Author: Christophe Theron
Date: 11:04:35 04/21/01
Go up one level in this thread
On April 21, 2001 at 03:26:00, Fabio Barrettone wrote:
>>OK, then your numbers indicate with more than 90% reliability that SelSafety=12
>>is better than the default setting (SelSafety=6) under the conditions of your
>>experiment.
>>
>>Thanks a lot for this information. That's something I need to think about for my
>>next version.
>>
>>I hardly dare to ask if, by chance, you would have more time for a similar
>>experiment against a different program...
>>
>>
>> Christophe
>
>
>Sure, I'm currently doing the same test against Crafty 18.07. The only problem I
>have now is testing against not-winboard engines since I don't have the time of
>making the moves manually (I have only one computer).
>
>For this I've asked if the Chessbase version of Tiger has a ct.ini (I think so),
>in this case I could test this theory against Fritz, Junior, Hiarcs, Nimzo etc.
>also, with a sort of reliability (400 games are enough, I think). Well, the
>computer at home makes all the job while I'm at work ;-)
The ChessBase version of Tiger has EXACTLY the same engine, so yes it has a
CT.INI to tweak.
Whatever number of games you play, there is an error margin associated with it.
If the error margin is below the score differences, then you know the match has
a meaning. If not, you need to play more games. That's rather simple actually.
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.