Re: computer-go: Ko and solving 5 x 5 perfect play

[Don Dailey <drd@xxxxxxx>]

To build a database, you do  not have worry about infinite loops.  You
can  ignore ko  by  making all  moves  legal.  Since  you are  working
backwards  in layers,  you will  not encounter  ko or  infinite loops,
because  your only  concern is  to score  positions that  haven't been
scored.  Each unscored position  will eventually connect to a position
that has  already been scored.   It's possible that this  will require
that we  take care  not to  reference positions that  are at  the same
layer (or depth) of the table during the build procedure.

So the  database can built  without (or with little)  consideration of
Ko, the question is, "does this make the database invalid?"

I'm not sure, I need to think about this some more. 

Don



   Date: Mon, 21 Oct 2002 09:54:59 +0200
   From: Robert Jasiek <jasiek@xxxxxxxxxxxxxxxxx>
   Organization: jasiek
   X-Accept-Language: en
   Content-Type: text/plain; charset=us-ascii
   Sender: owner-computer-go@xxxxxxxxxxxxxxxxx
   Precedence: bulk
   Reply-To: computer-go@xxxxxxxxxxxxxxxxx


   Andrew Derrick Balsa wrote:
   > I mean that (without any proof of whether this is true or false) my gut
   > feeling is that finding a solution for the problem of perfect play for
   > any board size can be made independent of the set of rules used to avoid
   > infinite loops (kos).

   The feeling is wrong.

   > Or in other words, that I would rather try to find solutions for perfect
   > play with a simple set of ko rules

   It depends on what you call simple:)

   > and having solved that,
   > worry about what happens to my solution if I change the rules of ko,

   Possible, OC. The changes can be severe.

   --
   robert jasiek