Re: computer-go: speculative introspection

[Heikki Levanto <heikki@xxxxxx>]

William Harold Newman <william.newman@xxxxxxxxxxxxxxxxx> wrote:

> Actually, I'd be interested even to know how the brain quickly
> recognizes solid connections from A to B:

>  . . . . . .
>  A X X X X X
>  . . X . X .
>  . . . . X .
>  . . . X X X
>  . . . X . .
>  . . . X X .
>  . . . . B .

This may well be the slower part, as all the pattern matching I speculated
about can happen in parallel. Unfortunately this kind of problem does not
parallelize very well.

I know from my own experience (a mere 5kyu), that as the endgame approaches,
I start scanning the borders between black and white living groups. I have
no other way, but to start in one place and (mentally) run my finger along
the edge until I find a big-looking place, or if I have time, until I have
found the best-looking place.

Luckily in most go positions there is one reasonably easy way from A to B,
and the side branches are usually short and can be skipped quickly. Some
times one may have to check two alternative routes, but that is not so
common. I believe we do them sequentially.

-H



-- 
Heikki Levanto     LSD Levanto Software Development   heikki@xxxxxxxxxxxxxxxxx
               "In Murphy we Turst"