Re: [gnugo-devel] Weakness calculation

[Gunnar Farneback]

Inge wrote:
> When I think of it some more, maybe it is not so strange after all. 
> In the dragon2_data struct, we have the following two declarations:
> 
>   int moyo_size;                    /* size of surrounding influence moyo, */
>   float moyo_territorial_value;       /* ...and its territorial value        
> */
> 
> Maybe we can get what we want by just saying that the moyo is defined
> as moyo_size - moyo_territorial_value.

No, I don't think that is a useful quantity at all.

> The inside and the outside is also not always that well defined,
> especially not for a group in the center. What is strange, though is
> that moyo_size is an int. Shouldn't that be a float as well?

The definition of moyo_size is the number of vertices in the
surroundings of the dragon (including the "inside") with influence
strong enough to qualify for moyo classification, which clearly makes
it an integer. The moyo_territorial_value is roughly an expectation of
how many points of territory the surrounding influence area will end
up as.

/Gunnar