Re: computer-go: Bad Tactical move in MFG

[Daniel Bump <bump@xxxxxxxxxxxxxxxxxx>]

> The way this issue can about is I remember hearing that a single stone stone
> surrounded on three sides with it's only liberty two spaces away from the
> edge was provably dead, but I don't see that.  That is how the files I sent
> start out, but we see that black lives.  Did I make an incorrect move for
> white somewhere?  Is is it not true that black is provably dead in that
> starting configuration?

Obviously your B stone was at the 5-3 point and after B descends, W
pushed it towards the corner. W apparently wasted a move connecting
unnecessarily at 'a'. Omitting this move W wins quickly.

|......       [Your diagram.]
|.OOa..
|OXXO..
|.XO...
|XXO.OO
|OXXOOX
|O.XXX.
 ------

Alternatively W can also push B the other way. After a few moves it
looks like this:

|..D...             B's only hope is to clamp at A. W connects at
|.A....             B, B connects under at C and W continues at D.
|COB...             B can thrash never make eyes.
|.XO...
|.XO...
|.XXO..
|.OO...
|......
|......
|......
 ------

Dan