Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: a mistake of Fruit2.2Uri in CEGT (bug or another problem?)

Author: Uri Blass

Date: 14:00:49 10/19/05

Go up one level in this thread


On October 19, 2005 at 16:39:03, Uri Blass wrote:

>On October 19, 2005 at 16:25:14, Christian Koch wrote:
>
>>>Do you plan to give some evaluation improvements in the next update
>>>that are not related to this problem or do you plan to give only update for
>>>tablebases support.
>>
>>How should the evaluation improvements look like? Please post some suggestions
>>with reliable examples.
>>
>>Only tablebase support is not good enough for a next free release?
>
>I did not complain
>I only asked about it.
>
>I do not claim to know how should evaluation improvements look like(except maybe
>evaluation of some basic endgames like KQ vs KP and even if the program use
>tablebases it is still productive because the program cannot evaluate tablebases
>positions at every node) but I know that Fabien plans to improve the evaluation
>based on an email that he sent me.
>
>I do not know if he had success in it.
>I do not know if he know if he is going to have success in it in the time of the
>next release.
>I do not know if he plans to implement it in the next update in case that he has
>success in it.
>
>I am sorry that some people consider every question that I ask as a personal
>attack.
>
>I could ask Fabien by email but I think that it is information that may be
>interesting to other posters so I did not do it but maybe it is better that I
>ask question by email because every question that I ask here is understood as
>personal attack by at least one of the CEGT testers.
>
>Uri

I can add that I have more ideas for evaluation improvement but I prefer to post
nothing about it because I am not sure if they work and I believe that fabien
probably has better ideas than me for evaluation improvements.

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.