It was not. This tournament was in no way, shape or form a Candidates Tournament, though the organizers wanted it to be.

<The competitors in this championship were the eight top players from that time. This game is from the second championship. In the first one the title was shared between <Fine>(USA) and <Keres>(estonia) with third going to <Botvinnik>(Russian Federation).>

Not sure what you mean here. There was only one AVRO tournament that all these people played in, which was won by Keres on tiebreak over Fine. This game is from that tournament, not from some subsequent tournament.

1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 d5 (4…0–0 or 4…c5 are more flexible) 5.a3! Bxc3+ 6.bxc3 c5 7.cxd5 exd5 8.Bd3 0–0 9.Ne2.

This settled in theory as the Botvinnik variation of the Rubinstein system. Today it is well known that if Black defines his pawn structure in the center prematurely, it provides White with better chances to seize the initiative. But in those days both opponents were wandering in ’terra incognita’ – the routes that were to develop in the Nimzo-Indian Defense were just being formed.

9…b6
(Keres’ idea; 9…Nc6 10.0–0 Re8 11.f3! was also played)

10.0–0 Ba6.
“The d3-bishop is Black’s main headache, and he is striving for the bishop exchange” (Botvinnik).

By the standards that dominated the minds of chess players in the 1920s and 1930s, Black has absolutely nothing to complain about. Good pawn structure, comfortable development, and no tangible weaknesses. White has no concrete threats. However, Botvinnik foresaw that after f2–f3 and e3-e4-e5 White could get an powerful attack on the kingside. Hard to believe!? Well, at least you are thinking along the same lines as the great Capa…

11.Bxa6.
Subsequent experience has shown that the immediate 11.f3 is more accurate – after 11…Bxd3 12.Qxd3 Re8 13.Ng3 the white queen gets to d3 anyway, and the knight b8 has yet to come out. For example: 13...Nc6 14.Bb2 (14.Ra2!? Rc8 15.Re2) 14…c4 15.Qd2 Qd7?! (sharper is 15…b5 16.Rae1 a5 17.e4 b4! 18.e5 Nd7 19.f4!) 16.Rae1 h5?! 17.e4 g6 18.Bc1! Nh7 19.Qh6! Re6 20.f4! Ne7 21.f5 with a menacing attack (Kasparov – Ivanovic, Niksic 1983).

11...Nxa6 12.Bb2?!
Alas, pioneers are doomed to mistakes: 12.Qd3! was necessary – move order is always important for a deep strategic plan. However, it is so easy to gradually, move by move, improve old lines (as is so often done nowadays), and so much more difficult to invent a new long-term strategy!

12...Qd7 13.a4
(forced: now 13.Qd3 was not good in view of 13...Qa4! sealing White’s queenside)

13...Rfe8?!
“A strange mistake for Capablanca!” (Botvinnik). Indeed, after the obvious 13...cxd4 14.cxd4 Rfc8 followed by Rc4 and Rac8 White would have to forget about the breakthrough in the center and switch over to the defense.

14.Qd3 c4?!
“This is a serious positional mistake. Black must have thought that White would not be able to advance the e-pawn and Black’s advantage on the queenside would tell. Capablanca counted on the Na6–b8–c6–a5–b3 maneuver, as it is hard to defend the a4-pawn after that. However, the e3-e4 breakthrough turned out to be highly effective. It was both sufficient and necessary to play 14…Qb7”. (Botvinnik)

Capablanca’s decision has a clear historical background: the 50-year-old ex-World Champion in his long chess career had never come across the term “hidden dynamic factors”. So his unique intuition and rich experience failed to send any danger signals. Capa wanted to win and launched, as he later said, “the fight of two intellects”:

15.Qc2 Nb8 16.Rae1
(this is braver than 16.Ba3 Nc6 17.Bb4=; 16.Ng3!?)

16…Nc6
(16…Nh5!?)

17.Ng3 Na5.
If 17...Ne4 then 18.Nh1! — and White’s plan works out anyway, for example: 18...f5 19.f3 Nd6 20.Ba3 g6 21.Ng3 and after the inevitable e3–e4 the weakened position of the king is a source permanent headache for Black.

18.f3 Nb3
(both sides are quite happy: the a4-pawn is lost, but the long-awaited central advance is also ready)

19.e4 Qxa4.
This miserable pawn has played the important role of bait for the shark!

20.e5.
Interestingly enough, this very position, but with the bishop on a3, emerged in the game Botvinnik — Alexander (radio match USSR — Great Britain, 1946) and after 20.Qb2 a5 21.e5 b5 22.Bd6 (22.exf6! b4 23.Qf2 is even better) 22…Re6 23.exf6! Rxd6 24.fxg7 b4 25.Re5! Rae8 26.f4! Qd7 27.Qe2 Rde6 28.f5 Rxe5 29.dxe5 White was clearly better.

20…Nd7
(but not 20…Nc5? 21.Re2!)

One way or another, the clouds are growing thicker over the black king: the white knight aspires to f5 and d6, and besides, there is the threat of a devastating f-pawn march. But Capablanca could not be frightened by these shadows on the horizon: the staunch supporter of economy in chess was always famous for his amazing ability to defend with minor forces while the core of the army was engaged elsewhere.

21...g6 22.f4 f5!
(forcing the opening of the e-file and thus an exchange of rooks) 23.exf6 (there is no other way to reach the king)

23…Nxf6 24.f5 Rxe1 25.Rxe1 Re8!
A tactical trick (indirect defense of the f6-knight: 26.fxg6 hxg6 27.Rxe8+ Nxe8) lets Capa reduce White’s attacking potential. In case of 25...Rf8? 26.Qf4! White’s attack is unstoppable: 26...Qa2 (26…Qd7 27.Re6! etc.) 27.fxg6! Qxb2 (27...hxg6 28.Qg5) 28.g7! Kxg7 29.Nf5+ Kh8 30.Qd6 Rf7 (30...Kg8 31.Qg3+) 31.Qxf6+ mating.

26.Re6! (the only chance)

26…Rxe6.
There is no other way: 26...Kg7? 27.Rxf6! Kxf6 28.fxg6+ Kxg6 (28...Ke7 29.Qf7+ Kd8 30.g7) 29.Qf5+ Kg7 30.Nh5+ Kh6 31.h4! Rg8 32.g4 Qc6 33.Ba3! with mate to follow.

27.fxe6
(it was only by a miracle that the modest f-pawn wasn’t consumed in the flame of attack, and now it is having a dream career!)

27…Kg7 28.Qf4 Qe8
(the queen must return: 28...Qa2? 29.Nf5+ with mate in five - 29...gxf5 30.Qg5+ etc.)

29.Qe5.
Culmination of the historical battle. Which of the sleeping pieces from this forlorn flank will come round first – the black b3-knight or the white b2-bishop?

29…Qe7?
“Meeting White’s aspirations, but Qe7 was inevitable anyway. For example, after 29…Na5 30.Bc1! (with the threat 31.Bh6+ or 31.Qc7+ and 32.Bh6) Black is forced to play 30…Qe7, and White could carry out the same combination as in the game” (Botvinnik). But fifty years later the magazine “Chess in the USSR” (#3/1988) published V.Goldin’s analysis. He found the only defense — 29...h6! For example:
1) 30.h4 Na5! 31.Bc1! Qe7 32.Bg5! hxg5! (after 32...Nc6 33.Bxf6+ Qxf6 34.Qxd5 Qxh4 35.Qd7+ Ne7 36.d5 Kf6 37.d6 Qxg3 38.Qxe7+ Ke5 White strikes a decisive blow 39.Qg7+!) 33.hxg5 Nc6 34.gxf6+ Qxf6 35.Qxd5 Ne7 36.Qd7 (36.Qxc4 Qf4! 37.Ne2 Qf5=) 36…Qg5! 37.Kf1 (or 37.d5 Qe3+ 38.Kf1 Qf4+ 39.Ke2 Qe5+) 37...Kf6 38.Qxa7 b5 39.Qd7 Qd5=;

2) 30.Ba3 Qd8 31.Qf4! (31.h4 Na5! 32.h5 Nc6 and Black’s position is sound enough) 31…b5 (31…Na5?! 32.Bc1!; 31…Qc8? 32.Qd6!) 32.h4 a5 33.Qe5! b4 34.cxb4 c3! (34…Qe7 35.h5! axb4 36.hxg6!+–) 35.b5! c2 36.Ne2 Nd2! 37.b6 Nc4 38.Qc7+ (the computer also hints at 38.Qf4!? Nxb6 39.Bc1 Kg8 40.Qxh6 Qe8 41.Nf4 Nc4 42.Qg5 Ne4 43.Qxd5 Ned6 44.Qg5 with a significant advantage. — G.K.) 38…Qxc7 39.bxc7 Nb6 40.Nc3! Nc8! (40…g5 41.h5! Nxh5 42.Nxd5+–) 41.Kf2 (41.g4 Nxg4 42.Nxd5 Nf6 43.Nxf6 Kxf6 44.d5 Ke5 45.e7 Nxe7 46.d6 c1Q+ 47.Bxc1 Kxd6 48.Ba3+ Kxc7 49.Bxe7 Kd7 50.Bf8 Ke6 51.Bxh6 Kf5 52.Kf2 Kg4 53.Bg5 a4=) 41…g5 42.h5!? (42.hxg5 hxg5 43.Ke2 Kg6 44.Kd2 Kf5 45.e7 Ke6 46.Kxc2 Kd7 47.Bc5 Kc6 48.Kb3 Ne8 49.Ka4 Nxc7 50.Nb5 Ne8 51.Kxa5 Kd7=. — G.K.) 42...Nxh5 43.Nxd5 Kg6 44.Ke3 Ng7 45.e7 Kf7 46.Kd2 Ne8 47.Kxc2 Ke6=, “and Black captures the c7-pawn”.

However, a-man-plus-computer tandem immediately refutes 46...Ne8? by 47.Bc5! Ke6 48.Nb6 Nxb6 49.Bxb6 Kd7 50.Bxa5 h5 51.Kxc2 h4 52.Kd3 g4 53.Ke2+–. The only remaining way is 46…Ke6, but even here life is not that easy for Black: 47.Ne3 Ne8 48.d5+ Kd7 49.Nc4 Kxc7 50.d6+ Nexd6 51.Bxd6+ Kd7 52.Kxc2 Nxd6 53.Nxd6 Kxe7 54.Nf5+ Kf6 55.Nxh6 Kg6 56.Ng4 Kf5 57.Ne3+ Ke4 58.Nf1 a4 59.Kc3 a3 60.Kb3 Kd3 61.Ng3 Ke3 62.Nh5, saving the g-pawn and winning.

30...Qxa3
(30…Qe8 31.Qc7+ Kg8 32.Be7 Ng4 33.Qd7 etc.)

31.Nh5+! gxh5. Or 31…Kh6 32.Nxf6 Qc1+ 33.Kf2 Qd2+ 34.Kg3 Qxc3+ 35.Kh4 Qxd4+ 36.Ng4+! 32.Qg5+ Kf8 33.Qxf6+ Kg8.

Mikhail Moiseevich recalled: “While I am sitting there thinking about the most convincing move order, Capablanca saunters along the scene and seems to be quite cool-headed. Euwe comes up to him: “So?” and Capa with an expressive gesture shows that anything’s possible, obviously hoping that I am watching this conversation. A brilliant practitioner, he used this last psychological chance: trying to convince his exhausted opponent that the position was still unclear, — what if my tension results in some accidental mistake?”

34.e7.
34.Qf7+ Kh8 35.g3! is also winning (White’s king hides on h3), but not 35.e7? Qc1+ 36.Kf2 Qd2+ 37.Kg3 Qg5+ 38.Kf3 Nxd4+! 39.cxd4 Qg4+ with a perpetual.

34...Qc1+ 35.Kf2 Qc2+ 36.Kg3 Qd3+ 37.Kh4 Qe4+.
“Capa answered at once — I was to realize that my opponent was absolutely sure of a favorable outcome of the game. Leaving his queen on c2, d3, and e4, he hoped that White would avoid the queen exchange in the time-trouble and agree to the perpetual”. (Botvinnik)

38.Kxh5 Qe2+.
38...Qg6+ doesn’t save the game in view of 39.Qxg6+ hxg6+ 40.Kxg6 and 41.e8Q(R)#.

39.Kh4 Qe4+ 40.g4
(simpler than 40.Kh3)

40…Qe1+ 41.Kh5.
There are no more checks, and Black resigns. The hobbled knight on b3 is a mute reproach to Black’s plan, launched still in the middlegame. The hall burst into a storm of applause; this is a rare phenomenon, as before it was only Euwe who was honored with such a reaction from the Dutch audience…

It is clear that after 20...Nd7 Black is almost lost, having the Queen, Ra8 and Nb3 out of play, and no serious protection for King and castle , while White keeps its Rooks already well placed, a strong e5 advanced Pawn, and Ng3 + Queen ready for the attack.

Instead of 21. Qf2?! White could play the stronger 21. Nf5!, and it is difficult for Black to find a defense.

If a) 21... Nbc5; 22. Qd2! Nd3 (22...Nb3 23. Qg5 g6 24. Ne7+ and Nxd5 etc.) 23. Ra1, Qc6 (23...Qb3? 24. Rfb1! and Black’s Queen + Re8 have serious troubles) 24. Ba3! and Black has no defense (if 24...Qc7 25. Qg5, g6 26. Ne7+, Kh8 27. Nxd5 Qd8 28. Be7!, Qc8 29. Bf6+ etc. if 24...Kh8 25. Nd6 Rf8 26. g4! followed by f4, f5 and White wins. if 24…Qe6 25. Nd6, Rf8 26. Nb5! Rfc8 27. Bd6 + - ).

If b) 21...g6 22. Nd6, Re6 (22...Re7 23. Qf2! followed by 24. Qh4 gains a tempo 22...Nbc5 23. Qd2! Nd3 24. Ra1! Qb3 25. Rfb1 ) 23. g4! Rf8 (23...Nbc5 24. Qd2, Nb3 25. Qg2!, Raf8 26. f4! and Black's position collapses) 24. Ra1! Qc6 25. Rxa7 followed by f4-f5 and White wins.

There could be several other sub-variations, but plan is always the same, after 21. Nf5! Black has no viable defense, while White is pushing forward f4 and g4 pawns, and crushes Black's position.

21.Qf2?! is too slow, allowing 21...g6!, and slowing down White's attack, whereas after 21.Nf5! Black cannot send away the strong Knight by pushing g6, without fatally weakening its castle.

Of course no doubt that 31. Ba3!! was a great and correct combination, and we have to admit that it was very difficult to calculate everything during a real game (chapeau Botvinnik!).

But the game is not at all a perfect one even from White side, as White could win before, may be this is the reason why Bobby Fischer was reluctant to indicate this game as the best among the ones played by Botvinnik.

If a) 21... Nbc5; 22. Qd2! Nd3 (22...Nb3 23. Qg5 g6 24. Ne7+ and Nxd5 etc.) ...>

If Black plays 24...g6 25. f5 is devastating too.

<If b) 21...g6 22. Nd6, Re6 (22...Re7 23. Qf2! followed by 24. Qh4 gains a tempo 22...Nbc5 23. Qd2! Nd3 24. Ra1! Qb3 25. Rfb1 ) 23. g4! Rf8 (23...Nbc5 24. Qd2, Nb3 25. Qg2!, Raf8 26. f4! and Black's position collapses) 24. Ra1! Qc6 25. Rxa7 followed by f4-f5 and White wins.>

Let's look at the following variation: 21.Nf5 g6 22.Nd6 Re6 23.g4 Nbc5 24.Qd2 Nd3 25.Ra1 Qc6 ...

click for larger view

(I think I am just using you moves, especially for White, though I do suggest the Nc5-d3 and Qa4-c6 for Black.) White to move.

The Black position in the diagram seems quite defensible and I am not sure that White has enough for the pawn.

---

Btw, the Nd6 can become a liability. Here is an arguable, but fun variation:

26.Ba6 N7xe5!? 27.dxe5 Rxe5 28.Rab1...

White is now fighting to either extricate the stranded knight or at least minimize the material damage.

28...a6 29.Rfd1 b5 30.f4 Re6 31.f5 Rxd6 32.Bxd6 Qxd6...

click for larger view

May be Black could try 26...Nf6!? but after 27. h3! the only way for Black to prevent 28. f5 is to sac the exchange for two pawns 27...Rxd6 28. exd6, Ne4 29. Qg2 Nxb2 30. Qxb2, Qxd6 31. Rf3! and it seems now that the two Rooks and the attack on f file give White a clear advantage. I checked also the exchange sac before, 26...Rxd6!? but conclusion seems very similar, Black is helpless against the attack on f-g files, and the two Rooks (pawn a7 is weak too), the two pawns are not a compensation enough, in such a position.

If Black can try - after Nf5/Nd6 to slowing a bit White's attack on f-g files just through an exchange sac - but without a true compensation, as White's attack is keeping on anyway - now this seems to me the evidence that 21. Nf5! is very much stronger than 21. Qf2.

If you think that Capa - as pointed out above - in the real game missed the strong defence 29...h6!, with good drawing chances, now you have to admit that 21.Qf2?! can be a weak and slow move, and White had a better plan.

Personally, I still like the iron logic of Botvinnik's 21.Qf2. Also, I think that the possibilities of the Black position should not be underestimated. For the alternative,

21.Nf5 g6 22.Nd6 Re6 23.g4 Nbc5 24.Qd2 Nd3 25.Ra1 Qc6 <26.f4>

I indeed have been thinking of the x-change swap

26...Rxe6 27.exd6 Qxd6

click for larger view

Although nominally equal in material, with Black having N+2P for the rook, I think that Black can look forth with optimism: The extra pawns on the Q-side are free to move up the board, Black has the squares e4 and d3 as a basis of operation for his pieces, especially his knights, and the Nd3 promises to be enough pain that White may consider returning the x-change.

After 28. f5 may be Black could try a) 28...Nf6 or b) 28...Nxb2. If a) 28...Nf6 29. Ba3, Qd7 (to avoid 30. fxg6) 30. h3 is very bad for Black, in view of Qg5 or Qh6

If b) 28...Nxb2 (probably better)29. Qxb2, Nf6 30. fxg6, fxg6 31. Qe2, Ne4 (31...Re8? 32. Rxf6! ) 32. Qe3 with the idea Rf4 and Raf1 it seems White still keeps a strong attack against f file and Black castle (Qe3 can go to h6 etc.)

You said about the "iron" logic of Botvinnik (anyway, in a real game, even Botvinnik sometimes made mistakes), but the question is: are we sure that in such position White can hesitate and allow 21...g6 after 21. Qf2? If the analysis from Goldin is right, and 29...h6! could draw the game, it is strange that White had no better ways to change its advantage into a win, and probably 21.Nf5 was a more active move.

I personally like very much more the game Lilienthal - Capablanca, should I indicate a game in which White exploited with no hesitation the uncommon mistakes in Capa's career...