----- Original Message ----
From: Darren Cook <[EMAIL PROTECTED]>
> Am I confused in my understanding that a weakness of MC evaluation is
> that due to its random play it will miss sequences where there is
only
> one winning move at each play? ...
> This was exactly the topic I tackled in this article:
> http://dcook.org/compgo/article_the_problem_with_random_playouts.html
> The conclusion is that random playouts are weak, and you need to use
> heavy playouts. But, as Remi explained, when you do that you need to be
> careful not to penalize good moves (self-atari in a nakade) when
> penalizing bad moves (self-atari generally).
Well, Amen to that!
I've been thinking about this a bit. The machinery to detect a true nakade on a
move-by-move basis during playouts may be too expensive; it might so slow down
playouts by one or two orders of magnitude.
But what about doing a top-level analysis of the board just before starting
playouts, offering some hints
to the more interesting moves? During the middle game, and especially during
the endgame, top-level
analysis can provide excellent guidance regarding good and bad plays, I would
think.
Such guidance has to be fairly subtle, however; it often must take the form of
"if he plays here, do this, if there, do that."
It also must not be simple pattern matching. If a group has only three enclosed
liberties, playing the center of three is a good idea. If it has a second eye,
such a play might not be useful. Factor in ko threats and threats to cut, etc.
____________________________________________________________________________________
Never miss a thing. Make Yahoo your home page.
http://www.yahoo.com/r/hs_______________________________________________
computer-go mailing list
[email protected]
http://www.computer-go.org/mailman/listinfo/computer-go/
