Re: computer-go: perfect play

[Roland <goprog@xxxxxxx>]

Hello Vlad,

Tuesday, September 26, 2000, 3:46:58 AM, you wrote:

VD> *** First, I think it is obvious that perfect play is not possible without
VD> knowledge of the full game tree. This knowledge can be explicit or implicit

That is the big question, it is not that obvious as it first seems to
be. I think there are two possibilities:
- we mathematically prove that perfect play is only possible with
full game tree search. In this case the best we can hope to find is
some suboptimal solution.

-we prove that we can find the perfect move without needing the full
game tree. Then still we have to find the method to find this move and
hope that this method is simpler than full game tree search.

VD> There is actually an alternative to that. Someone might discover a smart
VD> analogy with some mathematical model where things can be proven using
VD> mathematical methods (just about the way it is done for the simple endgame
VD> in "Mathematical Go Endgames"). In my opinion, at the present this
VD> alternative is also out of reach for us, non-omniscient people. But
VD> developing such a theory is at least possible for a human being, so all hope
VD> is not gone :-)

RIGHT.

VD> Thus it is beyond anyone's power to prove anything about the strength of the
VD> perfect player.

Well, someone has presented a quite interesting prove based on
statistics and standard deviation...


-- 
Best regards,
 Roland                            mailto:goprog@xxxxxxxxxxxxxxxxx