gnugo-devel
[Top][All Lists]
Advanced

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

Re: [gnugo-devel] atari_atari:17


From: bump
Subject: Re: [gnugo-devel] atari_atari:17
Date: Wed, 20 Nov 2002 20:53:06 -0800

> I believe atari_atari:17 is a bad test.  The reason is that there is not a
> combination attack against any particular group.  S10 is an attack on
> either S12 or Q10.  After w S10, b T10, w Q9, b S9, the S12 group is safe.
> Or, after w S10, b q9, the Q10 group is safe.  When written as a genmove
> test or attack_either test, gnugo gets the right answer.  The following
> patch comments out the testcase.  Should it be rewritten somewhere else?

>From the comments to atari_atari:

/* atari_atari(color, *move) looks for a series of ataris on
 * strings of the other color culminating in the capture of
 * a string which is thought to be invulnerable by the reading
 * code. Such a move can be missed since it may be that each
 * string involved individually can be rescued, but nevertheless
 * one of them can be caught. The simplest example is a double
 * atari. The return value is the size of the smallest opponent
 * worm. 

The position in atari_atari:17 fits this definition exactly.

Dan




reply via email to

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