gnugo-devel
[Top][All Lists]
Advanced

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

Re: [gnugo-devel] test cases


From: Arend Bayer
Subject: Re: [gnugo-devel] test cases
Date: Wed, 8 Sep 2004 14:45:05 +0200 (CEST)



On Tue, 7 Sep 2004, Evan Berggren Daniel wrote:

> On Tue, 7 Sep 2004, Evan Berggren Daniel wrote:
> 
> > These are some test cases from a game I played against TheGNUGo last night.
> > It managed to make a fairly large number of mistakes, and ended up resigning
> > after the last of them.  CVS passes test 430, so perhaps it is a result of
> > Arend's escape patch? (TheGNUGo is running that, right?)
> > 
> > As always, comments are very welcome.
> 
> Of course, it would help if I included the test cases...

I will add it to CVS, with minor changes:

> Index: regression/kgs.tst
> ===================================================================
> RCS file: /cvsroot/gnugo/gnugo/regression/kgs.tst,v
> retrieving revision 1.8
> diff -u -d -r1.8 kgs.tst
> --- regression/kgs.tst 4 Sep 2004 06:25:43 -0000       1.8
> +++ regression/kgs.tst 7 Sep 2004 19:55:48 -0000
> @@ -177,4 +177,59 @@
>  370 restricted_genmove white H11 G12
>  #? [G12]
> 
> +loadsgf games/kgs/evand-TheGNUGo.sgf 8
> +380 reg_genmove black
> +#? [P8|P12|Q12]
Commented this one out, P9 seems ok to me. (P8 has bad cutting aji.)

> +loadsgf games/kgs/evand-TheGNUGo.sgf 80
> +440 reg_genmove black
> +#? [R18]
While GNU Go shouldn't miss R18, the game move should kill the white
stuff at the top, so it's ok with me.

> +loadsgf games/kgs/evand-TheGNUGo.sgf 176
> +470 reg_genmove black
> +#? [G14|F14]

I changed this to
loadsgf games/kgs/evand-TheGNUGo.sgf 178
470 reg_genmove black
#? [G13]

Thats _very_ urgent and a very typical GNU Go mistake.

Arend




reply via email to

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