RE: computer-go: Live or Die

["Sheppard, Brian" <BSheppard@xxxxxxxxxx>]

> The algorithm is correct under the following two assumptions:
>
>1. The set of eye patterns is complete and the attributes for center,
> border, etc. are correct.
>2. If there is some attacking or defending move that works, then there is a
>pattern that suggests one that works.

Thanks for the algorithm, Tim.

Can you tell us the "degree of correctness" of these two assumptions, both
within your program as it stands today and in the theoretical limit of your
current approach?

Thanks,
Brian