Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: To James Walker re: F5.32 v H7.32

Author: blass uri

Date: 23:37:33 08/03/99

Go up one level in this thread



On August 03, 1999 at 16:23:40, James T. Walker wrote:

>On August 03, 1999 at 13:37:00, eric guttenberg wrote:
>
>>Jim,
>>
>>Some time ago you posted the results of some 240 games played between
>>Fritz 5.32 and Hiarcs 7.32 on separate computers showing a surprising
>>edge for Fritz..i.e. 40 games at g/15 were 27-13 in favor of F5.32.
>>
>>Today I am informed by blass uri that in subsequent games you played
>>in which Hiarcs was provided a higher hash setting, the results were
>>much different, almost, but not quite, in line with some of single-
>>computer results posted by Tania.
>>
>>Could you please post the results of those additional games, and the settings?
>>
>>eric
>
>Hello eric,
>I looked up the stats and here is what I found.  The g/15 time control had Fritz
>ahead after 40 games by 27-13 as you mention.  I believe this was mainly caused
>by the lack of hash memory for Hiarcs 7.32.  After correcting this problem Fritz
>still won the 74 more games I played at that time control by a score of 39-35.
>Of course this is not enough games to separate the two.  So at g/15 the total is
>66-48 favor Fritz mostly due to the score in the first 40 games where Hiarcs was
>"starved" for hash memory.
>I have more games at blitz and here is what I found.  At game/5 min the total
>score is 407.5 to 380.5 in favor of Hiarcs 7.32.  I also played 50 games at
>game/3 min and the score is 26.5 to 23.5 favor Hiarcs 7.32.
>The games were all played on two computers using auto232 connection.  The
>computers were a PII-333 and a K6-2-350/400.  At some point I overclocked the
>350 to 400 Mhz.  I gave both progams an equal amount of games on each computer
>to help compensate for the difference in speed.  I recently upgraded my K6-2-350
>to a K6-3-450 and Hiarcs seems to gain more from the change than Fritz.  I have
>not run any test yet using the K6-3-450.
>As an added note I have played only 12 games at "standard" time controls between
>the two and Fritz leads 9.5 - 2.5.  I can't explain this except that it is a
>small number of games and will probably even out with more games.
>Jim Walker

It is interesting to see if there are stupid blunders of Hiarcs in these games.

I remember that in one game Hiarcs did a stupid blunder against Cstal when my
Hiarcs could not reproduce the blunder.
The game I am talking about was played manually so it was not a problem of the
autoplayer.

It is possible that Hiarcs has problems with some machines.

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.