🚀 go-pugleaf

In soc.culture.jewish on Fri, 12 Sep 1997 16:09:43 -0600 Helge
Moulding <h...@slc.unisys.com> posted:


>Judson D. McClendon wrote:
>> You apparently haven't studied the original article 

>You are right. I read it on the web, and my understanding of the
>author's procedure may be flawed. However, nothing you've
>written so far has helped me figure out what, if anything, I've 
>missed.

>> 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.

>That is a test I looked for, but did not find. The authors seem
>to have relied on scrambling the names or the text. If they did
>compare the frequency with which they could find the names "near"
>the correct dates as opposed to "near" some arbitrary dates, I
>never read about it. Can you point out the passage to me?

>I merely found the test to be one of being surprised to find that
>it was possible to construct lots of instances of common names
>using a text that contained lots of instances of the same common
>names, and then finding that these instances often were "near" 
>the birth or death dates associated with real people who actually
>had these common names.

If the math is done right, all of this is taken into consideration.


>If you'll look at the table of names, you'll find that most names
>have a Hebrew version in common with some other name. I suspect 
>that there are millions of Jews who were given these common names,
>and that their importance to Judaism has nothing to do with the
>likelihood of being able to construct their names and dates using
>ELSs from the Torah. This is even easier when you notice that the
>Hebrew ways of spelling a given date are at least as varied as the
>ways to spell these names.

Here, too, if the math is done right all this is included.  If they
chose different formats or sets of formats for the dates and then ran
the statistics for each one and chose the set of formats that gave the
best results, that would invalidate things, but they chose the set of
all normal date formats in advance and ran those statistics from that
premise.   I haven't seen a written description of Witzum's experiment
but today at a bookstore I notice that Drosnin had about a 6 or 7 page
abstract with formulas.  Didn't have time to read it though.  I wonder
if Drosnin makes clear enough that he didn't run such analyses on his
codes.

> And then add in the point that the 
>method used by the authors to chose their names is about as 
>haphazard as looking for a name in the phonebook by pointing at
>random. 

>In any case, you seem to be convinced that the presence of these
>codes means something. That opinion is probably off topic for half
>of the groups to which this thread is crossposted, so let's leave
>it out of the discussion, ok? What's on topic is the question if
>the discovery of these "codes" is in fact a surprising statistical
>anomaly. 

>I am saying that given the parameters, the codes are about as 
>unusual as finding two people with the same first names and 
>birthdates in New York City. The authors' haven't convinced me
>otherwise.

Your example of New York City show how it really is necessary to read
the mathematics.  No one here has even started to remotely scratch the
surface of how that was done.
>-- 
> Helge "Statistical certainties are flukes!" Moulding
> mailto:h...@slc.unisys.com                Just another guy
> http://www.geocities.com/Athens/1401     with a weird name

P&M

mei...@erols.com

        I miss many posts so please e-mail if you want me to answer.

          Judge all men favorably. Mishna Avos 1:6 from Lev 19:15,
B'tsedek tishpot amitecho. You shall judge your neighbor with righteousness.