RE: computer-go: Engineering (was: Most simple Go rules)

["Mark Boon" <tesuji@xxxxxxxxx>]

> -----Original Message-----
> From: owner-computer-go@xxxxxxxxxxxxxxxxx
> [mailto:owner-computer-go@xxxxxxxxxxxxxxxxx]On Behalf Of Don Dailey
> Sent: Friday, June 29, 2001 6:13 PM
> To: computer-go@xxxxxxxxxxxxxxxxx
> Cc: computer-go@xxxxxxxxxxxxxxxxx
> Subject: Re: computer-go: Engineering (was: Most simple Go rules)
>
>
>
>    From: "Mark Boon" <tesuji@xxxxxxxxxxxxxxxxx>
>
>    In any ruleset there must be an arbiter who gives the final score. It's
>    necessary since one player might miscount (something a
> computer wouldn't
>    do?). I think the arbiter giving the score is very natural.
> How do you get
>    the score using Tromp/Taylor rules when playing online?
>
> I don't have any problem  with your protocol,  because it has the nice
> feature that it's optional.  I think your protocol actually works well
> for Tromp/Taylor.  But that's  because I believe you  are exaggerating
> the convience of Japanese scoring.
>

You're right that I'm exagerating a bit. But I do prefer the Japanese rules
in practise (for humans, that is). If I somehow always magically had exactly
180 stones in my bowl at the start of a game (and not lose any during),
without needing some weird contraption, I would probably prefer the Ing
rules. I believe most of the really bad stuff, like point fractions in seki,
have been fixed.

These two rule-sets are always highly criticised here for their theoretical
flaws. But they are used all around the world by millions of people who are
all not even aware of all their intricacies. And still they manage to bring
a game of Go to an end, each and every time, without any major problems.
Discussing rules in theory is fine by me, but when we're discussing
practical rules to play by, be it for computers or humans, any theoretically
sound rule that makes it more complicated to finish a game, even if that
complication is minor, is useless.

The complication of the Chinese rules is that one needs to count much larger
amounts, something humans are very bad at. The Ing rules have the problem
they add a hundred moves to a finished game, which I find a problem
especially for online systems. The Tromp/Taylor rules need to continue the
game beyond where most people are willing to go, and the Japanese rules have
the problem of (protentially) negotiating over dead stones. The scheme I
propose will make all these rule-sets usable for online systems with the
exception of the Ing rules.

What it all comes down to is a reasonably reliable program to decide which
stones are dead. At the moment I'm using Goliath, even though it's not
without flaws. And there may already be programs out there that can do it
more reliably than Goliath. But at least developing a program that can
reliably do just that is very feasable.

Oh, by the way, I don't know what the "pass-for-ko" rule is in the Japanese
ruling, but I doubt I'll ever need to. It's certainly no reason for me to
assume the Japanese rules are unusable.


>    No matter what you guys argue, I'm absolutely convinced that
> if people have
>    the option of playing to the bitter end, or they get the (same) result
>    several dozen meaningless moves before that, a vast majority
> of people is
>    going to opt for ending the game early. I'm trying to give
> both options: if
>    there's a potentially problematic situation you can always
> play it out. For
>    the other 95% of the cases it should be possible to finish earlier.
>
> This is the strength of your end game arbitration protocol.
>
>
> Don
>