Author: Christophe Theron
Date: 13:37:29 03/09/02
Go up one level in this thread
On March 09, 2002 at 12:09:54, Kurt Utzinger wrote:
>A great list showing the chess people very clearly that a great number of games
>is required to say something about difference in playing strenght between two
>programs. I was lucky to see that my assumption to play at least 100 games,
>better even 200-300 games, was rather well.
>Kurt
And it all depends on the accuracy that you need.
I think it's the most difficult to explain. You have to explain that more games
are needed for a better ELO measurement, that less games are needed if the
opponents are of really different strengths, or if you are not looking for a
high reliability in the result, and so on...
You cannot give a simple answer like: "just play X games and you are done".
Christophe
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.