Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: King Night Bishop - 50 Moves

Author: Steve Coladonato

Date: 05:38:00 04/26/99

Go up one level in this thread



On April 23, 1999 at 16:12:53, KarinsDad wrote:

>On April 23, 1999 at 14:21:38, Steve Coladonato wrote:
>
>>Steven,
>>
>>Dave Gomboc clued me in.  I have to apologize, but I really don't know who's who
>>yet.  Anyway, the move count I had from memory isn't too far off from the
>>current tablebases.
>>
>>Perhaps you can answer me this.  If a computer were on the losing side of the
>>KBN v K endgame, would the computer run to/towards the wrong corner at every
>>opportunity?
>>
>>Steve
>
>If the program has tablebases, it will attempt to get as far away from mate as
>possible at all opportunites (this means the best move on the board, regardless
>of whether that is towards the opposing corner). If it does not have tablebases,
>then it's moves may appear almost random (with the exception that it will
>normally not back away from the minor pieces) until it gets to a point where it
>can recognize the mate threat within it's event horizon.
>
>KarinsDad :)

I set up the following position:
/4k3/8/8/8/8/8/8/2B1K1N1/ { pieces on original squares }

Fritz 5.32 played randomly.  It actually offered a draw when on the winning
side.
Shredder 3 asked for the Ken Thompson CD and played as you noted.

It's pretty impressive when you think about it.  I don't know the in's and out's
of tablebases but the people who have researched and built them had to have put
a tremendous amount of work into it.

Thanks for taking the time to explain some of this to me.

Steve



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.