Author: Richard Pijl
Date: 05:21:48 12/05/02
Go up one level in this thread
On December 05, 2002 at 08:03:52, Jouni Uski wrote: >On December 05, 2002 at 07:51:02, Gian-Carlo Pascutto wrote: > >>On December 05, 2002 at 07:19:51, Jouni Uski wrote: >> >>>Probably, but no excuse for pro program! >> >>Bullsh*t. >> >>AFAIK Shredder normally puts out a BIG warning when using >>the autoplayer that you should never run with incomplete >>tablebases. >> >>So effectively the tester in question must have knowningly >>run it knowing it was crippled. Very bad points for SSDF >>here. >> >>Some programs have a workaround that makes the ugliest >>appearances of this problem go away. But that doesn't >>fix the problem - it still occurs everywhere else in the >>search tree. I think it's better to warn the user there >>is a problem than to hide it. >> >>-- >>GCP > >I disagree. Program should handle this queen promotion without 8 GB of TBs! >What a waste of data storage. > >Jouni If you (well, Tony in this case) is using incomplete TB's, you will have to know which ones you need to have. For instance, you need to have all of the TB's dealing with promotion of pawns. Normally the queen ones are enough, but there are numerous examples where that fails as well. So using kbpkb requires that you have kbqkb and the others (kbrkb, kbnkb and kbbkb as well, just as you will need the 4 stone TB's to handle the captures. This is a well known problem of EGTB's. I consider this an error that is about as big as using 128Mb hash tables on an 128Mb machine. Most professional programs will warn you about it but allow you to do it anyway if you insist. This raises the question: Is the EGTB set used the same on all machines that SSDF use to test programs? Richard.
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.