Author: Harald Faber
Date: 08:32:44 05/01/00
Go up one level in this thread
On May 01, 2000 at 11:26:48, Ernst Walet wrote: >On May 01, 2000 at 02:59:08, blass uri wrote: > >>On May 01, 2000 at 01:54:30, Ernst Walet wrote: >> >>>On April 30, 2000 at 10:46:55, Harald Faber wrote: >>> >>>>This position is a nice test of endgame knowledge for programs: >>>> >>>>Which program(s) avoid 1.Bxh7?? g5+!! 2.hxg6+ Kg7 3.Rg8+ Kh6 and draw, white has >>>>no chance to enter the black fortress. >>>> >>>>[D]7R/5kpp/8/7P/4BK2/8/8/3r4 w - - am Be4h7;bm Kf4e5 Kf4g3;id >>>>Sax-Kovacevic,Sarajewo 1982; >>> >>>This is what Fritz6a analyses overnight after making the moves above. >>> >>>[D]6R1/7B/6Pk/8/5K2/8/8/3r4 w - - 0 1 >>> >>>Still it sees a win for white. Can you line out how the draw is achieved? >>> >>>Ernst. >> >>Fritz does not see a win for white. >>The best score for white was 4.91 after 12 seconds. >>I guess that Fritz cannot win against itself and it is a proof that fritz does >>not see a win for white and that Fritz can tell you how the draw is achieved. >> >>If I am wrong then please post the win for white when you give Fritz to play >>against itself at tournament time control. >> >>Uri > >I assumed a win because of the score. > >Ernst. This is always the problem: People look at the evaluation instead of the board... Like written here thousands of times before: the eval is not God's word, it exists to help the program find the best move.
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.