Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Goliath Light is stronger than Hiarcs7.32!?

Author: blass uri

Date: 12:52:41 08/07/00

Go up one level in this thread


On August 07, 2000 at 15:36:17, Terry Ripple wrote:

>On August 07, 2000 at 14:40:41, blass uri wrote:
>
>>On August 07, 2000 at 14:24:18, Peter Skinner wrote:
>>
>>>On August 07, 2000 at 13:13:23, blass uri wrote:
>>>
>>>>On August 07, 2000 at 12:44:29, Peter Skinner wrote:
>>>>
>>>>>On August 07, 2000 at 03:15:06, blass uri wrote:
>>>>>
>>>>>>On August 07, 2000 at 01:35:22, Jouni Uski wrote:
>>>>>>
>>>>>>>According to latest SSDF list Goliath Light is stronger than Hiarcs7.32.
>>>>>>>Congratulations to Michael. Hope, that Goliath also goes to London WC!
>>>>>>>
>>>>>>>Jouni
>>>>>>
>>>>>>according to the ssdf list we cannot know which program is better.
>>>>>>1 elo difference is not significant.
>>>>>>
>>>>>>Uri
>>>>>
>>>>>I fully agree. Also I se on the list that Fritz 5.32 is still rated higher than
>>>>>Hiarcs... I simply can not see how, as everytime I have matched the 2 programs,
>>>>>Hiarcs kills Fritz 5.32... Hiarcs always seems to win in the endgame...
>>>>
>>>>What is the time control of your match?
>>>>
>>>>Is it 2 hours/40 moves?
>>>>
>>>>Uri
>>>
>>>Yes, in the 2/40, Hiarcs beats Fritz 5.32 almost every game. Even in quicker
>>>time controls, Hiarcs seems to best Fritz all the time. Rarely dropping a game.
>>
>>I am not surprised if it wins in quicker time control because hiarcs is known to
>>be good at blitz(it never lose on time when Fritz5.32 can lose on time)
>>
>>My tests (not against Fritz5.32) showed that hiarcs has problems in retaining
>>the hash tables at longer time control and it may do errors that it does not do
>>if I give it to analyze positions with no learning.
>>
>>What is your hardware?
>>
>>Uri
>
>What would happen if you turned the learning off for the long time controls,
>would this then help Hiarcs?
>
>Terry

I do not think that it would help hiarcs because my experience is that turning
learning off does not prevent the hash tables retaining problem.

When I play against hiarcs I get different result relative to results based on
search without playing even when I turn learning off.

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.