gnugo-devel
[Top][All Lists]
Advanced

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

Re: [gnugo-devel] Semeai


From: Gunnar Farneback
Subject: Re: [gnugo-devel] Semeai
Date: Mon, 28 Jul 2003 20:51:53 +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)

Inge wrote:
> Notice that the owl defence is at T18 (correct) but the semeai defense
> is at M14 (wrong)!  I let Gnu go do detailed trace using -t -t , and
> the analysis never even tried T18 as a semeai defense move.

The semeai analysis partially uses the same move generation as the owl
analysis but it also tries some other moves (typically liberties of
the opponent) and sometimes changes the move ordering. Node limits can
also cause a move found by owl never to be tested by semeai.

> Now, my question is: wouldn't it be natural to try the owl defense
> move as the semeai defense move first in any semeai?  Or have I missed
> something? 

It's probably a reasonable idea but it might have issues with e.g.
caching. Feel free to try to implement it.

/Gunnar




reply via email to

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