Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: CadaquƩs Tournament: Tiger - Hiarcs 7.32, 10.5-9.5

Author: James T. Walker

Date: 06:31:23 01/14/00

Go up one level in this thread


On January 14, 2000 at 09:18:47, James T. Walker wrote:

>On January 14, 2000 at 07:24:42, Enrique Irazoqui wrote:
>
>>On January 14, 2000 at 07:19:37, Jouni Uski wrote:
>>
>>>One note: even if  this tournament is 40/40 - it is played in P500 - so
>>>this means almost same as TOURNAMENT 40/2h in 200 MMX PC!! So no excuses type:
>>>my program is better with longer time controls...
>>>
>>>Jouni
>>
>>I keep saying that in my opinion there is no significant difference between
>>40/40 and 120/40. I also know that some people that always tested at 40 in 2 are
>>thinking about switching to 40/40. Maybe there are programs like CST and Diep
>>that need really long time controls, but for the rest it is roughly the same, I
>>think.
>>
>>Enrique
>
<snip>

I forgot to mention, I have also discovered that I can manipulate the ratings by
selecting which opponts are played and how many games.  By properly selecting
opponents and number of games played I can make any of the top 5 programs finish
in first place by a few points.  The SSDF apparently does not consider this when
creating their matches.  For example I tested Rebel vs Fritz 5.32 and Fritz 6
and they both beat Rebel by 8-2.  Then against Junior 5 Rebel won by 13.5 to
6.5.  Junior 5 plays pretty close to Fritz.  If all the top programs don't play
the same opponents the same amount of games then you are affecting the final
ratings because they will do better vs some opponents than others.  The SSDF is
apparently assuming that the opponent does not matter, only the rating of the
opponent.
Jim Walker



This page took 0.01 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.