The first edition of Keres' best game was indeed published in 1961/1964, but that only included his early games. A second volume came out later, and the final volume (which includes games as late as 1962) in 1969. The original work from 1961 obviously could not have included all of these games. The version I have ("The Complete Games of Paul Keres" translated and edited by Harry Golombek) purports to include everything in the three volumes.

I am not certain that my indication that the combined work is from 1969 is correct; I'm just citing the date from my volume. What I really want to do is to distinguish Keres' original book about the 1948 World Championship Tournament from his later analysis in his book containing his best games.

I didn't know about the 1961 Estonian version of the first volume, so perhaps my use of "1969" is sloppy. So long as I am able to convey the changes over the years in Keres' thoughts about this game, I think I have accomplished the task I set for myself.

21. Ng2

Solid, but hardly the most effective way to combat Keres' idea: 21. e4!

Of all the commentators, only Golombek deigned to mention this strong pawn advance, and that only to denounce the move.

According to Golombek, 21. e4 would lead to an advantage to Black after 21...gxN 22. exB exd5 23. Bd2 Qf6.

Actually, White would be at most slightly worse if play followed Golombek's line. But all that is beside the point, since it is hard to believe that Smyslov would have played the wimpy 23. Bd2? rather than the strong 23. Bh3! White's light square Bishop would then have considerable scope, and Black's King would be vulnerable (the very thing Keres said was problematic about his 20...g5?!). White's King would, it is true, also be exposed, but Black's pieces would not be poised to pose much of a threat to the White monarch.

In sum, 21. e4 looks clearly best. It was the first move that occurred to be when I first played over the game, and the silence about this option in most of the commentaries is strange.

21... f5

click for larger view

"!"--(Kmoch)

Kmoch was still handing out "!" marks and lauding Keres' play, claiming Black was now clearly better.

While, having said "A" (20...g5) Keres pretty much had to say "B" (21...f5), especially after Smyslov failed to play 21. e4, the situation was hardly as clear as his comments would suggest.

More realistic was Euwe:

"It is hard to decide whether White is better here."

But by far the best evaluation of the situation in the above-diagrammed position appear in the later commentary by Keres (i.e., the one I have perhaps inaccurately been referring to as his "1969" commentary) in which he said, in part:

"Black has been able to prevent the advance of e4 by the greatest efforts, and has had to create in consequence some marked weaknesses in his position. Moreover, White is in a position to concentrate the fire of his Knight, Bishop, and Rook on the e4 and thus can eventually force through his e4. For these reasons Black must seek counter-play as quickly as possible and best prospects of this are afforded by the Queen-side since fewer White pieces are present there."

22. Bf2

In both his commentaries, Keres continues to stress the idea that White is and should be preparing for the right moment to play e4. The question, that he discusses at some length (I heartily recommend reading both of Keres' accounts of this game, which I am only describing/quoting in part).

Keres candidly recognizes the difficulty of the position and his uncertainty of the correct plan. Notably, even with help from both Fritz and Stockfish, it is extraordinarily hard to decide exactly how White can best proceed (after his failure to play 21. e4 before Black got in 21...f5). To give a sample of his comments on the plan Smyslov adopted:

"With the text-move White does defend himself against the possibility of c5, but now Black gets counter-chances on the queenside with the a5 thrust." (Keres' 1949 commentary)

"It is naturally hard to say whether the plan devised by White is the best, but in any case it is in keeping with the demands of the position..." (Keres later commentary)

Keres discusses some of the proposed alternatives to 22. Bf2, and properly concludes that none are significant improvements on the text>

(A) Levenfish suggested 22. Bc3, but Keres' 22...b6 ("threatening an eventual c5") leaves White no better than in the game, and in fact Black can do even better with 22...c6 and probably also with 22...Nb6 or even 22...Nf6. Simply put, 22. Bc3 does not yield any meaningful edge or strong prospects.

(B) Romanovsky gave 22. Ne3 as best, but this--as Keres noted in his 1949 commentary--allows Black counterplay via 22...Qg7, not to mention the availability of other strong plans for Black after 22. Ne3 beginning with 22...Nb6 or 22. c6 or 22...Qf7.

I had initially considered 22. a4 to forestall operations by Black on the Queen-side, but Black then seems fine after either 22...Qg7 or 22...Qf7.

The more I studied the position the more I liked Smyslov's 22. Bf2, which left the position as follows:

click for larger view

22... Nf6

Nobody raised a peep about this move, but the engines don't like it and it seems like mainly a rote attempt to gang up on e4 rather than attend to the other demands of the situation . But the game is so complex that it is difficult to decide what is best. Stockfish prefers 22...c6 and Fritz likes both that and 22...Qf7.

In any case, after 22 moves, the clocks read;

Smyslov 1:38
Keres 1:42

The position was now:

click for larger view

23. Ne1!

A subtle move that reflects Smyslov's understanding of the power his Knight can wield on d3 (which gives him access to both e5 and c5. In addition, the move prepares to allow the White f1 Bishop to move to g2 or h3.

23... a5

Before analyzing the situation, let us first tune in to the Kmoch fantasy that Black is in the ascendancy:

"With the strategic square e4 under firm control and White's minor pieces inactive, Keres has the better game. With this (23...a5), he takes the initiative."

Returning to planet Earth (in which White is somewhat better), let's hear what Euwe and Keres had to say about the text:

"Exploiting the absence of the White dark-square Bishop." (Euwe)

"After the placement of White's Bishop on f2, Black has no prospects of an attack on the King-side...Instead the action began on the queenside by the text-move constitutes annoying counterplay, and directs White's focus away from the preparation of his main plan e4..." (Keres, 1949)

"...The diversion begun by the text-move on the Queen's wing is unpleasant enough for White and at least forces him to abandon his main plan, the e4 advance." (Keres, 1960's commentary)

Keres, in his later commentary, considered the alternative 23...g4, but rightly rejected it because Smyslov would likely then play Bg2 followed by Nd3 and leave Black with even more serious weaknesses on the King-side.

Keres in both commentaries--and in his play-- was still fixated on the possibility of e4 by White. But there was also Nd3 by White to worry about, and thus best for Black was probably 23...Bc4 so he could take off the Knight after 24. Nd3 and thus avoid incursion on e5 or c5.

The text (23...a5) left new issues to be considered on the queen-side and new plans to consider. As Keres aptly put it in his 1949 commentary:

"The game now becomes especially interesting, and the excitement is raised even more by the fact that both players at this point had relatively little time left for the following moves [through move 40--KEG]."

The position after 23...a5 was:

click for larger view

24. Nd3

"?"--(Golombek)

Consistent with his plan, but perhaps the situation had to be reevaluated after 23...a5. Golombek and Kmoch, probably correctly, preferred 24. bxa4, and Keres apparently came round to their way of thinking in his later commentary. In theory, both the text and 24. bxa5 are strong, but the latter move would raise practical problems for Black: 24. bxa5 Ra8 (not 24...Nc4? which, as Kmoch pointed out, would get crushed by 25. e4!) 25. Qc5 [Keres in his 1960's commentary gives 25. Nd3, which is also good and would leave White with a clear edge after 25..Rxa5 26. Be1 (or 25...Rdc1 which would perhaps be even better than Keres' move)] Nd7 and White is certainly better, but not after 26. Qb4? as given by Golombek and Kmoch and certainly not 26...Qxc7?? which--as Kmoch noted--would lose instantly to 26...Rdc8) but rather via 26. Qc2!

The balance of the commentary by Golombek after 26. Qb4? is so error-laden that I will spare readers on this site from detailed discussion of it. Meanwhile, Kmoch, who was still engaging in the fantasy that Black was better, simply ways that after 26. Qb4? White had "counter-play." In fact, after the clearly inferior 26. Qb4? White was not worse.

In any case, let's return to the actual game after 24. Nd3:

24... Ra8

As Keres noted only in his later commentary, Black should here play 24...axb4.

After the somewhat inferior 24...Ra8, the position was:

click for larger view

25. Bg2

Again rigidly adhering to his plan. But--as only Keres seems to have noticed (and then only in his 1949 commentary), 25. bxa5 was correct. After 25...Rxa5 26. Be1 White is much better placed, and also better placed than after the text which failed to exploit Keres' second-best 24th move.

25... Ra7

"Black does not want at all to eliminate the possibility of [a pawn trade] but would like to double his Rooks before the exchange on b4." (Keres, 1960's commentary)

But, in his 1949 analysis, Keres, quite properly in my view preferred 25... axb4 as a means of frustrating Smyslov's plan.

For some reason, both players shunned the trade, as happened yet again on Smyslov's 26th move in the following position (after 25...Ra7:

click for larger view

26. Re1

"The attempt to force e4 is doomed to failure, as Black is able to concentrate an unusually large number of pieces on the vital square. Again correct is 26. bxa5 Rxa5 27. Be1 Ra7 ([27...Ra6 is more accurate--KEG] 28. Bb4 (even stronger for White here would be 28. Ne5--KEG]." (Golombek) (see Euwe similarly)

Despite my quibbles in his line, I believe that Golombek's concept on this move is superior to Smyslov's 26. Re1. Of course, Golombek did not have a clock ticking when he gave his analysis (and, needless to say, neither did I).

26... Qh7

"!"--(Keres--both commentaries)(Kmoch)

"A good move by which Black increases his control of e4 and, in the event of White continuing with Nc5 followed by e4, will have a troublesome pin on that square after a pawn exchange." (Keres, 1960's commentary)(See similarly Keres' 1949 commentary)

"Otherwise the complications arising from 27. e4 fxe4 28. fxe4 either Knight takes pawn 29. Nc5 can become dangerous for Black." (Kmoch)

After 26...Qh7, the position was:

click for larger view

27. b5

"?"--Everybody except for me.

Everyone and his/her aunt/uncle has denounced this pawn sacrifice by Smyslov:

"An unsound pawn sacrifice, possibly based on the misapprehension that Black dare not accept it for fear of White's next move. He should abandon all notions of achieving e4..." (Golombek)

"This appears to be a miscalculation although it may be that Smyslov, recognizing that his game is bad anyway [There he goes again. In fact, White had the better chances both before and after this move--KEG] makes the move to raise complications." (Kmoch)

"An unnecessary pawn sacrifice. White is better advised returning to [the] plan with Rec1 and Be1." (Euwe)

"This pawn sacrifice is not justified by anything, and gives Black at least an extra pawn as compensation for the difficult position [take that--Kmoch--KEG]...the text-sacrifice gives White nothing..." (Keres 1949 commentary)

"27. b5 is a baffling move." User <NBZ> on this site.

Of the denouncers, I come out closest to:

"By means of this pawn sacrifice White at last carries out his idea...but he has, however, to pay too highly for it all. In the first place Black will now have a pawn more and in the second this advance no longer possesses the force it was once supposed to have..." (Keres, 1960's commentary)

If indeed, White had something much better, I might join the chorus. But none of the critiques of Smyslov's move had any super alternatives.

Golombek suggests forgetting about playing e4 and trying 27. Reb1. But then Black replies 27...axb4 and White has certainly not increased his advantage.

Euwe preferred 27. Rec1, but yet again Black is in decent shape with 27...axb4.

Keres gives what is probably the most solid line yielding White the superior position: 27. bxa5 Rxa5 28. Rec1. But after 28...Rda8 Black has reasonably good counter-play.

Again rigidly adhering to his plan. But--as only Keres seems to have noticed (and then only in his 1949 commentary), 25. bxa5 was correct.>

The position after 27. b5 was:

click for larger view

A tough one to evaluate, as the differing assessments of the commentators makes clear. These differences notwithstanding, everyone except Kmoch seemed to think that the game remained tense and very much in the balance.

27... Nxb5

Accepting the pawn sacrifice, a choice not commented up in any prior analysis of the game with which I am familiar. But this could have led to trouble for Black had Smyslov taken advantage of what it gave him (i.e., if Smyslov had not been in dreadful time trouble). Black could, and should have declined the offered pawn and worked on his own counter-play with 27...a4.

This left:

click for larger view

28. Qc5

So far so good for Smyslov.

28... c6

Pretty much forced.

29. a4

Again well-played by Smyslov.

29... Nd7

The only move that seems to hold for Black.

Given the time-pressure situation, the last few moves by both sides were remarkably accurate. But then, inevitably, the clock began to take its toll, and Smyslov began to weaken (as who would not in such a mind-bogglingly complicated configuration, the position now being:

click for larger view

30. Qc2

Smyslov probably still had the edge after this move, but he, like the commentators missed an even stronger move: 30. Qc1! which eyes the King-side while still keeping the Queen for now on the c-file. Play could then have continued: 30. Qc1 Nd6 (forced) 31. h4! gxh4 (probably best) 32. Qf4 (the key to this line) Nf7 33. Qxh4 and White, though still down a pawn, clearly has the superior chances.

30... Nd6

Forced.

click for larger view

31. Ne5

"?"--(Golombek)(Keres)(Wade--Whitely--Keane)

Kmoch, who was still operating under the delusion that Black was winning, passed over this move without comment. Euwe said that 31. Nc5 was "also a good move.' But Golombek and especially Keres (in both his commentaries) rightly condemned the text and identified 31. Nc5 as best. But even these superb commentators missed the only line that could have given White an edge. After 31...Nf6 (best) White should not play 32. e4 (as suggested by Golombek and Keres) but rather either 32. Qb1 or 32. Rab1 pressing hard on the b-file. By contrast, after 32. e4, White's edge would be gone: 32...fxe4 33. fxe4 b6 34. Rac1 (given by Golombek and Keres and probably best) bxN 35. exB QxQ 36. RxQ exd5 37. Rxc5 and now either 37...Rde4 (Golombek) or 37...Rc8 (Keres) leaves Black somewhat better placed.

After White's inaccurate 31. Ne5, the position was:

click for larger view

31... Nf6

"!"--(Keres, later commentary)

"With the text-move, Black once more protects the e-4 point and now stands clearly better." (Keres, 1949 commentary)

But is Black really better here? He is a pawn up, true, but there are clear weaknesses on his Queen-side, and his backward e-pawn continues to provide options for White.

In his 1949 commentary, Keres correctly pointed out that the superficially attractive 31...f4? would be terrible for Black after 32. QxQ+ KxQ 33. gxf4.

As Keres point out in his 1960's commentary, White--if he still fancies the chance to play e4, he must retreat his Knight to d3 and thereby concede that his 31st move was bad.

32. Rac1 Rae8
33. Nd3

"The Knight once more aims for c5." (Keres)

click for larger view

A look at the position suggests that is was probably time for Smyslov to stop fretting about getting to play e4 and try to force matters on the Queen-side. 33. Qd3 or 33. Qc3 looks like a good start. After the text, Black could have prevented 34. Nc5. But should he?

33... Rab8

"?"--(Kmoch)

"A serious mistake which allows White to carry through the advance of the e-Pawn for which he has been yearning. Correct is 33...b6! to prevent 34. Nc5." (Kmoch)

In fact, Keres was well-aware of the merits of 33...b6, but had other ideas and was willing to allow Nc5.

"...Black regards the ensuing complication *after 34. Nc5) as not dangerous, for him at any rate." (Keres, 1960s commentary).

34. Nc5

"!"--(Kmoch)

"With the strong threat of e4." (Kmoch)

34... b6
35. e4

"!"--(Kmoch)(Keres, 1960's commentary)

"Smyslov...obtains attacking chances with the text-move."

In his 1949 commentary, Keres opined that 35. Nd3 might have been preferable, but doubtless had recognized that this would give Black time to play 35...b5 with either b4 or f4 to follow depending on whether Smyslov put his Knight on e5 or c5.

After 35. e4, the real fun begins (at least for those of us to sit back and analyze and not have to make crucial decisions while the vanishing remaining time on the clock ticked away, the position now being:

click for larger view

35... fxe4

35...bxN? 36. exB is strongly in White's favor. Black seeks to soften the break-through by swapping Queens." (Kmoch)

Keres: 2:17.

click for larger view

36. Nxe4

"?"--(Kmoch)

Smyslov 2:21

Kmoch considered the text to be a blunder and said that 36. fxe4 would "give White excellent chances."

Golombek said the text was best and that Black would obtain a won game after 36. fxe4.

Keres, in his 1949 commentary said that it was "difficult to say whether the endgame after 36. fxe4...offered White better prospects of a draw."

In his later commentary, Keres hedged his bet a bit and said that:

In view of the time trouble this capture [36. Nxe4] is far more dangerous for Black than going over to the endgame by 36. fxe4...which would offer Black the better prospects..."

As I will attempt to show, Smyslov's move--at least theoretically--was best and left him at no disadvantage, while 36. fxe4 would have left him with a tough ending to hold.

My assessment of the text will be born out in my commentary on the balance of the game (which I think will reveal that 36. fxe4 was not the reason Smyslov lost). For now, let's consider the position after the move that Kmoch and Keres preferred, 36. fxe4.

Play would then continue 36...bxN 37. exB QxQ 38. RxQ. This would have left:

click for larger view

As noted above, Golombek thought that Black here would have had a won game, Keres thought that Black would be better, and Kmoch thought White was better and would have "excellent chances." Who was right?

We can dispose first of Kmoch's analysis, which assumed that Black would here play 38...c4? Now Black would indeed be lost after 39. dxc6, especially if he continued with one of the moves Kmoch considered: (a) 39...Rbc8 which gets smoked after 40. Rxe6; or (b) 39...Rdc8 which falls to Kmoch's 40. g4 (actually, 40. Rxe6 would be stronger still). The only chance for Black in this line would lie in 39...Re8, but Black would still be in major trouble after 40. Rec2, since 40...c3 loses to 41. Re3 and otherwise White just grabs the Black pawn on e6.

A better chance for Black is 38...cxd5, but then, as Keres pointed out in his 1949 commentary, White gets all sorts of counterplay with 39. dxc5 ("!"--Keres) Nde4 40. Bd4 [or even 40. c6--KEG] leaving the game very much in the balance.

If there is a way for Black to make progress, it must be 38...exd5. Play would then likely continue 39. Rxc5 Nde4. Here, Keres and Golombek only consider 40. Rxc5, but in fact White can almost certainly hold after the superior 40. Rxa5. If instead 40. Rxc5, Black gets the better (though materially equal) endgame after 40...Rdc8 as Keres pointed out in both his commentaries on this game.

in the above line, Golombek claimed that Black wins after 40...Rb2, but White would then be fine with 41. Rf1. And even if White played the weaker 41. Be3 as given by Golombek, I see nothing resembling a winning line for Black.

The conclusion: had White played the much recommended 36. fxe4, he would--against best play--have had to fight fora draw in a complex position with his clock running down.

That being said, let us return to the game after Smyslov's actual 36. Nxe4:

click for larger view

The game was now highly complicated, and it is hard to decide who is better. But not for Kmoch.

36... Qg6

"Once again, Black has the advantage although the consequences os his mistake [on move 33 he means--KEG] are by no means eradicated; White has counter-chances which he didnt have before." (Kmoch)

click for larger view

37. Qe2

Beginning here, and though not mentioned by any of the other commentators, Smyslov began to wilt under the time pressure. Here, he could have at tied up Keres' kingside and gotten his dark-square Bishop to g3 with 37. g4.

The text was certainly not a losing move, but the first of four second-best moves by Smyslov before the move 40 time control.

37... Rb7

Keres was also in time trouble, and this may explain why he avoided the theoretically better (but more complicated) 37...b5.

After 37...Rb7, the position was:

click for larger view

38. Nc3

Stronger was 38. Qa6, as well as 38. NxN and perhaps even 38. g4. But with the clock ticking, Smyslov's move was a decent simpler choice.

38... Bc4

click for larger view

Smyslov: 2:25
Keres 2:25

Five minutes to make two moves would be an eternity for Reshevsky. But for most mortals, it would constitute serious time trouble in this sort of position.

39. Qb2?

Had Smyslov had a bit more time to think, I am confident he would have played the much stronger 39. Qd2. But with the clock running, these sorts of small errors are inevitable.

39... b5

"Finally Black tries to realize his advantage on the queenside. But during time-trouble he has placed his pieces unfavorably, so that White once more obtains dangerous counter-play. It was however difficult for Black to convert his edge by other means, as White was threatening the very troublesome 40. Ne4."

More than a decade later, Keres had another thought:

With the text-move Black seeks to exploit his Queen-side pawn preponderance, but in so doing he has to allow his opponent dangerous counter-chances. Perhaps, therefore, it would have been better to prepare this advance by playing 39...Rf8 so as to meet 40. Ne4 with 40...NfxN 41. fxN Rf7, etc."

But if 39...Rf8, Smyslov could have responded 40. d5! with excellent play and arguably the better position.

All in all, Keres' actual 39...b5 looks clearly best, and left:

click for larger view

40. axb5?

Smyslov's other second-best moves in time trouble were--all in all-not so terrible. But this was a lemon. There was still time to correct his last move with 40. Qd2.

40... cxb5

Smyslov here sealed his move in the following position:

click for larger view

Golombek considered that Keres had a won game here, but the other commentators came to the more reasonable conclusion that, while Black was certainly for choice (he was of course a pawn up), a tough fight loomed, especially given all the tactical chances at Smyslov's disposal and given his chance to explore those possibilities over-night.

41. Ne4

"!"--(Keres)

"The sealed move, which threatens to win a piece by 42. NxNd RxN 43. RxB." (Golombek)

"The sealed move, which will still yield White good chances. He is already threatening to win a piece [see Golombek--KEG], while the move Qa3 may also be important in certain circumstances (after the Knight exchange) (Euwe)

"Whit'es adjournment move, which causes Black the greatest difficulties...Moreover, an eventual Qa3 or Nc5 is also threatened, that is why the following exchange is forced..." (Keres)

41. f4 may initially look appealing, but--as Keres went on to explain...41...Rf7 would be an effective retort.

This is not to say that 41. Ne4 was a panacea for all of Black's troubles, just that it was the best practical over-the-board chance, and perhaps even the best theoretical chance, for White to create counter-play to offset his pawn minus.

As it turns out, the pawn minus was not destined to be a key factor, since Keres had cooked up a sacrificial line of his own.

41... NdxN

"Virtually forced." (Keres)

41...NfxN would connect White's d and e pawns and would have been less effective (though Black would still be somewhat better).

41...Rf7, as Golombek noted, would been a catastrophe for Black after 42. NxNd RxN 43. RxB! bxR 44. Qb8+ and White wins at least a piece and the game.

41...Bd5 would also have been inferior to the text, but Golombek's claim that Black would then "get the worst of it" is wrong: 42. NxNd RxN 43. Rc8+ Kg7 and Black is still somewhat better, especially if White plays Golombek's awful 44. Qa3 (which would be answered by b4 leaving Black with excellent winning prospects.

42. fxN

click for larger view

In this position, Keres sprung his surprise:

42... Ng4

"?"--(Golombek)

"?!"--(Kmoch)

"!"--(Euwe)(Keres)

This bold sacrifice of a piece by Keres drew a variety of reviews.

"Brilliant but unsound...Keres' move is a sacrifice which enjoys practical success, but only because the normally imperturbable Smyslov allows himself to become rattled for once into playing at least two bad moves." (Golombek)

"A doubtful combination." (Kmoch)

"Typical of Keres. The first thing he does is go for tactical chances,without taking consideration of the material balance. The sacrifice connected to the text move will in any case yield Black a draw, with good winning chances if White slips up, wheres other moves, although preserving Black's material advantage, allows White's Bishop pair to swing into action." (Euwe)

"After thorough home analysis, Black came to the conclusion that the sacrificial combination initiated by the text-move offers him te best practical chances...With the text-move, Black embarks upon a surprising combination, which it is very difficult for White to find correct defense against." (Keres, 1949 commentary)

In his later commentary, Keres added that:

"The point is that Whie now not only threatens 43. RxB, but also the very strong 43. d5!, when, at the cost of a second pawn, all kinds of lines of action would be opened for his Bishop and Rooks."

As a purely theoretical matter, best play is 42...Rf7. But that was hardly an sort of easy win and it wouldn't have subjected Smyslov to the severe problems posed by the text, which did in fact topple the brilliant rising star. After 42...Rf7, play would likely continue 43. d5 exd5 44. exd5. Nxd5 [Not Golombek's 44...Bxd5 as given by Golombek and Keres in his later commentary after which 44. Red1 ( and not Golombek's 44. Qxb5? which gets crushed after 44...BxB 45. KxB Ng4) and White has excellent chances to survive despite Black's two extra pawns] 45. Red1 after which, though Black's two extra pawns might give him a theoretical win, White's Bishops are plenty scary and Black will have his work cut out for him (even if Keres' claim that White here has no advantage is a bit of an overstatement.

Also no real improvement, as Keres pointed out, was 42...Rbd7 43. d5 exd5 44. Bb6 because, to quote Keres again: "The White pieces suddenly become extremely active and I am not at all convinced that Black's extra pawn would ensure him an advantage.

In light of the foregoing, Keres' move was at a minimum an excellent practical choice that apparently dook Smyslov by surprise, the position after 42...Ng4 being:

click for larger view

43. RxB

Smyslov had nothing better than to grab the Bishop and exploit the pin on the b-file, even though this meant he would have to face Keres' attack. He might as well have a piece in hand.

As Golombek pointed out, a defensive move such as 43. h3? would get crushed by 43...NxB 44. QxN [not 44. KxN Qf6+--KEG] Rbd7 (44...a4 of 44...Qf7 or 44...g4 would be even stronger than Golombek's move--KEG) 45. Red1 Qg7 [45...a4 or 45...g4 were again even better--KEG] 46. e5 Qf8 [46...a4 again is yet again stronger--KEG] 47. Qe3 Qf5 [and even here, a4 is a superior choice] 48. Rd2 Rf8 and even here Black has a likely win.

After 43. RxB, the position was:

click for larger view

43... Qh5!

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

The key to Keres' combination.

"This threatens not only 43...Qxh2+ but also to regain the piece by 44...NxB 45. KxB Rr7+" (Kmoch)

Keres reports that Smyslov, not having expected the combination unleashed on Keres' 42nd and 43rd moves, now used up the majority of his remaining time.

Of course, and as pointed out by Keres, 43...NxB would allow 44. Rc3. If then Black played 44...Ng4?, White would even have winning chances with 45. Rc6 (far better than Keres' 45. h3 which at best yields equality).

After 43...Qh5, the position was:

click for larger view

This proved to be the pivotal moment of the game. And it was here that Smyslov faltered under Keres' relentless pressure:

44. Rc2?

"?"--(Golombek)(Kmoch)(Euwe)(Keres)(Wade-Whiteley-Kean- e)

"White loses his way in an intricate and difficult position." (Golombek)

"The final mistake in this game of ups and downs" (Kmoch)

"This is insufficient, as the rest of the game shows." (Euwe)

"After long thought, Smyslov decides to hang onto the extra pieces, but this is the decisive mistake and quickly assures Black a winning attack"

The only move with even a chance to save the game was, as each of the commentators pointed out, was 44. h4! The play to follow would have been tense and complicated almost beyond belief:

44. h4 NxB 45. Rf1 (best--everything else loses: 45. Rc6 falls to 45...Nd3; 45. Rc3 loses to 45...Nd1!; 45. QxN loses to either 45...bxR or 45...Rf7; 45. KxN gets crushed by 45...Rf7+ or 45...Qf7+; and 45. Qb3 bxR 46. QxR gxh4 47. Qb6 Rf8 48. Qxe6+ Rf7 49. Qd5 QxQ 50. RxQ Kg7 with an easily won endgame.

Returning to 45. Rf1, everyone considers 45...Nd1 after which White survives: 46. Qb3! bxR (the only winning chance since if 46...Ne3 47. Rc8) 47. QxR Ne3 48. Qc7 and White's pressure is probably sufficient to survive despite Black's extra pawn.

If there is anything approaching a win for Black after 44. h4 Nxb 45. Rf1, it must come from 45...Ng4! A possible line then would be 46. Bh3 Qg6 47. d5! Qg7 48. Rc2 gxh4 49. Rf4 h5 50. Qc1! with good counter-play for White and excellent drawing chances.

After 44. Rc2?, by contrast, White was almost certainly lost:

click for larger view

Now, Keres went to town:

44... Qxh2+
45. Kf1 Rf7

Now, Smyslov faced another crucial decision if he were to retain any even practical hope of staving off impending defeat:

click for larger view

46. Ree2

Rightly called forced by Keres, though Smyslov had little chance to save the game at this point even with this accurate move.

I had thought that 46. d5 might have given Smyslov some hope, but then 46...exd5 47. exd5 Qxg3 is brutal.

46... Qxg3

"Threatening 47...Nh2+ 48. Kg1 Nf3+ 49. Kf1 Nh4 50. Bh1 Qh2." (Golombek)(see Euwe, similarly)

This left:

click for larger view

47. Qc3

"?"--(Golombek)(Euwe)

"Meekly submitting to his fate." (Golombek)

"Desperately trying to stop that terrible Pawn by getting the King Bishop out." (Kmoch)

"This quickly leads to a farlry easy win for Black." (Euwe)

"This loses quickly, but there was no longer any saving move." (Keres)

Much ink saw spilled concerning Smyslov's chances with 47. e5. but Keres cut through all of this by pointing out that Black then wins with 47...a4. Even better and more devastating is 47...b4! as given by Golombek and Euwe. Play would then likely go 48. d5 Nh2+ 49. Kg1 Nf3+ 50. Kf1 Nh4 51. Be4 and now even stronger than 51...exd5 as given by Golombek and Euwe is 51...Qh3+ 52. Ke1 Rf4 53. Kd2 RxB! 54. RxR (what else?--KEG) Qf3 55. Rd4 QxB+ 56. Kc1 Qe3+ 57. Kb1 Nf3 58. Rd1 Rxd5 59. Rc8+ Kg7 60. RxR Qe4+ 61. Kc1 QxR and with FOUR passed pawns for the exchange Black has an easy win.

In short, the by now shell-shocked Smyslov's 47. Qc3 was little short of surrender, but everything else would have lost as well.

In any case, after 47. Qc3, the position was:

click for larger view

47... QxQ

As Keres pointed out, he could also have won with 47...NxB 48. QxQ Nxe4+ 49. Qf3 RxQ+ 50. BxR Ng3+ followed by 51...NxR leaving him with four pawns for the White Bishop in an endgame that is probably winnable but requiring some work. The text was much simpler.

48. RxQ Rdf8

48...b4 also wins here for Black.

After 48...Rdf8, the position was:

click for larger view

49. Rcc2

49. Rc5 or 49. Kg1 might have prolonged the game, but the result was no longer in doubt,

49... NxB

49...RxB+ was perhaps most accurate, And 49...b4 seems to win easily as well. However, the win would have become more problematic, as Keres pointed out, if he had played 49...a4 since Smyslov could then have replied 50. Bh3, though even there Black should win: 50...NxB 51. Bxe6 Nd3+ 52. Kg1 b4 53. Rc6 b3 54. Bc4 Kg7 55. BxN a3 56. e5 Rf3 57. Bg6+ Kh7 58. Be4 Rf1+ 59. Kh2 R8f4 and Black's four passed pawns overwhelm White's two passed pawns and Bishop.

50. RxN RxR+
51. RxR RxR+
52. KxR a4!

click for larger view

"White can no longer stem the advance of [the a-pawn." (Keres)

53. Bh3

"The intended defense. Alas, it only prolongs White's suffering." (Kmoch)

53... Kf7
54. d5

click for larger view

54... exd5

54...a3 would, as Keres pointed out, have been simpler.

55. Bd7

If 55. exd5 a3 and the Black pawn Queens.

55... Kf6

55...a3 also does the trick.

56. Bc6

click for larger view

56... dxe4

56...a3 or 56...Ke5 are simpler still.

57. Bxb5 a3

0-1

click for larger view

<KEG:...My assessment of the text will be born out in my commentary on the balance of the game (which I think will reveal that 36. fxe4 was not the reason Smyslov lost). For now, let's consider the position after the move that Kmoch and Keres preferred, 36. fxe4...>

I guess this is just a quibble but I don't agree that Keres preferred 36 fxe. In fact he said:"...this capture (36 Nxe4 as played in the game) is far more dangerous for Black tha going over to the endgame by 36 fxe".

Complicated game. Think mindset had much to do with decisions in this battle.

First, and maybe most importantly, believe Smyslov gives away any real Catalan advantage with his, guessing experimental, 16. Bf1 and following idea. Think by 21...f5, he knows he didn't get what he wanted and is probably slightly worse and looking to equalize.

But afterward, Keres seems to look for an advantage haphazardly with moves like ...Ra7 and ...Qa7. Gives Smyslov a chance to regroup, though at the cost of a pawn, to 30...Nd6.

Then it looks like Smyslov, rightly or wrongly, goes for complications. Guessing going for counter play being a pawn down. Think he might of missed 38...Bc4 which gains an important tempo and dislocates the Q. Now likely figures Blacks got to be better.

Keres pushes the initiative, so we get to 43...Qh5. Smyslov knows he probably botched the opening and has been on the back foot for most of the game. Now he's worse and looking for the best/any chance to save it. He certainly considered 44. h4 but probably assessed it as being worse likely losing. So he chose liquidation, a piece up with connected center pawns. Again worse and probably losing but a piece is a piece.

<....Before analyzing the situation, let us first tune in to the Kmoch fantasy that Black is in the ascendancy....>