Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: another question

Author: Bas Hamstra

Date: 12:50:48 07/07/03

Go up one level in this thread


On July 07, 2003 at 14:45:15, Peter McKenzie wrote:

>How did you get the position from the end of the PV?  With hash table cutofs it
>is sometimes impossible to know the entire PV, so how did you solve this
>problem?

I try to keep perfect PV's. In case of a hash cutoff I try to reconstruct the PV
immediately from hash and store it. At shallow depths nearly 100% of my PV's are
correct till the end. If the PV was not complete I skipped the learning for that
step. I assumed the PV to be correct, if the evaluated score of the last PV
position matched the PV score. I stored the PV's and the PV score after each
search termination, so it's an easy check then.

>I think the knightcap experiment returned a complete position every time they
>returned an evaluation, but this seems somewhat extreme!


Best regards,
Bas.



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.