Author: Robert Hyatt
Date: 08:18:17 09/17/03
Go up one level in this thread
On September 16, 2003 at 17:47:06, Russell Reagan wrote: >On September 16, 2003 at 17:11:03, Robert Hyatt wrote: > >>That's certainly one way to do this... And that's the basic idea, that some >>"characteristic" of the board is bad or good, not with respect to the color, >>but with respect to which side the computer is playing. > >Ah, so I guess one very obvious example is opposition in the endgame, if I >understand correctly. No. Opposition is for the side on move. IE if our kings have one square between them, both on the same file, then whomever has to move does not have "opposition" and has to give way. That is a function of the position and who is to move, not which side is the computer and which side is not. > >The difference is that if black has doubled pawns (and you penalize for this), >then black will carry that penalty regardless of which side is to move. The >opposition bonus (if you used this) would depend on which side is to move, and >so it isn't really "static". It isn't exactly "static" in your context, but it would be scored the same whether the computer has the opposition or the opponent does. That is different from asymmetric evaluation. One idea... in king safety. If the computer and opponent castle kingside, then opening a file is equally bad/good for _both_ sides. Either your program had better understand piece cooperation _very_ well, or else you make the open file worse for the computer than for the human so that the computer won't allow it to be opened. > >So symmetrical evaluation is a pure "static" evaluation where the side to move >doesn't matter, while an asymmetrical one will include a bit of dynamism because >it will evaluate some things based upon the side to move. Is this correct?
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.