Author: Uri Blass
Date: 06:07:58 07/15/01
Go up one level in this thread
On July 15, 2001 at 08:40:28, Thorsten Czub wrote: >On July 15, 2001 at 07:36:37, Uri Blass wrote: >>The problem is that it is possible that Nimzo8 was lucky to get positions that >>it understands. > >the way it plays is the way the strength gets realized. >the scores, the mainlines, the game. > > >>I agree that seeing games give more information then seeing only results but the >>problem is that people do not know to get the right conclusions from this >>information. > >which conclusions do you get from 500 game results ? >none. >does it tell you what to change ? the weaknesses ? >no. >it only tells you how much. I agree that it does not tell you the weaknesses and the best way to learn about weaknesses and changes in the program is to look at the games to compare both versions on the same positions > > >>I believe that 500 results give more information then 10 games even if the >>people know to get the right conclusions from watching games. > >500 times throwing bones into the air gives 500 times the information >1-0, 1/2, 0-1. > >you have 500 events, and get 3 different results. >200 times 1-0 >200 times 0-1 >and 100 times 1/2. > >and now ? >what conclusions do you get ? I get a rating for the program based on the results and I can compare it with rating of other programs or the rating of the previous version of the same program. I can learn nothing about the style of the program and about the content of the change without watching games but I can learn if there was a positive or a negative change more than what I can learn from small number of games. 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.