Author: Mark Young
Date: 21:28:30 04/11/98
Go up one level in this thread
On April 11, 1998 at 17:32:08, Thorsten Czub wrote:
>----------------------
> 10-9 for CSTal.
>
>If Moritz says CSTal has 2250 ELO on a 200 Mhz machine, and Fritz5 is
>NOT able (as you can
>see from the games and the results) to beat it although it has ELO 2589
>(2589-2250 = 339 ELO
>stronger !!).
>It should not play that lousy against a 339 ELO weaker program.
>Also as a program with 180.000-200.000 NPS it should not play those
>games against a program
>doing 3500 NPS.
>Also as a program "outsearching" all other programs it should not
>outsearch a very selective
>and slow program.
>
>Conclusion:
>1. Fritz5 is not that strong.
>2. playing manually gives different results (not games) than playing
>with the ChessBase-
>Autoplayer system. All ChessBase-Autoplayer system games are worthless.
>
>Alternative conclusion:
>1a. as CSTal gets almost 50 % against fritz5, it has an ELO 2589 too. Or
>even better.
>New Leader of ssdf-list (and in opposite to the results of ssdf-guys and
>Enrique and Moritz we
>played anything manually - so no cheating with a special device...) is
>therefore Chess System
>Tal.
>
>:-))
>
>
>I will later present my data (so far) of OTHER opponents playing vs.
>Fritz5.01. Here Fritz5 is also
>losing nice.
>
>
>You asked for data. Here is a few data. More follows later (it takes
>time to play 40/120 games if
>you have to do it manually. I guess anybody knows this from old days
>before we had auto232-
>device).
>
>Have fun :-)
------------------------------------------------------------------
1. You can not compair NPS with other programs. They search at diffrent
rates. But may come up with the same move at the same time level.
2. You may be getting funky results because you are running Fritz 5 on a
slow computer. It is possible that Fritz 5 may have a greater ratings
gain then other programs get from a faster cpu. You say that your
computer get about 180 knps to 200knps. I tested fritz on a much faster
computer. I am getting well over 100knps more in my testing. Since Fritz
is a searched base program this could explain alot. What do you think? I
have seen no other date about fritz 5 running on slow computers. Fritz 5
may be the best program running on the faster intel chips but not the
best on slower computers. If your results and my results are correct
there can be no other explaination.
Mark Young.
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.