Author: Sune Fischer
Date: 16:33:36 01/16/02
Go up one level in this thread
On January 16, 2002 at 07:41:28, Graham Laight wrote: >It has occurred to me that it is wrong to evaluate a position in terms of >relative pawns (the "de facto" standard - whereby an evaluation of 2 means that >you're approximately the equivalent of 2 pawns ahead). > >This means that many aspects of evaluation have to be squeezed into a dimension >which is not appropriate at all. You can create a map between the two if you like. It shouldn't be a problem to make a table that could take a given "pawn score" and translate it into a probability or vice versa. However, if the pawn score is higher, then the probability score will also be higher (assuming it is a good program with lots of positional understanding as well, and not just a simple material evaluator of cause). The map will be monotonic, i.e. its derivative will have one sign only. So, no matter how you choose to look at it, if move A is better than move B with pawn eval, then it will also be the preferred move with a probability eval. Result will be that the search will find the same move either way. >A better way would be to evaluate "winning probability". If a position was a >draw, the value would be 0.50 (or 50%). If the player should win 3 out of 4 >times, the eval should be 75%. If the player must win from here, then the >evaluation should be 100%. You would need three probabilities, one for draw and two for white or black winning (they should sum to 1 of cause), it is not simpler IMO. >It seems strange when you think about it that all programmers have chosen to >adopt the traditional "pawn equivalence" standard. I don't see any particular strength to this method, other than it may seem more natural to some people. >-g -S.
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.