Thanks Dave, but you're answering the wrong question. I'm not asking which super ko cases exist, but rather which ones can result in infinite random play. The two stone ko you mentioned is resolved with random play because other moves exist that break the cycle (even though they are poor moved, random play will play them)

I'll read through the links to seeif any apply. I did see my triple ko case when looking at it quickly.

Sent from my iPhone

On Jul 22, 2008, at 11:53 PM, "David Fotland" <[EMAIL PROTECTED] games.com> wrote:

There can be more than 3 kos in a cycle. There are some pathological cases of loops involving captures of two-stone groups, but I've never seen this in
a real game.

Here are some example odd positions:
http://www.cs.cmu.edu/~wjh/go/rules/bestiary.html
http://www.goban.demon.co.uk/go/bestiary/molasses_ko.html

David

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:computer-go-
[EMAIL PROTECTED] On Behalf Of Jason House
Sent: Tuesday, July 22, 2008 8:19 PM
To: computer-go@computer-go.org
Subject: [computer-go] Super ko in random playouts

I now tracked down another super ko bug.  I'm curious if anyone has
worked out which infinite cyles can occur in random playouts that avoid
eye-fills and suicides.  Additionally, how do people handle this type
of situation in playouts?  I believe libego checks game length and
assigns "no result" if the game is too long.  This certainly seems
simple enough to do...

A single ko will either be a capture when one side takes it or else
it'll be legal to fill the ko when the other side passes from no legal
moves left.

A double ko can end up with one side owning both ko's, so either the ko
will naturally do a capture, or filling a ko will be legal.

A triple ko occurred in the attached game.  It's 3 ko's between two
eyeless groups.  Each time a color can move, it owns 1 out of 3 ko's,
and has only one legal move option, take the only legal ko (according
to simple ko rules)

Can any other triple ko situations occur?  What about more complex ko
situations?  I think the triple ko is the only case, but I have not
rigorously proven it.

_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Reply via email to