Re: computer-go: Seki

[Nick Wedd <Nick@xxxxxxxxxxxxx>]

Jeff Massung <jmassung@xxxxxxxxxxxxxxxxx> writes
>Okay, seki question:

< snip >

>My question is this:  are both setups "seki" until white plays at 'a' (in 
>the second setup), and what is the "proof" (in algorithm form) that says 
>that 1 is seki and 2 isnt (if 2 definitely is not seki).

In both cases, the outside surrounding wall has cutting points.  This
affects everything.  Let's get rid of these cutting points:

>1.
 X X O X X
 X O O X X
 X O X X X
 X O O a b
 X X O O O
 . X X X O
 . X . X X

This is seki.  If either player is daft enough to play at a (or b), it
will lose its group.  Otherwise, it stays as seki.


2.
 O X a O O X O
 O X b O X X O
 O X X X X O O
 O O O O O O .
 . . . . . . . 

In this position, there is a simple device whereby White can kill the
black group.  Black can do nothing to prevent this, even if it moves
first.  If this is left on the board at the end of the game, and the
players are competent, the black group is scored as dead.

But if White is a computer program and so not able to state that it
knows the black group is dead, and if Chinese (area) counting is in use,
White would be well advised to play at a after the dame-filling, to make
it perfectly clear that the black group is dead.

You ask "is this group seki before white plays at a?"  If the players
are good, the answer is "no, black is dead".  If their competence is
doubtful, the answer is "it depends on what moves they are going to
make".

Nick
-- 
Nick Wedd    nick@xxxxxxxxxxxxxxxxx