RE: computer-go: Most simple Go rules

[Tom Cooper <coopert@xxxxxxxxxxxxxxxxxx>]

-----Original Message-----
From:	Robert Jasiek [SMTP:jasiek@xxxxxxxxxxxxxxxxx]
Sent:	Thursday, June 28, 2001 8:56 AM
To:	computer-go@xxxxxxxxxxxxxxxxx
Subject:	Re: computer-go: Most simple Go rules



<snip>

All known arguments against superko do not maintain any 
justification if considered carefully. The only exception is
preferring traditional precedental rulings for the sake of
adoring traditional precedental rulings as such. Clearly,
computer programs' AI and history are not sufficiently 
developed for such a traditional culture!




I think that this may be over stating your case somewhat.
There are some people who dislike superko on aesthetic
grounds.  To say that these are not justified is subjective.
Perhaps you are using a more precise definition of argument.

How about this for an argument?  I would like (if I had more
time and talent) to write an arbitration programme of the
type discussed earlier in this thread.  I would like this programme
(and the machine it runs on) to be provably correct.  This may
be more than I can hope to do but I think it might have its
uses: for automatically
running go competitions with significant prize money for instance.

I don't think that this is possible with the superko rule because of
the large amount of storage that would be required to remember
which positions had already appeared.

Worse than this, I don't think it would even be possible (with superko)
to make an arbiter that wasn't vunerable to malicious players whose
goal was to try to make it missbehave.

I suspect that these considerations would seem too theoretical to
someone who likes the superko rule.  Also, I am aware that automatically
ruling a draw for a repeated position is vunerable to the same objection.
Nevertheless, I think that this is a valid objection to superko even if it is
not as convincing as some other objections to other rules.

Tom.