To: cel, ftl, koods, ponzio
Subject: card trick
--text follows this line--
I wrote some code to simulate the card trick
to see how often all ten initial card choices converge to the same 
final card, and I'm getting about 
	32% when the face cards each count as 10,
	84% when the face cards each count as 1,
	80% when the face cards each count as 2, and 
	80% when Jack = 1, queen = 2, king = 3.
	(and 100% when every card counts as 1)
(Based on several thousand shuffles each.)

How does this sound?  (Not good enough for lecture?)

I haven't tried to see how many of the 10 initial choices fail to agree with 
the top card (or with the majority) in the bad 20%
(i.e., to see how badly the trick fails when it fails, 
or to see how changing the restriction of choosing from among
the top 10 cards affects the confidence).

Steve