Re: [gnugo-devel] trying some OWL tuning

[Arend Bayer]

Portela Fernand wrote:

> The problem I'm looking into is 13x13.tst:20
>
> At first sight (the Regression failure html views are just great!), it seems
> that the moves connecting the F6 and G5 stones are largely overvalued. With
> some tracing, it is possible to see that GNU Go considers the dragon L5 as
> WEAKLY_ALIVE (?!) and hence appreciates greatly to connect it to a healthier
> one.
>
> Noticing the big lunch, I started looking at the code and found 2 spots in
> dragon.c where I could make a small testing change in order to make the
> engine more confident about this dragon.

Lunch is not what you think it is, and it is GNU Go's fault you
misunderstood :-) Lunch is supposed to be a critical worm whose capture
would guarantuee an eye. I think the problem here is that the worm at K3 isn't
considered dead. (This is due to a misanalysis by small_semeai() here.)

If I understood it correctly, Gunnar is working on an eye patch that
would (among other things) teach GNU Go the difference between a half
eye and a place where there is two eyes if we move first but none
otherwise. These should places should be treated identically to a big
lunch, so maybe it is worth reconsidering big lunches then.

Arend