[Top][All Lists]

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

Re: [gnugo-devel] Proposed owl_determine_life() improvement

From: Trevor Morris
Subject: Re: [gnugo-devel] Proposed owl_determine_life() improvement
Date: Sun, 03 Nov 2002 10:08:26 -0500

At 07:04 AM 11/3/2002 +0100, Nando wrote:
>handtalk:2  FAIL   Accidental. With --owl-node-limit 1200, the test
>                   succeeds. Another way to deal with it is to raise
>                   a bit the value of D1401 (I tested it has no
>                   other impact than fixing this fail). Not sure
>                   it's the "right thing" to do though.

Tweaking pattern values to improve test results is something that I've
certainly done, and I don't think anyone would object.  I think it's even
a good idea.  As someone (Arend, I think) pointed out, there are a lot
of patterns with the same value, and this doesn't really make sense.
In particular in owl_defendpats, there are 40 patterns with value(35),
including D1401.

Related to this, I think automated tuning of these values is a fine idea.
Likewise, I think automated modification of the owl pattern databases
based on known life & death problems would be useful. 


reply via email to

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