Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How Greedy is your Program? Good Positional Test!

Author: Shaun Brewer

Date: 03:54:49 04/16/99

Go up one level in this thread


On April 15, 1999 at 23:15:08, odell hall wrote:

>Hi
>
>
>The below position is taken from the Book Secrets of Chess Analysis by Jan
>timman.
>
>3q1rk1/1p3p2/r2p2pb/P1pP1n1p/2N2P2/1QPP4/6BP/1R3RK1 w - - id Geller - Spaasky;
>
>
>28. Be4!!!  Timman remarks "White appreciates the strength of the strongpoint f5
>and defers the capture on b7, which would give black the chance for a promising
>exchange sacrifice.  28. Qxb7, Rxa5 29.Nxa5, Qxa5 and now 29.Qb2(to be able to
>defend important weak points from d2 or c2).
>
>
>I tested rebel on this position and it simply could not resist the pawn on b7!
> However hiarcs6 after 82 sec Chose  28. Bh3!! which I think is about the same
>as Be4.  In my opinion this is a excellent positional test for computer programs
>to see just how well they understand position. I suspect that only the strongest
>programs will be able to resist the juicy pawn on B7.  I also Tested MChess 8
>which chose Be4 after one minute thought and +76 evaluation.

Are program avoiding the capture because they have better positional
understanding (in this position) or are the searching deep enough to see the
risks?

Can people provide ply depths when the program they use switches.

Thanks Shaun



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.