Ok, point, I was being too clever.  Can I save face by claiming that the number 
of games is countably infinite as long as you require that all the games be of 
finite length?

I concede that if you allow for games which never terminate, then my method 
will 
not enumerate them.

I should probably stop now before I annoy the Real Mathematicians among us.
--gvc

Eugene Deon wrote:
> *That's like saying you can enumrate the real numbers by starting with 
> the first decimal digit and building a tree with 10 children per node 
> off into infinity.  *
> 
> *I'm quite sure I can find a unique game of scrabble for every real 
> number.  And the decision tree is trivial.  Both players exchange 7 
> every turn.  One decision.  It's what's on the rack that makes it 
> different.*
> 
> 
> Even allowing a game with an unbounded number of consecutive exchanges, 
> it's
> still countably infinite, because you can enumerate not just all the 
> games, but
> all the infinite moves in these possibly-infinitely -long games by a
> breadth-first scan of the decision tree starting with turn one.
> 
> Now, if you want to figure in the length of time taken for each move, 
> you're
> getting somewhere, maybe, until you run into the quantization of time (the
> nature of which has not yet been plausibly postulated to date).
> --gvc
> 
> 
> 
> 

Reply via email to