Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Crafty 19.13 bug? Mate in 1 position evaluated as draw

Author: Slater Wold

Date: 14:41:28 06/03/04

Go up one level in this thread


On June 03, 2004 at 17:28:14, Peter Skinner wrote:

>On June 03, 2004 at 16:35:40, Robert Hyatt wrote:
>
>>On June 03, 2004 at 16:28:33, Robert Allgeuer wrote:
>>
>>>On June 03, 2004 at 15:32:14, Günther Simon wrote:
>>>
>>>>On June 03, 2004 at 14:48:55, Robert Allgeuer wrote:
>>>>
>>>>>
>>>>>In the following endgame position Crafty 19.13 (I have tried both the compiles
>>>>>by Slater Wold and Dann Corbit) analyses this as a draw with score 0.00. Crafty
>>>>>19.6 on the other hand correctly detects the mate in one.
>>>>>
>>>>>Winboard, hash 96 Mb, Athlon 2400+, all 5 piece EGTBs installed.
>>>>>
>>>>>[d]8/6P1/8/8/2b5/8/p1k2p2/K7 b - - 0 90
>>>>>
>>>>>Robert
>>>>
>>>>Hi Robert, which dates have your versions?
>>>>Even if you had tried both named compiled ones, it's still
>>>>possible that those are first releases of Crafty 19.13,
>>>>which had a tablebase bug.
>>>>
>>>>Günther
>>>
>>>I am aware that there was this bug and consequently 2 different 19.13 versions,
>>>therefore I downloaded after that, so I am pretty sure I have the corrected
>>>version.
>>>The compile of Slater Wold has been downloaded on 13.5.2004, the one from Dann
>>>Corbit just a week ago.
>>>Can anybody reproduce the result (with 5 piece EGTBs) or does Crafty 19.13
>>>detect a mate for everybody else and it is a problem just on my machine? Corrupt
>>>EGTBs cannot be the reason, because 19.6 uses exactly the same EGTB files.
>>>
>>>Robert
>>
>>
>>I can not reproduce this.  I ran with 6=piece tables.  mate in 1.  With only 5
>>piece tables, mate in 1.  With _no_ tables.  Mate in 1.
>
>I as well can not reproduce this error at all. I have tried with 6,5,4,3,no
>tablebases, and I get mate in 1 consistantly.
>
>Someone else emailed me from the Crafty site about a week ago with a similiar
>postition where Slator's for some reason couldn't find the mate, yet mine had no
>problems.
>
>I am wondering if his beta compiler is doing something funny to the code..
>
>Peter

All of my versions mate.  He even stated that he tried it with Dann's compile,
and got the same result...



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.