home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!pipex!bnr.co.uk!uknet!acorn!armltd!dseal
- From: dseal@armltd.co.uk (David Seal)
- Newsgroups: sci.math
- Subject: Re: Marilyn Vos Savant's error?
- Message-ID: <10914@armltd.uucp>
- Date: 16 Dec 92 17:22:12 GMT
- References: <1gj5grINNk05@crcnis1.unl.edu>
- Sender: dseal@armltd.uucp
- Distribution: sci
- Organization: A.R.M. Ltd, Swaffham Bulbeck, Cambs, UK
- Lines: 108
-
- In article <1gj5grINNk05@crcnis1.unl.edu> burchell@cse.unl.edu (David
- Burchell) writes:
-
- >I ask the net Gods: Can this be correct??
- >
- >From ``Ask Marilyn'' by Marilyn Vos Savant, Parade magazine, December
- >13, 1992.
- >
- >You have a hat in which there are three pancakes. One is golden on
- >both sides, one is brown on both sides, and one is golden on one side
- >and brown on the other. You withdraw one pancake and see that one side
- >is brown. What is the probability that the other side is brown?
- >
- >---Robert H. Batts, Acton, Mass.
- >
- >It's two out of three. The pancake you withdrew had to be one of only
- >two of them: the brown/golden one or the brown/brown one. And of the
- >three brown sides you could be seeing, two of them also have brown on
- >the other side.
-
- The usual problem with probability questions comes up yet again... People
- almost never state exactly what is being randomised or what the distribution
- is, leaving it up to the reader to make assumptions. In this particular
- case, you need an assumption about the distribution of the pancake chosen:
-
- Assumption 1: the three pancakes are chosen with probability 1/3 each.
-
- and one about which side of the pancake you see:
-
- Assumption 2 - version A: you saw a random side of your pancake, with the
- two sides being chosen with probability 1/2 each.
-
- Assumption 2 - version B: the words "one side is brown" are actually
- intended to mean that you look at both sides and see that at least one
- side is brown (then conveniently forget whether both sides are!).
-
- Assumption 2 - version C: the brown/golden pancake is always placed under
- the hat brown side up. When you pulled your pancake out, you did so by
- sliding it out, in such a way that you were certain to see the top side
- first.
-
- Using assumption 1 and assumption 2(A), the stated solution is correct.
- There are 6 *equiprobable* possibilities:
-
- Pancake chosen Side seen Colour seen
- ----------------------------------------
- B/B Top B
- B/B Bottom B
- B/G Top B
- B/G Bottom G
- G/G Top G
- G/G Bottom G
-
- Knowing that we saw a brown side tells us that one of the first three cases
- is correct, and gives us no further information. These three cases are
- therefore still equiprobable, so the chance that the other side is brown is
- 2/3.
-
- OTOH, using assumption 1 and assumption 2(B) gives us an answer of 1/2. The
- equiprobable cases are now:
-
- Pancake chosen Side seen Brown seen?
- ----------------------------------------
- B/B Both Yes
- B/G Both Yes
- G/G Both No
-
- Knowing that a brown side has been seen now tells us that we've got one of
- the first two cases, which gives a probability of 1/2.
-
- Assumption 1 and assumption 2(C) also give an answer of 1/2, since there are
- now three equiprobable possibilities:
-
- Pancake chosen Side seen Colour seen
- ----------------------------------------
- B/B Top B
- B/G Top B
- G/G Top G
-
- The observed facts tell us that only the first two are still possible.
-
- Or if you want a silly case, try:
-
- Assumption 2, version D: this is just like assumption 2(C), except that the
- brown/golden pancake is always placed under the hat golden side up. Now the
- possibilities are:
-
- Pancake chosen Side seen Colour seen
- ----------------------------------------
- B/B Top B
- G/B Top G
- G/G Top G
-
- and the probability wanted is 1, since the observed facts exclude all but
- the first possibility. Of course, you don't need to use probability for
- this: under these assumptions, you can deduce logically that the other side
- of the pancake is brown.
-
- So to summarise: the error is really in the problem, in that it doesn't
- state important facts about how the "experiment" was set up. There's no real
- error with the stated solution, though it would have been nice if the
- assumptions being made about the problem were stated explicitly rather than
- having to be read between the lines.
-
- David Seal
- dseal@armltd.co.uk
-
- All opinions are mine only...
-