Author: Ricardo Gibert
Date: 20:07:03 01/07/00
Go up one level in this thread
On January 07, 2000 at 20:38:22, Tim Mirabile wrote: >Hard to say. Endgames like this require bits of very specific knowledge plus >knowledge of when to look out for exceptions. As a human I can chunk the >position and evaluate the win in a few seconds - i.e.: Nxe3 Kxe3 and now > >- passed pawns separated by one file tie down the black king and can be stopped >but not captured > >- connected passed pawns will tie down the white king and can be stopped but not >captured > >- the healthy pawn majority on the g+h files will decide the game because three >connected passed pawns can promote against the enemy king without help Not true, but here they _do_ get help in the form of a temporizing move by black. Example: 1...Nxe3 2.Kxe3 Kc6 3.a5 g6 4.Kf2 h6 5.Ke3 g5 6.h5 g4 7.Kf2 and black would be in zugswang if not for the availability of the temporizing move Kc5. It is important to note that if 6.hxg5 hxg5 7.Kf2 g4 black would win without the need to temporize, so in a slightly different position, the choice of exchanging or not could be decisive. You're right that the position can be evaluated quickly by a human, but you must be aware of all of the subtleties to do it successfully. You got lucky. Right evaluation, but your "chunking" was off. Summing up: -Black stops whites pawns (K vs 2P) and _can_ tempo with his king. -White "stops" blacks pawns (K vs 3P) and _cannot_ tempo with his king. Therefore, black wins. BTW, there is a trick worth noting: [D]8/1k6/1P6/P1P5/6p1/4p3/4Kp2/8 b - - 0 0 1...f1Q 2.Kxf1 g3 black wins. If 1...g3 2.Kf1 white wins. But if... [D]8/1k6/1P6/P1P3p1/4p3/4Kp2/8/8 b - - 0 0 and here the analogous trick does not help e.g. 1...f2 2.Kxf2 g4 3. Kf1 temporizing for the win. As you can see, chunking is fraught with pitfalls and one must have a lot of knowledge and experience to do it correctly and even then a mistake can be made. Berliner himself trips himself up in his book where he talks about chunking. So you are in good company. Chunking should be checked by concrete calculation to be safe. > >And now the exception: > >> 11 285 33.50 2395723 Nxe3 Kxe3 Kd6 a5 Kc6 Kf2 h6 a6 Kb6 c5 Kxa6 c6 Kb6 > >In this PV instead of ...a6, ...h5 draws, as the pawn majority becomes defective >(backward pawn). Of course, g6 instead of h6 wins. > >On January 07, 2000 at 20:15:53, Will Singleton wrote: > >>In a recent ICC game (PostModernist - Amateur), black can easily win with Nxe3. >>My program wanted to play f2 (and did), leading to a draw. >> >>I wonder if there's a way to quickly figure out that getting rid of the knights >>is better here. >> >>[D]8/6pp/4k3/8/P1PKp2P/4Np2/6n1/8 b - -
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.