Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Crafty MPC tests (long post)

Author: Robert Allgeuer

Date: 13:37:19 10/18/03

Go up one level in this thread


On October 18, 2003 at 13:20:16, Russell Reagan wrote:

>On October 18, 2003 at 12:00:09, Robert Allgeuer wrote:
>
>>Conclusion:
>>===========
>>
>>My tests indicate that the overall playing strength of Crafty 18.15 remains
>>more or less unchanged by the addition of Multi-ProbCut. However, the
>>characteristic of the engine changes significantly due to ProbCut: Even though
>>nominal search depth is increased by one to two plies, tactical strength is
>>severely reduced.
>>Furthermore with ProbCut match results become more unpredictable and
>>inconsistent: Apparently there are types of opponents against which ProbCut
>>works very well and results in significantly improved results, but there are
>>also other opponents (the tactically stronger ones?) where ProbCut has exactly
>>the opposite effect.
>
>Hi Robert,
>
>I did some testing also, and I experienced basically the same thing. MPC Crafty
>seemed to perform a bit weaker all around for me, but when it came to blitz time
>controls it stood no chance. For instance, I have tested Crafty 19.03 against
>Ruffian at 1/0 and 2/0, and while Ruffian wins the matches, it is somewhat
>close. Around 55% to 45%. When I tested MPC Crafty in 1/0 and 2/0 (and even
>longer, I think up to 20/0) Ruffian basically butchered it. It got better as the
>time controls got longer though.

I guess only more test games at longer time controls would tell the truth, but
even without doing this I think that MPC is not a real improvement as it seems.

Robert



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.