Author: Robert Hyatt
Date: 16:22:52 06/03/00
Go up one level in this thread
On June 03, 2000 at 11:45:34, Kim Roper Jensen wrote: > >I wondered, if you have 2 chess programs, thats identical in every way, and elt >them play against each other for x number times ( x -> +oo ) then the score >should be approx. 50% > >then we take program 1, and remove 1 thing in the evaluationcode and again let >them play against each other, what is the score then ?? It can win significantly. Think about it. Say you know nothing about an outside passed pawn. You clone yourself. Each game you play, both you and your opponent know everything the other knows. Some wins will be luck, etc. But there is nothing to give one side an 'edge'. But teach one program about outside passed pawns and you will begin to see that program pull ahead, because it knows something that the other doesn't, and that is _all_ that makes them different. > >And then remove one more thing again ?? > >I mean, if you just take one thing out of the eval. function will, the original >complete program just win a "little" more times, or what ?? > >I dont mean some bizarre endgame-code like how to mate with 3 knigths, but >doubling of pawns, etc. > >I hope for some response, with regards > >Kim Jensen > >PS. i think ccc is one of the best forums for chess-programmers, keep it up :)
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.