>> I.e. it makes the diagonal directions more important, compared to moves
>> in a straight line.
> 
> I try to make that apparent in the presentation, and maybe also in the
> name I use - it creates circle-like structures on the square grid. In
> other words, increments in gridcular metric approximate increments in
> the classical Euclidean metric. At the same time, the increments provide
> quite fine granularity in the area covered, which is also useful in the
> usual application - matching of variable-sized patterns.

But why is that better for go? Have you (or anyone) compared each way,
and this gives some quantitative improvement?

It seems to me that for variable-sized patterns that simple manhattan
distance fits go better.

Darren


-- 
Darren Cook, Software Researcher/Developer

http://dcook.org/gobet/  (Shodan Go Bet - who will win?)
http://dcook.org/work/ (About me and my work)
http://dcook.org/blogs.html (My blogs and articles)
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go

Reply via email to