Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: table-based SEE or "evaluation in rebel (hanging pieces)"

Author: martin fierz

Date: 06:58:15 11/27/03

Go up one level in this thread


On November 27, 2003 at 09:24:52, Steve Maughan wrote:

>Martin,
>
>>am i missing something here, or is this just another inaccuracy that such a
>>table-based SEE
>
>It's an inaccuracy of the system.  I think the reason that Ed is using this
>system is that he's generating the attack info as part of the evaluation which
>is carried out at every interior node - hence the exchange evaluation is "free".
>
>>(like not resolving xray attacks)?
>
>Ed does account for some xray attacks by incrementing the number of attacker
>without changing the piece mask e.g. if a bishop attacks a square though a queen
>the attacker count will be 2 but the only piece mask that is set will be the
>queen mask.
>
>Hope this helps,
>
>Steve

hi steve,

thanks for the explanation. i asked because i also generate all attacks at every
node, and since this is very expensive (at least the crappy way i programmed
it...) i would like to do something useful with the information :-)
of course i'm already using the attacks for something, else i wouldn't be doing
it at all, but i need to use it for more, the way it slows me down!

i also have to figure out how to produce x-ray attacks properly, i don't do that
yet.

BTW, how is monarch doing? your webpage says you are rewriting it in C, but it's
been saying that for a long time?!

cheers
  martin



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.