|Jun-07-10|| ||ozmikey: In Bronstein's book it says that Black can't play 16...Qe7 because of 17. e5, but what happens after 17...Bxe5 18. Re1 Rb5? Sure, Black's rook looks a bit weird, but it looks more or less OK for him. Certainly better than surrendering a pawn for nothing...|
|Jun-07-10|| ||vonKrolock: <16...♕e7 17. e5, but what happens after 17...♗xe5 18. ♖e1 ♖b5?>|
click for larger view
Then 19.♗c6! ♖a5 ▢ 20.a3!, and black can not play bxa3 because of b2-b4! etc 20...b3 or 20...♕d6 seems to be better candidate moves, but white will win at least the exchange in the sequence
|Jun-07-10|| ||ozmikey: Interesting, thanks. I wonder whether Black can play (after 20. a3) 20...bxa3 21. b4 Bf5!?|
|Jun-07-10|| ||An Englishman: Good Evening: My, this brings back happy memories of childhood, studying this very Knight ending. The last 30 moves of the game are quite remarkable. Stahlberg thought the 4:3 ending could be drawn as it would be with Rook vs. Rook. Indeed, Taimanov avoided the Knight exchange for the same reason.|
The entire last phase revolves around the attack and defense of the f7 square (and not so much the pawn that sits on it, interestingly). White's entire plan consists of clearing a path for his f-pawn, and after all these years, it's still fun to watch him at work.
Just out of curiosity, has anyone ever turned this endgame over to the silicon monsters for their final verdict?
|Feb-21-15|| ||Howard: In response to the most recent comment, computer analysis has revealed that Taimanov could have saved himself quite a bit of trouble with 33.Nxa6 ! It would have won rather easily, in fact.|
|Feb-21-15|| ||keypusher: <An Englishman> <Howard: In response to the most recent comment, computer analysis has revealed that Taimanov could have saved himself quite a bit of trouble with 33.Nxa6 ! It would have won rather easily, in fact.>|
Yes, but it is not so easy to see why. Both Bronstein and Najdorf, in their tournament books, recommended 33.Rc5, on the grounds that it would have restricted Black's knight and helped preserve the a-pawn. An engine confirms that 33.Rc5 would have been stronger than Taimanov's 33.Ra5.
Taimanov, Bronstein, and Najdorf probably rejected 33.Nxa6 because it forces the exchange of the last set of queenside pawns, which is the last thing White wants: 33.Nxa6 Ra8 34.Nc7 Rxa3. But now 34.Rd8+ Kg7 35.Ne8+ Kh6 (35....Kf8 36.Nf6+ Kg7 37.Rg8+ Kh6 38.g4 and Black will get mated) 36.Nf6 and Black is forced to seek the exchange of rooks: 36....Ra8 37.Rxa8 Nxa8 38.Nd5 and the Black knight has no moves!
click for larger view
This line is so hopeless for Black that the engine prefers 33....Rc8, just going two pawns down with a lost position. But the line is hard for a human being to see, which is why three GMs missed it.