[gnugo-devel] I think I've found a bug.

[Daniel Olszewski]

Hi.

I have worked a bit with gnugo source code using it for my own purposes (I  
was trying to develop efficient program playing atari-go). Anyway during  
my tests I've encountered strange situation. I got two symmetrical layouts  
that are absolutely different evaluate by gnugo (I attach sgf files).  
Interesting is that, besides gnugo determines different moves for both  
situations, also life and death evaluation is slightly different (checked  
with -T option).
I'm curious if you already know this problem (if it is a bug at all) and  
what is the reason?

Best regards.
PS. Sorry for my English.
--
Daniel Olszewski

----------------------------------------------------------------------
PS. >>> http://link.interia.pl/f19a6

Human-Computer2(7x7).sgf
Description: Binary data

Human-Computer22(7x7).sgf
Description: Binary data