Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Draw Detection by Move Repetition Procedure -- Comments

Author: vladan

Date: 14:38:54 08/01/04

Go up one level in this thread


On August 01, 2004 at 14:32:28, Rémi Coulom wrote:

>On August 01, 2004 at 14:19:21, Rémi Coulom wrote:
>
>>So, TCB needs 357057 nodes to complete ply 12, while your programs needs
>>8524217! Also note that the right move is found at depth 7, with the right
>>evaluation.
>>
>>Rémi
>
>I looked more closely at your figures, and another reason for the inefficiency
>of your search is probably that you search too many nodes in the quiescence
>search because of checks. TCB currently searches no checks in the QS. I know
>that some good engines do. It seems to be a good idea to search checks only at
>the first level of the QS. That is particularly good right after the null move.
>It helped passed versions of TCB to solve mate combinations faster. You probably
>search way too many checks in your QS. It might be a good idea to limit them
>somehow.
>
>Do you perform check extensions ? If you don't, you should.
>
>Rémi



Hi Remi,

Yes, Axon generates millions of nodes per second but it is not selective enough.
There are many new techniques that must be tested. The implementation of the
good selective techniques is the probably the most important problem in computer
chess. Test results in paper are generated by the old test version of Axon;
there are many improvements in new versions.

We have complex Qseracher, not only captures, but also check and promotion
moves…

Also, our experience is that simple Qsearcher decrease search stability.



best regards,

Vladan

(Axon programmer)




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.