[Top][All Lists]

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

Re: [gnugo-devel] more testcases

From: Arend Bayer
Subject: Re: [gnugo-devel] more testcases
Date: Sat, 21 Dec 2002 00:12:56 +0100 (CET)

On Tue, 17 Dec 2002, Evan Berggren Daniel wrote:

> On Mon, 16 Dec 2002, Evan Berggren Daniel wrote:
> > >From two nngs games today.
> >
> > Three from the game against aburry, and the rest from a game against thor
> > (NR).  Gnu Go takes 9 stones from thor and loses by 85 points, due to a
> > number of mistakes.  These should probably be reviewed by someone stronger
> > than I am for correctness.
> >
> > Thanks
> >
> > Evan Daniel
> A couple minor changes.

I've looked through these, and commited parts of it. A lot of tests were
in situations unclear to me, and I have omitted them for now, as I
didn't have time to investigate the positions in depth.

Generally speaking, it is probably most productive if you omit
situations that are unclear to you. If you are not sure about the best
move, it is probably too hard for GNU Go anyway.
Also, the [!H3]-kind of condition should best be omitted. They've always
caused high regression noise, and with Gunnar's restricted_genmove we
can now do better.


reply via email to

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