click for larger view

<19...Bxf4!!> This tactical stroke is already decisive.

Keres annotation of the game is:

"Black must avoid the other enticing sacrificial opportunity 19 ... KtxP; 20 PxKt, QxP, wince then White could force an exchange of Queens by 21 QK6ch, K-R1; 22 QKt-B3."

click for larger view

White to move

1) -1.06 (32 ply) 20.Nh3 Bd6 21.Nf2 Bf5 22.Nf1 c5 23.dxc5 Bxc5 24.Be3 Bd6 25.Qe2 Bd7 26.Re1 Bb5 27.Qd1 Qf5 28.Qg4 Qf7 29.Nd2 Ne5 30.Qf4 Qe6 31.Qg5 Nd3 32.Nxd3 Bxd3 33.Bf2 Qf7 34.Qe3 Kh8 35.Nb3 h6 36.Nc5 Bxc5 37.Qxc5 Be4 38.Qe3 Qh5 39.Qe2 Qh3 40.Be3 Kh7

2) -2.93 (31 ply) 20.Ndf3 Bd6 21.Qe3 Bf5 22.Bd2 h6 23.Rf1 Bg4 24.Rf2 hxg5 25.Nxg5 Qxf2+ 26.Qxf2 Rxf2 27.Kxf2 c5 28.dxc5 Bxc5+ 29.Be3 Be7 30.Nf3 a5 31.b3 Kf7 32.h4 Bd6 33.Ng5+ Kg8 34.c4 dxc4 35.bxc4 Ne5 36.Ne4 Nxc4 37.Nxd6 Nxe3

3) -3.27 (31 ply) 20.Ngf3 Bd6 21.Qe3 Bf5 22.Kg2 Bg4 23.Ng1 h5 24.Ndf3 h4 25.Bd2 h3+ 26.Nxh3 Bxf3+ 27.Kg1 Bg4 28.Ng5 Qf5 29.Re1 Qc2 30.b4 Ne7 31.Rc1 Qxa2 32.Qd3 Bf5 33.Qe3 Be4 34.Re1 Nf5 35.Qe2 Nxg3 36.hxg3 Bxg3 37.Nxe4 Bxe1

6.0 minute analysis by Stockfish 9 v010218

= = =

<19.Qxe1 Bxf4 20.gxf4?>

click for larger view

Black to move

1) -5.96 (27 ply) 20...Nxf4 21.Ndf3 Ne2+ 22.Kg2 h6 23.h3 hxg5 24.Nxg5 c5 25.a4 cxd4 26.cxd4 Qxd4 27.Bd2 Qxb2 28.Rc1 Qe5 29.Rc3 Be4+ 30.Nxe4 Qxe4+ 31.Kh2 Qe5+ 32.Kh1 d4 33.Rxc7 d3 34.Rc1 Qh5 35.Kg2 Nxc1 36.Bxc1 Qf3+ 37.Kh2 Qe2+ 38.Qxe2 dxe2

2) +1.90 (27 ply) 20...Qxf4 21.Qe6+ Kh8 22.Ndf3 Qf5 23.Qxf5 Bxf5 24.Ne1 h6 25.Ngf3 Be4 26.Nd2 Bf5 27.Kf2 Nh4 28.Kg3 Re8 29.Nef3 Nxf3 30.Kxf3 g5 31.Nf1 Kg7 32.Be3 a5 33.Re1 a4 34.h4 gxh4 35.Bxh6+ Kf7 36.Rxe8 Kxe8 37.Ne3 Bd3

In my book, this was a well-earned victory by Keres (who of course never played this line again).

This game was played just after the tournament had moved from the Netherlands to Moscow. Keres had scored 4-4 to this point, and was tied for third place with Smyslov behind Botvinnik at 6-2. As Keres noted, he needed a win here to get back into contention. Indeed, he scored two wins and a draw in his first three games in Moscow, and pulled to within a point and a half of Botvinnik before they sat down to play their Round 15 (of 25) game. This game was therefore of important sporting interest (though Botvinnik soon pulled away from his rivals).

Euwe had lost five and drawn three before coming to Moscow. This loss--coupled with his demolition at the hands of Botvinnik in the following round--left little doubt that Euwe was destined to remain in the cellar in this event.

1. e4 e5
2. Nf3 Nc6
3. Bb5 a6
4. Ba4 d6

As in Round 1 against Euwe, Keres employed this deferred variation of the Steinitz Defense to the Ruy Lopez. Euwe got the edge in that earlier game (before getting blown away by Keres), and Keres had prepared a major surprise for the instant game.

5. c3

As played by Euwe in the earlies game. It was also played by Smyslov in this 11th Round in his game against Reshevsky (but with a much better result for White).

5... f5

click for larger view

Keres had played the more usual 5...Bd7 in the above-cited first round game. Reshevsky played 5...Ne7 against Smyslov. The text is the so-called "Siesta Gambit" which, as Golombek noted, was introduced in master play by Capablanca at Budapest 1928.

"Normally such an early attacking attempt from Black is rather suspect, but here, when White has made the unnecessary non-developing move 5. c3, this continuation comes into serious consideration and yields Black quite good prospects." (Keres)

Theory now rates White as definitely better after 5...f5. But, as Keres further noted:

"Owing to my position in the tournament I was compelled to play for a win in this game at all costs, so as to retain any chance whatsoever of the first place. For this reason, the opening variation suited my purposed admirably."

6. exf5

"...the line recognized as correct by most theoreticians"--(Golombek).

6... Bxf5

click for larger view

7. d4

According to Keres, this was at the time deemed the strongest continuation for White. Keres noted, however, the suggestion of 7.0-0 by Panov. Keres called this idea "worthy of consideration" and "could become a threat to the whole variation chosen by Black," since if then 7...Bd3 White can simply play 8. Re1. Since Keres' annotations, 7. 0-0 had indeed replaced 7. d4 as the best continuation and probably the only one to offer White a way to retain any meaningful edge.

7... e4!

click for larger view

This could almost be called the refutation of 7. d4. The position is now complex and double-edged, which is probably just what Keres wanted. The real key to Keres' strategy, however, was not revealed until his 9th move.

For now, note the placement of the White Bishop on c1. It was to remain a sorry spectator for the balance of the game, only stirring on move 25 by which time White was completely busted.

Already, Euwe had to make an important decision . Unlike Keres, who was obviously prepared to the hilt. He selected:

8. Ng5

This is what Horowitz played against Fine at Syracuse 1934. It was also played by Steiner in his loss to Kashdan at the 1931 Prague Olympiad. The move, as pointed out by Golombek, threatens 9. Qb3.

Back in 1909 in the game cited by <sudoplatov>, Capablanca played 8. Qe2. Golombek claimed that this is "good for Black," but with 8...Be7 [Marshall's move, perhaps 8...Qe7 was strongest] Capablanca was OK until after 9. Nfd2 Nf6 he erred with 10. h3 (10. f3 seems to lead to equality). Marshall now had the edge, but Capablanca still achieved a draw.

In his 1931 game against Steiner, Capablanca tried the inferior 8. d5 and was worse after 8...exN 9. dxN b5 10. Qxf3 BxN 11. Bb3 Bg6 losing a piece. Capablanca did have some compensation, and once again managed to draw.

Capablanca was Black against Steiner at Budapest 1928 where White played 8. Bg5, which once again passed the edge to Black: 8...Be7 9. Nh4 (perhaps 9. BxB would have been better) Be6 10. BxB NgxB [even better would have been 10...QxB. the text let White off the hook] 11. Qh5+ g6 with approximate equality (and, not surprisingly, Capablanca managed to win).

The critical variation here, pointed out later by Keres, is the piece sacrifice 8. 0-0 exN 9. Qxf3. In this line, according to Keres: "White would retain a lasting initiative with good attacking prospects, but it goes without saying that such a method of playing is not to everyone's taste."

Romanovsky tried 8. Ng1 against Bondarevsky at the 1945 USSR Championship, but this line hardly looks attractive (and not surprisingly Romanovsky lost that game).

Perhaps White's best option was 8. Nfd2, though so far as I am aware that move has never been played.

After Euwe's actual 8. Ng5 the position was:

click for larger view

8... d5

The "main line of the Siesta Variation" (Keres).

Perhaps the most prudent line here for Black is 8...Be7. Golombek condemned this alternative as leading to "much the better game for White." But this assessment was based on 9. 0-0 BxN which does give White the better chances after 10. Qh5+. But Black looks fine here after the superior 9...b5 10. Bc2 d5.

9. f3

"The main line of the Siesta Variation and supposed to be good for White"--(Kmoch).

click for larger view

It was in the above position that Keres unleashed:

9... e3!

"!?"--(Kmoch)

click for larger view

"This pawn sacrifice is the only way to keep the counter-attack alive"--(Golombek)

"A positional gambit. Black attacks the Knight and is threatening 10...h6 11. Nh3 BxN, apparently with serious effect"--(Kmoch)

"The intention behind the pawn sacrifice is simple: 10. BxN h6 11. Nh3 BxN and Black has compensation. But White does not have to take the pawn at once and the working-out of this idea leads to very complicated positions. "--(Euwe)

Actually, the text had been introduced, as previously mentioned by Fine back in 1934, and had been played in several Soviet tournaments:

"This pawn sacrifice is not new but in this game Black essays the attempt to rehabilitate its dubious reputation...In practice 9...e3 is undoubtedly Black's best chance since 9...exf3 10. 0-0! [and not 10. Qxf3 Qe7+ 11. Be3?? as given by Golombek--KEG] and 9...h6 10. fxe4 give White the better [much better--KEG] game"--(Keres).

10. f4

click for larger view

"The engines like White's 10. f4 but Euwe was caught off guard."--(<cehertan>)

10... Bd6

click for larger view

"Black's main objective in this position is NOT protecting the e-pawn...but to keep White's knight away from the important e5-square. The text-move initially wins a tempo as White is forced to protect his f-pawn"--(Keres)

11. Qf3

"?"--(Kmoch)(Keres)

"...not entirely satisfactory for White, since it does no more than secure equality"--(Golombek)

"Too gullibly played. White hurries to capture the Pawn without considering what Black must have intended with his sacrifice. White plans to capture with check, then bring his Knight via f3 to e5. But the plan costs valuable time"--(Kmoch)

"This move cannot be considered adequate. First of all, White loses valuable time...and secondly the queen deprives the knight of the important f3 square from where it could transfer to e5"--(Keres)

The above comments notwithstanding, the text hardly warrants a "?" White is certainly not worse after this move. Given that Euwe has been taken by surprise, it is unfair to trash him for making sure his head remained above water. There were, however, stronger options.

(A) 11. Bxe3. A solid safe choice, though not much an improvement on the text. Play might then continue: 11...Qe7 12. Qe2 Nf6 13. Nf3 Bg4 and now not 14. h3 (given by Golombek and Keres) but 14. 0-0 with a small edge for White.

(B) 11. Qe2--A solid choice not mentioned by any of the commentators I have seen. White looks better in the endgame that would result from 11...Bxf4 12. Bxe3 BxB 13. QxB+ Qe7 14. Bxc6+ bxc6 15. QxQ+ NxQ 16. 0-0 h6 17. Nf3/

(C) 11. 0-0 [Kmoch's choice] which likely results in a small edge for White: 11...b5 (rather than Golombek's 11...h6 or Kmoch's 11...Bxf4) 12. Bb3 Nf6 13. Bxe3 0-0,

(D) 11. Qh5+ --probably best. Play might then have continued: 11...Bg6 12. Qf3 Qf6 13. Qxe3+ Nge7 14. 0-0 with a clear advantage--though hardly decisive--for White.

Given that none of the alternatives yields anything overwhelming for White, Euwe's choice, which was also played by Horowitz in his 1934 game against Fine and by Robert Byrne in his 1977 draw against Westerinen at Geneva, should not come in for any major criticism. it left the position as follows:

click for larger view

11... Qf6

"!"--(Kmoch)

"11...Bd7 had been played in Horowitz--Fine Syracuse 1934 where White went on to win [albeit after a major blunder by Fine--KEG]. 11...Qf6 ws Keres' improvement" -- (<plang>)

"...much stronger than the 11...Bd7? that Fine played against Horowitz"--(Kmoch)

I agree with Kmoch, but even stronger than Horowitz' 12. Qxe3+ would have been 12. Qxd5 Qe7 [practically forced] 13. Bb3 0-0-0 14. Qf7 with a major bind on the Black position.

12. Qxe3+ Nge7

click for larger view

"In this position Black, with his excellent development and taking into account the many weaknesses in his opponent's position, has attained full compensation for his pawn. White must now play very carefully if he does not wish to get the disadvantage"--(Keres)

13. BxN+

"?"--(Golombek)(Keres)(<plang>)

"An anti-positional capture that increases the strength of Black's attack. It gives Black three further advantages: (1) two Bishops; (2) more open lines; (3) a stronger center"--(Golombek)

"Difficult to understand. White has a bad game anyway [??--KEG], and by exchanging this Bishop he only accentuates the severe weakness of his white-colored squares"--(Kmoch)

"This exchange is undoubtedly weak. First it improves Black's central position, and secondly it weakens the light squares in White's position which enables Black to generate a strong attack"--(Keres)

"13. BxN+? was a serious error because it (1) weakened White's position on the White squares; (2) strengthened Black's center; and (2) gave Black the two Bishops"--(<plang>)

While there is considerable merit to the above quoted comments and while 13. BxN+ was almost certainly not the strongest move available to White, the criticism of Euwe by these commentators strikes me as over the top. For one thing, Euwe most certainly did not have a "bad game" either before or after 13. BxN+. Among other things, White had an extra pawn. A more balanced assessment was:

"This exchange has an advantage (doubling of the Black pawn formation) and a drawback (White allows Black to have the Bishop pair"--(Euwe).

For what it's worth, both Fritz and Stockfish rate the position as (0.00) after the text. While in such positions I do not take the evaluations by my silicon friends as gospel, it is grossly wrong to say that the text did anything more than forfeit Euwe's chance to retain an edge. After this exchange, Black's compensation for his pawn minus was adequate. But he was still down a pawn, and that surely counts for something. As will be seen, Euwe was not at serious risk of losing until his disastrous 15th move.

The above being said, the above-quoted commentators are all surely correct that 13. 0-0 was far superior. To follow Keres' suggested continuation: 13...0-0 [13...0-0-0 would be better, since White should get much the better game after 13...0-0 14. Nf3 [Keres notwithstanding, 14. Nd2 is probably even stronger] Be4 and now now both Keres' 15. Ng5 and 15. g3 leave White with much the better chances

Back to the actual game:

13... bxB

click for larger view

14. 0-0

Contrary to the analysis by Golombek and Euwe, 14. Nf3 was also fully sufficient for equality. If now 14...0-0, and as pointed out by Keres, White can just respond 15. 0-0.

Golombek and Euwe find fault with 14. Nf3 based on their notion that Black would then get the better game with 14...BxN 15. RxB Qg6. But this is clearly wrong and 14...BxN would make Black's position hazardous. Golombek gave the continuation 16. Ra1 Qxg2 17. Rg1 Qh3 18. Rxg7 0-0-0 which is fine to this point. But Golombek then has White blundering with 19. Ng5? which loses to 19...QxQ+ 20. BxQ Nf5. Golombek, however, overlooked the far superior 19. Rg3 after which White--is anyone--is better.

Thus, both Euwe's 14. 0-0 and 14. Nf3 are fully satisfactory for White here.

14... 0-0

click for larger view

This, as it turns out, was the crucial position in the game. As Euwe later pointed out, White is fine after 15. Nf3. Black's best then would have been 15...Be4 after which White can simply play 16. Ng5 either leading to a draw by repetition if Black plays 16...Bf5 or to equality after 16...Rae8 17. NxB dxN 18. Qxe4 Nd5 19. Qd3 [rather than Golombek's awful 19. Qf3? Bxf4 20. BxB NxB after which White would be in serious difficulties].

But instead of 15. Nf3 and equality, Euwe--perhaps fretting about his undeveloped Queen-side--played the disastrous (and losing):

15. Nd2??

click for larger view

How bad was Euwe's 15th move?

"...the text is too slow and suffers from the defect of temporarily weakening f4"--(Golombek)

"This leads to disaster"--(Kmoch)

"The text move should...be regarded as the decisive mistake"--(Euwe)

"A mistake that leaves White in a lost position...a speedy catastrophe follows; White is no longer able to close the e-file in time, and this enables the black rooks to penetrate into White's position decisively...The decisive mistake after which Black obtains an irresistible attack against his opponent's undeveloped position"--(Keres)

15... Ng6

"!"--(Kmoch)

"Decisive"--(Kmoch)

"Not only the black queen but also the knight makes good use of g6"--(Euwe)

click for larger view

16. g3

Insufficient, but so was everything else.

If instead 16. Ngf3 Black wins with 16...Bxf4 17. Qf2 Rae8 18. Re1 Be4

If 16. Nb3 (which Golombek called "The only other way of protecting f4, then 16...Nxf4 17. RxN QxN 18. Qg3 QxQ 19. hxQ BxR and wins.

And if 16. Qg3 then not 16...Nxf4? 17. RxN BxN 18. QxB which leaves White with two minor pieces for the sacrificed win and excellent winning chances but Kmoch's crushing 16...h6!

Simply put, White is lost.

16... Rae8

"!"--(Keres)

"It is clear that White's undeveloped position must sooner or later collapse"--(Keres)

This left:

click for larger view

17. Qf2

White had nothing better. As both Golombek and Keres pointed out, 17. Qf3 loses a piece after 17...h6.

17... Bd3

As most of the commentators pointed out, Black wins more spectacularly with 17...h6 18. Ngf3 Bxf4! since if White then tries to grab a piece with 19. gxB then 19...Nxf4 20. Kh1 (everything else is even worse) Re2 annihilates the White position. But since the more prosaic text wins without the need for difficult over-the-board calculations, Keres' move was the most practical way to win.

18. Re1

click for larger view

18... RxR?

The commentators on this game have allowed this lemon to pass with virtually no discussion. The only comment on this move was by Euwe:

"Black could also have decided the game with 18...h6 19. Ngf3 BxN 20. BxB Nxf4."

Euwe was correct that 18...h6 wins for Black. But he--apparently like everybody else--got everything else wrong.

For starters, he missed Black's best line: 18...Nxf4 and now White can choose his poison: (A) 19. RxR (as good as anything) RxR 20. gxN Re2 21. Qh4 Bxf4 22. Ndf3 BxN 23. BxB QxN 24. Qg3 (forced) QxQ+ 25. hxQ Rxb2 leaving White two pawns down in the ending where Bishops of opposite colors cannot save the day; or (B) 19. Nde4 Ne2} 20. QxN dxN 21. Qg2 h6 22. Nh3 e3! 23. Bxe3 Be4 24. Qf1 (what else?) QxQ+ 25. RxQ Bf5 26. RxB (else Black wins a piece) RxR and wins with his exchange plus.

Secondly: While Black can still win with 18...h6 ; after Euwe's 19. Ngf3 [19. Ndf3 is better but still loses to 19...hxN 20. Nxg5 RxR+ 21. QxR Nh8 (en route to f7)] 19...Bxf4? blows the win [19...Nxf4 is better, though the win now is less than clear] after 20. gxB Nxf4 21. Kh1 Nh3 22. Qg2 RxR+ 23. NxR Nf2+ 24. Kg1 Qe6 25. NxB NxN 26. Nf1 (forced) and White can probably untangle himself with good chances to save the game.

But thirdly, and most importantly, Keres' 18...RxR+ doesn't win:

19. QxR

click for larger view

And here, Keres played what seems to be the culmination of his attack:

19... Bxf4

"!"--(Golombek)(Euwe)(Keres)

"!!"--(Kmoch)(<whiteshark>)

"A beautiful stroke, after which White's position collapses like a house of cards"--(Golombek)

"A pretty (though not particularly deep) combination that crushes White's badly developed game"--(Kmoch)

"This sacrifice breaks the back of the white position. No wonder: Black has all his pieces at hand, whereas White's queenside is still undeveloped"--(Euwe)

"...immediately decisive"--(Keres)

But is this pretty move actually "decisive"?:

click for larger view

Knights are the trickiest pieces in chess, and here the White Knights can work a miracle with: 20. Nh3! (a move apparently not considered by any of the commentators) 20...Bd6 (the only winning chance for Black) 21. Nf2 which would have left:

click for larger view

Here, amazingly, White seems to hold (so great to have Stockfish and Fritz--I'm sure I wouldn't have found this on my own).

Euwe, playing over the board played the "obvious":

20. gxB?

The only alternatives considered by the commentators are all hopeless:

20. Qe6+ QxQ 21. NxQ Be3+ 22. Kh1 Rf1+ 23. Kg2 (since 23. NxR allows mate in one with Be4 checkmate) Rf2+ 24. Kh1 RxN and it's game over; or

20. Ngf3 Bh6 (superior to Kmoch's 20...Bd6 which probably also wins).

But after Euwe's 20. gxB the position was:

click for larger view

20... Nxf4!

Cute!

click for larger view

21. Ndf3

Finally beginning to develop his tangled Queen-side, but this is too late. In fairness to Euwe, however, nothing else would have helped.

If instead 21. Qe5, then, as Golombek demonstrated, Black annihilates the White position: 21...Ne2+ 22. Kg2 [22. Kh1 would only have delayed the inevitable by a few moves--KEG] Qf2+ 23. Kh3 Bf5+ (23...Nf4+ is even quicker--KEG]

Also hopeless is (once again as per Golombek] 21. h4 Qf5 followed by 22...Qg4+ (Keres' 21...h6 also wins)

If 21. Ngf3 and--contra Kmoch--21...Qg6+ wins easily: 22. Kf2 Nh3+ (or even 22...Qg2+ since--Kmoch notwithstanding-- after 23. Ke3 Qg4 cleans house)

Finally, and as per Keres: if 21. Qh4 Ne2+ followed by 22... Be4+.

If sum: White is busted:

21... Ne2+

click for larger view

22. Kg2

22. Kf2 would have only slightly have prolonged the game.

22... h6

click for larger view

"Winning both the White Knights"--(Kmoch)

"It is all so very simple: the knight on g5 cannot budge and will be lost"--(Euwe)

"Winning back the sacrificed piece with a lasting attack; White's position is lost"--(Keres)

23. Qd2

"Against other Queen moves simply 23...hxN is sufficient"--(Golombek)

So too after 23. h3 or 23. h4 or 23. Bd2.

23... Qf5
24. Qe3 hxN

click for larger view

25. Bd2

Finally moving this Bishop, but much much too late.

25... Be4

click for larger view