Re: computer-go: Go4++: Passing in tsume searches.

["Michael Reiss" <mick@xxxxxxxxxxxxxxxxx>]

> Hope I understands this correctly
> 
> what about this situation
> 
>   6 . . . . . . . .
>   5 . . * . . . . .
>   4 * * * o o o o . 
>   3 o o o * * * o .
>   2 * . o * . * o .
>   1 . * o . * * o .
>     A B C D E F G H
> 
> White wants to pass and lives but must answer to B2 with A1 to live. 
> This is of course bad for black but does the search terminate as black 
> has an attack move left?
> Maybe you rule this out in your search without looking any further but 
> I dont think all life and death algos does.

My program would not judge this to be a tsume problem because it is 
not neatly bound by one colour. I would judge this to be more like a 
semeai problem. However it can be converted to an analogous tsume
problem like this:

   6 . . . . . . . .
   5 . . . . . . . .
   4 * * * * . . . . 
   3 o o o * * * . .
   2 * . o o o * . .
   1 . * o . o * . .
     A B C D E F G H

If o is to go 1st then its first move must be pass. Then * may play
B2 now o's 2nd move must be A1. This is a perfect example where
the defender must play a pass move which is not the final move.
This answers the second of my two questions perfectly. Than answer
is YES and my original guess was wrong. Thanks very much for 
finding this example.

Now, can anyone find a situation where *two* defender passes 
are required?

Mick.