🚀 go-pugleaf

"Judson D. McClendon" <ju...@mindspring.com> wrote:

>[regarding]  the original article in Statistical Review. 
>To put it simply, the significance is not just that the names and
>birth/death dates were found, because one would expect to do so in a
>sufficiently large text.  The significance of the statistical finding is
>that there was an enormously significant correlation (on the order of
>100,000 to 1) that the birth and death dates were found in 'close
>proximity' to names to which they belong, but not to the names with which
>they did not belong.
>
>The actual correlation is somewhat analogous to selecting several dozen
>prominent people, and for each person writing the name, birth date and
>death dates on three identical pieces of cardboard and placing all of the
>pieces into three bowls, one bowl for names, one for birth dates and one
>for death dates.  Then you repeatedly draw one piece of cardboard from each
>bowl and find that every time you are drawing a name and the dates which
>match that name!  

I am a layman and not an expert in probability and statistics.  I know
enough to be dangerous, and in this field it is easy to go astray. 

The above characterization is close to the results as presented by
Witztum and Rips.  They defined a metric to assess the "closeness" of
a name to the dates of birth and death, and they found that the metric
was significantly less for the correct permutation.  The referees were
statisticians, untrained in Jewish scholarly texts, and as such did
not see any flaws in the study.  The results appear to be of extremely
low probability, but there were multiple variants of each name and
date.  More important is Witztum and Rips' selection of a metric,
which is far from an obvious choice.  If you multiply the probability
obtained by the numbers of variant readings accepted and by the even
larger number of feasible metrics, the results are not going to be as
surprising as the paper indicates.  

>This situation could only happen if somebody messed with
>the order of the cards to make it happen.  And that is the point, that no
>such correlation would be present unless some 'intelligence' arranged it
>so.  And this particular 'intelligence' knew the future, because all the
>people in the test were born many hundreds of years after Genesis was
>written.

Low probability events do not necessarily imply "intelligence" with
knowledge of the future.  Consider a lottery, where the odds against
any particular person winning could be astronomical.  But somebody has
to win!  

In the case of the paper under question,  the experimenters appear to
have been honestly motivated to find some result of seemingly low
probability.  But the assumptions are not always clear.   I suspect
that the results of this study are analogous to the case of an archer
who fires an arrow and then paints the target around the spot where
the arrow landed.  In my admittedly less than expert opinion,  the
study was sincerely intended and unwittingly fallacious.
--
David  J  Ellis
92 Wilson Drive / Framingham, MA 01702
d...@mkitso.ultranet.com