endgame (Was [computer-go] Re: Should 9x9 komi be 8.0 ?])

[A van Kessel]

> the thing that got me thinking about this is that i've never seen an MC
> player really play out a ko fight.  (or perhaps they are in their own cryptic
> MC way that i can't see).

Well this could easily be solved by *always* investigating
moves that take (or create) a ko.

This of course will force the program to actually *recognise*
the ko's.
A sloppy way to "always investigate" would be to "always try moves that capture".

A similar approach can be used for ladders.
Perhaps snapbacks can be handled this way, too. Not sure.

Some "classical" programs return "special values" in their
evaluation function, such as "can win, but needs ko(s)".
This outcome can be used to steer the upper parts of
the programs.

IIRC there is some stuff about this on Dave Dyer's website.

HTH,
AvK