Not to quibble, but please note, for future reference, that '...' before a move indicates that the move was made by Black. '58.Ke1' would mean that White played Ke1, whereas '58...Ke1' would indicate that Black had played Ke1. Thus the above should read "if now 58.Ke1" and "which leaves only 58.Kf2 Kf4 etc."

Regrettably, and unusually, there is a misleading error in <Once>'s post. After his second diagram, he states that Black can win by either ...Kg3 or ...Ke3, after White retreats from f2. Not if W retreats to f1: then only Kg3 will work, and it only works because of the possibility of capturing on h3. Without the h-pawns on the board, Black can triangulate forever and never be able to promote the f-pawn.

Now it is the same position from the diagram, with White to play and lose, as described above. Funny, but I struggled with this until I realized that Black could easily triangulate, i.e. move the king over three squares while White shifts his king between two squares, thus "losing" a move.

JfP, thanks for your post. We should also point out the black cannot win if the h-pawns are on h2 and h3, instead of h3 and h4. Only if this is known can one understand Kasporov's maneuvering between moves 42 and 45, which force Alburt to advance 46.h3 .

I don't know about the whole game, but White's last chance to draw might have been 46.g3, instead of h3. The position at this point is way too difficult for me to solve, but maybe others can do it, with an engine if necessary?

Moves other than 53...Ke5 and ...Kd5 are wasted moves. They don't throw away the win, but they only win after B discovers the correct plan! (I had a typo in my first attempt, and deleted it. In case you saw e4 and d4, instead of e5 and d5)

Did Kasparov know the ending was winning after he forced the Qs to come off? He must have thought so, at least. Awesome.

SuperPatzer77

After all that, I still had errors: I meant 53...Ke5 and ...Kf5, of course.

http://www.365chess.com/tournaments...

Game Collection: USSR First League, Ashkhabad, 1978

The concept is broad, and it's called coordinate square theory. As I said, this is a fairly simple example of it, and triangulation is sufficient to win it. But there are some fascinating studies in which mapping is really essential: With more pawns, one often has to remember, e.g., that if the defender's king moves to c8, the attacker must move his to f5 to preserve the win.

http://en.wikipedia.org/wiki/Lev_Al...

Thanks for the link - I either never saw it, or completely forgot it.

kingscrusher: There seems to be a great improvement to 12...Qa5 in the form of Nd4

click for larger view

So it implies a potential double pawn sac :

Bxc5

And we have Nxe2+ with Qc8 in mind :)

See something that violates our rules? Blow the whistle and inform a moderator.