26.Qd4?! (26.d6!? White's a bit better)

26...Rbc8?! (26...e3!? with some advantage), now 27.Bd1!? led to a serious advantage for White, but in the game 27.h3?! led it to Black.

(Zeitnot stage?!) 34.g3? losing move, but Black missed 34...Rf8!?

In fact White was still in big trouble after 37.Rb1?! (37.Nc5!?), but <Hesam7> 43...Ra8? (after zeitnot) slipped away every thing.

"After the opening, I created a maze of complications which, if White played correctly, could have got me into deep trouble. But Karpov failed to find the right plan, and by a series of tactical strikes I achieved a won position. The, under time pressure, I overlooked the continuation that would have clinched it. All the same, from the position we adjourned at I was counting on a win, though my sealed move turned out to be second rate. Having calculated that however White responded, I would have a clear win, we suddenly discovered, just two hours before the game was due to resume, that White in fact had a strong answer. We began frantically searching for other possibilities, but didn't have enough time to find a precise solution.

Returning to the game without any clear idea of how to continue it, I was unable to find my way out of the situation and so lost the initiative. As a result, we drew.

click for larger view

<Kasparov writes that this endgame is objectively a draw, but that White will have to play very carefully to obtain it.<>>

Black should try to penetrate with his rook but must also do so without allowing too many pawn exchanges. However, I think White should, without too much difficulty, be able to prevent Black's rook from coming in.

<48.Bc3 Rb8 49.Bb4 Rd8 50.Ke2 a3 51.Bc3 f6 52.Bb4 Kf7?!>

"Kasparov criticizes this move as the king is poorly placed here later (the possibility of a knight check on d6 indirectly protects the pawn on f5). Instead he suggests 52...Kh7 with the idea that if White tries to use the same defensive idea as in the game, he will lose the kingside pawns and probably the game as well. 53.Bc3 Rb8 54.Bb4? (54.Nb4 Rb5 55.g4 Rb8 56.Kd3 Ra8 57.Na2 Ra4 58.Bb4 Kh6 59.Bd2+ and White has built up an impregnable fortress.) 54...Rb5! 55.g4 Rb8 56.Kd3 Rd8+ 57.Kc2 hxg4 58.hxg4 Rd4 59.Bxa3 Ra4 60.Kb3 Rxg4 61.Bc1 Rg3+ 62.Nc3 Rf3 63.Kc2 Rxf5 64.Kd3.

Kasparov writes that it is difficult to assess this endgame as there is no relevant material to compare it with, but Black would probably have excellent practical winning chances. I managed to find one position from the World Championship match between Steinitz and Zukertort - see next game." (Esben Lund)

<53.Nc3 Rb8 54.Na2 Rb5 55.g4 Rb8 56.Kd3 Rd8+ 57.Kc4 Rd1 58.Bxa3 Ra1 59.Kb3 Rh1 60.gxh5 Rxh3+ 61.Nc3 Rf3>

The problem with the king on f7 becomes evident in the line 61...Rxh5 62.Ne4 Rxf5? 63.Nd6+.

<62.Bc1 Rxf5 63.h6 g6 64.Ne4 Rh5 65.Bb2> 1/2

Lund continues with the Steinitz-Zukertort fragment and concludes that the rook and pawns most likely will win, and then returns to the position with after White's 64th move from Kasparovs analysis of <52...Kh7>.

However, his conclusion is confusing. First he writes: <The conclusion is that, instead of <52...Kf7?!>, <52...Kh7> probably leads to a win for Black with correct play.<>>, and then a few lines further down: <To recap, the game would still be a draw with precise play, but after <52...Kh7> Karpov would have to use a different defensive plan than in the game, as indicated in the notes - see the line beginning with <54.Nb4!>.<>>

Well, I admit, a lot of detail to pinpoint a minor mistake, since all in all Lund's lucid explanations are thorough and easy to understand.