Re: computer-go: Most simple Go rules

[Robert Jasiek <jasiek@xxxxxxxx>]

David Fotland wrote:
> I guess that makes me a pessimist :)

:)

> I haven't read tromp/taylor for a while.  Do they require that all dead
> stones be removed
> before passing?

This is their strategic consequence because they score once
alternation ends with two successive passes.

> If so, then I expect that no human uses them.

At the moment this is roughly right. Some centuries ago it was
standard to play out everything.

> Doesn't it seem
> strange to ask programs to play go using a procedure that no human
> uses? 

It is almost the same as what humans use. Humans (if we
neglect the intricacies of Japanese style particularities
for the moment) tend to use an additional phase of
optional agreements. For humans this is nothing but a so
called shortcut to removing so called dead stones by
continued alternate play. Humans imagine all the strategies
of removing by alternate play, i.e. they play out 
hypothetically. If they agree on their strategic evaluation,
then they remove the so called dead stones without
demonstrating a typical alternating continuation of playing
it out. If the disagree, then they not only imagine what a
typical alternating continuation of playing it out would look
like but they perform one. Either humans do imagine the same
as a computer does by playing it out or they actually do the
same as a computer does by playing it out.

The only difference is a modern tradition of human belief
that an agreement tends to be more convenient than an
actual performance. Need computers develop the same sort of
belief in convenience? I don't think so.

> I'm trying
> to make a program that plays go (by which I mean the game that people
> play), not some
> other game designed for computers, even if it is very similar :)

The game is so similar that it is strategically the same
and differs only in a view on convenience.

> Do you really think you will get the Ing foundation or the Japanese go
> association
> to change the rules they use for computer go?

Yes!

I have met Mr. Yang of the Ing Weiqi Educational Foundation and
everything he could tell me about Ing rules was "But they are
in that rules booklet!" With such a weak understanding of the 
quality of the own rules I expect a soon change in politics.

Concerning the Japanese go association, I do not really know
about any details but even they have become aware of 
discrepancies in their rules. Since those are more apparent in
computer go than in human play, we can expect changes for CG
much faster than for human Japanese tournaments.

--
robert jasiek