I ran some tests but I don't have access to the specifics right now
(my computer is still in the box from moving, only got internet at
home yesterday). Simply always capture didn't gain any strength.
Capturing with low probability (like 5%-10%) depending on the number
of stones seemed to gain a little bit, but it was never enough for me
to be sure I wasn't looking at noise.
Capturing / defending a neighbour of the last move gained a little bit
I believe. But nothing came close to using the ladder-heuristic, which
is defend a stone in atari if it can escape. Capturing the last move
gains little for me, as it is covered in the first case. (I.e. the
ladder heuristic will escape a stone in atari by capturing the last
move.)
David Fotland said he has a low probability on capture, but I don't
think he ever gave specific numbers.
Mark
2009/4/27 Łukasz Lew <lukasz....@gmail.com>:
> 2009/4/28 Mark Boon <tesujisoftw...@gmail.com>:
>> On Sun, Apr 26, 2009 at 11:54 PM, Łukasz Lew <lukasz....@gmail.com> wrote:
>>> Hi,
>>>
>>> Have any one of you tried uct + capture heuristic?
>>> Is it stronger? How much?
>>>
>>
>> >From memory I'd say it was winning about 80% against plain UCT. But
>> only capture if it can escape (which means it can make three
>> liberties.)
>
> Thanks. Any additional rules in your experiment? For instance no
> capturing stones that can't escape by this definition.
> Do you know what happens if there's no such rule (always capture) ?
>
> Lukasz
>
>>
>> Mark
>> _______________________________________________
>> computer-go mailing list
>> computer-go@computer-go.org
>> http://www.computer-go.org/mailman/listinfo/computer-go/
>>
> _______________________________________________
> computer-go mailing list
> computer-go@computer-go.org
> http://www.computer-go.org/mailman/listinfo/computer-go/
>
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
