Author: blass uri
Date: 08:59:40 06/24/00
Go up one level in this thread
On June 24, 2000 at 10:56:13, Mogens Larsen wrote: >On June 24, 2000 at 09:13:46, blass uri wrote: > >>This is the worst way to test because if you use deterministic programs you will >>get the same games again and again. >> >>You may also get wrong results because some programmers did not include some >>knowledge about opening because opening book covers this knowledge most of the >>time. > >Your first objection is true, the second one isn't. Results from that kind of >experiment can't be described as wrong. It can be described as wrong if the target is to see which program is better in chess(remember that opening book is part of the program) >The knowledge parameter has an influence >on the determination of strength during a chess game, no matter what stage of a >chess game we're talking about, so stop repeating that nonsense. Some knowledge is less important if you have opening book. The programmers assume that their program will use an opening book so playing without opening book is playing a game that many programmers did not prepare to play. You have the right to do it(I also did it when I tested Junior and told the programmer about problems that I found). I agree that it is relevant also for playing chess because the opponent may take the program out of book and it may do similiar mistakes(this is the reason that I believe that testing programs without opening book is important) but I do not think that the results of the games tells you about the strength of the program. It is possible that a problem that happen only in 1 out of 30 games when the sides play with books happens almost every game when the sides play without books so you get a wrong picture about the real rating of the programs. Uri
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.