Re: computer-go: perfect play

["Vlad Dumitrescu" <vladdu@xxxxxxxxxxx>]

> > "Knowledge of the full game tree" does not necessarily involve having
all of
> > it available at any time. Should the search be depth first, one could
stop
> > when finding a winning move, yes. But the point is that one has to be
> > prepared to search the whole tree anyway, because there might not be any
> > winning move at all.
> You are about to re-invent A-B-cut.
:-) yes. The point was that in the worst case, pruning doesn't work and one
has to search all variations anyway.

Of course, when talking about the full search tree for the game of Go,
pruning by a factor of 359, or even 359E10 is only of theoretical
importance, since the remaining tree is too huge to handle anyway...

> And you don't need any fancy evaluation function, you only need to
evaluate end positions.
yes, if one only relies on the full search.

But it would be much more fun to have the perfect evaluation function, and
need not do any search at all!