Re: computer-go: perfect play

["Vlad Dumitrescu" <vladdu@xxxxxxxxxxx>]

> Whew, my question about the difficulty of chess and go and distance of
play from a
> perfect go-omniscient entity aptly coined "GoGod" (which I did not know at
the time)
> has sparked a very interesting debate.
<snip>

A very interesting argument!

Just one note: the "handicap against GoGod" issue was not as you perceived
it - it's "how many stones will secure the victory against GoGod for all but
the most novice beginners (that might kill their own groups)?" Then 1 stone
is obviously not enough! :-)

As for emulating human play, I don't agree completely. There are (at least)
two main reasons to develop a game-playing program: to solve the game (and
then you are right!), or to gain insight into how the human brain works (and
then even tic-tac-toe is interesting, even if the outcome is known).

I make a correction for an earlier affirmation I made: it is true that an
empty board is either a winning position for black, for white or a draw
(komi excluded). This is true for any deterministic game with complete
information - so in fact any of them are boring for the perfect player. The
conclusion is then that GoGod and ChessGod are playing bridge!? ;-)

Or maybe when GoGod meets GoDevil over a beer and some chips, they play Go
on a board of infinite size - that should give them enough to do!

Luckily for us mere mortals, even if it were proved that the perfect play is
always a win for one of the players (given that we use komi), from here to
actually discovering the perfect moves there's a huge, step to take.

cheers,
Vlad