[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] Re: some answers and ask for help
From: |
David G Doshay |
Subject: |
Re: [gnugo-devel] Re: some answers and ask for help |
Date: |
Wed, 8 Sep 2004 22:08:34 -0700 |
I agree with the idea that major improvements in strength
are not going to come in the late endgame. Proper ordering
from about the first monkey-jump can make a big difference,
though.
I just wanted to point out that the endgame is where I have
found the greatest number of bugs in SlugGo, when SlugGo
makes mistakes that the underlying GNU Go does not. So, I
do think that it would be a very good exercise for Eric to start
on the endgame, particularly considering that he says that
he is not that strong a player. The endgame is far easier
to understand than the interaction of strategy and tactics
in the middle game.
But then again, I have no idea what an AI Planner is or how
it works. I think in terms of algorithms and data structures.
Cheers,
David
On Sep 8, 2004, at 9:07 PM, Eric wrote:
--- Xavier Combelle <address@hidden> wrote:
One point is there is already quite good heuristic
in endgame move. Even
if your planar can give
few more points, it's not a great succes in making
gnugo stronger. So, I
believe that improving
the endgame will not improve the level of gnugo.
At the opposite, it seems that strategic-tactic glue
can improve
significally
the level of any computer program.
Wow, you ask hard questions ;-) I will think about
this in the upcoming days and I will get back to you
on it. Okay?