Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fritz6a major Endgame BUG !

Author: James T. Walker

Date: 08:04:07 04/10/00

Go up one level in this thread


On April 10, 2000 at 09:10:00, Robert Hyatt wrote:

>On April 10, 2000 at 03:04:45, Bernhard Bauer wrote:
>
>>On April 09, 2000 at 22:56:04, Robert Hyatt wrote:
>>
>>>On April 09, 2000 at 20:08:41, Laurence Chen wrote:
>>>
>>>>Thanks for your output. Your data demonstrates that Fritz 6 has some kind of bug
>>>>when accessing the tablebases. But it managed to win.  Perhaps, Robert Hyatt can
>>>>explain or someone, I always thought when a chess engine accesses tablebases and
>>>>finds a distant mate in XX moves, it will play the best move to mate.  Fritz 6a
>>>>shows a very odd behavior, as seen the score of the game provided below, it
>>>>plays moves which extends the mate, noticed that from move 100 to 105, the mate
>>>>announcement keep jumping up and down between mate in 16 and 17.  I take this to
>>>>be a bug in the Fritz 6a.  I don't see this kind of behavior in my other chess
>>>>engines. So what's your opinion?
>>>>Laurence
>>>
>>>
>>>first suggestion, since you downloaded files, check the sizes between what
>>>you have and what is on my machine. They must match _exactly_.  Not one byte
>>>bigger or smaller, otherwise that file is unusable.
>>>
>>>even a one byte change will change mates into draws and so forth.
>>>
>>>
>>
>>A download error may not always be recognized by the user - and there
>>are many files. I once had such a corupted file which made trouble.
>>So perhaps it would be nice to have a tool to check the egtb file sizes.
>>It's a pain to do it by hand.
>>
>>Kind regards
>>Bernhard
>
>I believe Eugene's code does this if the files are compressed.  IE a short/long
>compressed file should produce an error.  But not a file where a single byte
>was changed.  his tbdecode.exe program will validity check the file (I am not
>sure how thoroughly it checks, however).

CSTal-2.03 seems to check the tablebases before running.  If it finds something
wrong it will NOT run.  The problem with CSTal-2 is that there are a couple of
tablebases which cause it to not run no matter if they are good.
Jim Walker



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.