Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: CM 6555 vs CM6000-4

Author: Mark Young

Date: 03:20:11 12/14/98

Go up one level in this thread


On December 14, 1998 at 05:18:25, Didzis Cirulis wrote:

>On December 14, 1998 at 04:48:20, Mark Young wrote:
>
>>On December 14, 1998 at 04:04:05, Didzis Cirulis wrote:
>>
>>>On December 14, 1998 at 03:08:52, Mark Young wrote:
>>>
>>>>On December 14, 1998 at 02:15:55, Didzis Cirulis wrote:
>>>>
>>>>>Hi,
>>>>>Pentium 200MMX, 30 min/game, standard Hash table size (by mistake, as I forgot
>>>>>to change this back after I reinstalled Chessmaster). 44 games played
>>>>>
>>>>>                 W     L     D      Score
>>>>>CM 6555         16    11    17      24.5
>>>>>CM 6000-4       11    16    17      19.5
>>>>>
>>>>>CM6555 still looks to be the best.
>>>>
>>>>I have my doubts, when I jumped to game in 30 min CM6555 starts to have a big
>>>>drop off. My games in 15 min are like your game in 30 min. and at game in 15 on
>>>>my computer CM6555 was still doing well. So you will need to test at game 1hr.
>>>>to confirm my findings.
>>>>
>>>>I played some games at 60/90. Test 4, and default and SS=10 are doing much
>>>>better so far. And the longer the time controls get the better default is
>>>>looking on a P II 400. I think it my be best. But more testing will tell.
>>>>
>>>
>>>So, Mark, what is your proposal? To play a tournament games? 40/120? It will
>>>take tame a lot. But Ok, I can play 40/120 on my home PC while I am at work. But
>>>Is for sure, those tests should be coordinated somehow, all work has to be
>>>distributed among the group of interested testers,and all results must be
>>>gathered into one place. Otherwise this all testing is not very effective.
>>>MY proposal is as follows: I can offer my web page as an "easy to access" place
>>>for everyone to see the latest results, and me :-) as one who could upload them
>>>there upon receiving from you. This applies to 40/120 games only. Maybe later
>>>the results of CM vs other opponents may be added there as well.
>>
>>Something like this will have to be done, if we are to find the best long time
>>control settings. What plays strong at fast time controls seems to be playing
>>weaker at longer time controls. This will make the job of finding good settings
>>much harder, if we can not count on the settings being strong at fast and slow
>>time controls. But if the point is to find just stronger fast setting, the way
>>we are doing it will work fine.
>>
>Yes, but this is not an answer to my proposal, Mark.:-) I said I am crazy enough
>to undertake some coordination work, and people may send in their test results.
>Are you ready to take part in this?

Sure, I test programs everyday. I have a spare computer I can help test settings
on. Lets get some people together, and chose some settings to test.
>
>Didzis



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.