[Top][All Lists]

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

Re: [gnugo-devel] Opponent follow-up value and group connection

From: SP LEE
Subject: Re: [gnugo-devel] Opponent follow-up value and group connection
Date: Tue, 13 May 2003 14:59:28 -0700

I'm still running the regression test. I got an unexpected fail on
trevora.tst 300 (correct B8 got B7), so far.

About the threshold value, I don't know where I can put a variable. For the
moment I think a constant is enough. At least my problem is solved :).

SP Lee

----- Original Message -----
From: "Evan Berggren Daniel" <address@hidden>
To: "GNU Go development" <address@hidden>
Sent: Tuesday, May 13, 2003 2:02 PM
Subject: Re: [gnugo-devel] Opponent follow-up value and group connection

> On Tue, 13 May 2003, SP LEE wrote:
> > This is a more careful treatment for larger dragons. Please ignore my
> > previous patch. The fix here is like "forget the shape factor if either
> > attacking or defending a dragon of more than 20 points is involved."
> > kind of play can be seen even in professional games.
> Interesting idea, I like it.  Two comments:  you probably don't want a
> hard threshold like this.  I think you want to completely ignore it for
> large dragons, and reduce the impact for mid-sized dragons, and keep it
> entire for small dragons.  Something like a linear scaling between 8 and
> 20 point dragons, or maybe something more complex.  Secondly, do you have
> regression results for the patch?  Is the regression suite lacking in
> testcases in this area (I think it is, but I'm not sure...)
> Evan Daniel
> _______________________________________________
> gnugo-devel mailing list
> address@hidden

reply via email to

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