On Mon, Feb 12, 2007 at 11:20:43AM -0500, Weston Markham wrote:
> I think that you are essentially correct.  However, this is only going
> to affect a small number of games where two different moves are
> exactly tied for the best winning percentage, after many playouts.
> Even if the underlying probabilities are exactly the same, you can't
> really expect this to happen much.

I am not sure I agree. In the end of each game there comes a time when
the winner is already certain, or very neaqrly so. Even a MC player will
see this at some point. When that point comes, there is no direction in
pure MC play, except to avoid the worst blunders.  The winning program
is happy to let a tail be cut off here, a small group die there, and a
territory be invaded here, if none of this shakes its unfaltering faith
in its victory. Conversely, the loosing program doesn't even try all
these tricks, as it sees it looses anyway. Both of them play pure
random. It doesn't affect the result, usually. It can be that both
players have misjudged something, and having wasted the winning margin,
that something may turn out to be valuable anyway.

Alas, I don't have my own MC player coded (haven't even started), so I
can not make the experiment myself. If someone here would like to try,
I'd like to hear of it.


Heikki Levanto   "In Murphy We Turst"     heikki (at) lsd (dot) dk

computer-go mailing list

Reply via email to