Re: [gnugo-devel] gnugo 3.4 problems

[bump]

> > I think that indirect kos are quite common in semeai situations
> > but less so where only one group is at stake, though there is
> > the complication that there could be internal ko threats. If
> > there is only one group, usually one fights a ko to gain an
> > eye. Such a ko is direct since whoever wins it can immediately
> > settle the issue.
> 
> That is not always the case. See for example ld_owl:177. We also have
> at least one very good example in the regressions that I haven't
> managed to locate. I'll keep an eye out for examples.

Here is a simple idea that might be sufficient.

We would track two things: first, who gets to take the ko
first (as in our existing KO_A, KO_B classification), and
second, how many extra tenukis the person losing the ko
gets. (This could be negative.)  This data would be used in
alpha/beta search of move trees involving ko. It would be
necessary to get the extra parameter into the Hashnode data
field.

Dan