Re: [gnugo-devel] Inge's patch (fix)

[Inge Wallin]

Gunnar wrote:
> The difference between genmove() and genmove_conservative() is exactly
> whether empty threats should be carried out or not. 

I know that.

> The gtp function
> gg_genmove uses genmove_conservative() because it's more natural not
> to carry out empty threats in a regression setting. 

I didn't know this.  That explains a lot.  I worked with standard
command line switches and that works as it would in a game using ASCII
mode, for instance.  Does this mean that Gnu Go might play differently
on NNGS (using the GTP client) than in a personal game (using cgoban
and GMP)?  If so, I don't think that is a good thing.  If it is only
for regressions, then fine, but during games, the behaviour should be
the same. 

> My patches have nothing to do with this. In retrospect it's easy to
> see that your patch introduced a bug in genmove_conservative:

Yes, the bug was mine.  But then again, I told you so in my mail.

> And indeed your bugfix does solve the problem also with
> filllib:15. Run the filllib test suite and you will see that it all
> passes again.

Beleive me, I have run the tests many times.

> In these particular cases it's easy to switch color by increasing the
> move number by one since there's a pass move next in the game record.
> You would probably have been better helped by a switch which causes
> the LOAD_AND_ANALYZE mode to use genmove_conservative() instead of
> genmove().

But in the general case it is not that easy.  I found it useful, and I
think that I could have used it a number of times in the past too.

        -Inge