 Kf4 10.Nh5+ Kxe4 ll.Nxf6+ Kf5
Black doesn’t do any better with 11…Kf4: 12 .KbS e4 13 .Nxe4 Kxe4 14.Kc6 g3 1S.Kxc7 Ke3 16.Kxd6 Kf2 17.KeS Kxg2 18.d6 Kf3 19.d7 g2 20.d8=Q g1=Q
21.Qd3+ when “White forces the trade of Queens (21…Kf2 22.Qd4+ or 21…Kg2
22 .Qg6+), leaving himself with an extra baby Queen on c4.
Mter 11.. .KfS Black intends to push his e-pawn to the ends of the Earth. How can the Knight deal with this threat?
12.Ne8!
“White’s super-Knight goes in the opposite direction, seeking a quick lunch before dealing with the e-pawn.
12…e4 13.Nxc7 e3 14.Nb5 Kf4
PART EIGHT – ENDGAMES F 0 R MASTERS ( 2 200-2 3 99) 405
He didn’t like the look of 14…e2 15.Nd4+ followed by 16.Nxe2.
15.Nc3 Kg3 16.c5, 1-0. An impressive endgame that shows us just how light on its feet a Knight can be.
Our final example gives us more fancy hoof-work by a Knight.
Diagram 411
8
7
6
5
4
3
2
a b c d e f g h
Bernstein- Suesman, U.S. Championship 1938
\Vhite to move
White’s superior pawn structure and centralized pieces make an obvious impression. However, White has an idea here that allows him to add one more advantage to his quiver of plusses: he can create a crushing Knight vs. a poor Bishop scenario.
l.Bxa6! bxa6 2.a5!
The point. White clears away the one defender of the c5-square, which will turn into a wonderful home for white’s Knight.
2 …b5
This lets White rule the position with an iron grip thanks to the holes on b4 and c5 (lovely resting places for white’s Knight and King). Like it or not, Black should have tried either 2 … bxa5 or 2 …Kb7.
3.Ng4!
By using tactics aimed at the pawns on f6, f7, and c6, White forces his Knight onto the dream c5-square.
3 …f5 4.Ne5
Threatening a fork on c6.
3 …Kc7 5.exf5 exf5
Ugly, but S…BxfS 6.Nxf7 was even worse.
6.Nd3 Kb7 7.Rhel Rhe8 8.Nc5+ Ka7
406 SILMAN’S COMPLETE ENDGAME COURSE
Diagram 412
8
7
6
5
4
3
2
a b c d e f g h
An experienced player will always head for a position like this. A Knight this good (along with black’s passive King and the target on a6) will bring home the full point almost every time.
9.Kc2
The famous German player Fritz is gung ho for 9.Ka3 f4 10.Rxe8 Rxe8 ll.d5 cxd5 12.Rxd5 Re2 13.Rd7+ Kb8 14.Nxa6+. However, I think Bernstein’s choice is by far the more “human” move (Note that “human” doesn’t necessarily mean “best.” It’s just a good, practical decision). The idea is simple: He intends to move his King to d2 and then swap off all the Rooks. The resulting minor piece endgame is winning for White, while if Black retains a Rook he will allow White to invade the 7th rank. What’s particularly nice about this plan is that there is no risk involved-White doesn’t want to allow Black any counterplay.
9…f4+ 10.Kd2 Bf5
The alternative, 10…b4, lets White decisively invade on the 7th rank: ll.cxb4
Rxd4+ 12.Kc3 Rxdl 13.Rxdl Re3+ 14.Kc4 Re2 15.Rd7+ Ka8 16.Rc7 Rxg2
17.Rxc6 Rxh2 18.Rxa6+ Kb8 19.Rb6+ Ka8 20.b5 h5 21.Rd6 and it’s all over.
ll.Rxe8 Rxe8 12.Rel
Diagram 413
8
7
6
5
4
3
2
1
a b c d e f g h
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 407
12…Rg8
Though 12 …Rxel 13.Kxel is indeed losing for Black, picking the right plan might be harder than one might suppose. In my view, White can make this easy on himself if he realizes that he’s a King up-black’s King is frozen to the defense of a6. Thus, all White has to do is find a way to penetrate into black’s camp with his King: 13 …h5 14.Kf2 h4 15.g3 fxg3+ 16.hxg3 hxg3+ (16…h3 17.g4 is easy) 17.Kxg3 Bc8 (Passive play offers no hope either: 17 …Bg6 18.Kf4 Bb1
19.Ke5 Ba2 20.f4 Bc4 2l.Kf6 Bd5 22.Ke7 f5 23.Kd6 Be4 24.Kc7 Bf3 25.Ne6
Bd5 26.Nd8 and it’s time for Black to give up) 18.Kf4 Kb8 19.Ke5 Kc7 20.f4 f5
2l.d5 cxd5 22.Kxd5 and Black is out of options.
13.Re7+ Ka8 14.Re2 Bc8
14…Ka7 15.Ke1 followed by Re7+ and Kf2 is a trivial win.
15.Nd3 Be6 16.Nxf4
This not only wins a pawn but it also defends g2. Now white’s Rook is free to roam.
16…Bc4 17.Re7 Rd8 18.Nd3, 1-0. The position after 18… Bxd3 19.Kxd3 c5 20.Rxf7 b4 21.Rc7 is completely lost.
The King is an extremely strong piece, but many players lose sight of this fact due to the necessity for it to hide during the middlegame. Experienced tournament competitors, though, are well aware of the need to let their King flex its muscles once an endgame is reached. In fact, many games are saved or won when one side manages to activate their King, while the other King sits smugly at home doing nothing.
Though the “Use your King in an endgame!” concept is given regular lip service from class “D” on up, few realize just how important it is. Hopefully the examples that follow will boil your blood and make your want to reach for your King and kick it into dance mode.
Diagram 414
a h c d e f g h
Lilienthal- Smyslov, Leningrad 1941
Black to move
The Bulgarian chess saying, “I make your pawns look like your grandmother’s teeth,” seems a good way to describe black’s structure-his pawns are weak and many of them appear to be doomed.
Though things look bad for Black, he’ll be able to save himself by following two classic endgame concepts:
In the endgame, make your King a fighting piece! Keep your Rook active at all costs.
408
P A R T E I G H T – END GAME S F 0 R M A S T E R S ( 2 2 0 0 – 2 3 9 9 ) 409
l…g5 2.Rxh7 Rxa2 3.Rh6+ Ke5 4.Rxc6
No better was 4.h4 g4 5.h5 (5.Rxc6 Ke4 6.Rxc5 f4!, =) 5 …Ke4 6.Rf6 Rb2 (6…Ral+, =) 7.h6 Rxb3 8.h7 Rbl+ 9.Kg2 Rb2+ with an instant draw since lO.Kfl allows 10…Rh2.
4…Ke4!
Black’s King, which is extremely active, has turned into a berserk fighting unit of death, while white’s pathetic leader is trapped on the back rank, begging for its life.
5.Rxc5
White is three pawns up, but one King down.
5 …f4!
Sacrificing another pawn to ensure the safe advance of the King.
6.exf4
6.Kfl Ral+ 7.Ke2 f3+ 8.Kf2 Ra2+ is another way to make a draw.
6…KB
Diagram 415
8
7
6
5
4
3
2
1
a h c d e f g h
Now black’s King is safe from checks while White is threatened by immedi
ate mate.
7.h3
Suicide could still occur by 7.h4 Ral+ 8.Kh2 Ra2+ 9.Kh3?? g4 mate.
7…Ral+, ‘h-‘h.
Our next example shows Larry Christiansen enjoying superior King position, then sacrificing a pawn to make it even better!
410 SILMAN’S COMPLETE ENDGAME COURSE
Diagram 416
a b c d e f g h
D Gurevich- Christiansen, U.S. Championship 2000
Black to move
Black’s a pawn up and has the better King position, but his d3-pawn is vulner able and there are no obvious breakthroughs happening any time soon. The way Black turns this into a forced win is very instructive.
l…Bd5! 2.Bbl
Trading would leave White with a losing Queen endgame after 2.Bxd5 Kxd5
3.Kxd3 g4 4.Kc3 h5 5.Kd3 a6 6.Kc3 Ke4 7.Kc4 Kf3 8.Kc5 Kg2 9.Kb6 Kxh2
10.Kxa6 h4 ll.b5 hxg3 12.b6 g2 13.b7 gl=Q 14.b8=Q+ Kg2 15.Qa8+ Kf2
16.Qf8+ Ke2.
2 …d2!
Black sacrifices his extra pawn to increase the domination of his King. This, in turn, will allow black’s King to make contact with (and eventually win) the b-pawn.
3.Kxd2 Be4!
Ending the threat of 4.Bxh7 by offering up another King and pawn end
game.
4.Ba2
Now the exchange of Bishops is even worse for White than it was before:
4.Bxe4 Kxe4 5.Kc3 g4 6.Kc4 Kf3 7.Kb5 Kg2 8.Ka6 Kxh2 9.Kxa7 Kxg3 10.b5
Kf3 ll.b6 g3 12.b7 g2 13.b8=Q gl=Q+ 14.Ka6 Qal+ 15.Kb7 Qb2+ 16.Kc8
Qxb8+ 17.Kxb8 h5.
4…Kd4
Black’s King is a dominating piece. Next on his agenda is:
–Place all his pawns on dark squares so they will be immune from the white Bishop’s intentions.
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 411
…. Force the b-pawn to advance to b5 so the c5-square will become available to the black King. He will then surround and win it by .. .KcS mixed with the Bishop moving to the fl-a6
diagonal.
Diagram 417
8
7
6
5
4
3
2
a b c d e f g h
5 .h4 h6 6.hxg5 hxg5 7.Be6 Bd5
Giving black’s King access to the c4-square.
8.Bd7
As always, the trade of Bishops leads to a lost King and pawn endgame: 8.Bxd5
Kxd5 9.Kd3 g4 10.Ke3 Kc4 11.Kf4 Kxb4 12.Kxg4 aS.
8…Kc4 9.b5 Kd4 10.Ke2 Kc5 ll.Ke3 Bc4 12.Ke4 Bxb5 13.Be6 a5
Black has won the b-pawn and should now win the game. However, care must be shown: White would love to pick up the pawn on g5 and then sacrifice his Bishop for black’s remaining pawn.
14.Bb3 Kb4 15.Bdl Ba4
Nothing wrong with this, though 15…Bc4 16.Ke3 Bb3 17.Bg4 Kc3 would lead to white’s quick resignation.
16.Bg4
Here Black suffered a meltdown: 16…Kc3 17.Ke5 Bb3 (17 …Bc6 wins) 18.Bd7 g4 (18…Kb4 was a winner) 19.Kf4 Bdl 20.Be8 Kb4 21.Ke3 Ba4 22.Bh5 Bd7
23.Kd2 Kb3 24.Kcl Ka2, 1h- 1h. Apparently he forgot that white’s King could scoot back to the queenside and take advantage of the fact that the a-pawn is a “wrong colored rook-pawn” (If white’s King can reach a1 he could make a draw by sacrificing his Bishop for the g-pawn.).
Sadly, he had an immediate win after 16.Bg4 by 16… Bc2+ 17.Ke3 (17.Kd4 a4
18.Be6 a3 19.Ba2 Bb3 20.Bbl Bg8 with …Kb3 and …Kb2 to follow) 17 …Kc3
18.Bd7 a4 19.Bxa4 Bxa4 20.Ke4 g4 21.Kf4 Bdl and White must resign.
412 SILMAN’S COMPLETE ENDGAME COURSE
Most young players today aren’t familiar with Salo Flohr. That’s a shame, since he was a powerful positional player and was also blessed with fantastic technique. In our next game, Flohr uses both his pieces (his King and Rook) to the maximum. His Rook seems to dance rings around its Black counterpart, while his King almost single-handedly rules the board.
Diagram 418
8
7
6
5
4
3
2
1
a h c d e f g h
Flohr- Petrov, Semmering 1937
White to move
A glance might convince some players that Black is doing well; he has a Rook on the 7th rank and a passed a-pawn. However, Black is actually in serious trouble. How can this be? How can a white advantage be logically explained? The answer is simple: White’s pawn structure is superior, and his King will prove to be far more active than black’s (which is too far away from the kingside to defend its pawns there)!
l.Ra5!
A very nice Rook maneuver. The threat against d5 forces black’s reply.
l…Rb5 2.Ra2!
Now black’s Rook is denied the 7th rank and f2 is no longer under attack. This frees the white King for a triumphant march into black’s kingside.
2…g6
It’s possible to push the a-pawn, but after 2 …a5 3.Kf4 Ka6 4.Ke5 Rbl 5.Kxd5
Kb5 6.Kd6, White wins: 6…a4 7.Ke7 Kb4 8.d5 Rdl 9.d6, etc.
3.Kf4 f6
Trying to prevent white’s King from breaking through.
4.g4 hxg4
4…a5 5.gxh5 gxh5 6.Kf5 wins material.
5.Kxg4 Rbl
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 413
It seems that White threatened h4-hS, and this prevents it. However, what if Black ignores the “threat” and instead plays the logical S •• .aS, simply shoving his passed pawn? It turns out that white’s threat wasn’t h4-hS at all, but rather f4-fS! Thus, after S …aS 6.f4! Black’s busted: 6…Re3 7.fS! gxfS+ 8.Kf4 and wins.
6.Ra5!
There’s that wonderful Rook maneuver again!
6…Rb5 7.Ral!
White’s 1st rank is now more important than his 2nd rank. The way white’s Rook zips to aS then backtracks to whatever rank is crucial at that moment is nothing less than striking.
7 …Rb2 8.Kf3
Threatening to win black’s g-pawn by Rg1.
8…Rb6 9.Rgl f5
Black has defended g6, but now the eS- and gS -squares have been opened up for white’s King.
10.Kf4
Diagram 419
8
7
6
5
4
3
2
a b c d e f g h
10…Re6 ll.Rhl+ Ka7
Or 11…Rb6 12.Ra1 Rb2 13.f3 Rb3 14.RaS! RbS 1S.Ra3 (Amazing! This same maneuver has been used to control the 2nd rank, then the 1st, and now the 3rd.) 1S …Rb2 16.KeS Rh2 17.KxdS Rxh4 18.e4 fxe4 19.fxe4 and white’s two connected passed pawns will win the day.
12.h5 Re4+ 13.Kg5 Rg4+
I think Black could have put up more resistance by 13 …gxhS! 14.KxfS h4!
with a rather odd position.
414 SILMAN’S COMPLETE ENDGAME COURSE
Diagram420
a b c d e f g h
The variations concerning 15.Kg5 are extremely complex and outside the scope of this discussion. However, I suspect white’s best chances can be found with 15.Kf6 when Black has:
__… 15…h3 16.Rh1 Rh4 17.Ke6 Rh5 18.f4 Kb6 19.f5 Kc7 20.f6
Kd8 2l.e4! dxe4 22.d5 and Black’s dead, though I’ll let you analyze a bit and convince yourself of that.
15 …a5 16.Rb5 Ka6 (16…a417.Rb4!) 17.Rxd5 a418.Rc5
Kb6 19.Rc8 Kb5 20.Rb8+ Kc4 2l.Ra8 Kb5 (21…Kb3 22.d5 wins for White) 22.Kf5 Re7 23.Rh8 Rf7+ (23…Kc4 24.Rxh4 a3 25.d5+ Kxd5 [25 …Kb3 26.d6] 26.Rd4+ Kc6 27.Ra4 Rf7+
28.Ke6 Rxf2 29.Rxa3 Kc7 30.Rd3 with a winning position Nalimov TableBase claims mate in 38!) 24.Ke6 Rxf2 25.Rh5+! Kc4 26.Rxh4 a3 27.d5+ Kb3 28.d6 a2 29.d7 al=Q (Worse is
29…Rf6+ 30.Kxf6 al=Q+ 3l.Rd4 Qa6+ 32.Ke7 Qb7 33.e4 when White wins: 33…Kc3 34.Rd5 or 33…Qc7 34.Rd6! Kc4
35.e5 [threateningKe8] 35 …Qb7 36.Ke6 and Black must resign) 30.d8=Q Qa6+ 3l.Ke5 when White is clearly better.
Amazing stuff, and I barely scratched the surface. This should give you a real wake up call as to just how complicated endgames can be!
14.Kf6 gxh5 15.Kxf5
White’s plan is clear: he wants to create two connected passed pawns in the center.
15 …Rg2 16.Ke5 Rg5+
16… Rxf2 let’s White achieve his goal after 17.Kxd5.
17.Ke6 h4 18.Rhl Rh5 19.f4
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 415
Diagram 421
8
7
6
5
4
3
2
1
a h c d e f g h
White didn’t manage to get his two connected passed pawns, but it turns out that this single passer will do more than both of black’s rook-pawn runners. Why is one pawn better than two? Because it’s being escorted by white’s King, while black’s monarch is still on a queenside vacation.
19…Kb6
19…aS 20.£5 a4 21.f6 is easily winning for White.
20.f5 Kc7 21.f6 Kd8 22.Rfl! Rh6
22 …Ke8 runs into 23.Rb1 Kd8 24.Rb8+ Kc7 25.£7.
23.Kf7!, 1-0. The threat is Kg7 followed by f6-f7 and f8=Q. A beautifully played endgame by the great Flohr.
I’m finishing this section with a famous King and pawn endgame by Ru binstein, who is widely thought to be one of the greatest endgame players of all time. The reason for its inclusion is obvious-one rarely gets to see such a dominating King!
Diagram 422
8
7
6
5
4
3
2
a b c d e f g h
Cohn- Rubinstein, St. Petersburg 1909
Black to move
416 SILMAN’S COMPLETE ENDGAME COURSE
White has just played his Rook to cl, which is a horrible mistake since it al
lows Black to trade into a winning King and pawn endgame.
l…Rxcl 2.Kxcl
Why is this endgame winning for Black? Aside from his superior pawn struc ture, the answer is simple: Black$ King will dominate white$! In general, you don’t want to rush into a King and pawn endgame if your opponent’s King can take up a far superior position to your own.
2 …Kf6
Many players blindly rush their King to the center, which (I must admit) is usually the right thing to do. However, in this case black’s superiority is on the kingside, so he makes a dash to that embattled area where it will target the pawns on h2, f2, and f3 for extinction.
3.Kd2 Kg5 4.Ke2
White’s King tries to shore up his weaknesses on the kingside, but the vast superiority of the black monarch will eventually drag him down. However, trying to create a pawn race also ends in dismal failure: 4.Kd3 Kh4 5.Kd4 Kh3
6.Kc5 Kxh2 7.Kd6 Kg2 8.Kc7 h5 (Black can gain a tempo with 8…b5, but it’s not really necessary: 9.Kc6 h5 10.Kxb5 h4 ll.Ka6 h3 12.Kxa7 h2 13.b5 hl=Q
14.b6 Qal 15.b7 Qxa3+ 16.Kb6 Kxf3 17.b8=Q Qb4+ 18.Kc7 Qxb8+ 19.Kxb8
Kxf2) 9.Kxb7 h4 10.Kxa7 h3 ll.b5 h2 12.b6 hl=Q 13.b7 Qal 14.b8=Q Qxa3+
15.Kb7 Qb4+ 16.Kc7 Qxb8+ 17.Kxb8 Kxf3 and it’s time for White to resign.
4…Kh4 5.Kfl Kh3 6.Kgl
Diagram 423
8
7
6
5
4
3
2
1
a b c d e f g h
Just in time! However, the difference in King placement is glaringly apparent.
6…e5
A good move that stops the f3-pawn from safely moving to f4. Thus, the normally desirable 6…b5 (fixing white’s queenside pawns and squirreling away a useful tempo move via …a6) would be met by 7.f4.
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 417
7.Khl
Other possibilities:
-7.e4 (not very challenging) 7 …g5 8.Khl h5 9.Kgl h410.Khl
g4 11.fxg4 Kxg4 12 .Kg2 h3+ 13.Kg1 Kf3 and wins-analysis by
Hans Kmoch.
Speelman and Mestel say, “The best chance was 7.a4 b6 8.b5. Now the winning plan used in the game wouldn’t work as Black doesn’t have the tempo-gaining …a6 in reserve (this is explained in the note to white’s 12th move). However, he can
still win by another method: 8…f5 9.Khl g5 lO.Kgl h5 ll.Khl
h4 12.Kgl e4 13.fxe4 fxe414.Khl (14.f3 exf3 15.e4 g4 16.e5 g3) 14…Kg4 15.Kg2 h3+ 16.Kgl Kf3 17.Kfl g4 and wins.”
7 …b5
It’s always useful to fix enemy pawns, especially if you are able to retain a tempo move like …a6.
8.Kgl f5 9.Khl g5 lO.Kgl h5 ll.Khl
White is completely helpless and must sit tight and hope Black can’t find a way to break through.
ll…g4
Diagram 424
a b c d e f g h
The deathblow finally comes and, not surprisingly, there isn’t a defense. Of course, 11…h4, as in the note with 7.a4, was also winning.
12.e4
This makes things easy. Far trickier is 12.fxg4 hxg4 13.Kgl f4 14.exf4 exf4
15.Khl
418 SILMAN’S COMPLETE ENDGAME COURSE
Diagra sm 425
8
7
6
5
4
3
2
1
a b c d e f g h
Trickier than one might imagine
Hans Kmoch wrote that 15…g3?? 16.fxg3 fxg3 17.hxg3 Kxg3 was winning, and a glance might make us agree with him since what can stop black’s King from marching over to the queenside and eating white’s remaining pawns? Speelman and Mestel, though, showed this to be an illusion. White draws: 18.Kgl Kf3
19.Kfl Ke3 20.Kel Kd3 21.Kdl Kc3 22.a4! (The key defensive trick. Taking
on a4leaves Black with two rook-pawns. Of course, if we move things over one file [WH: K-el, P-b3, c4; BL: K-d3, P-b7, c5], Black would win easily-Silman)
22 …a6 23.axb5 axb5 24.Kcl Kxb4 25.Kb2, =.
Once again, our friends Speelman and Mestel show us the correct way for
Black to continue from diagram 425. 15 …f3 16.Kgl Kh4 and now:
17.Khl Kg518.h3 gxh3 19.Kh2 Kg420.Kgl Kf421.Kh2
Ke4 22.Kxh3 (22.Kg3 h2) 22 …Kd3 23.Kg4 Ke2 24.Kg3 and here we must return to Speelman and Mestel’s comment in the analysis to 7.a4: “Now the winning plan used in the game wouldn’t work as Black doesn’t have the tempo-gaining …a6 in reserve.” This is the position they were referring to. Mter
24.Kg3 we have a Trebuchet, meaning that whoever has to move their King loses. Fortunately for Black, he has that all important tempo move 24…a6 available, when White must res1gn.
17.Kfl Kh5! (Very nice! Black wants his King on g5 after White plays Ke1. That way the flexible black Monarch retains the choice of going after h2 via …Kh4, or penetrating in the center via …K£4-Silman) 18.Kel Kg5 19.Kfl (Not what he wanted, but 19.Kd2 loses immediately to 19…Kh4 20.Ke3
Kh3 21.Kf4 Kxh2 22.Kxg4 Kg2-Silman) 19…Kf4 20.Kel Ke4
21.Kd2 Kd4 22.Kc2 Kc4 23.Kd2 Kb3 24.Ke3 Kxa3 25.Kf4
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 419
Kxb4 26.Kxg4 a5 and wins-analysis by Speelman and Mestel. I (Silman) should add that this is a nice illustration of the old “opposite corner rook-pawn rule” where whoever promotes first wins since he covers the queening square of the other pawn: 27.h4 a4 28.h5 a3 29.h6 a2 30.h7 al=Q, etc.
12…fxe4 13.fxe4
No better is 13.fxg4 hxg4 14.Kgl e3 15.fxe3 e4 16.Khl g3-analysis by
Kmoch.
13…h4 14.Kgl g3 15.hxg3 hxg3, 0-1. The finish could be: 16.f4 exf4 17.e5 g2 18.e6 Kg3 19.e7 f3 20.e8=Q f2 mate.
Passed pawns are important in all endgames, but they have the greatest impact in Queen endings. Why is this the case? Let’s look at the following series of
Kingless diagrams.
Diagram 426
R
7
6
5
4
3
2
a h c d e f g h
8
7
6
5
4
3
2
Diagram 428
Diagram 427
a b c d e f g h
a b c d e f g h
In all these diagrams, we have a position where each side has the same piece (Knight vs. Knight, Bishop vs. Bishop, or Rook vs. Rook), White is a pawn up, and the defender’s piece is blocking the enemy passed pawn. In each case, the blockade of the passed pawn can’t be broken.
Let’s compare the following Kingless diagram with the previous three:
420
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 421
Diagram 429
8
7
6
5
4
3
2
1
a h c d e f g h
Here’s White can break black’s blockade by Qb5/Qb4, Qc5, and Qc7.
This series of four diagrams has shown us that, in positions with Bishop vs. Bishop, Knight vs. Knight, or Rook vs. Rook, without Kings or other pieces or pawns helping, a firm blockade can’t be broken. However, in a Queen ending, this defensive blockade can easily be torn down by the stronger side’s Queen and pawn.
RuLE
Passed pawns have the greatest impact in Queen endgames because a blockade of an advanced passer is very difficult, if not impossible, to maintain.
The following examples should “brutally” push this rule home.
Diagram 430
8
7 RuLE
6 In a Queen endgame,
5 a passed pawn’s value is dependent on how
4
advanced it is. When both
3 sides have passed pawns,
2 the player with the fastest passer (even if he’s a pawn or two behind) often has
a h c d e f g h
a marked advantage.
Black to move, draw
Though Black is no less than four pawns ahead, white’s single pawn on a6 is so powerful that Black is forced to scramble for the draw (1...Qg5+ forces perpetual check).
422 SILMAN’S COMPLETE ENDGAME COURSE
Understanding the rule connected to diagram 430 allows you to save posi tions that appear lost, simply because you’ll know that it’s not a material count that matters, but rather the creation and subsequent speed of a passed pawn.
Diagram 431
8
7
6
5
4
3
2
a b c d e f g h
White to move and draw
White is two pawns down, apparently for nothing. All of black’s pawns are safe, his King is secure, and white’s queenside majority seems to be going nowhere fast. But, as has already been stated, material often takes a back seat to passed pawns in Queen endgames. Since white’s only plus is his queenside majority, he must create a passed pawn in that sector, no matter what the price!
l.Qd4+ Kg8
White’s task is even easier after l…Kh7?? 2.Qf4! when it’s not clear how Black can save the game: 2 …Qb6 (2 …Qxf4 3.gxf4 d4 4.Kgl intending b6 is an easy win for White, while 2 …Qe8 3.Qc7 is also grim for the second player.)
3.Qxf7+ Kh6 4.Qf8+ Kh7 (4…Kg5? 5.Qf4 mate) 5.Kg2 d4 6.Qf7+ Kh6 7.Qb7
Qc5 8.b6!, winning.
2.b6!
Diagram 432
8
7
6
5
4
3
2
a b c d e f g h
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 423
Sacrificing a pawn (which puts him three behind!) in order to create a far advanced passer. Also adequate is 2.Qc5 with the idea of Qc6-b7.
2… axb6 3.Qa4 Qa7
Avoiding the losing race that occurs after 3 …d4? 4.a7 Qa8 5.Qa6 d3 6.Qxb6
Kh7 7.Qb8 Qf3 8.a8=Q Qxf2+ 9.Qg2, 1-0.
4.Qe8+ Kg7 5 .Qc8
Suddenly it’s clear that white’s a6-pawn-soon to be unblocked by Qb7-is light years ahead of black’s whole armada.
5 …b5
Realizing that the tide has turned, Black bails out and plays for perpetual check.
6.Qb7
The most striking way to force the draw. Of course, 6.Kg2 is a bit less extreme:
6…Qd4 7.Qb7 Qe4+ 8.Kgl h4 9.gxh4 Qh4+, =.
6…Qxf2+ 7.Khl Qxg3
Diagram 433
H
7
6
5
4
3
2
a b c d e f g h
Six pawns to one, yet Black must play for the draw!
8.a7 Qel+, 1h-1h.
Since blocking a far advanced passed pawn is often impossible in a Queen endgame, the defender often resorts to perpetual check as his main “survival mode” weapon. To combat this, there are two basic ways for the stronger side to avoid the perpetual and try for the full point.
–… Duel Purpose Queen: The Queen escorts the pawn down the board and simultaneously covers the only road available to its King.
424 SILMAN’S COMPLETE ENDGAME COURSE
….. Doubled Pawn Protective Cover: The doubled pawns around the stronger side’s King act as a wall, blocking any and all checks.
Let’s look at each of these concepts in action.
REMEMBER
The Queen has the ability to control
the queening square of a passed pawn while simultaneously protecting its King.
Diagram 434
a b c d e f g h
Duel Purpose Queen
White to move
White’s monster b-pawn will win the game unless Black can generate some counterplay against white’s King. Black to move would allow 1…Qd1+ (1…Qel +
2.Kh2 Qg3+ 3.Kgl Qel+, etc. is an immediate draw) 2.Kh2 Qd6+ followed by
3 …Qxb6. However, the whole “King safety” problem can be taken care of by one simple move.
l.Qc7!
White’s Queen becomes master of the h2-b8 diagonal (which means that when white’s King moves to h2 it will be immune from harm) while simultane ously forcing through the b-pawn by b7 and b8=Q (the fact that it attacks f7 also doesn’t hurt!).
l…Qel+ 2.Kh2 Qa5
The only chance.
3.Qxf7+ Kh8
Far easier for White is 3…Kh6 4.Qf4+ g5 (4…Kh7 5.b7) 5.Qd6+ with b7 to follow. By staying on the h2-b8 diagonal, White leaves Black no hope of saving the game.
4.Qe8+ Kh7 5.Qe7+
And not 5.b7?? Qc7+ followed by 6…Qxb7. On e7 the Queen covers the c7- and e5-squares, making it hard to check along the sensitive h2-b8 diagonal.
5…Kh6 6.Qxh4+
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 425
Though hardly necessary, why not pocket a free pawn and then return to the same position a little bit richer?
6…Kg7 7.Qe7+ Kh6 8.b7 Qa7 9.Qf8+ Kh7 10.b8=Q, 1-0. Note how the new Queen takes up the task of defending the diagonal.
Diagram 435
8
7
6
5
4
3
2
a h c d e f g h
Doubled Pawn Protective Cover
White to move
In this position white’s doubled f-pawns form a perfect “anti-check seal” against the attentions of the black Queen. Note that if we took the doubled pawn on f3 and made it “healthy” by placing it on h2, Black would be able to annoy his opponent via checks on d5 and/or e4.
Since the King is safe, White can go about the business of promoting his
pawn.
l.c6, 1-0. Black can’t bother white’s King or stop the c-pawn’s decisive advance.
Let’s see an illustration of these ideas at the highest level.
Diagram 436
8
7
6
5
4
3
2
a h c d e f g h
Rubinstein-Capablanca, St Petersburg 1914
Black to move
426 SILMAN’S COMPLETE ENDGAME COURSE
Black’s a solid pawn down, his c-pawn is loose, and white’s Queen is more actively placed than black’s. One might think that a superb endgame player like Rubinstein will reel in the full point, but Capablanca is aware of the principles illustrated in this section and knows that he must use his queenside pawn major ity to create a passed pawn or face certain defeat. If this calls for the sacrifice of another pawn or two, so be it!
l…b4!
Immediately striving to make a passed pawn.
2.c4!
This is white’s best winning try. In the actual game White played 2.Qxc5 when, after 2 …bxc3 3.Qxc3 Qbl+ 4.Kh2 Qxa2 (the powerful passed a-pawn will soon force White to take a draw by perpetual check) 5.Qc8+ Kh7 6.Qf5+ g6 7.Qf6 a5 8.g4 a4 9.h5 gxh5 10.Qf5+ Kg7 ll.Qg5+ Kh7 12.Qxh5+ Kg7, the game was agreed drawn.
An instructive position arises that beautifully illustrates our “a passed pawn is
more important than material” philosophy after 2 .cxb4 Qxb4 3.Qxa6 c4 4.Qa8+ Kh7 5.Qe4+ Kg8 6.Qe8+ Kh7 7.Qxf7 c3.
Diagram 437
8
7
6
5
4
3
2
1
a b c d e f g h
Even though White is ahead by three big pawns, the c-pawn is too strong White must take a perpetual check.
2 …Qc8!
Simply bad are 2 …a5?? 3.Qd8+ Kh7 4.Qxa5 and 2 … b3?? 3.axb3 Qxb3
4.Qxa6.
Tempting but inaccurate is 2 …Qe4? 3.Qxc5 Qbl+ 4.Kh2 Qxa2 5.Qxb4 Qxf2
6.Qe7! with a winning position.
This leaves us 2 …Qa 7 which, apparently, should (along with 2 …Qc8) prove adequate: 3.g4 (3.Qd8+ Kh7 4.Qa5 Qe7 5.Qxa6 and now 5 …Qxh4 6.Qb5
Qe7 is okay, but I prefer 5 … Qe4, =.I’ll let you figure out the variations here!)
PART EIGHT – ENDGAMES FOR MASTERS (2200-2399) 427
3 …aS 4.h5 a4 and, as always, the passed pawn will prove to be equivalent to a small nuke:
Diagram 438
8
7
6
5
4
3
2
a b c d e f g h
S.gS b3 6.g6 Qa8 7.axb3 a3 8.Qd7?? (8.gxf7+ Kxf7 9.Qg6+ Kg8 10.h6 Qa7
11.h7+ Kh8 12.Qe8+ Kxh7 13.Qh5+ is a draw by perpetual check and would be the sane way to play for White) 8…a2 9.gxf7+ Kf8 and Black wins.
The idea behind 2 …Qc8 is twofold: Black intends both 3…Qf5 with …Qb1
and …Qxa2 to follow, and also 3.. .aS followed by 4…a4 when a passed pawn will soon be born.
3.Qb6
This freezes black’s a-pawn and, apparently, ties black’s Queen down to the defense of both a6 and c5.
3 …Qf5!
Refusing to buy into the “defend your pawns passively” myth.
4.Qxa6 Kh7!
Diagram 439
8
7
6
5
4
3
1
a b c d e f g h
It’s interesting how Black, now two pawns down, can play a quiet move like this and still assure himself of good chances. The idea behind 4…Kh7 is clear:
428 SILMAN’S COMPLETE ENDGAME COURSE
he wants to avoid an unpleasant back rank check after the Queen eventually penetrates to b1 or c2.
5.Qa7
As usual, allowing Black to create an advanced passed pawn is a bad idea:
5.Qb7 Qbl+ 6.Kh2 Qxa2 7.Qxf7 b3 8.Qh5+ Kg8 9.Qd5+ Kh7 10.Qxc5 b2 and
White should bail out with perpetual check.
5 …f6!
Another calm move, telling White that he has no good way to improve his position. Also possible was 5…QeS!? with the usual idea of giving up lots of pawns to create an advanced passed b-pawn.
6.Qe7
Also fine for Black is 6.Qa4 Qbl+ 7.Kh2 Qb2 with annoying threats.
The most dangerous move is probably 6.e4 when 6…Qxe4?? 7.Qxc5 is win ning for White. However, 6…Qe5! 7.Qf7 Qal+ 8.Kh2 Qxa2 gives Black enough counterplay to draw.
6…Qbl+ 7.Kh2 Qxa2
Diagram 440
8
7
6
5
4
3
2
1
a b c d e f g h
Another monster passed pawn
8.Qe4+ Kh8 9.Qe8+ Kh7 10.Qh5+ Kg8 ll.Qxc5 b3 when, as we’ve seen time and again, the strength of the passed pawn, mixed with a Queen’s ability to torment the enemy King, should allow Black an easy draw. (Some of the varia tions were culled from those by Paul Keres in the classic The Art of the Middle Game.)
Capivara subserviente do Xadrez
