Author: Uri Blass
Date: 17:47:28 12/30/05
Go up one level in this thread
On December 30, 2005 at 20:29:44, Uri Blass wrote: >On December 30, 2005 at 19:16:07, Ernest Bonnem wrote: > >>Who can give an answer? >>Vas? >>Ernst? >>anybody else knows? > >I think that it is trivial. > >It is dangerous to play with an untested version. > > >Rybka beta was tested heavily by the CEGT team and other people. >Rybka Preview is a new version and still was not tested enough. > >A more interesting question is why the opponents resigned against rybka and did >not continue until mate because it is known that rybka has a bug that may cause >it problems to win won positions because it does not care about distance to >mate. > >The bug was no problem in CEGT because the interface adjudicates the result >based on evaluations but it could be a problem in real games. > >Uri To be more specific The question is if the opponents decided that they do not like to try to take advantage of rybka's bugs because they felt it is not sportivic to get points in that way or that there was a rule that games can be adjudicated based on evaluation of both programs or evaluation of humans. Note that I remember a case when xinix lost on time a game when it had 3 queens against a pawn because it did not care about the fastest mate and in that case the opponent did not want to resign so resigning is not obvious even when you do not know that the opponent has a bug. 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.