Re: [computer-go] SlugGo approach: GNU vs.Goliath

[Gunnar Farnebäck <gunnar@xxxxxxxxxxxxxx>]

Mark wrote:
> On the positive side, there are quite a few cases where the program plays a
> much better move than the original program. I have included a game played
> against Go++ as an example. Goliath plays Black.

The game record seems to say that Go++ is black.

> Until about move 60, Black's play is excellent, much better than
> Goliath would play normally and it has built a large lead already.
> But after that it starts to make occasional big mistakes. Mistakes
> the original Goliath would never have made, and these were all
> caused either by the horizon effect or by an 'accident' elsewhere in
> the sequences. The end-result is disastrous. Other games I tried so
> far show a similar development.

I made some observations about how SlugGo's play compared to ordinary
GNU Go on the GNU Go development list in August. See
http://lists.gnu.org/archive/html/gnugo-devel/2004-08/msg00078.html

You may also be interested in some other parts of the thread, starting at
http://lists.gnu.org/archive/html/gnugo-devel/2004-08/msg00060.html

/Gunnar
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/