Re: Bayesian boxes and Independence of Scales

From: Marchal <>
Date: Tue May 11 00:31:50 1999

GSLevy wrote:

>I agree that the implicit distribution of m is essential in determining if
>should switch after making our first choice. Interestingly, if we assume a
>logarithmic distribution and compute the expected value of the content of
>other box (the one not selected), we find that it is identical to the
>of the selected box:
>Expected Value = (1/2) (Log(m/2) + Log(2m)) = (1/2) (Log(m) - Log2 + Log(m)
>+ Log2)
> = Log(m)
>Which implies that the logarithmic distribution is the one assumed by common
>sense: don't switch. In other words, the value that any particular number is
>equal to m is equally likely, and INDEPENDENT OF SCALE. This independence of
>scale has some intriguing connection with the MWI and the issue that the
>probability of survival after death is extremely low but FROM THE POINT OF
>VIEW of the observer it must be equal to one.

It seems to me that this remark is very interesting.
I would nevertheless be very
pleased if you could give some more precise hints about the
connection between the mentionned independance of scale and MWI (and the
*immortality* issue).


 Bruno MARCHAL Phone : +32 (0)2 6502711
 Universite Libre Fax : +32 (0)2 6502715
 de Bruxelles Prive : +32 (0)2 3439666
 Avenue F.D. Roosevelt, 50 IRIDIA, CP 194/6
 B-1050 BRUSSELS Email :
 Belgium URL :
Received on Tue May 11 1999 - 00:31:50 PDT

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:06 PST