[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] trevor_1_15.2
From: |
Trevor Morris |
Subject: |
Re: [gnugo-devel] trevor_1_15.2 |
Date: |
Wed, 21 Nov 2001 14:26:40 -0500 |
At 10:34 PM 11/20/2001 +0100, Gunnar Farneback wrote:
>Trevor wrote:
>> http://www.public32.com/games/go/trevor_1_15.2
>> - removed "strategic effect".
>>
>> The strategic effect seems to cause almost as many problems as it solves.
>> Here's a list of FAILURES (35) / PASSES (26) with removing the effect.
>> There are a lot of fixmes in this section of code. Is this effect
>> really necessary? If so, it probably needs an overhaul.
>
>It certainly has weaknesses, but yes it's definitely needed. Without
>it you'll for example get no value (except an occasional point of
>territory) for cutting or connecting groups unless the cut or connect
>does kill or save one of the groups.
>
>It may be worth reminding that the regressions almost only cover
>positions where GNU Go has made mistakes in the past. Thus there are
>many examples where the weaknesses in the strategical evaluation cause
>bad moves, but very few examples where the strengths of the
>strategical valuation cause good moves.
Indeed, I ran the trevor_1_15.3 based tests with the strategical
valuation turned off, and there were a number (~8) of serious
mistakes that GNU Go made, just on a 9x9 board that it didn't
make with the strategical values on. On second thought , I should
have probably left those tests in, but I removed them.
In general, I assume that I should err on the side of submitting too
many tests, rather than too few?
Not that I need to say it, but please withdraw this patch submission.
-Trevor