Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Also Fritz 8.0.0.23 and Crafty 19.03 miss a mate in one!

Author: Uri Blass

Date: 00:10:34 07/26/03

Go up one level in this thread


On July 26, 2003 at 03:00:00, Uri Blass wrote:

>On July 25, 2003 at 19:44:43, Sune Fischer wrote:
>
>>On July 25, 2003 at 19:24:21, Eran wrote:
>>
>>>On July 25, 2003 at 17:50:21, Ernst Walet wrote:
>>>
>>>>This afternoon my account on the chessbase server played a game against Rebel
>>>>12.  To my astonishment Rebel missed a mate in one, and got mated right away.
>>>>Any thoughts on this, Ed?
>>>>
>>>[D]8/8/6p1/6k1/3p2P1/4b2Q/4Bq2/7K w - - 0 151
>>>
>>>If Fritz 8 or Crafty 19.03 plays the same game instead of Rebel 12, Fritz or
>>>Crafty will play the same problem as Rebel 12. Please see below.
>>>
>>>So, do Fritz and Crafty have a bug for a long time?
>>
>>I suspect most engines has that bug, including mine.
>>I've known about it for over a year, but have decided not to fix it because I
>>can't find a way to do it.
>>
>>The reason is the engine makes the 50th move, checks the 100 ply counter and
>>returns the draw score.
>>
>>To fix it, the engine needs to check for check mate _before_ it checks the
>>counter, but most engines can't do that staticly, they have to search another
>>ply to see the mate.
>
>No
>There is a more simple solution.
>
>If counter=100 the engine check for checkmate before it decides that it is a
>draw.
>
>If counter<100 you do not need to do it.
>
>I do not do it in that way and I check for checkmate at every node first.
>Thanks to my legal move generator I do not need to search another ply but only
>to generate a move if I want to verify no checkmate.
>
>Uri

I can add that even if you need to search then searching one ply in your draw
detection function only in case that the counter equal to 100 is not a big
price.

Uri



This page took 1.23 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.