12. b5

"There is no good way to maintain the Knight Pawn on its present square." (Horowitz)

True, but 12. bax5 Qxa5+ 13. Bc3 Qc7 14. Bb4 was about as good as the text. Both 12. b5 and 12. bxa5 suffer from the same problem, the resulting isolated White a-pawn. This was not the fatal weakness many of the commentaries on this game suggest, but it was a problem that would haunt Botvinnik for much of what followed.

12... Nb6

Both sides had used considerable time to reach this point [Botvinnik 1:02; Reshevsky 1;06 according to Horowitz].

13. Bd3 cxb5
14. Bxb5 Bd7

click for larger view

"Black's development proceeds apace with critical gain of tempi. If White now exchanges Bishops, Black's Knight will establish itself on c4." (Golombek)

"Thus Black has completed his development." (Horowitz)

"Black's position is clearly better. He has a lead in development and White has weaknesses on the queenside." (Euwe).

"White is vulnerable to...a4...Na4 or a later...Nc4." (Soltis)

15. Qb3

"?!"--(Kasparov)

Most of the commentators agreed that the text loses time. Keres and Kasparov both claimed that an immediate 15. Qd3 (using one move rather than two to bring the Queen to this square) was best. Golombel and Euwe feared 15...Na4 in response. Keres and Kasparov responded that White would not be so bad after 15...Na4 16. Bc1. But 15...Rc8 is a stronger response to 15. Qd3, and after 15...Na4, 16. BxN BxB 17. Nc3 is a better line for White.

Only Horowitz defended the text, claiming that Botvinnik wanted to "tempt" Resheveky to play 15...a4 since this would preempt a later Qa4 by Black. But if this was indeed Botvinnik's plan, it was an unusually obtuse effort by this master strategist. The loss of time coupled with allowing Reshevsky to dominate the Queenside looks horrible.

Surprisingly, none of the commentators mention White's best move: the simple 15. Bd3. This avoids loss of time and gets the Bishop to a nice square both for defense and for some possible later King-side operations.

15... a4

"!"--(Kasparov)

Time was becoming a more serious issue, if Horowitz's figures were correct:

Botvinnik: 1:17
Resheveky 1:21

16. Qd3 Ra5

click for larger view

"A timely maneuver to compel White to abandon his hold on b5." (Horowitz)

"In addition to excellent development, Black has achieved a position where he had good prospects of utilizing the weak light squares on White's queenside." (Keres)

As with much of what has been written about this game, the above comments--though incisive--miss the forest for the trees. The best chance for Black to exploit his advantage lay in 16...BxB (17. QxB Nbd5 leaving White with an uphill defensive task on the Queen-side) or 16...Rc8

The text gave Botvinnik an opportunity to solve most of his problems.

17. Nc3

As Golombek and Horowitz aptly noted, this move was an implicit admission that White's 11th move was a lemon.

But what about 17. BxB?

Keres, Kasparov, and Moran all claimed that 17. BxB would run into trouble after 17...QxB 18. Bc3. But they all fail to mention 18. Rb1 which seems to avoid the problems that arise after 18. Bc3 and yields good counterplay--including on the Queen-side for White.

17... Qe8

"With this move Black finally conquers the b-5 square, although at the cost of tempo." (Keres)

Both Flohr and Euwe suggested 17...Nfd5, but Keres' 18. 0-0 then seems to get White out of trouble.

I had thought 17...Nbd5 (not mentioned by any of the commentators) was best, but after 18. BxB (pretty much forced) QxB 19. Rb1 Botvinnik would be nearly out of the woods.

In fact, Black has no way to make significant progress here. The text leaves Black with only a modest edge. I see nothing much better. Stockfish plays 17...Bb8 here, but White then seems fine awith 18. Rb1 BxB 19. NxB).

18. BxB QxB

click for larger view

19. 0-0

As Keres and Kasparov point out, 19. e4 would be answered by 19...Rc8, at which point, as Kasparov further notes, 20. e5 loses a pawn to 20...Bxe5!

Botvinnik could, however, have played the slightly more accurate 19. Nd2 here.

19... Rc8

click for larger view

Somewhat overstating his case, Kasparov says that Black had now been outplayed in both the opening and the middlegame. The one compensating fact, he continues, was Botvinnik's lead on the clock which, according to Horowitz, showed:

Botvinnik: 1:32
Reshevsky: 1:44

20. e4

"?"--(Keres)(Moran)

"?!"--(Kasparov)

Opinion was hotly divided on the merits of this advance.

Golombek and Horowitz thought it the best chance:

"White's only counter-chance lies in this central thrust..." (Golombek)

"Conceding that Black has won out on the Queen-side, Botvinnik acts with dispatch on the other wing." (Horowitz)

Keres and Kasparov thought otherwise:

"...the action in the centre initiated by the text-move worsens the situation even more,since after the opening of the centre, Black's pieces gain an even broader field of action." (Keres)

"A pseudo-active move, which merely helps Black to widen his expansion on the queen-side, whereas White is not able to create anything real in the centre." (Kasparov)

Kasparov and Moran suggest 20. Rb1 Nc4 21. Bc1 as a better line for White. but after 21...e5 Black definitely seems better.

Keres (while also--mistakenly in my view--supporting the notion of 20. Rb1) presented what looks to me to be the best assessment:

"It was time to take up a long-term stubborn defense by 20. Nd2. It seems that this would have entailed White to defend his vulnerable position sufficiently."

For what it's worth, both Fritz and Stockfish are with Keres' above analysis here, and think that White's position is not all that bad.

20... Nc4

click for larger view

"A magnificent square for the Knight." (Euwe).

21. Bc1 e5

"!"--(Keres)(kasparov)(Moran)

"...in order to destroy White's hold on the centre." (Golombek)

"Even here...Botvinnik has been crossed. The move he wished to prevent is now enforced." (Horowitz)

"Black opens up the whole center and coniderably increases the activity of his pieces." (Keres)

"A very strong move, forcing White to determine his pawn structure." (Kasparov)

The uniform acclaim bestowed on this move suggests even the esteemed commentators on this game were looking at prior published analysis, since none of them even mentions 21...b5, which seems most consistent with Black's Queen-side play (and is supported by my silicone friends).

After Reshevsky's 21...e5, the position was:

click for larger view

22. Rd1

"?!?--(Kasparov)

Botvinnik here faced a difficult decision. Should he play the text or 22. d5.

Keres and Kasparov opt, on balance, for 22. d5, warts and all. Glombek and Horowitz preferred the text, fearing that by locking the center Black would have an easier time on the Queen-side, which would--in their view--more than compensate for White's protected passed D-pawn.

Keres said that 22. d5 was Bovinnik's "best practical chance," but thinks Botvinnik's position would still not be satisfactory.

Only Kasparov provided any variation to consider play after 22. d5. He says that locking the center was White's best choice, saying that despite Black's undoubted edge the win would still be "a long way off."

But Kasparov's line looks flawed:

22. d5 b5 (22...Rac5 is the crucial line and probably the only way for Black to create serious winning chances against best play. Kasparo'vs 22...b5 looks clearly inferior) 23. Nh4? (this makes trouble for White, who would likely have a defensible position with 23. Nd2, and probably also with 23. Re1 or 23. h3. After 23. Nh4?, Black is again definitely worse) 23...b4 (this is strong, but perhaps the simple 23...Nxa3 24. BxN BxB 25. RxB b4 26. Rxa4 RxR 27. NxR QxN is the best winning chance, Black's advanced passed b-pawn being here far stronger than White's passed d-pawn) 24. axb4 Bxb4 25. Na2 Bf8 (25...Bc5 is better) 26. Bg5? (Awful. White probably has a defensible, albeit still somewhat difficult, game with 26. Nf5 or 26. Nc3) Ne8 27. Nf5 Ned6 after which Black is indeed better, but he would be better still with 27...h6 or 27...f6.

In sum, not even Kasparov has demonstrated that 22. d5 was noticeably--or even at all--better than Botvinnik's Rd1 as a theoretical matter. Speaking practically, by keeping the center fluid, Botvinnik left more for Reshevsky to ponder. Given the latter's time crunch, on balance I think Botvinnik's move was the best practical chance, as indeed it proved to be.

In any case, after Botvinnik's 22. Rd1, the position was:

click for larger view

22... exd4
23. Qxd4?

This recapture has passed without comment in all the commentaries I have seen on this game. But it looks really bad, and 23. Nxd4 was the only move I even considered. Fritz and Stockfish also play the seemingly obvious 23. Nxd4 here. After Botvinnik's 23. Qxd4, the position was:

click for larger view

23... Qe6

"!"--(Kasparov)

Really? White appears to have a good retort to the text. I agree with Golombek's assesmemt of 23...Qe6, to a point:

"Good enough to maintain his advantage, but even stronger is 23...Qc7. threatening the deadly Ne5."

But even Golombek misses 23...Qe8, the most difficult move for White to meet here.

After 23...Qe6, the position was:

click for larger view

24. Ra2?

This move, like Kasparov's doubtful 23. Qxd4, has passed without criticism. The only comments I have found on 24. Ra2 are:

"It is necessary to protect the f-pawn since Black was threatening Bc5 followed by an eventual Ng4." (Golombek)

"White has no decent moves left to play." (Euwe)

But, in part thanks to Reshevsky's incorrect 23rd move, Botvinnik looks basically OK with 24. Qd3 (getting the Queen off the diagonal where Botvinnik put it one move ago, if now 24...Bc5 Botvinnik could just play 25. h3 or perhaps even better 25. Nd4). Even if Black played the stronger 24...Qe8, White can probably hold with 25. Nd2 (drawing the sting from th Black c4 Knight).

After 24. Ra2?, Botvinnik was now truly in trouble (on the board, even if not on the clock):

click for larger view

24... h6!

"...deprives the white pieces of the g5-square..." (Keres)

This was also, as Keres went on to discuss, a useful type of move to be able to play when in time trouble.

25. h3

Somewhat weakening. 25. Re2 or maybe 25. Kf1 (to get the King off the vicious black-square diagonal) were probably better, but in any case Botvinnik was in trouble.

25... Ra6

"White supports his Bishop in order to play Ne5." (Golombek)

"To protect the Bishop at d6 and free the Knight at c4 for action." (Horowitz)

"Threatens 26...Ne5!" (Moran)

"It is here that the Rook is now needed most of all." (Kasparov)

The Bishop sacrifice beginning with 25...Ne5 loses, as demonstrated in the following line given by both Golombek and Soltis: 26. QxB NxN+ 27. gxN Qxh3? (this is terrible, but even the "better" 27...Qb3 is hopeless for Black after 28. e5) 28. Qg3 QxQ+ (28,,,Qh5 or 28...Qe6 are no real improvements) 29. fxQ RxN 30. RxR+ (the key to this line) Kh7 31. Bd2 with a winning Rook skewer,

Black also probably wins with 25...Rac5, but the text (25...Ra6 looks best).

According to Horowitz, the clocks now showed frightening time trouble:

Botvinnik- 2:14
Reshevsky--2:20

The position was now (after 25...Ra6) was:

click for larger view

26. Nd5

"!"--(Keres)(Kasparov)(Moran)

"An incorrect pawn sacrifice which, however, wins the game! Any passive continuation allows Black to play Ne5 and win comfortably." (Golombek)

"Botvinnik takes a firm stand. Otherwise he must succumb to the weakness of the a-pawn." (Horowitz)

"White sacrifices a pawn, hoping to confuse his time-troubled opponent in the upcoming complications. In this position who could have believed that after the next seven moves it will be Black who is forced to resign." (Keres)

The position was now:

click for larger view

26... Nxe4

Euwe and Kasparov both claimed that Reshevsky could equally well have declined the offered pawn and won with 26...NxN 27. exN Qf5, but this seems to give Botvinnik new life with 28. Nh4.

27. Re2

A gallant effort by Botvinnik. As Kasparov has pointed out, 27. Nf4 was no better and would lose to either: (A) 27...BxN 28. BxB Qf5 (even better here is 28...Nf6); or (B) 27...Qf5 (strongest--KEG) 28. g4 (or 28. Nd5 Ra5 29. Rc2 Rac5) Ne5 (28...Ng5 is even better--KEG) 29. NxN (Kasparov's proposed 29. Ne1 here is very weak and gets crushed by 29...Nxg4! 30. hxN Qxg4+ 31. Nfg2 Bc5 32. Qd7 Bxf2+ [32...Nxf2 also wins--KEG] 33. RxB QxQ 34. RxQ NxR) QxN 30. QxQ BxQ and Black should win because of his extra pawn and White's still weak isolated a-pawn.

27... f5

click for larger view

28. g4

"!"--(Kasparov)

Theoretically, 28. Nh4 is best (as any computer will verify). Golombek's characterization of the text as "unsound" is correct in an ideal world. But 28. Nh4 would also lose in that ideal world. Botvinnik's move was designed to--and did--present complex problems for his time-troubled adversary.

Most of the commentators are agreed on this; it was "do or die" (quoting Horowitz) for Botvinnik, and passive/positional play would not save him. As Soltis notes, after the text, 29. gxf5 winning a piece was threatened. This is just the sort of threat one would rather not have to meet with the clock running down.

After 28. g4, the position was:

click for larger view

28... Bc5?

"?"--(Golombek)(Euwe)(Soltis)(Kasparov)(Moran)

"??"--(Keres)(Horowitz)

As everyone and his uncle has pointed out, 28...Ng5:

(A) 29. RxQ NxN+ 30. Kg2 NxQ 31. RxN fxg4 32. hxg4 (or 32. Rxg4 Kf7 33. Reg6 Ne5 34. Rxg7+ Ke6 35. Nf4+ Kf5 36. Ne2 NxR 37. hxN+ Ke4 38, Rxb7 Rc2 39. Kf1 Kd3 (even better than Kasparov's 39...Rac6) Kf7 33. Ree4 (if 33. Re1 b5) b5.

(B) 29. Qd3 NxN+ 30. QxN fxg4 31. hxg4 Ne5 32. Qf5 (everything else is worse--KEG) QxQ 33. gxQ Qac6 34. Bf4 Nf3+ 35. Kf1 BxB 36. NxB Rc1 37. Re8+ RxR(e8) 38. RxR Re4 with an easily won ending.

(C) 29. NxN (what Kasparov calls a "sudden attack" and the only move that "forces Black to be on the look-out--but which objectively is the worst of White's options) QxR 30. Nf6+ gxN 31. Qd5+ Kh8 (any other move leads to immediate check-mate) 32. Nf7+ Kh7 33. Qxf5+ Kg7 34. Bxh6+ KxN 35. Qd7+ Qe7 36. QxR Qe4 37. Rd4 Nb6 38. Qc3 Be5 after which, to quote Kasparov, "White's attack peters out" (and he is down a piece--KEG)

Golombek claimed a win here with 28...Ne5, but Kasparov has completely refuted this motion.

Euwe claimed a win with 28...Qf7, but after 29. gxf5 Nf6 30. NxN+ QxN 31. Qe4 Black's winning chances seem to be gone.

The only winning move other than 28...Ng5 is what Kasparov aptly calls the "computer move": 28...Ra5. All I can say for this move (which has been exhaustively analyzed by Kasparov) is that I sure didn't consider it but Fritz and Stockfish found it in a second. It seems to have little to do with chess played by humans, and could hardly have been found by Reshevsky operating with seconds left on his clock.

After Reshevsky's actual (and unfortunate) 28...Bc5?. the position was:

click for larger view

Even after 28...Bc5? Reshevsky had much and better chances--and perhaps even a theoretical win:

29. gxf5

click for larger view

Horowitz claimed that White much now win a piece. In actuality, however, Botvinnik's move still left--with best play--Black with Rook and two pawns for Knight and Bishop, and at the very least excellent theoretical winning chances.

But all that meant nothing because: (i) Reshevsky, as the sequel shows, was shocked by the turn of events; and (ii) Reshevsky was nearly out of time.

I will discuss how the above position was transformed into a loss for Black in just three awful moves by Reshevsky in my next post on this game.

29... Qxf5?

As just about every commentator has pointed out, Reshevsky should have played 29...BxQ. Contrary to what Horowitz claimed, White does not win a piece. Rather, after: 30. fxQ Bxf2+ 31. RxB NxR 32. KxN Rxe6, the position would be:

click for larger view

Black here has Rook and two pawns for Bishop and Knight and has some chances to win. He is certainly better.

But after 29...Qxf5?, the position was:

click for larger view

Black now is sunk. Reshevsky's move did pose a small trap (as noted by Kasparov). Had Botvinnik here carelessly played 30. QxB?? then Black wins with 30...RxQ 31. Ne7+ Kh7 32. NxQ Nc3 and the Knight fork dooms White to defeat.

But Botvinnik was too wily to fall for that, and played the winning:

30. QxN(e4)

click for larger view

Reshevsky now had nothing better than 30...QxQ 31. RxQ Nxa3 (or 31...b5 32. Rd3) 32. Be3 Bd6 33. Nb6 leaving Botvinnik up a piece for a pawn in an ending he should easily have been able to convert.

Reshevsky, in terrible time trouble, however, played something even worse:

30... Qxh3

click for larger view

What happened now was truly bizarre, and so far as I can determine has only been spotted by <wwall> on this site.

Botvinnik had an easy win with 31. Nf4. Instead, he played (perhaps because he was also short of time):

31. Nh2?

click for larger view

Suddenly, Reshevsky had a chance to salvage the game with 31...Qh5! (e.g., 32. Rd3 Bf8 33. Rg3 Bf8 34. Rg3 Rd6 35. Ne7+ BxN 36. QxB and Black with two pawns for a Bishop in a complicated position might have at least practical chances. But:

31... Rcc6??

<wwall> calls this the "losing move." One can argue about whether Black had any real chance to save the game with 31...Qh5. But the text definitely was a loser:

32. Nf4!

click for larger view

With the Black Knight about to fall (if Black tries to defend the Knight with 32...Qc3 or 32...Qb3 he gets mated after 33. Rd8+), the game was over.

Reshevsky is said to have lost on time at this point. <keypushr> may well be right that Reshevsky, seeing all was lost, may have let his time run our (he was quite capable of making nine moves in well under a minute). But the game was lost in any event.

1-0

With this win, Botvinnik was off and running. After he beat Keres in the next round, he was a point ahead of the field after the first lap, and he never looked back on his way to the title.

Botvinnik should be given some credit for posing difficult problems for Reshevsky when he was very short on time.

<KEG:Of course, had Reshevsky won this game, it is possible that Stalin or others might have instructed Smyslov to lose another game to Botvinnik, and Keres might have lost his final round game to Botvinnik had the World Championship been at stake at that point. >

I am far from convinced that any of the games at the 1948 World Championship were thrown. But there has been a lot written about whether Keres was instructed to lose games to Botvinnik and some circumstantial evidence to support this.

While I have reader nothing specific about Smyslov, several sources--most notably Kasparov in Volume II of "My Great Predecessors" report that between the 2nd and 3rd laps of the tournament, Botvinnik was summoned to a meeting with the secretariat of the Communist Party Central Committee who were allegedly concerned that Reshevsky would overtake Botvinnik and win the title. Kasparov goes on (at p. 161) to quote Andrei Zhdanov as having asking Botvinnik: "How would you view it if the other Soviet players were to lose to you deliberately?" Kasparov says that according to Botvinnik, he initially rejected the offer, but then went on to say "Vry well, let's leave the question open--perhaps this won't be needed."

It is hard to know how seriously to take any of this. I was certainly not claiming that I believe there were any games thrown. But I was--in the language from my post you quote--playing "what if," and speculating about Reshevsky's chances of becoming World Champion had he won this 4th round game in the first lap against Botvinnik. Having stated that a win here would have given Reshevsky real chances, I thought it only fair to point out that my speculations would be meaningless if Smyslov or other Soviet players were ordered to lose to Botvinnik.

click for larger view

It's your turn to move, and you have a choice.

You can either play 1. Rd5 with the idea of Ra5, attacking and defending at the same time.

Or you can go 1. Rd3 with the idea of Ra3, defending from a3, with the option of Ra1.

Which one do you choose?

If you are citing this position as evidence that Keres was throwing games to Botvinnik, I should note that Hans Kmoch, with all the time in the world to analyze this position, said that Black's "threat forces White to retreat his Rook to a passive position" (i.e., play Rd3).

If so fine an analyst as Hans Kmoch could get this wrong when not under time pressure, I suppose even Paul Keres could get it wrong over the board if--as he later reported--he was in time trouble.

I am not saying that there aren't troubling bits of information that suggest something sinister was afoot at the 1948 World Championship Tournament, only that I have seen nothing which I consider remotely conclusive. I am, however, in the process of going through the games of this tournament and reading more about it. Perhaps I will change my mind when I examine the tournament further.

* 28...Ng5: -1.71 (40 ply). OK, not crushing but sufficient to get into a superior endgame which at that level means probably winning. That said, Reshevsky was in time trouble so anything could have happened.
* 28...Ra5: -1.40 (39 ply). The above-mentioned computer-like move.
* <28...Bc5>: -1.11 (45 ply). Just 0.6 off the best move so maybe deserves "?!" more than "?"

<29...Qxf5>: +3.20 (47 ply). The first big mistake, should have played Bxd4.

<30...Qxh3>: +10.92 (35 ply). Second big mistake, should have played Qxe4. Now Botvinnink misses the crushing 31.Nf4.