click for larger view

<andrea volponi: 21...Qxc5! -Nxf7 d3+ -Kh1 Qc2 -Nh6+ Kh8 -Nf7+ Kg8=>

There's also

1) =0.00 (33 ply) <21...gxf5> 22.Rxf5 Ne5 23.Rdf1 Rf8 24.Bxd4 Qa3 25.Qxa3 bxa3 26.Bc3 Rb8 27.Nf3 Nxf3+ 28.R5xf3 Bxc3 29.Rxc3 Rb5 30.Rf4 Rb1+ 31.Rf1 Rb5

and the game continuation is clearly worse:

3) +1.19 (32 ply) <21...Nxc5?> 22.Qh3 h5 23.f6 Bxf6 24.Rxf6 Nd7 25.Rd6 Qxg5 26.Rxd7 Rxe4 27.Bxd4 Rbe8 28.Rf1 R4e7 29.Bf2 a5 30.Qd3 c5 31.Rd5 Re5 32.Qb5 Rc8 33.h4 Qf5 34.Rxe5 Qxe5 35.Rc1 Qd6 36.Qc4 Rd8 37.Be3 Qd5 38.Qxd5 Rxd5 39.Rxc5 Rxc5 40.Bxc5

In reviewing the game, it is fascinating to read the commentary by contemporary sources (which I am only quoting in part--the entirety should be read to get the full flavor).

1. c4 Nf6
2. Nf3 c6

"Smyslov chose the Grunfeld Defense in the tournament. But in this game he chose the Slav Defense...But in fact the game soon takes on the character of a Grunfeld Defense." (Keres)

3. Nc3

Showing he does not intend to employ the Reti system for which it is necessary to have the diagonal a1..h8 clear for the sweep of the Queen's Bishop after b3." (Golombek).

3... d5
4. e3

"The text-move gives rise to theoretically less studied positions, but hardly offers White better prospects than 4. cxd5." (Keres)

4... g6

The game now transposed into the Schlechter Variation, which may be regarded either as a branch of the Grunfeld Defence or as a variant of the Slav Variation of the Queen's Gambit." (Golombek)

An irregular way of reaching a comparatively unfavorable line of the Gruenfeld. 4...e6 is routine and probably as good as any." (Horowitz)

click for larger view

Horowitz was certainly correct that 4...e6 was more usual, but the text was hardly bad and makes perfect sense for someone comfortable with Gruenfeld set-ups for Black.

5. d4

"A very unusual route has taken us to the Slav Queen's Gambit." (Euwe)

5... Bg7

click for larger view

6. cxd5

"White has to choose here between this clear-cut exchange line and the maintenance of tension in the center by 6. Qb3." (Golombek)

"...White usually proceeds with 6. Bd3 or 6. 0-0. The text-move is a novelty." (Keres)

In fact, the text had been played at least 11 times before, and was apparently championed in the early 1920's by Saemisch. 6. Be2 and 6. Bd3 were more usual and look better.

6... Nxd5

As all of the commentators agreed, 6...cxd5 was better, but provide insight into Smyslov's idea in playing the text:

"Smyslov wishes to retain the option of c5 later on. In doing so, however, he gives up on the center. 6...cxd5 is stodgy but solid." (Horowitz).

"With 6...cxd5 he can obtain easy equality; the text is more ambitious since it avoids symmetry...but it suffers from the defect of allowing White some control of the centre." (Golombek)

"6...cxd5 is simpler, although even then White gets some slight initiative after 7. Bb5+ [As Keres correctly points out, 7. Bd3 would be better in this line]." (Golombek)

"Smyslov takes the first chance to stear the game into what for him are familiar tracks, and by so doing he justifies White's previous exchange...[which Keres ultimately decided was not best]...6...cxd5 is stronger...White should [then] proceed with 7. Bd3..." (Keres.

The text followed Colle-Rabinovitch, Baden-Baden 1925, and was later played by Yusupov and Korchnoi (as a senior in 2006).

The position after 6...Nxd5 was:

click for larger view

7. Bc4

click for larger view

"A Keres move...He hopes to overwhelm his opponent by fluid piece play in the early middle game..." (Golombek)

7. Bd3 was the main alternative.

7... 0-0

Smyslov could also have played 7...NxN immediately.

8. 0-0 b6

"?"--(Golombek)(Euwe)(Keres) (Wade--Whiteley--Keene)

"This move is wrong since it weakens Black's Queen-side and does nothing towards the solving of the problem of the center."(Golombek)

"This move which deprives Black's Knight of its natural retreating square is a mistake..." (Keres)

"Unnecessarily weakening his position." (Euwe)

Despite the nearly universal condemnation of this move, it was not all that bad and had little to do with Smyslov's ultimate defeat.

The alternative moves suggested by the commentators that criticized 8...b6 are for the most part no real improvement. Keres and Wade-Whiteley-Keene claim that 8...Nb6 was best, but White would be in at least as good shape as in the game after 9. Bb3.

If 8...Be6 (suggested by Horowitz) White then would have a major advantage after 9. Ng5 (much better than Golombek's 9. Qb3).

The only real improvement on 8...b6 was 8...NxN immediately. Golombek claimed that this would be bad since Black has lost a tempo after 8...NxN 9. bxN c5, but: (a) 9...Qc7 is better for Black here; and (b) Black is better even on Golombek's line as compared with the other suggestions.

After the much maligned 8...b6 the position was:

click for larger view

9. Qb3

As all of the commentators seems to have agreed, this is strong but 9. e4 was probably even better. Both moves give White a small but real edge:

"This forces Black to exchange on c3, but for this purpose 9. e4 was even stronger." (Keres)

"Even stronger is 9. e4." (Euwe)

9. Qb3 left the position as follows:

click for larger view

9... NxN

"Virtually forced, since White was threatening 10. Ne5." (Euwe)

Black was basically OK after the text, and none of the alternatives to 9...NxN were at all appealing:

"If 9...Be6 10. g5." (Golombek)

"If 9...Ba6 10. NxN cxN 11. Bxd5 wins a pawn." (Golombek) [10...BxB is better than Golombek's move, but even in this case White would be much better after 11. Nxe7+ QxN 12. QxB White would be a pawn up with excellent winning chances.

"9...Nc7 10. Ne5 Be6 11. f4 would yield White a strong attack." (Keres) [Keres' line is flawed. After 9...Nc7 either 10. Rd1 or 10. e4 would be best and give White a clear edge. After Keres' 10. Ne5 Black should simply play 10...BxN].

10. bxN

click for larger view

10... Ba6

"Black must exchange the Bishop since otherwise White would be ablt to initiate a strong King-side attack." (Keres)

None of the alternatives to 10...Ba6 was any better:

"10...Bg4 allows White a formidable attack by 11. Ng5." (Golombek).

10...Nd7 and 10...Qc7 were the best alternate choices, but neither was an real improvement.

Perhaps the best justification for the text was that given by Horowitz:

"Black's strategy is to cut down the wood and, in consequence, the attacking chances. This appears sound, particularly against such a formidable attacking player as Keres."

11. Ba3

There was a hot difference of opinion on the merits of this move. Golombek loved it:

"An excellent attacking move which both prevents Black from freeing his game by Qc8 followed by BxB and Qa6 and deprives him of the possibility of e4." (Golombek).

Keres disagreed:

"Not best, since the Bishop stands unfavorably on a3 after either c5 or b5 followed by a5. It was therefore better to play 11. a4 with attacking chances on the queen-side or 11. e4 and a further development of the Bishop on the c1..h6 diagonal."

But after either of Keres' alternative moves, Black seems just fine after 11...BxB.

My view was best put by Euwe:

"11. BxB NxB e4 or 12. a4 [or perhaps better still 12. Rd1--KEG] is more effective."

After 11. Ba3 the position was:

click for larger view

11... BxB
12. QxB

click for larger view

12... Re8

"Intending to continue with 13...Qd5." (Euwe)

"Black wants to play 13...Qd5 and for this reason protects his e-pawn." (Keres)

Since Keres' next move prevented 13...Qd5, some commentators rejected the text as pointless:

"...White stops [13...Qd5] by his next move and he might as well have played 12...b5 at once." (Golombek)

But White's position after 12...b5 13. Qb3 looks even better than what occurred in the game, since Black' Queen-side weaknesses--especially on the Black squares--are now even more pronounced.

David Bronstein suggested the heroic remedy of 12...c5, but as Keres pointed out, after 13. dxc5 Qc7 [Bronstein's trick] 14. Qe4 bxc5 15. Bxc5 (much better than Keres' 15. Rab1) Nc6 16. Bd4 White ends up a (somewhat weak) pawn to the good with much the better chances. In this line, giving up the White Queen for two Rooks via 15. QxR Nc6 16. QxR+ gives White no advantage.

In sum, Smyslov's move (12...Re8), though not conferring anything like full equality, was probably best.

The position after 12...Re8 was:

click for larger view

13. e4

Keres' aggressive intentions (he might also have tried 13. Qb3 to make room to get his c-pawn rolling) were clear and he now dominated the center. Smyslov still had a defensible position, but he now faced an unpleasant choice:

click for larger view

13... b5

"More or less forced." (Horowitz)

"Otherwise Black would be unable to continue his development." (Euwe)

"Black is already forced into such weakening maneuvers, since by other means it would not be possible for him to develop his Knight." (Keres)

While the text must have been noxious to so talented a positional player as Smyslov, it did give him a defensible position while the alternatives all seem worse.

13...Qc8, discussed and rejected by Golombek, does not look attractive after 14. Ng5 or 14. Rfd1 (both better than Golombek's 14. e5 b5.

13...e5, as Golombek correctly noted, "only facilitates White's attack (after 14. Nxe5 BxN 15. dxB Rxe5 16. f4).

13...Qc7, as pointed out by Horowitz, could lead to trouble after 14. e5.

And 13...Nd7 sacrificing a pawn looks questionable at best.

Thus, by default, Smyslov's 13...b5 was likely best, and left the position as:

click for larger view

14. Qb3 Nd7

click for larger view

15. c4

"!"--(Keres)

"Played at just the right moment. This prevents Black from posting his Knight on c4 via b6."

"Black was threatening 15...Nb6 with counter-chances." (Euwe)

Thwarting Black's Knight from reaching c4 which would yield Black good counter-play..." (Keres)

As it turns out, Smyslov had more than adequate defensive resources after the text, and 15. Rad1 was perhaps the only theoretical way to retain an edge for White. But the text--and Keres' subsequent energetic play--must have been difficult to deal with over the board. Keres' theme of restricting the play of Black's Knight (which reminds me of strategies employed in some of the game of Alpha Zero) was fascinating, and the fact that the complications proved too tangled for even so great a defensive player as Smyslov to untangle more than justifies Keres' attacking scheme.

The position after 15. c4 was:

click for larger view

Smyslov now faced the problem of how to deal with Keres' scary-looking pawn-phalanx and impending King-side attack.

15... Rb8

As every commentator on the game has noted, 15...c5 was best here. White would then have many plausible options, none of which appear to yield any edge to White:

(A) 16. cxb5 was analyzed extensively by Keres, Euwe, and Golombek. It does not lead to any edge for White after 16...cxd4 (although Keres claims a "small positional advantage" for White because of his extra Queen-side pawn). Black's resources now include a break with ...e5 of occupation of the c-file.

(B) 16. Rad1 was analyzed by Horowitz and Keres. The latter concedes that after 16. Rad1 cxd4 17. Nxd4 bxc4 18. Qxc4 White has only a "somewhat freer but hardly much better position." Horowitz' analysis was flawed. He claimed that after 16. Rad1 b4 [playable but inferior to Keres' line] White could not play 17. dxc5 because bxB 18. c6 would lose to 18...Nc5. But White is fine in this line with 18. Qb5 (instead of Horowitz' 18. c6??).

(C) 16. Rfd1, analyzed by Euwe, leads to better chances for Black after 16...cxd4 (16...Bxc4 may be even better) 17. Nxd4 bxc4 18. Qxc4 Rc8.

(D) 16. e5, Fritz' choice, which is interesting but seemingly getting White at best equality after 16...b4 17. Bb2 e6.

In sum, 15...c5 was best and would have given Smyslov at least an even game. After Smyslov's actual move, Black faced problems, the position now being:

click for larger view

Keres, however, did not find the best line:

16. Rad1

"?"--(Euwe)

"With 16. cxb5, White could have retained a large advantage [an overstatement--KEG], whereas the text move gives rise to unclear positions in which the issue will have to be decided by tactical moves." (Euwe)

While I agree that 16. Rad1 was theoretically best (Fritz and Stockfish both endorse it), I would predict that if Keres were here today to respond he would still argue for the text. The game does quickly become a tactical brawl, and I suspect this was quite to Keres' liking, the position after 16. Rad1 being:

click for larger view

At first blush, the most notable feature of the position seems to be White's pawn phalanx in the center. But as will be seen, Keres had looked deeper into the position and saw how this feature could be transformed into a powerful King-side attack. The sequel would prove to be thrilling.

What I of course meant to say at the end of my last post was that while I agree that 16. cxb5 was probably theoretically correct (and endorsed by Fritz and Stockfish), Keres would probably still opt for 16. Rad1 for the reasons indicated.

Picking up the game after 16. Rad1

16... Qa5

"!"--(Euwe)

Quite correct. Now Black threatens 17...bxc4." (Euwe)

"The Best choice. He threatens to capture on c4, since the White Queen must protect the Bishop." (Golombek)

"With the text move, Black threatens 17...bxc4 and if 17. cxb5 then 17...Rxb5 followed by ...c5 with equality." (Keres)

As Keres goes on to explain, Black would have had problems had he played 16...bxc4 immediately since after 17. Qxc4 the weakness of the Black c-pawn would remain.

16...b4 was probably inferior to the text, but it was not "hopeless" as Golombek claimed. Golombek reached this conclusion only via a flawed analysis: 16...b4 17. Bb2 c5 18. d5 Qc7 19. BxB KxB 20. Qb2+ f6 21. Ng5. So far so good. But Golombek's 21...Qe5? is a blunder (and leads to the big edge he claims for White following 22. Ne6+ Kg8 23. Qc2 [even better than Golombek's proposed 23. Qd2]). But after 21. Ng5 in Golombek's line, Black should simply play 21...Nf8 with a slightly inferior but likely tenable game.

In any case, Smyslov's 16...Qa5 was best [Stockfish's 16...Qc8 may also be adequate, though too passive for my taste] and left the position as follows:

click for larger view

Black has now achieved at least equality. But Keres claimed that White still was better:

"It may seem [as it does to me--KEG] that Black is over the worst, but now a sharp tactical struggle takes place where Black is punished for his previous play..." (Keres)

17. c5

"!"--(Keres)(Wade--Whiteley--Keene)

"Now 17. cxb5 runs into 17...Rxb5." (Euwe)

While the above comment by Euwe seems to support Keres' 17. c5, he went on to temper his enthusiasm:

"The text move may be positionally bad, but it does open the possibility of combinations that will be very dangerous for Black.

The position after 17. c5 was:

click for larger view

"White had to calculate the outcome of this move very accurately, since Black now gets the opportunity to demolish the White pawn centre with the ...e5 thrust. The outcome of the game depends on whether Black is able to carry out this advance without being punished." (Keres)

As Keres notes, the struggle from here did become a tactical thicket. As will be seen, even the great commentators from whom I have quoted liberally to this point failed to recognize all the tactical nuances in the next several moves, and it was only on this site that key points have been spotted by such superb analysts as Honza Cervenka, babakova, and whiteshark. While I have some small proposed corrections to their contributions, they have blazed the way toward arriving at the true assessment of this fascinating game.

17... b4

This is best. As Golombek pointed out in his commentary on this game, trading Queens via 17...Qa4 would lead to trouble for Black. Golombek's analysis is flawed in several places, but his overall assessment of 17...Qa4 seems correct: 17...Qa4 18. QxQ bxQ 19. d5 [This allows White to retain some advantage, but 19. Rc1 or 19. Rfe1 or 19. h4 all look even stronger] Nf6 [Weak. Black should play 19...cxd5] 20. Rfe1 [20. dxc5 is much stronger] cxd5 21. exd5 Rbd8 [21...Red8 has to be better] 22. d6 [22. Re5 is more flexible and gives White good winning chances, whereas after Golombek's 22. d6 Black seems fine after 22...exd6 23. RxR+ NxR 24. cxd6 Bf8 25. Rd4 Bxd6 26. g3 Bc7 27. Rxa4 Bb6 with an even endgame[.

After 17...b4, the position was:

click for larger view

Now Keres had to decide whether to retreat his Bishop to b2 or to c1.

18. Bb2

"?"--(Golombek)

"It is hard to decide whether 18. Bc1 is better here. In the two main variations, the White Bishop is certainly more effectively posed on c1." (Euwe).

"A case where the natural-looking move happens to be wrong. Correct is 18. Bc1 and if [then] 18...e5 19. Ng5 Re7 20. f4 exd4 21. f5 where White threatens Nxf7." (Golombek).

Golombek's line looks like best play for both sides, and yields a complex position in which both sides have exciting chances and prospects look about even.

The final word on 18. Bb2, however, must go to Keres. His analysis is profound (for his full remarks, please consult Keres' fantastic book on the 1948 World Championship tournament, an excellent part of any chess lover's library]:

"An inaccuracy which,however, was very difficult to forsee at this point, taking into consideration the complexity of the following complications. It later become clear that the Bishop would support the following attack better from c1, from where it protects the important g5 square."

Keres' remarks notwithstanding, the position after 18. Bb2 was hardly bad for White, and the game remained a complicated unbalanced struggle in which both sides had chances:

click for larger view

18... e5!

Thrust and counter-thrust. Smyslov attacks Keres' frightening-looking pawn phalanx and looks primed to seize the initiative.

click for larger view

"Black must count-attack in he centre before White's threatened King-side attack gets home." (Golombek)

"Black has calculated well; White's pawn chain can be broken and will fall. But he reckons without his King, who unfortunately [for Smyslov] is dragged onto the scene." (Horowitz)

But Keres was not daunted, and while acknowledging the power of Smyslov's counterplay, noted that:

"...White had already foreseen these difficulties wheil playing c5 on his 17th move and now proceeds in sacrificial style by giving up all the protection of his centre pawns in an attempt to break the defensive wall surrounding Black's King." (Keres)

19. Ng5

"!"--(Horowitz)(Keres)(babkova)

click for larger view

"With a few deft strokes Keres switches the play to the important wing. Here Smyslov is ill-prepared for defense." (Horowitz).

As will be seen, Horowitz has overstated the case. Smyslov in fact was quite prepared for the assault Keres had in mind. His loss in this game was the result of later inaccurate play (probably in time trouble).

But the fact that Keres' attack was not necessarily a winning one against best play in no way detracts from the beauty of his conception. The matter was best put by <babakova> many years ago on this site:

19... Re7

Sufficient for approximate equality but not best. Better was 19...Rf8.

Golombek and Euwe argued in their respective commentaries on this game that 19...Rf8 was bad. But this conclusion in each case was based on poor analysis.

They both get off on the right foot with (following 19...Rf8): 20. f4 exd4 21. f5 Nxc5 (not 21...Qxc5 22. Nxf7! and White wins) 22. Qh3.

So far so good.

But here Golombek analyzes only the disastrous 22...h6?? which gets crushed by 23. f6!. The correct Black move--identified by Euwe--is 22...h5

Euwe now says that White has three good lines, none of which yield any advantage to White:

A) The flashy-looking 23. Nxf7?, which loses after 23...RxN 24. fxg6 RxR+ 24. RxR Rf8 leaving Black a solid piece ahead;

B) 23. g4? which also leads to an inferior position for White after 23...Ne6! or even 23...Rbe8;

C) 23. Bxd4 BxB+ 24. RxB Qxa2 25. Qg3 Qb3 with at least equal chances for Black.

Euwe overlooks what is probably White's best move: 23. fxg6, which also only leads to approximate equality with 23...fxg6 24. Qg3 RxR+ 25. RxR Rd8.

In sum, 19...Rf8 was at least as good as the text.

After Smyslov's actual move (19...Re7), the game was very much in the balance:

click for larger view

In this position, Keres unleashed the attack he had been planning, and the game became a complex tactical slugfest:

20. f4

"!"--(Keres)(Wade, Whiteley, Keene).

"!!"--(Horowitz)

click for larger view

"Rapid mobilization. One threat, among others, is 20. Nxf7 RxN 22. fxe5)." (Horowitz)

"...With a strong double-pawn sacrifice, White starts to exert strong pressure on the f-file, and especially on the f7 point..." (Keres)

But Keres correctly tempers his enthusiasm:

"Even though later analysis proved that Black has sufficient defensive resources against this attack, White's strategy should be considered to be fully justified....the attack ensures White at least equality...[and]...the complexity of the sacrifices that occur makes it very difficult for Black to find the right defense."

20... exd4!

Careful and excellent defense by Smyslov.

As Golombek correctly pointed out, 20...h6 would get crushed by 21. Nxf7! RxN 22. fxe5 Rbf8 (22...Nxe5 23. dxN is also insufficient to save the game for Black) 23. RxR RxR 24. e6 Rf8 25. exN+ and wins. In this line, 24...Re7 is slightly less disastrous than Golombek's 24...Rf8, but also loses after 25. exN+).

The position after 20...exd4 was:

click for larger view

21. f5

"!!"--(Horowitz)

"A surprise. Smyslov only expected 21. Qh3 (an attacking move that does look more solid from a strategic point of view) and because of time-trouble failed to adapt to the changed situation." (Euwe)

The climax of the game had now been reached,the position now being:

click for larger view

21... Nxc5?

"?"--(Golombek)(Euwe)(Keres)(Wade-Whiteley-Keene)

"The decisive error. Black should not relinquish control of f6 under any circumstances." (Euwe)

The decisive mistake after which White's attack breaks through." (Keres)

As I will attempt to show, the text (21...Nxc5?) was indeed a (non-necessarily fatal) mistake. Smyslov had two better moves, each of which with best play would have led to at least equality for Black. Moreover, and as everybody other than <babacova> and <Honza Cervenka> on this site seem to have missed, Smyslov still had chances even after the inferior text move and was not necessarily lost until his (til now overlooked) mistake on move 23.

Let's start with the two good defensive moves Smyslov could have played instead of 21...Nxc5?

A) 21...Qxc5

With the exception of Horowitz (who mistakenly thought White still had a win), all of the commentators concluded that Black would have been OK after 21...Qxc5.

The crucial line after 21...Qxc5 was 22. Nxf7!

Let's dispose of the alternatives first.

(i) If 22. Rc1 (which Horowitz called "decisive"), Black would win with 22...Qd6 rather than Horowitz' horrible 22...Qb6 which would lose to 23. fxg6 (and not Horowitz' 23. Nxf7? Nc5 with equal chances [rather than Horowitz' awful 23...d3+?]).

(ii) 22. e5
Euwe claimed that this offered White "very good chances," but as Keres showed, Black would emerge with far the better game after the Queen sacrifice 22...Qxe5!, i.e., 23. Bxd4 QxB+ 24. RxQ BxR+ 25. Kh1 Nf6 "and Black has a good game with a Rook, Bishop, and two pawns for the Queen." (Keres)

The correct move for White here (analyzed by Keres and Euwe and by user <andrea volponi> on this site) was: 22. Nxf7!. As these fine analysts have shown, equal chances now result from: 22...d3+ 23. Kh1 Qc2! [first found by Keres] 24. Nh6+ Kh8 25. Nf7+ ! [Not 25. QxQ? dxQ 26. RxN c1 (Q)] with a draw by perpetual check (since 25...RxN loses to 26. QxR QxB 27. QxN and White is a Rook ahead).

As the above analysis shows, 21...Qxc5 would have saved the game for Black. But only babacova and White Shark on this site have discovered an alternative drawing line (quoting the moves identified by White Shark): 21...gxf5! 22. Rxf5 Ne5 23. Rdf1 Rf824. Bxd4 Qa3 25. QxQ bxQ 26. Bc3 Rb8 27. Nf3 NxN+ (27...Ng6 also seems sufficient--KEG] 28. R5xN BxB 29. RxB with equality.

My only correction to White Shark's fabulous analysis comes after his 29. RxB where he suggests 29...Rb4 (where 29...Rxe4 seems a simple route to a draw) 30. Rf4 (30. Rxa3 would seem to give White some small winning chances) Rb1+ 31. Rf1 Rb5 (31...Rb4 looks a little better).

BRAVO White Shark!

But now let's return to the actual game after Smyslov's inferior 21...Nxc5?:

click for larger view

Is this position a win for White? Prior to the contributions by babacova in 2005 and Honza Cervenka in 2011 on this site, the verdict seemed to be that Black was now lost in light of the upcoming brilliant combination by Keres. But--as it turns out--things are not that simple.

22. Qh3

click for larger view

"Sledge hammer blows. Black has no time to consolidate." (Horowitz).

22... h5!

Forced. As several commentators have shown, 22...h6 loses to 23. f6! Bxf6 (if 23. Nxe4 24. NxN RxN 25. fxB; if 23. hxN 24. fxB Kxg7 [24...f5 loses to 25. Qh8+] 25. Bxd4+ with mate to follow soon.

23. f6!

This left:

click for larger view

In the position with which I ended my last post, Smyslov played:

23... Bh6?

"...a howler" (Honza Cervenka)

As Keres quickly demonstrated, Black is quickly dispatched after the text. Did Smyslov have anything better.

Outside of the contributors on this site, only Keres suggested any improvement for Black here (the other commentators simply concluded that Smyslov was dead lost and celebrated Keres' brilliant wrapping up of the game from this point).

Keres' suggestion, as he acknowledged, gave Black no real chances:

A) 23...Ne6

click for larger view

This is indeed "more stubborn defense" (to quote Keres) than Smyslov's effort, but "even so, White wins": 24. NxN RxN 25. fxB c5 (25...Rxe4 26. Rxd4 is better than Keres' line, but insufficient to save the game--KEG] 26. Rxf7! KxR 27. Rf1+ Ke7 28. Rf8 (28. Bc1 and 28. Qf3 are both stronger than Keres' line, which also wins albeit not as decisively--KEG] RxR 29. gxR(Q)+ KxQ 30. QxR leaving White with a Bishop for two pawns and a theoretical win.

There are, however, two stronger lines:

(B) 23...Bxf6

click for larger view

This line noted on this site by barbacova in 2005 may not save the day, but it is a far better try than Smyslov's move or Keres' suggestion:

24. RxB Nd7 25. Rd6 QxN 26. RxN RxR 27 QxR c5 28. Rf1. Although Black has three pawns for White's extra Bishop, White's pressure on f7 is probably sufficient for a win. BUT, in babacova's line, Black may yet be able to save the game with 26...Rxe4 (instead of 26...RxR), e.g., 27. Bxd4 Rbe8. Here, Black again has three pawns for White's extra Bishop but also has counter-play and at least practical drawing chances.

The best line of all, however, is Honza Cervenka's:

C) 23...Nxe4!

click for larger view

This proposed move would have posed severe problems for White, e.g. 23...Nxe4! 24. fxR QxN 25. Qd7 Qe3+ 26. Kh1 Nf2+ 27. RxN QxR 28. Qd8+ Kh7 [obviously not 28...RxQ? 29. exR(Q)+--KEG] 29. QxR [Not 29. e8(Q)?? RxQ 30. QxR QxN leaving Black with four pawns for the exchange and a easy win] Qe2 30. Rg1 Qxe7 leaving Black with four pawns for White's extra Rook and decent drawing chances in the complicated ending to come.

In sum, Smyslov had at least practical--and probably theoretical--chances to save the game until his poor 23rd move.

But none of the above happened. Instead, as noted at the beginning of this post, Smyslov played 23...Bh6? This left the position as follows:

click for larger view

24. fxR BxN

click for larger view

25. Qf3

"!"--(Keres)(Wade/Whiteley/Keene)

"Keres does not let up for a moment. Every move is a picture." (Horowitz)

Without in any way meaning to detract from the beauty of Keres' move and attacking scheme, the claim (advanced by Keres) that the simple 25. Qg3 would have been a mistake and allowed Smyslov to escape is wrong. Keres gave the following line in support of his claim: 25. Qg3 Be3+ 26. Kh1 Re8. So far so good. But from here White can win with 27. Qe5 (the immediate 27. Rxd4 also wins) Qxa2 28. Rxd4 Qxb2 (28...BxR gets Black quickly mated after 29. BxB) 29. Qf6!

But now let's return to Keres' pretty 25. Qf3:

click for larger view

25... f6?

"Smyslov took 24 minutes over this move without being able to find a saving move." (Golombek)

"The buzz of analysis stirred up by his game seemed to amuse Keres, apparently because he had no doubt in his own mind about the soundness of his combination. Smyslov, who was not amused, stared at the position for twenty-eight minutes [or four minutes longer than Golombek reported] before he played 25...f6." (Kmoch/Horowitz).

Smyslov was indeed lost here. To make matters even worse, after taking so much time on his 25th move, Smyslov was now in terrible timepressure.

The "best" 25th move at Smyslov's disposal was probably 25...Qxa2. But this too would fail after 26. Qg3 Be3+ [Keres' line; 26...Re8 was "better," but would lose to 27. QxB) 27. QxB [a little Queen sacrifice in Keres' notes!--KEG] QxB [27...dxQ would lead to annihilation after 28. Rd8!--KEG] 28. Qf4.

Similarly, 25...Be3+ would--as Keres showed--also lose after 26. Kh1 Qxa2 27. QxB [another Queen sacrifice!--KEG].

After 25...f6, the position was:

click for larger view

26. Bxd4 Nd7

26...Ne6 or 26...Kh7 would at best have prolonged the game.

27. h4

"!"--(Wade/Whiteley/Keene)

"!!--(Horowitz)

A pretty finish.

click for larger view

1-0

Smyslov's resignation was not premature. Apart from the fact that (according to Kmoch and Horowitz) he only had two minutes left on his clock to make 13 moves), his position was hopeless.

If 27...Bxh4, White wins with 28. Qf4 (ye old nasty fork)(Golombek, Keres)(Wade/Whiteley/Keene) or with 28. Qh3 (another brutal fork)(Horowitz) or--perhaps with the even more brutal 28. Qb3+ Kg7 29. Qh3 (or 29. Qe6).

If 27...Bh6 28. Qb3+ Kh8 29. Qe6 (Or 29. Qf7).

All in all, a fine performance by Keres.

On <cehertan>'s <....It's always been the case that even super GMs have difficulty defending against unpleasant surprises...>: