[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] NNGS crash
|
From: |
Gunnar Farneback |
|
Subject: |
Re: [gnugo-devel] NNGS crash |
|
Date: |
Tue, 01 Oct 2002 21:53:50 +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) |
Dan wrote:
> This would break nngs1:21. M6 in that position is a very
> bad move.
On the other hand is the replacement move at M3 hardly the right way
to play in that position either. With K3 in place the stones become
more than a little over-concentrated.
> We could of course make it an antisuji or (better) give it a large
> negative shape value.
Even better would probably be to add some influence intrusion pattern.
> When we did the stable release 3.2 we disabled the assertion
> that makes this crash. In other words we accepted multiple
> replacements.
Not multiple replacements but rather an arbitrary replacement,
depending on which one happened to match first. Sure, we could do that
also for the development versions but I can't say it feels
satisfactory.
/Gunnar