Bayes not Occam! (was: feeding hungry people)

[Henry Mu]

Thank you. (Really) I stand corrected. (Truly) I am the weakest link.
(Not exactly) Goodbye! (for the moment)

Henry

-----Original Message-----
From: owner-pynchon-l at waste.org [mailto:owner-pynchon-l at waste.org] On
Behalf Of rhaenda
Sent: Monday, October 15, 2001 6:29 PM
To: pynchon-l at waste.org
Subject: Bayes not Occam! (was: feeding hungry people)

Bayes theorem is really what is in play here, not Occam.  Bayes theorem
treats conditional probabilities in terms of the subjective
probabilities of
simple events.  Without data, the Razor cannot cut.

Let's call E the event that someone is killed by a mine while trying to
get
a food package.  The original assertion (call it the M hypothesis) boils
down to a belief that the media would not report the story of E given
that E
actually occurred and that E was observed by the media.  A different
assertion is that the media would not get the story if E occurred
because of
the low probability that E would have been observed by the media.  Both
of
these subjective probabilities in turn depend on the probabilities that
the
event actually occurred, another subjective probability.

The truth value of the entire proposition depends entirely on the
beliefs of
those evaluating the proposition and not at all on facts until and
unless
the event E actually occurs.  Here's the interesting part:  Suppose that
E
occurs and that the M hypothesis is correct.  Then the event will not be
reported.  On the other hand, if E does not occur, then it will not be
reported and the M hypothesis cannot be rejected.  The M hypothesis may
be
considered true as long as there is no data regarding the probability of
E.
It may also be considered false with the same degree of confidence.



----- Original Message -----
From: <MalignD at aol.com>
To: <pynchon-l at waste.org>
Sent: Monday, October 15, 2001 4:22 PM
Subject: Re: Re: feeding hungry people


> <<In the face of uncertainty based on limited facts, apply Occam's
Razor.>>
>
> This makes no sense. Occam's razor suggests that the simplest logical
> solution, in the absence of contravening evidence, is likely to be
correct.
>
> What is the application here, to a case of someone speaking of his own
> imaginings as if they were true, as if they had already occurred?
>
>