[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: computer-go: perfect players



On Wed, May 02, 2001 at 02:19:42PM -0400, Fant, Chris wrote:
> I think it is useless to talk about a perfect player as one that can read
> minds.  That will never be approached by human or machine.

I don't think it is a black and white as that. We can have the "clean"
perfect player, that always finds the best move in a position. (two such
ones would agree on the result without playing a stone. Boring!) 

Or we can have one that makes observations and assumptions on the style of
the opponent, and finds a move most likely to win against the current
opponent. (Such one might play on in a lost position, if the opponent had
shown that he can make a mistake in complex fighting, and direct the game
towards such).

Or we can have one that takes the history of the opponent into account. (I
think chess (on the highest level) is played a lot like this.)

Or, hypothetically, we can have one that actually reads the mind of the
opponent, sees what he has missed, and plays accordingly.  (I agree, we
won't be seeing them in the foreseeable future).


I am sure there can be variations in between...

-Heikki


-- 
Heikki Levanto  LSD - Levanto Software Development   <heikki@xxxxxxxxxxxxxxxxx>