Re: computer-go: Most simple Go rules

[Nick Wedd <nick@xxxxxxxxxxxxx>]

In message <3B39C922.22148FA7@xxxxxxxxxxxxxxxxx>, Erik van der Werf
<E.vanderWerf@xxxxxxxxxxxxxxxxx> writes
>Nick Wedd wrote:
>> 
>> It is quite possible for a reasonable game on a 19x19 board to take over
>> 1000 moves (500 each).  Such a game might involve an early molasses ko
>> (see http://www.britgo.org/rules/molasses.html#mol) with four moves
>> being played in the molasses ko for every one that progresses the rest
>> of the game.
>> 
>> I see no reason to prevent such games.
>> 
>
>On the same page it states that for such cases in Japan 'if it appeared
>in a game, it would be judged as a "no result"'. 

Only if the play were such that the position repeated.

>The way I see it there are 3 solutions:
>1 play on and decide by super/meta-ko
>2 draw (if EACH player has placed more then ... stones)
>3 Have a third pary (judge) declare no result (and replay the game)
>
>For humans I prefer 3, for computers 2. 

Infinite looping is a problem.  It can be addressed in any of the ways
which you list.

Games that take 1000 moves or more are not a problem.  I am strongly
opposed to having special rules, either for humans or for computers,
that do not address any problem but just make the ruleset more
complicated.  A rule that ends the game after some specified finite
number of moves is an example.

Nick
-- 
Nick Wedd