Author: Will Singleton
Date: 09:52:17 03/31/99
Go up one level in this thread
On March 31, 1999 at 07:04:32, James Lank wrote: >IM Larry Kaufman wrote a fascinating article on the evaluation of material >imbalances in the March Chess Life. To make a long article short, he did some >database crunching and came out with new numerical values for the pieces, when >taking into consideration other factors (for example the value of bishop over >knight is greater when the bishop is part of the bishop pair, or Q+N is worth >slightly more than Q+B, but that advantage disappears if there are also rooks on >the board, etc.) > >Some of this has been common knowledge among stonger players (i.e. most stonger >players know the value of having a pawn + bishop pair for the exchange), but >some of his conclusions were startling to me (a correspondence master); perhaps >not so startling to IM and GM level players. > >This strikes me as useful knowledge as a chessplayer, but it also seems that it >would be invaluable information to put into chess programs. > >Has this concept been implemented to any great degree in top level programs? > >Jim Lank The concept? Do you mean, are the values of pieces static or fluid, or rather, have his specific ideas been implemented? The values assigned to the pieces are just starting points. They are continually modified throughout the game, using different methods. Will
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.