Re: [computer-go] Perfect sub-tree move path (was Brute Force...)...

[Don Dailey <drd@xxxxxxx>]

>   However,  this  same pressure  acts  on  player  B in  exactly  the
>   opposite way.   Player B's  goal will be  to ensure he  chooses the
>   move  that requires  player A  have  the maximum  number of  future
>   choices  committed  to player  A's  limited  memory.   Player B  is
>   playing  to  maximize Player  A's  fallibility.   Player  B is  not
>   playing to win, lose, or even to tie while player A is successfully
>   remaining on the  "perfect" play track.  Player B  is merely trying
>   to  push player  A into  a board  state for  which player  A (being
>   fallible which implies forgetful) cannot recall the "perfect" move.

>   If  player  B eventually  succeeds  in getting  player  A  to make  an
>   "imperfect" move, the game moves into an state where the outcome is no
>   longer guaranteed, and returns to probabilistic.

Yes, the  game would  change into a  different kind of  game.  However
modern chess  is already something like  that.  Many games  are won or
lost before the players meet over the board.  Great players are always
looking for the "theoretical novelty" that they may only get to play 1
time.  It doesn't  have to be winning, it just has  to surprise and be
difficult to refute over the board.  Later it will either get properly
refuted, or avoided.

The best players would have to work much harder to be best.  Years ago
a natural talent could be great without working too hard.

- Don
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/