Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How you judge in your program: "opening", "middle game", "end game"

Author: Amir Ban

Date: 05:18:23 01/29/00

Go up one level in this thread


On January 29, 2000 at 03:08:33, Christophe Theron wrote:

>
>Tiger uses different algorithms for each stage of the game.
>
>That means my search, pruning, extensions set and knowledge set are different in
>every stage of the game.
>

Of course, but the real question here is where is the phase of the game decided.
A true node processor would decide the phase for each node individually, a true
preprocessor would do it once at the beginning of the search. It's also possible
to take an intermediate position about this.

Fritz 5.32 behaves like a preprocessor for some phase transitions. This places a
real cap on its ability to understand some things.

A preprocessor also cannot use the hash table left from previous searches (well,
it can, but what it's not clear what the result is), because the premise behind
using a transposition table is that the evaluation of a position with fixed
search depth is a constant that does not depend on the search environment.

Amir


>I display this on the screen in order to amuse the kids that pretend to know
>everything.
>
>And it works. ;)
>
>
>
>
>    Christophe




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.