Comparitive Experiment: Matching a Nonsense Term with Baal Zephon

With respect to the word (pronounced Shoah) for Holocaust, while it is true that it occurs at an 50-letter Equidistant Letter Sequence (spacing) or ELS where the rabbi indicated, it is also true that at this same interval (either forwards or backwards), it occurs in 33 other places in Torah.  If the interval was 51 letters rather than 50, it occurs 24 times.  At a spacing of 52 letters, it occurs 25 times, and at 53 letters, it occurs 34 times. If we search between skips -100 and +100, we find it 2,411 times.  Between skips 1,000 and -1000, it pops up 24,941 times.  Indeed, 4-letter words that like shoah (which do not contain rare letters) are so easy to find at an ELS, that they often fill a computer screen.  As such, their use is generally of very limited value.

Further, even when a term has eight letters (and ELS terms are rarely longer), they are only of statistical value when paired with a-priori terms - that is, terms specified before the computer search was begun.  But do they prove anything when they are so paired?  Again, the question can not be intelligently answered without a lot of experience, or without a statistical analysis.

Let us, for example, create as an axis term a nonsense word composed of only two letters - alef and bet, but have the two repeated 4 times.  In English our axis term looks like this: ABABABAB.  How many times does this occur at an ELS in (wrapped) Torah between skips -150,000 and +150,000?  The answer is 48 times.

Normally I start out by leaving the program at a default matrix size of 80 columns by 50 rows with the row split function enabled.  If we do this here, we obtain 14 matches of ABABABAH with Baal Zephon (and another two matches of ABABABAB with ABABABAB).  How good are these meaningless matches of a nonsense term with a real place?  For ABABABAB with Baal Zephon, we find that all 14 matches were indeed with an open text Baal Zephon.   In area (rows times columns), the sizes of the matrices were as follows:

If we were to increase the allowed area to more rows and/or more columns, the number of matches would grow.  If we decrease the allowed matrix area, the numbers of matches obviously shrinks.  Which of the above ELS finds for ABABABAB will still have a match if we turn off the row split function of CodeFinder, and what will the areas of the matrices look like if we still leave the program in the default setting of 80 columns and 50 rows?  There are now only five matches for Baal Zephon (with improvements in area shown in blue).  Results are as follows:

So, as we can see, it is not that hard to obtain some nice matches of a nonsense axis term like ABABABAB with a term that occurs only three times in the open text.  It is also interesting that the best match (smallest matrix area - just 102 letters) occurred with the minimum ELS of ABABABAB.