home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.math:17068 rec.puzzles:7983
- Path: sparky!uunet!wupost!waikato.ac.nz!canterbury.ac.nz!math!wft
- Newsgroups: sci.math,rec.puzzles
- Subject: Re: Marilyn Vos Savant's error?
- Message-ID: <BzDpn8.AI0@cantua.canterbury.ac.nz>
- From: wft@math.canterbury.ac.nz (Bill Taylor)
- Date: Thu, 17 Dec 1992 01:04:19 GMT
- References: <1992Dec15.052211.24395@CSD-NewsHost.Stanford.EDU><1992Dec15.160000.3714@cs.co <MARTIN.92Dec15232645@lyra.cis.umassd.edu>
- Organization: Department of Mathematics, University of Canterbury
- Nntp-Posting-Host: sss330.canterbury.ac.nz
- Lines: 50
-
- In article <MARTIN.92Dec15232645@lyra.cis.umassd.edu>, martin@lyra.cis.umassd.edu (Gary Martin) writes:
-
- |> Take a bag with a bazillion pennies and a bazillion nickels. Randomly
- |> draw 100 coins and put them in a much smaller bag. Randomly draw
- |> 99 coins from the smaller bag. Having observed that 99 of them are
- |> pennies, what is the probability that the remaining coin is a nickel?
- |> If it were a nickel, then 99 times out of 100 you would have seen it
- |> already, so having seen 99 pennies would have been a freak occurrence.
- |> But if it were a penny, then the event you just observed would be a
- |> certainty. We have two hypotheses to choose from: one which makes
- |> our observation inevitable and one which makes it a freak accident.
- |> Which hypothesis is more likely?
-
- If this last line is meant to imply it's overwhelmingly more likely that we had
- 100 pennies in the small bag, this is yet another boo-boo in this most boo-boo
- infested of all hoary threads.
-
- The fact is, we were overwhelmingly more likely to have had just 99 pennies in
- the small bag to begin with, than 100. The probabilities are almost exactly
- binomial(100 , .5) , in fact, if a bazillion is as big as it sounds.
-
- The probability of getting an observed 99 pennies from a 100-penny bag is much
- bigger than getting 99 pennies from a 99-penny bag, of course; but this exactly
- cancels the other bias.
-
- P(100 pennies in small bag | at least 99) 100
- ---------------------------------------------- = ---
- P(just 99 pennies in small bag | at least 99) 1
-
- and...
-
- P(observe first 99 coins are pennies | 100-penny bag) 100
- ----------------------------------------------------- = ---
- P(observe first 99 coins are pennies | 99-penny bag) 1
-
- These combine by the usual conditional probabilty calculation to give
-
- P(unobserved coin is a penny | 1st 99 are pennies) = 1/2 ,
-
- which is the common sense result, deducible in a few seconds. This all has little
- to do with the original problem, however, where (in effect) the 99 observables
- are chosen deliberately rather than randomly.
-
- -------------------------------------------------------------------------------
- Bill Taylor wft@math.canterbury.ac.nz
- -------------------------------------------------------------------------------
- Galaxies - results of chaotic amplification of quantum events in the big bang.
- Free will- the result of chaotic amplification of quantum events in the brain.
- -------------------------------------------------------------------------------
-
-