Re: computer-go: Perl Module for next move.

[Daniel Bump <bump@xxxxxxxxxxxxxxxxxx>]

> Ascii Interface
> Even if you do not have CGoban installed you
> can play with GNU Go using its default Ascii
> interface. Simply type gnugo at the command
> line, and GNU Go will draw a board. Typing
> help will give a list of options. At the
> end of the game, pass twice, and GNU Go will
> prompt you through the counting. You and GNU
> Go must agree on the dead groups--you can
> toggle the status of groups to be removed,
> and when you are done, GNU Go will report
> the score. 

This is a reliable way to get the score but it requires
human intervention. To get the score automatically, you
can use --mode gtp and the command `final_score.'

> What discourages me is the fact that gnugo needs human input in deciding
> live/dead status. If this was a problem only rarely encountered I guess I
> could live with it. So I'm asking, how good is gnugo in determining
> live/dead status? Can it be run without user intervention? I'm not afraid to
> get my hands dirty and hack through the code if needed to make small
> modifications in order to achieve complete automation.

If there is no seki on the board the chances are that GNU Go can
get you a reliable score. But GNU Go may play correctly (filling
liberties, then stopping at the right time when there are no more moves)
yet not understand clearly whether seki stones are alive or dead. Thus
the reported score may be wrong. And I think there are cases not
involving seki where the reported score may be wrong.

The module aftermath.c is Gunnar Farneback's attempt to get a
reliable score even when there is a seki. We want to have
reliable scores before we release 3.0.

As to how often you get a wrong score, Heikki Levanto has some
results about that. Around April 25 he played 100 games between
2.7.231 and 2.6 and found 8 games where they reported different
scores. Either 2.6 or 2.7.231 gave a wrong score, and as I recall
it was 2.7.231 in many cases. This percentage is too high.

There have been revisions to aftermath.c and I do not know
how reliable it is at this time. I think the answer is,
not completely.

Daniel Bump