Author: Uri Blass
Date: 12:20:08 09/22/03
Go up one level in this thread
On September 22, 2003 at 14:16:15, Pete Waring wrote: >On September 22, 2003 at 14:00:26, Drexel,Michael wrote: > >>Fritz,Ruffian and Goliath evaluate a repetition of moves always as 0.00, >>the others not. >> >>Michael > >Thanks Michael. I don't understand why an evaluation of 0.00 for this would >ever make any sense, but that was definitely what was happening. > >Pete evaluation of of 0.00 make sense because most of the programmers of chess engines do not care about blunder check and it was only a feature that was added by the people who designed the GUI. From the engine point of view 0.00 saves time because the engine does not need to search after repetition and usually repeptition is a bad idea when the engine has a better position because the engine does not blunder. There are possible 2 solutions: 1)to search first repetition and not to return 0.00 but in this case you will often search additional nodes for nothing because if the engine evaluates the position as better and did not blunder then it is clear that repeating the position is not going to help and in most cases engines do not blunder. 2)to return something different and not 0.00 and in this case there is a problem what to return and if you return the evaluation that you remember than the program may prefer to play repetition that does not make sense when it see no way to progress its evaluation and it is possible that the best thing is to lose a pawn and avoid repetition because if you do not do it next move you have no choice and need to do second repetition that is a draw. Uri
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.