Author: Tony Hedlund
Date: 10:05:22 12/06/02
Go up one level in this thread
On December 06, 2002 at 00:53:51, Harald Faber wrote: >On December 05, 2002 at 03:58:01, Tony Hedlund wrote: > >>Take a look at game 21. Shredder had a winning endgame, both programs showed >>mate in 9. But suddenly the game ended in a repetition draw! A bug in Shredder? >>Tony > >- This behaviour has been noticed with several engines before. >- It has ALWAYS been caused by incomplete tablebases. >- In this case AFAIK it is caused by the incomplete "Fritz endgame turbo". >- That is the reason why I call this "Fritz endgame turbo" a piece of shit. >- The question arises: WHY does SSDF use the Fritz endgame turbo? IIRC SSDF only >tests chess programs kind of "out of the box". The Fritz endgame turbo is no >part of Shredder 6/7 nor any other CB chess program. Why do you use it? You provide the answer yourself in your last sentence. >- Why aren't the programs tested by using their provided package? So when there >are some 4- and 5-man tbs on the released CD, install them for that specific >program. It's done. >- The problem with incomplete tbs is well known, also for you. I wasn't aware that Shredder had this bug. Yes, I agree with Bob that it *IS* a program bug. >Why don't you >change your test setup to either complete, none or provided tbs? I've now changed the setup. I've removed all 5-men tb's with pawn's. In fact the same setup as Tiger have. >OK, finally I have to admit that the conditions are equal for all engines >running under a CB-GUI... but anyway, games like this kind of falsify results.
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.