Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Neverending story with incomplete tablebases

Author: Dieter Buerssner

Date: 11:30:49 08/17/03

Go up one level in this thread


On August 17, 2003 at 05:11:28, Johan de Koning wrote:

>Good point, though I'd prefer a [keep hash] button. Better is Yace's approach
>that Uri just mentioned, especially if it's controlled by a [keep summaries]
>button. Even better would be to be able to force positions and values.

I can do it in my debug versions of Yace (only in text mode, no convinient
feature available through any GUI). I see no easy way to make it usable from the
GUI. In UCI one can do many things, but this would not work well. The problem
is, I could have a button, like "Learn current position as good". But what
exactly is good? And current position is also not too well defined in the UCI
protocol (last position that was searched, should be enough, however). When I
want to give a score, this at first view also seems possible with UCI. Something
like: "Assume current position has score ..." where ... is selectable. But it is
very prone to misunderstandings. Often UCI GUIs will only send new settings when
something changes (and the GUI will not be able to figure out that current
position changed). So, when a user would set 2 differnt positions to the same
score, the GUI would not send this to the engine (because score did not change).
He could set one position to score 10.0 and the next to 10.01 - but that really
sounds like an awful hack.

The feature is also useful for some special purpose debugging.

Regards,
Dieter



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.