Author: Guido
Date: 02:47:00 03/11/02
Go up one level in this thread
On March 11, 2002 at 02:07:47, Uri Blass wrote: >On March 10, 2002 at 19:00:14, Guido wrote: > >>On March 10, 2002 at 13:08:08, Uri Blass wrote: >> >>>On March 10, 2002 at 12:01:21, Guido wrote: >>> >>>>Some time ago I discussed on CCC about this subject with Bob Hyatt and he very >>>>kindly demonstrated to me that the right solution to this problem consisted in >>>>the duplication of the positions where a pawn can be ep-captured, so a good >>>>solution for this goal can be the use of the first (a1...h1) and the last rank >>>>(a8...h8) of the chessboard, where pawns are not normally allowed. >>>> >>>>For the moment in my program I preferred not to adopt this solution, generating >>>>tablebases without ep capture, so my present question is: >>>> >>>>Does someone know if exist positions in the following endings: KPKP, KXPKP (X = >>>>Q, R, B, N), KPPKP where results are different if ep capture is or not >>>>considered? >>> >>>yes there are >>> >>>[D]1K6/2p5/8/1P6/8/8/8/k7 b - - 0 1 >>> >>>c5 is a draw when white cannot reply bxc6 >>> >>>Uri >> >>This case IMHO is not an example as the Black wins in 23 moves with Kb7 or Ka7, >>but surely there are other positions, for example with 5 pieces, where Black has >>to move the pawn because black King is blocked by white King and the other white >>piece. > > >I do not understand >black cannot play Kb7 because the black king is not close to b7 > >The point is that if black plays c5 the only way for white to win is by en >passant capture so if the special tablebases assume that ep is impossible they >will see that this position is a draw and be wrong because white is winning by >the chess rules. > >They will not cause a mistake in this position because black is losing but they >may cause a mistake in previous positions when transition to this position >happens. > >Uri Oops! You are right! I exchanged the positions of the Kings :-(. Now, with ep, Black loses (and White wins) in 12 moves. So this case and other similar demonstrate that it is necessary to consider ep capture in generating tablebases. I'll have to think about because it seems not so easy to keep it into account in my program. Thanks Guido
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.