Re: computer-go: Ko

[Peter Drake <drake@xxxxxxxxxx>]

Yikes, you're right!

Someone here had said something about this in their ko-detection 
routine.

Perhaps the correct statement is simply:

A move which is the first to ever occupy a given point cannot be a ko 
violation.

I'll have to fix my code...

On Tuesday, June 3, 2003, at 09:33  PM, Gunnar Farnebäck wrote:

> Peter Drake wrote:
> > Is the following statement true?
> >
> > If a stone played at a point is the first stone to ever occupy that
> > point, no subsequent full-board position can be identical to a
> > preceding position.
> Definitely not, it doesn't even hold for a basic ko. I.e., assume that
> this position has been reached without captures:
>
> ......
> ..OX..
> .O.OX.
> ..OX..
> ......
>
> Black takes the ko, playing the first stone ever at that vertex
>
> ......
> ..OX..
> .OX.X.
> ..OX..
> ......
>
> If O immediately recaptures we get a counterexample.
>
> /Gunnar Farnebäck
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/