Re: [gnugo-devel] two stone games

[Evan Berggren Daniel]

On Mon, 21 Jul 2003, Paul Pogonyshev wrote:

> Inge wrote:
> > Paul wrote:
> > > the bad thing is that break-in code indeed seems to be quite useless
> > > in at least games against 3.2.
> >
> > Isn't the break-in code basically a defense against something that 3.2
> > never tries?  Then it might very useful in games against humans, but
> > not against earlier versions of GNU go.
>
> maybe so.  but i thought break-in code was also supposed to _break in_.

Well, if without the breakin code turned on gnugo makes overplays that
leave weaknesses that it doesn't see, and never gets punished for it, and
with the breakin code turned on it plays slower, more solid moves, it's at
least possible the breakin code would make it look weaker against 3.2 (or
3.3.23 without breakin).  So perhaps we shouldn't be surprised that it
doesn't help against 3.2?  I agree though, it would be nice if it helped
against earlier versions.  I think, however, that play on nngs and replays
of games would suggest that the breakin code helps against people.

Evan Daniel