[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] A new owl testcase
|
From: |
Gunnar Farneback |
|
Subject: |
Re: [gnugo-devel] A new owl testcase |
|
Date: |
Sat, 30 Aug 2003 22:35:01 +0200 |
|
User-agent: |
EMH/1.14.1 SEMI/1.14.3 (Ushinoya) FLIM/1.14.2 (Yagi-Nishiguchi) APEL/10.3 Emacs/20.7 (sparc-sun-solaris2.7) (with unibyte mode) |
SP Lee wrote:
> This problem is from the game between gnugo and go++, posted recently by
> Max. Go++ didn't find the killing move. While I asked gnugo, it couldn't
> solve it either.
>
> A B C D E F G H J K L M N O P Q R S T
> 19 . . . O X X . . . . . . O . . . X . . 19
> 18 . . . O O X X . . X X X X O . O O X . 18
> 17 . . . . O O X . X O X O O . O . O X . 17
> 16 . . O + . . X . O O O . . . . O X . . 16
> 15 . . . . . O X . . . . . . . . O X . . 15
> 14 . . . . . . O . . O . . O . O X . . . 14
> 13 . O . . O . O . X . . . . . . X . . . 13
>
> White can simply play J18 then H19 to make a nakade.
Basically this is the well-known "cutting point inside eyespace"
problem, which causes GNU Go to systematically overvalue the eye
potential. However, in this position black can answer white J18 with
H16 to either escape with half the dragon or get ko for the whole
dragon.
I think we skip over this test case since we already have several
cleaner ones for essentially the same problem.
/Gunnar