Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Piece Values in the Endgame

Author: Vasik Rajlich

Date: 02:40:02 02/12/04

Go up one level in this thread


On February 11, 2004 at 17:16:00, GuyHaworth wrote:

>
>When I was taught chess, pre-history now, the piece values were:
>
>Pawn = 1 (without loss of generality)
>Bishop = Knight = 3
>Rook = 5
>Queen = 10.
>
>These figures may not be 'current thinking' so feel free to correct me there.
>
>
>However, my question is this.  If piece-values are assumed to be dependent only
>on the force on the board, what values should they be given for various forces
>in the endgame?
>
>
>g

This is a very badly-explored topic - although the commercial programs surely
have their own secrets. In my program, a piece is worth 3.75 pawns with all
pieces present, and 2.75 pawns with all pieces missing. (Everything is scaled
linearly in between - the extremes are of course impossible.) A rook gains in
value as pawns are traded (1/8th of a pawn for each pawn pair), while a knight
loses in value (1/16th of a pawn).

For a very good scientific article on this top, do a google search for "Larry
Kaufmann", he wrote an article on this topic targeting computer chess
programmers, you will find some interesting data there.

Vas



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.