Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hiarcs7.32's learning bug

Author: Lonnie Cook

Date: 00:43:48 07/03/00

Go up one level in this thread


On July 03, 2000 at 03:40:49, blass uri wrote:

>On July 03, 2000 at 03:05:35, Mark Young wrote:
>
><snipped>
>>I have found that the default setting are best with Hiarcs 7.32. That means
>>learning on, and selectivity = 5.
>
>Did you test it on tournament time control?
>
>I read one year ago that selectivity=6,7 is better at correspondence games but
>not at tournament time control but computers today are faster.
>
><snipped>
>>
>>When set up as a position test, Hiarcs 7.32 has no problem seeing that 43. Bxb6
>>is losing and finding 43. Qe8.
>
>I know
>The problem is when I  play against it and give it the position before axb4.
>
>It plays 42.axb4 and after Bxb4 needs a long time to see that 43.Bxb6 is losing.
>no position learning does not fix the problem.

learning in general from what I have seen is like a blind men touching the
proverbial elephant's trunk ... take it from there
>
>Uri



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.