[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] Patch: Invasion patterns and INVASION_MOVE
From: |
Gunnar Farneback |
Subject: |
Re: [gnugo-devel] Patch: Invasion patterns and INVASION_MOVE |
Date: |
Mon, 07 Oct 2002 21:49:22 +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:
> Unfortunately my laptop is very slow, so I haven't had opportunity yet
> to run the tests, but I ran the first batch and the only difference
> was one FAIL (version 3.3.8). I didn't manage to fix the test I was
> going for (13x13:20), because there is a huge strategic effect on the
> chosen move that makes it impossible to get it right at this time.
> However, I managed to get the valuation of the correct move up to
> something that is in the vincinity of the right score.
The total regression delta is one FAIL and one PASS. The failure,
nngs1:50, is due to some e patterns missing an O classification and is
easy to fix. The pass, nngs2:490, is a recent test case which is
properly solved by the patch, reenabling monkey jumps.
/Gunnar