Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Static Evaluation Function

Author: Ryan B.

Date: 13:14:55 04/07/05

Go up one level in this thread


On April 07, 2005 at 00:15:33, Pallav Nawani wrote:

>On April 06, 2005 at 10:59:26, Graham Laight wrote:
>
>>On April 06, 2005 at 10:32:43, Meg Ryan wrote:
>>
>>>Has anyone got any advice of writing a good position eval?
>>
>>* write a function that takes two chess positions and produces a value between 0
>>and 1 depending on how closely they match
>>
>>* match the given position with every position in the chessbase database,
>>scoring it as 1 if the player won, 0.5 if the game drew, or 0 if the player
>>lost. Use the position matching function as a weighting, and then sum your total
>>answers
>>
>>-g
>>
>>
>
>This is a bad idea for the following reason:
>In chess, some times even a square's difference can be the difference between a
>win and a draw or a loss. Your method will fail in such situations. And these
>situations are far more plentiful than one would expect.
>
>Pallav

And some squares are more important than others.  If anyone can explain an
algorithm to calculate the importance of each square for the situation I would
greatly appreciate it.

-Ryan



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.