Re: [gnugo-devel] MAX_LIBERTIES

[Paul Pogonyshev]

Nando wrote:
> > A more interesting question is how this change interacts with also
> > reducing MAX_LIBERTIES.
>
> It almost looks like you were expecting it (I wasn't) : the effect is
> actually amplified. I ran regressions twice with CVS (which includes
> the MAX_LIBERTIES reduction) and twice with the patch, and the worst
> case indicates an improvement of a bit more than 3%.

Interesting, I was expecting it going the other way round.  But this
is of course the case where it is pleasant to be wrong, 3% is a very
good result ;)

Paul