[computer-go] MoGo

[Erik van der Werf]

On 4/5/07, Sylvain Gelly <sylvain.gelly at m4x.org> wrote:
>
> > > Sylvain, could you run the same test on 7x7 to verify that there the
> > > 'correct' komi would be 9 (try 8.5 vs 9.5)? If MoGo wouldn't converge
> > > to 9 we probably shouldn't have much confidence in the generalisation
> > > of the above results for higher levels of play on 9x9 (or you could be
> > > on your way to discover an error in the human solution for 7x7 :-).
> > > Either way I think it would be interesting to know what comes out...
> >
> Ok, here are the results (again MoGo_3k against itself):
>
> 6.5
> 3399/5200 65%
> 7.5
>  2397/4800 50%
>  8.5
> 2603/5200 50%
> 9.5
>  1848/5200 35%

That looks quite good!

Assuming no seki's (with neutral intersections), which IIRC is in
agreement with the human solution, the winner on 7x7 follows from:

Black intersections > (7*7 + komi)/2

so for 7.5 komi black needs to get at least 29 intersections, and the
same holds for 8.5 komi. Consequently, your statistics confirm 9 komi
for jigo.

Thanks,
Erik