home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.math:14718 sci.physics:18607
- Path: sparky!uunet!think.com!ames!agate!physics.Berkeley.EDU!aephraim
- From: aephraim@physics.Berkeley.EDU (Aephraim M. Steinberg)
- Newsgroups: sci.math,sci.physics
- Subject: Re: Three-sided coin
- Date: 10 Nov 1992 18:57:13 GMT
- Organization: University of California, Berkeley
- Lines: 44
- Message-ID: <1dp0m9INNkq6@agate.berkeley.edu>
- References: <5413@daily-planet.concordia.ca> <1992Nov10.032643.10467@galois.mit.edu>
- NNTP-Posting-Host: physics.berkeley.edu
-
- In article <1992Nov10.032643.10467@galois.mit.edu> jbaez@riesz.mit.edu (John C. Baez) writes:
- >In article <5413@daily-planet.concordia.ca> mckay@alcor.concordia.ca (John McKay) writes:
- >>
- >> von Neumann was asked, when about to enter a taxi,
- >>
- >>"How thick should a coin be to have equal (=1/3) probability
- >> of landing on its head,tail,or edge?"
- >My best shot at answer invoked an idea from stat mech, namely,
- >equipartition. This says that the solution occurs when the *energy* of
- >standing upright equals the energy of lying on edge. Guess this means
- >the radius should equal half the thickness of the coin, or if you
- >prefer, the diameter should equal the thickness.
-
- Isn't this only true if one assumes that the number of microstates
- corresponding to each situation is the same? For example, a molecule
- in thermal equilibrium is not likely to be in the rotational ground state,
- simply because the degeneracy goes as 2J+1 (or I guess 2\nu +1 or some
- such is the standard notation), making higher J's somewhat more likely.
-
- I would have said that if you can somehow ignore all dynamics (or perform
- the coin toss in a viscous fluid, say, though I in no way claim this is
- _necessary_) the important thing is for the solid angle pointing along
- the edge to be the same as the solid angle along each side.
-
- Thus each side must subtend 4\pi/3 steradians, meaning that the angle from
- the axis of cylindrical symmetry to the ``diagonal'' must be
- arccos (1/3), and the radius divided by half the thickness must be
- the tangent of that angle (by the way, I drew a picture in front of me
- and did some trig-- von Neumann I'm not) or 3\sqrt{1-1/9} or 2\sqrt{2}.
-
- The radius should be \sqrt{2} times the thickness, goes my guess.
-
- Note that this would make the surface area of the edge \sqrt{2} times the
- surface area of each face, contradicting my original guess that the area
- determined the likelihood... I stick with the solid angle approach, though.
-
- Any experimentalists out there?
-
-
- --
- Aephraim M. Steinberg | "WHY must I treat the measuring
- UCB Physics | device classically?? What will
- aephraim@physics.berkeley.edu | happen to me if I don't??"
- | -- Eugene Wigner
-