Author: Uri Blass
Date: 08:26:48 01/02/05
Go up one level in this thread
On January 02, 2005 at 10:51:24, Robert Hyatt wrote: >On January 01, 2005 at 12:21:28, James T. Walker wrote: > >>On January 01, 2005 at 12:10:53, Matthias Gemuh wrote: >> >>> >>> >>> >>>> The >>>>problem is that pondering is part of the program. If you are trying to test >>>>which is best at playing chess then cripling all programs is not necessarily >>>>cripling them equally. >>>> >>> >>> >>>According to this logic, it is not allowed to test with Ponder=ON because >>>some programs may be crippled (i.e. those that score better with Ponder=OFF). >>>Kurt's tests are wonderful and much appreciated ! >>> >>>Matthias. >> >>Please allow me to download one program that scores better with ponder off. >>Where can I get it? If it does it has serious problems but I would like to test >>it anyway. >>Jim > > >Crafty certainly used to play significantly worse in non-pondering mode. More >than one person sent me sample games showing that its time usage was not very >good, and it was adjusted to make it better. Problem is, nobody working on a >chess program tests it in all the "odd modes" thoroughly, and I consider >ponder=off an "odd mode" since nobody would use that mode in a serious game... I think that you cannot generalize here. There are people who test mainly with ponder off and there are people who test mainly with ponder on. ponder off is not an odd mode if you have one cpu and want to test changes that are not changes in the time management. In ponder on games with one processor part of the time is not used effectively because programs ponder about moves that they do not play. In ponder off games all the time is used effectively so there is an advantage for ponder off if you want to test some change at longer time control.
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.