Re: computer-go: perfect play

["Vlad Dumitrescu" <vladdu@xxxxxxxxxxx>]

> I am suspicious that IMPLICIT here is a bit of a weasel word.
you are right. :-)

> All these insights act synergistically and lead to rigorous reasoning that
> allows us to ignore large parts of the full game tree.
Yes, strictly speaking. No, since those parts of the tree are proven to lead
to a certain result by implicitly playing the sequence. Besides, moves in
the area might become actual later, in the endgame for example, and points
can be lost or won even if a group is dead and buried.

> Of course these
> could all be covered by the phrase 'implicit representation of the game
> tree' (or perhaps be labeled as technicalities ;)), but then so could a
> lookup table that gave a perfect move for every legal position.
If the lookup table is obtained by searching the full tree and writing the
results down, then your assertion does not hold. Otherwise, it's more in a
gray zone...

As you say, it's difficult to describe the meaning of "implicit". From a
point, it becomes a matter of interpretation.

/Vlad