home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.math:16959 rec.puzzles:7933
- Path: sparky!uunet!stanford.edu!ames!think.com!news!columbus
- From: columbus@strident.think.com (Michael Weiss)
- Newsgroups: sci.math,rec.puzzles
- Subject: The paradox of the second ace (was: Marilyn Vos Savant's error?)
- Date: 15 Dec 92 11:12:55
- Organization: Thinking Machines Corporation, Cambridge MA, USA
- Lines: 65
- Message-ID: <COLUMBUS.92Dec15111255@strident.think.com>
- References: <1gj5grINNk05@crcnis1.unl.edu>
- <1992Dec15.012404.24027@galois.mit.edu>
- <1992Dec15.055832.26324@galois.mit.edu>
- NNTP-Posting-Host: strident.think.com
- In-reply-to: jbaez@riesz.mit.edu's message of Tue, 15 Dec 92 05:58:32 GMT
-
- In article <1992Dec15.055832.26324@galois.mit.edu> jbaez@riesz.mit.edu
- (John C. Baez) writes:
-
- 1) You draw 4 cards from a well-shuffled standard deck. Given that one
- is an ace, what's the probability that they are all aces?
-
- 2) You draw 4 cards from a well-shuffled standard deck. Given that
- one is the ace of hearts, what's the probability that they are
- all aces?
-
- The answers to problems 1) and 2) are NOT THE SAME.
-
- This is similar to the paradox of the second ace: a bridge player announces
- he holds an ace. What is the probability he holds another ace? In a later
- game, a bridge player announces he holds the ace of spades. What is the
- probability he holds another ace? The second probability is greater than
- the first.
-
- (Assume randomly shuffled decks, and all that.)
-
- Martin Gardner wrote into Marilyn Vos Savant with this paradox after the
- famous "Monty Hall" brouhaha.
-
- (Incidentally, John commented on how QM and probability are both
- non-intuitive. I'm too lazy to construct a bosonic version of the boy/girl
- puzzle ("I've got two photons, one of them is a boy...") but it shouldn't
- be too hard.)
-
- Here is a proof of the inequality of the two probabilities that
- avoids lots of grungy arithmetic:
-
- Let:
- a = number of hands having exactly one ace
- b = number of hands having two or more aces
- s = number of hands having the ace of spades and no other aces
- t = number of hands having the ace of spades and at least one more ace
-
- Then:
- prob(two or more aces | one ace) = b/(a+b)
- prob(two or more aces | ace of spades) = t/(s+t)
-
- Now a=4s, as you can show by a simple combinatoric argument. Also
- b<2t, as you can show by a slightly more involved combinatoric
- argument: let b = b2 + b3 + b4, where bi is the number of hands
- having exactly i aces. Also let t = t2 + t3 + t4, where ti is the
- number of hands having exactly i aces, one of them the ace of spades.
- Then
-
- b2 = 2 t2
- b3 = (4/3)t3
- b4 = t4
-
- as simple combinatoric arguments will show. So
-
- b = b2 + b3 + b4
- = 2t2 + (4/3)t3 + t4
- < 2( t2 + t3 + t4) = 2t.
-
- With a little algebra, and the weaker inequality b<4t, you can now show
- that the first conditional probability is less than the second.
-
- Incidentally, this argument does not use the fact that the deck contains 52
- cards, or the fact that a bridge hand has 13 cards. It does use the fact
- that the deck contains 4 aces. However, it can easily be generalized to
- the case of a deck of cards with r suits, for any r >= 2.
-