Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: When I use Fritz, it's analysis varies significantly?!

Author: blass uri

Date: 10:46:17 07/29/00

Go up one level in this thread


On July 29, 2000 at 13:08:13, Chessfun wrote:

>On July 29, 2000 at 12:18:53, blass uri wrote:
>
>>On July 29, 2000 at 11:02:23, Chessfun wrote:
>>
>>>On July 27, 2000 at 01:19:51, Mike S. wrote:
>>>
>>>>I think you just need to shut down your system completey between each test run.
>>>>
>>>>From the previous version Fritz 5.32, I don't remember any problem reproducing
>>>>test results. I've read somewhere, Fritz 6 has positions learning or something
>>>>like that now
>>>
>>>
>>>From an email I got from chessbase, "Fritz 6 has no position learning function."
>>>
>>>Thanks.
>>
>>Maybe they mean that it does not learn from previous games but not that it does
>>not learn from previous search.
>>
>>You can check it by the following experiment:
>>
>>give Fritz6a to solve a test position.
>>give it to play and goback and again give it to solve the same position.
>>
>>If it does not solve faster in the second time then it does not have learning
>>from previous search otherwise it has learning from previous search.
>>
>>Uri
>
>
>I'm not sure but that just sounds like retaining hash table search.

Retaining hash table search is exactly what I meant
Fritz5.32 does not do it and I guess that the poster who claimed that Fritz6
does position learning or something like that meant to retaining the hash
tables.

>There are no files created or changed within Fritz 6 so actual position
>learning can't take place, and If I shut down and restart it solves no
>quicker, therefore no position learning.

You do not need to shut down and restart.

I think that downloading another engine instead of Fritz6 and downloading Fritz6
again is enough.

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.