[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gnugo-devel] some more tests

From: Gunnar Farneback
Subject: Re: [gnugo-devel] some more tests
Date: Mon, 28 Jul 2003 21:01:38 +0200
User-agent: EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/20.7 (sparc-sun-solaris2.7) (with unibyte mode)

Max wrote:
> the problem is go++ loses a lot of time at the end of the game so i
> think the more fair way to compare both programs is to use truncated
> games letting them play say 200 moves in total ,gnugo playing level
> 12-13 and go++ level 5. At this stage ,the time will be almost equal.
> Then let Gnugo estimates the score . I would say this one will be
> often in favor of Go++

Are you saying that Go++ is playing both worse and substantially
slower than GNU Go in the endgame? That's certainly encouraging and
shows that the work put into the endgame by a number of people has
paid off.

> which is probably the stronger strategical program (at the current
> stage of gnugo devellopment ) .

No question about that. Strategical understanding is known to be a
weak spot of GNU Go currently.

Thanks for doing these tests. It would be nice if you could also post
a couple of sample games.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]