G Marco vs Duras, 1907

Smyslov and Reshevsky squared off five times at the 1948 World Championship Tournament, Smyslov winning one of these games while the other four were drawn.

In this first round match-up, both players played a line of the Ruy Lopez they both knew well. Indeed, Reshevsky's 1948 book on his best games includes his win at the 1944 U.S. Open against E.S. Jackson, Jr., in the very line he played in his first-round game against Smyslov.

Smyslov emerged from the well-played opening here (both sides coming well-armed for this battle) with the better position, and then after Reshevsky's doubtful 23...Ndx4 had excellent winning chances that he squandered on his 25th move.

Although in time trouble (as was usual for him) and in a difficult position, Reshevsky managed to come out of the middle-game complications a pawn to the good, but was--perhaps because of the little time left on his clock--unable to do anything with his edge and Smyslov was able to wriggle out of trouble and obtain a draw in a game he earlier might have won.

1. e4 e5
2. Nf3 Nc6
3. Bb5

The Ruy Lopez was popular at the 1948 World Championship, being played in 13 of the 50 game (26%).

3... a6

All thirteen of the Ruys played in this tournament continued 3...a6 4. Ba4. The Berlin Defense was apparently not then in fashion, and its ability to draw the fangs of the Ruy not then appreciated.

4. Ba4 Nf6

This was played in 8 of the 13 Ruys played in the tournament, the only five games featuring 4...d6 (the so -called Modern Steinitz Defense), including the other first round game between Euwe and Keres (won by the latter).

5. 0-0 Be7

The Open Ruy Lopez (5...Nxe4) was played four times in the tournament, and the text (the "closed" Ruy Lopez) was played four times.

6. Re1

Keres played 6. Qe2 here against Euwe in Round 6 of this tournament. The other three closed Ruys featured 6. Re1. Interestingly, all three of these games involved Smyslov, twice as White (here and in Round 9 against Euwe) and once as Black in Round 4, also against Euwe).

6... b5
7. Bb3 0-0

click for larger view

Both here and in Smyslov's 4th Round game against Euwe (in which Smyslov had Black the text rather than the more usual 7...d6 was played. To quote Hans Kmoch on these games: "A silent question: 'What do you think about Marshall's Gambit 8. c3 d5?' "

8. c3

Again quoting Kmoch: "The silent answer: 'I don't fear it at all , otherwise I would play 8. h3 or 8. a4 or 8. d4.' "

8... d6

"Closing the silent dialogue, indicating that he just wanted to (Smyslov's/Euwe's) opinion but had no real intention of playing Marshall's line." (Kmoch).

9.h3 Na5
10. Bc2 c5
11. d4 Qc7
12. Nbd2

click for larger view

Smyslov also reached this position as Black in his 4th Round game against Euwe.

12... cxd4

Repeating the line with which (as Black) he had defeated Jackson in the 1944 US Open. Smyslov as Black here played 12...Nc6 against Euwe.

Keres in his notes on this game says that Tchigorin (who invented this form of defense to the Ruy Lopez) here played 12...Nc6, but Flohr "invented" the text. Golombek in his commentary stated that both lines were mainstream at the time of this game.

13. cxd4 Nc6

13...Bb7 is a good and perhaps superior alternative.

14. Nb3

click for larger view

"Up to recently, White played 14. d5... The text move aims to maintain tension in the center." (Horowitz).

"Confronting Black with annoying problems." (Euwe)

14... a5

"White was threatening Bd2 followed by d5 and Ba5." (Reshevsky).

"The most aggressive continutation." (Golombek).

"14...Rd8 will not do." (Euwe).

But Keres saw matters differently here:

"Practice has shown that this early attack on the queenside is not justified, and usually only ends up weakening Black's pawn structure. The attempt 14...Rd8, to be able to carry out an eventual d5...d5 thrust, is also not recommended for Black due to 15. Bd2 (Smyslov-Keres, 1941). Therefore the simplest reply seems to be 14...Bd7, with which Black first tries to complete his development and only later initiate counterplay on the queenside. It is clear that in this case White would also achieve a small spatial plus, but it should not have any decisive importance."

15. Be3

"In Pinkus-Reshevsky, US Championship 1841, Pinkus played 15. d5. [With 15. Be3] White plans to defer this advance and keep the center fluid..." (Reshevsky).

For the reasons pointed out by Reshevsky, the text leaves White more options and seems best. By over-protecting the d-pawn, Smyslov is now able to answer 15...a4 with 16. Nbd2.

15... a4

The thematic follow-up to his last move.

16. Nbd2

click for larger view

"...White has been able to develop his queen's bishop to the defense of the d4=pawn, and at the same time clear the c1-square for a Rook...It is clear that if Black is unable to stir up tactical complications, he will soon be positionally lost." (Keres).

16... Bd7

"This whole variation has been much discussed and played in Russia..." (Golombek).

"It is evident that after the active preparation of an attack by the advance of the a-pawn, such a quiet developing move cannot be the best continuation. White now gains time to finish his development, is first to take control of the c-file and gains an advantage." (Keres).

While acknowledging Keres' point, it is hard to find a superior alternative. The main alternatives that have been tried are 16...Nb4 and 16...a3. Are either of these truly an improvement?

A) Golombek claimed that 16...Nb4 is best and says that Reshevsky only avoided this move because it had been thoroughly analyzed in Russia. But this seems clearly wrong, since Reshevsky played 16...Bd7 back in 1944 and--in his 1948 book called 16...Nb4 "distinctly inferior for Black." Keres likewise said that 16...Nb4 was ill-disguised. So who was right. Let's see: If 16...Nb4 17. Bb1 a3 18. Qb3 Qa5 (Reshevsky only considered the inferior 18...axb2 19. Qxb2 with a clearly superior game for White) 19. dxe5 (or Keres' suggested 19. Qxa3 QxQ 20. bxQ Rxa3 21. dxe5 dxe5 22. Nxe5 "and White maintains the extra pawn" as in Boleslavsky-Ragozin, 15th USSR Championship 1947 [Keres]--though Golombek curiously says that "Blackc still stands well") 19...dxd5 20. Nxe5 Be6 (20...Nd7 may be slightly better) 21. Qxa3 QxQ 22. bxQ Rxa3 where Golombek says that White's a2 pawn is doomed and that Black here has the better game. But after 23. Ndf3 White seems clearly better and the a2 pawn, at least for now, cannot be captured.

Thus, it appears that 16...Nb4 is no improvement on 16...Bd7

What about B) 16...a3? Reshevsky said this was "playable," but Keres, Euwe, Horowitz, and Golombek thinks White emerges with the far better position. Let's see:

16...a3 17. bax3 Rxa3 18. Qc1! (Keres' suggestion, also given as best by Horowitz and Euwe) leaves White with much the better game. Reshevsky gives 18. Bb3, but White also seems better even on this move after 18...Na5 19. dxe5 or 19. Rc1 (Reshevsky points out that 19...Nb1 does not win the exchange here because of 19...NxB!

Thus, it appears that 16...Bd7 is Black's best option. As will be discussed in later posts, Black's position is not as awful as Keres suggests, and Black has a tenable game.

17. Rc1

click for larger view

17...Rfc8

As Golombek pointed out, 17...a3 allows White to get the better game, but the simple 18. bax3 here (18...Rxa3 19. Bb3) is much better than Golombek's 18. b3. Golombek's denunciation of 18. Bb1 seems overdone, since White is still better after 18...axb2 19. Rc2.

After Reshevsky's 17...Rfc8, the position was:

click for larger view

18. Bb1

"Natural and strong; it hss also the incidental advantage of taking Reshevsky out of his own book, so to speak, by avoiding 18. Nb1 Qa5 19. a3 b4 (Jackson-Reshevsky, US Open 1944) when Black has the better game."

If, as Golombek knew of Reshevsky's 1944 game at the US Open, the Russian research team in 1948 was indeed formidable. Reshevsky's book featuring this game was not published until 1948 (probably after March 2 when the instant game was played). If Smyslov had seen Reshevsky's analysis, he would have known that he considered 18. Nb1 inferior and had recommended 18. Nf1.

The text looks OK to me. Stockfish likes Reshevsky's 18. Nf1, and Fritz prefers 18. Bd3 and 18. a3. These all look like reasonable choices to me.

With the text, Smyslov threatened to win a piece by 19. d5 (Golombek).

18... Qb8

"The comparatively awkward post foe the Queen is chargeable to the chosen defense." (Horowitz)

Actually, Black's position doesn't look all that awful to me. But on this move Reshevsky had to decide: Where to retreat the Queen?

Golombek and Romanovsky recommended 18...Qd8, but as Keres noted White could answer with 19. Nf1 "...and Black would have great difficulties creating counterplay, as the Knight is tied to the defense of the e5 pawn." Even better for White after 18...Qd8 is the more flexible 19. Qe2, putting pressure on Black' somewhat over-extended Queen-side.

Probably the text of 18...Qb7 are best for Black here.

19. Nf1

I still like 19. Qe2 here, but none of the commentators mention that possibility. Fritz likes 19. a3. Stockfish prefers 19. Qe2. In any case, it is clear that White is (at least) slightly better.

Euwe would say that I am understating White's advantage: "White has built up an excellent position; his pieces can move freely, whereas Back must continuously be aware of small and big threats and is virtually powerless to undertake anything himself."

19... Na5

This seems misguided and allowed Smyslov to trade Rooks and obtain a formidable advantage. 19...Qb7; 19...g6; and 19...Bd8 look better.

Smyslov's superior preparation can be seen by the times taken to this point: Smyslov: 0:38
Reshevsky: 1:03

This disparity soon became (much) worse for Reshevsky.

20. RxR+ BxR

"20...QxR is impossible in light of 21. dxe5 (21...dxe5 22. Nxe5)." (Euwe).

21. Bg5

click for larger view

This worked well in this game for Smyslov, but only because of second-best moves by Reshevsky (time pressure?). 21, Bd2 or 21. Ng3 (or maybe 21. Qe2) seem (theoretically) better.

21... h6

"Reshevsky took 45 minutes for this move, and his clock now read 1 hour 48 minutes, as opposed to Smyslov's 43 minutes [45 minutes according to Horowitz]." (Golombek)

"This weakening of the kingside was not necessary, but in any case Black has difficulties due to [sic] the threat of Nf1-e3-d5." (Keres)

Fritz and Stockfish prefer 21...Be6, but Keres denounced this in light of 22. Ne3 threatening Nd5. Keres certainly seems to have the better of this disagreement.

Keres's suggested 21...Ra7 may well be best. But White is still in a happier state after 22. Bc1 followed by 22. Ne3 with an eventual Nd5 soon to come.

Looking at the alternatives, Reshevsky's 21...h6 begins to look better and better to me. It left the position as follows:

click for larger view

22. Bh4

Horowitz preferred 22. Bc1 "...staring at the Black's h-pawn."

Golombek preferred the text:

"Stronger than 22. Bc1 since it is advisable to keep this square vacant for a major piece and so to gain command of the c-file."

22. Bd2, not mentioned by any of the commentators (or by Stockfish or Fritz which prefers the text) seems to give White the best of all worlds, i.e., eyes the h-pawn while keeping c1 vacant for a major piece to allow White to work on the c-file.

22... Nc6

22...Nc4, as correctly noted by Golombek, Euwe, and Keres, is no improvement, e.g., 22...Nc4 23. b3 axb3 24. axb3 after which White gets a clear edge after either Euwe's 24...Na3 25. Bd3 or after Golombek's 24...Nb6 (25. Ne3 or 25. Qe2).

Keres summed the situation up well:

"With the text-move, Black tries to relieve his position by exchanges...In any case, White maintains the better game and good attacking chances."

23. Ne3

Knight13 suggests 23. d5, but this works poorly after 23...Nb4.

The text threatens 24. Nd5. (Golombek).

After 23. Ne3, the position was:

click for larger view

23... Nxd4?

"?"--(Keres) (Wade--Whiteley--Keene).

"Gaining some freedom of action at the expense of a serous weakness in his Pawn position." (Golombek)

"Sooner or later, Black will have no choice but to releas the tension, but it is doubtful whether the best moment for this has arrived. One thing is certain, the text move allows White to build up a large advantage." (Euwe)

"...[Back] liquidates in the hope that simplification offers the best chance. In doing so, however, he fixes his own Queen Pawn as a target. In addition, the text is subject to an immediate tactical refutation..."

"Black's position was difficult, but the text-exchange does not ease it at all. Rather the opposite, as White now gets the chance to bring his Bishop on b1 into play, and gains a very strong attack." (Keres).

There is thus consensus that 23...Nxd4 was a mistake. But what should Reshevsky have played instead.

A) Keres disposes of 23...exd4? which--as he points out--loses to 24. Nd5!

B) Romanovsky recommended 23...Qb7. Horowitz and Golombek agreed. But Keres showed--while 23...Qb7 is certainly better than Reshevsky's move, it still leaves White much better after 24. dxe5 [24. Qd3 looks better still] dxe5 (24...Nxe5 appears a bit more accurate, but White is still on top with 25. Nd4) 25. Nd5.

C) Keres recommended 23...Ne8 ("...liquidating the annoying pin on the kingside"), but White still seems much better with 24. Bg3 or 23. d5.

D) Fritz and Stockfish both recommend 23...Qa7, but Black is still burdened with the inferior game after 24. dxe5 Nxd5 and now either 25. Qe2 or 25. Bg3.

In sum, Reshevsky had no way to equalize here, but his actual choice,as will be seen, was the worst of the plausible options and should have led to a rapid defeat.

24. NxN exN

This left:

click for larger view

25. Qxd4?

"?"--(Keres)(Euwe)(Horowitz)(Wade--Whiteley--Keens)

As very commentator on the game has noted, White can win here with 25. Nd5! Reshevsky would then have had two plausible moves (25...NxN and 25...Qa7, neither of which would have saved him:

A) 25...NxN 26. exN Bf8 {obviously not 26...BxB?? 27. Re8 mate). 27. Qd3 (Keres' 27. Qc2 would blow the win because of 27...f5! 28. Re8 d3 [not 28...Qb7? 29. Qc6! QxQ 30. dxQ and Black is toast] 29. Qxd3 [29. Qc3 and 29. Qc3 are stronger than Keres' 28. Qxd3, but are likewise insufficient to win] Qc7 [29...Qb7 is even better, but Keres' 29...Qc7 is also good enough] and now Keres' "30 g4!" gives White nothing after 30...Qf7 [a move Keres apparently missed]. 30. f3 was better, but would leave White with only a tiny edge. I initially thought 27. Bf6 would win, but Black seems to hold after 27...Bd7 (I had a lot of fun looking for a win after 27...gxB 28. Qd3 Bf5 29. QxB Bg7, but this in more problematic than after 27. Qd3. Returning to the better (and winning) 27. Qd3! and now" (i) 27...g6 (suggested by Horowitz, Golombek, and Euwe) 28. Qg3! Qc7 29. Bf6! Kh7 30. h4 "...with a decisive attack" (Euwe)(accord--Golombek). (ii) 27...f5 (better but insufficient) 28. g4! (but not Golombek's 28. Re6? BxR 29. dxB g6 30. g4 (30 e7 is somewhat better than Golombek's 30. g4 but hardly good enough to win) Qb7 31. gxf5 Rc8 32. Bc2 gxf5 33. Qxf5 Rc5 and Black is much better.

B) 25...Qa7 is also inadequate: 26. NxN+ BxN 27. BxB gxB 28. Qf3! Qe7 (28...Kg7 may be better, but it would not save the day) 29. e5 Bb7 (this move by Euwe is suicide, but the "better" 29...Rb8 also loses after 29. Qg3+) 30. Qg4+ (even more crushing than Euwe's 30. Be4.

In sum, the intermediate move 25. Nd5! would have led to a win for Smyslov.

After Smyslov's actual 25. Qxd4, the position was:

click for larger view

Reshevsky was no longer lost, but he was clearly far from out of the woods at this point.

25... Qa7

"Hoping for more exchanges" (Horowitz).

Reshevsky was also still in trouble on the clock: Smyslov: 0:50
Reshevsku: 2:07

26. Qd3

The commentators were unanimous that Smyslov was right to avoid the exchange of Queens here:

"The exchange of Queens would rob White's attack of its sting." (Golombek)

"White not only has pressure against the center, but also prospects of a king-side assault. That is why he declines the offer to exchange Queens." (Horowitz).

But was 26. QxQ really bad for White. After 26. QxQ RxQ 27. Rc1! (much better than Golombek's 27. Rd1 which lets Black off the hook via 27...Nxe4 [not Golombek's 27...Ne8 after which White would still have hopes of winning] 28. BxB RxB and Black should be fine) with strong pressure on the c-file and much the better--even if not still winning--game.

26... Be6

"!"--(Keres)

"Black was unable to defend the b-pawn, because among other things White was threatening the strong 27. e5 dxe5 28. BxN followed by Qh7+ and Nd5 with a winning position. The text-move not only thwarts this threat, but also neutralizes the other possibility 27. Bd5 NxN 28. exN BaB and Black equalizes [an overstatement, White is still better here, but Black seems to be able to hold]. White can indeed win a pawn by 27. Qxb5, but after 27...Rb8 28. Qe2 Qd4 29. Nd1 he has taken on a laborious defense; Black can continue either with 29...a3 30. b3 Bd8 [30...Rc8 is the way for Black to equalize here] or the immediate 29...Bd8, and obtain a permanent initiative for the pawn.

After 26...Be6, the position was:

click for larger view

27. Bg3

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

All of the commentators agree that the text was a mistake that threw away any remaining chances Smyslov had to play for a win. They all also agree that 27. Rc1! was best.

Is this correct.

It is certainly true, as will be seen, that Smyslov's winning chances were significantly reduced by the text. But was 27. Rc1 truly an improvement?

Golombek only considered 27...Rc8 as a reply, which most certainly loses: 28. RxR+ BxR 29. e5! dxe5 (29...Qc5 is no real improvement on Golombek's move) 30. BxN BxB (30...g6, giving up a piece, would avoid the immediate catastrophe, but obviously is hopeless) 31. Qh7+ Kf8 32. Qh8+ Ke7 33. QxB.

But instead of 27...Rc8? Black should answer 27. Rc1 with 27...b4! Black now has counterplay that should be sufficient to hold the game.

The best assessment of 27. Rc1 was that of Euwe: "With 27. Rc1 White might have retained some advantage."

Only Keres considered the move identified as best by both Fritz and Stockfish: 27. Nd5! Keres claimed that Black could now equalize with 27...NxN 28. exN BxB. But in fact White--though perhaps not winning--is still on top after 29. Qh7+ Kf8 30. Re3! Bg5 (everything else loses) 31. Re4 Bf6 (forced) 32. dxe6 Be5 (again Black's only move) 33. exf7 Qxf7 and now, despite Bishops of opposite colors, White has chances.

27. Nd5 probably is now a winning move for White, but it was the only real chance still at Smyslov's disposal.

After Smyslov's actual 27. Bg3, Reshevsky was very much back in the game.

27... Qc5

27...Qb7 or 27...Qa5 were better, but Reshevsky still had a defensible position after the text, which left the position as follows:

click for larger view

28. Nd5

"?"--(Keres)(Golombek)(Horowitz)(Wade--Whiteley--Keene- )(Resignation Trap)

"Still another point of transition, from good to indifferent." (Horowitz).

All of the commentators agree that 28. Rd1 was best. The problems with the text were well-expressed by Keres: "With this exchange White eliminates his best-placed piece and at the same time gives himself a pawn weakness on d5 which later require permanent protection."

28... BxN

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

"There is an exception to every rule. A Bishop is [usually] stronger than a Knight. Yet here, the sacrifice of the minor exchange (Bishop for Knight) is indicated. For one thing, the approach to Black's Queen Pawn along the Queen file is closed. Also, White's Pawn (at d5) will become a target." (Horowitz).

29. exB

click for larger view

29... Ra7

29...Bf8 or 29...Re8 were easier roads to equality, but the text hardly ruined anything much for Reshevsky.

30. Rd1

"The pawn weakness on d5 is already making itself felt: Black simply threatened to play 30...Qxd5." (Keres)

"What White [Smyslov] missed in his calculations was that 30. Bh4 fails to 30...Qb4!" (Euwe)

Euwe was correct that 30. Bh4 would doubtless have been answered by 30...Qb4, but this would only have led to a draw, not to any advantage for Black.

30... Rc7

click for larger view

Reshevsky was past the worst and should be able to hold the position at this point and escape with a draw. Just one problem, the clock:

Smyslov: 1:14
Reshevsky: 2:16

31. Bh4

"Better...was 31. Bf4 controlling c1, and preventing Black's following maneuver." (Golombek)

"...In order to thwart the threat of 31...Qxd5, it was necessary to play 31. Bf4, also enabling Rc1 in some lines..." (Keres)

31. Bf4 does indeed seem better than the text, but neither move gets White anywhere close to winning, and neither move gives Black much to play for. The real problem with the move is that Smyslov spent no less than 33 minutes considering what to play, thus losing a major portion of his advantage on the clock.

31... g6

Much ink was spilled in considering Reshevsky's choices here, but in fact none of the options considered were all that much better or worse than the text, which was likely best.

A) Golombek argued that 31...g5 (which looks fine to me) would have been a mistake. Golombek's analysis of 31...g5 was badly flawed: 31...g5 32. Bg3 Qxd5 [this move is OK, but 32...Qc4 is a simpler route to a draw] 33. QxQ? [This makes White's task more difficult. 33. Qe2 leads to an even game] 33...NxQ 34. RxN? [This should lose outright, White is still in decent shape--though inferior because of 33. QxQ--with 34. Kf1 or 34. Kh2 or 34. Bf5 or even 34. Be4]. Rc8+ 35. Kh2 RxB 36. Rxb5 Bf6? [Black wins with 36...a3! and probably also with 36...f5] 37. Bxd6 Rxb2 38. RxR BxR 39. g4 and Golombek's claim that Black "...should draw [but] has the worse of the ending owing to his Pawns being on black squares" is ridiculous, since the position is now a routine draw.

Thus 31...g5, like the text, was fine.

B) What about 31...Qxd5. Horowitz said of this move: "The immediate 31...Qxd5 is playable but there is no hurry. The game is indeed roughly even after 31...Qxd5 32. BxN QxQ 33. BxQ BxB 34. Bxb5. especially because of Bishops of opposite colors. If anyone is better here, however, it surely is White. Thus, 31...Qxd5 is no improvement on the text (31...g6).

As for the text move, I agree with Keres' assessment that it was stronger than 31...Qxd5 and that: "Now it is not Black, but rather White who must start fighting for a draw."

After 31...g6, the position was:

click for larger view

The clocks at this point read (according to Horowitz):

Smyslov: 1:47
Resheveky: 2:21 [or 2:25 according to Golombek]

32. a3

Keres correctly noted here that the text was better than 32. BxN BxB after which Black, despite the Bishops of opposite colors, would have the better chances.

Both Keres and Golombek thought that the text was best:

"White's attack has now completely disappeared and he has nothing better than this after which the advantage passes to Black." (Golombek).

"White cannot do anything to protect the d-pawn so he therefore clears the a2 square for his Bishop in order to create attacking chances against the g6 and f7 points." (Keres)

The text, as Golombek and Keres recognized,leaves Black better positioned. Better therefore was 32. Bg3 which would allow this Bishop to retreat to h2 in the event of 32...Nh5 and seems to leave White with approximately even chances. As will be seen, after the text Reshevsky could have obtained the advantage.

32... Qxd5

"?"--(Golombek).

The text promised little more than equality. Was there something better?

A) Golombek claimed that 32...Nxd5 was best:

"Black was, typically, in acute time trouble, having only five minutes for his remaining nine moves. He therefore misses 32...NxP! 33. Bg3 Nf6 when he would be a pawn to the good, with winning chances."

In fact, and contrary to the analysis of both Golombek and Keres, after 32...Nxd5 White can obtain near equality with 33. BxB (instead of 33. Bg3). [Incidentally, in Golombek's line, 33...Nf6 gives away Black's advantage (34. Bf4) and Black should rather play 33...Nb6 with much the better chances].

B) Both Keres and Wade--Whiteley-Keene claimed that 32...Nh5 was best. Fritz agrees. While this was indeed superior to Reshevsky's 32...Qxd5, Black's edge is quite minimal after 33. BxB RxB 34. Qd2.

C) The best move seems to be Stockfish's 32...b4. A likely continuation is 33. Bg3 bxa3 34. bxa3 Qc4 35. QxQ RxQ 36. Kh2 Nh5 37. Bd3 Rc3 38. Bb5 Rxa3 39. Rd4 NxB 40. fxN Bf6 where, despite Bishops of opposite colors, Black--being a pawn to the good--would have some winning chances.

After Reshevsky's actual 32...Qxd5, the position was:

click for larger view

33. BxN

Golombek trashes 33. QxQ as an option, but his analysis was awful

33. QxQ NxQ 34. Bg3?? [33. QxQ is hardly best here, but it need not be catastrophic if followed up with 34. RxN (provided White does not--after 34...BxB, play 35. Rxd6? Rc1+ but instead plays 35. Bd3) or 34. BxB. 34. Bg3??, however, should lose immediately via 34...Nb6 but not after Golombek's very bad 34...Nf6? after which White would be fine if he plays 35. Bf4 or 35. Bd3. Golombek manufactured a win for Black here, however, by having White continue with 35. Bxd6?? which loses to 35...Rd7!.

Smyslov's 33. BxN, though better than 33. QxQ, still left White a pawn down with work to do to get a draw despite the Bishops of opposite colors. Beat for White here was either 33. Bg3 or 33. Re1, though even here Black would be better with some chances to torture White for a while.

33... QxQ
34. BxQ

"After 34. RxQ Rc1+ 35. Kh2 BxB 36. Ba2 Ra1 37. Bd5 Bxb2 [37...Be5+ is even more crushing--KEG] Black wins." (Golombek).

34... BxB

"Thus, Black emerges from the liquidation with a Pawn plus." (Horowitz)

35. Bxb5 Bxb2

click for larger view

"For the first time Black is ahead." (Horowitz).

36. Rd3

"Better than 36. Bxa4 Bxa3 when White will have difficulty in drawing." (Golombek).

But see contra:

"Black would also retain the extra pawn with 36. Bxa4 Bxa3. But despite this the endgame is a draw." (Keres)

Keres' assessment is doubtless correct, though I cannot persuade my silicon friends.

A bad idea here, incidentally, is Horowitz' proposed 36. Rxd6 which makes White's task harder after 36...Rc1+ 37. Bf1 [Not 37. Kh2+ Be5+ and White can resign] 38. Ra6 Rxa3 39. Bc4 Ra1+ (much better than Horowitz' 39...Be5) 40. Kh2 Bd4 41. Rxg6+ Kh7 42. Bxf7 Bxf2 and White, though having won back his pawn, has work to do to draw.

36... Rc5

"?"--(Horowitz)

As both Horowitz and Resignation Trap on this site have correctly noted, Reshevsky had better chances to try to squeeze out a win with 36...Rc1+ 37. Kh2 Ra1 38. Bxa4 Bxa3. Nonetheless, with Bishops of opposite colors, I would place heavy odds that Smyslov would have managed to draw.

37. Bxa4 Ra5
38. Bb3

click for larger view

38... Bxa3

"Black could have retained the extra pawn with 38...Rxa3 but, of course, this would not have changed the final outcome."

38...Rxa3 would have been met by 39. Rf3 [39. Kf1 looks simpler to me--KEG] also with a draw." (Euwe).

"He can come down to an ending with Bishops of opposite color and a Pawn more by 38...Rxa3 39. Bb4 RxR [39...Ra1+ might leave White some work to do--KEG] etc., but the game is still quite drawn." (Golombek).

39. Rf3!

click for larger view

39... d5

"Or 39...Ra7 40. Bxf7+" (Golombek)[Accord--Wade--Whiteley--Keene]

40. Bxd5 RxB
41. RxB

click for larger view