Post-Election Analysis Of The Ariel Sharon Matrices

Author: Keith York

This article is property of the author and may not be reprinted or distributed without permission.

Posted: February 13, 2001

On February 2, 2001 I posted on this website an article 
( http://www.thebiblecodes.com/news/2001/january/Ariel Sharon.doc ) by Moshe Aharon Shak containing Bible code arrays showing that Ariel Sharon would win the February 6 election.  Those arrays had been the subject of some e-mails between him and myself, and the article had been submitted by Mr. Shak for publication on this website.  While we do not approach the Bible codes in exactly the same way, we do share a belief in many of the important principles governing codes research.  Now that Ariel Sharon has indeed been elected the new prime minister of Israel, I believe it is important to write an analysis of the main array in the article from my perspective.  While this analysis is specifically for the main array in Mr. Shak's article, it also serves an educational purpose by discussing some broader principles in Bible codes research. 

I must first address the issue of the predictive nature of the codes and of this array.  I have written before on this website that "information presented in a Bible code array must be independently verifiable."  Therefore, "the future cannot be predicted using the Bible codes.  Why?  Simply stated, an event is not independently verifiable until after it has happened."  How should the reader interpret this array which correctly predicted the winner in light of the above statement?  First of all, note that this situation had two fixed elements: the date and the candidates.  Moshe himself wrote to me that he would never have been able to find this array before the date of the election was announced.  Once the election date had been announced, a crucial piece of data to search for was in hand.  Secondly, after Netanyahu decided not to run, there were only two possible winners: Barak and Sharon.  (A third possibility was Peres, if Barak had decided to step aside at the last minute.  That is why Moshe searched for Peres in the codes as well.)  Thus two important variables were known beforehand, a situation very rarely present when dealing with future events.

Secondly, when I posted this article Sharon was leading Barak in the polls by double digits.  The prediction seemed like a "safe bet".  (Indeed, Sharon's margin of victory turned out to be the largest in any prime ministerial election in Israeli history.)   However, here is a key thing to remember.  A "safe bet" that hasn't yet happened is inherently different from something that has already occurred and whose outcome is thereby known.  Thus, however good the array looked, before the election it could not be called a valid Bible code array.  Only after the outcome was known could it be called a valid Bible code array.

Actually, the story is more complicated than that.  Moshe first e-mailed me a partial matrix and asked that I not discuss it with anyone else.  He explained how his matrix showed that Sharon would win.  He said, almost in passing, that there were two letters he was not showing me.  After I agreed that the matrix did indeed appear to show a Sharon victory, he wrote back saying that the two extra letters were lamed-aleph (meaning "no" or "not") in front of the ELS meaning "will be chosen".  This, he then claimed, showed that the array meant that Sharon would not be chosen.  This upset me, for I did not want to see Barak re-elected, nor from the then-current polls did it appear that Barak would be re-elected.  Finally, he e-mailed the rest of the material, explaining how "will be chosen" referred to Sharon and "will not be chosen" referred to Barak and Peres.  It was at this point that he said he would allow me to discuss his findings and that he would write up an article detailing them.

The moral of the above paragraph is this.  With incomplete information, a candidate array can lead to erroneous conclusions.  Without knowing how an event will turn out, there is no way to know if one's interpretation of the array is correct or whether additional terms might be there which would radically change the array's message.  Thus again the danger of trying to use the codes to predict the future is underlined.  This is a message that Moshe wanted me to impart to my site's visitors, and one with which I definitely agree.

Having dealt with the criterion that the information in a candidate array must be independently verifiable, I turn to a second criterion I use in my analyses.  This is that all the terms be strongly and definitely related to the subject of the array.  Here I must note that my own approach is very conservative (which is why my arrays typically contain only a handful of terms), whereas Moshe's is more expansive.  For example, if I had been researching this particular array and had seen surface text about being circumcised on the eighth day, I would have ignored it as not being relevant to the subject of the prime ministerial election.  However, Moshe knew something that I did not.  Ariel Sharon was born on a Tuesday and in accordance with Jewish law was circumcised on the eighth day, also a Tuesday.  Furthermore, the election was on a Tuesday.  Array 2.1 in his article shows how this all relates to the central term, the date 13 Shevat.  Thus Moshe's main array is not just about the election on 13 Shevat, but also about other relevant details of the "three kings" Sharon, Barak, and Peres, and the political situation prior to the election.  To repeat, my own approach would have been to find an array on just the election, its date, and its result, resulting in much fewer terms.  Moshe was able to tie together a number of themes in a large matrix and then show how smaller matrices highlighted individual themes.  This is not to say that one approach is right and the other is wrong.  It simply illustrates the differences in the techniques of two experienced codes researchers.  

A third criterion concerns the statistical significance of the terms as reflected in their matrix R-values.  The reader of my previous articles will know that I do not like ELS's in a matrix with negative matrix R-values, particularly if such terms vastly outnumber those with positive matrix R-values.  Moshe's main matrix has 49 ELS's consisting of 44 with positive matrix R-values and 5 with negative matrix R-values.  Though there are a few exceptions here to my "zero-tolerance" rule (and he explains some of these exceptions in his notes), Moshe shows that he agrees with my approach that an array should contain terms that have positive matrix R-values meaning that they are statistically significant.  It is important to remember that a matrix R-value is calculated from the probability that a certain ELS will appear in a matrix in the -N to +N skip distance range.  Since many of the terms are three to five letters in length, they would almost certainly be found by chance in the array if one allowed all possible skip distance ranges for each and every term.  However, one would be very unlikely to find these terms all together in an array with so many at such short skip distances as 1, 2, -3, 5, or 8.  It is this principle of near-minimality of skip distances that makes the consideration of matrix R-values so very important.

How does the matrix as is score when my Threshold calculation is applied?  As noted, there are 49 ELS's.  All terms will be analyzed, including terms I might have found questionable under my more conservative approach and including those with negative matrix R-values.  (Note that the effect here of including negative matrix R-values in the threshold calculation is to lower the final result.)  Out of the 49 ELS's and surface text terms listed in his main array, there are 38 different Hebrew words or phrases.  (A few of the words are found in multiple occurrences.)  [Actually, there are six cases where one term of a pair simply consists of another term with one added Hebrew letter.  An example is the pair 'Sharon' and 'in Sharon'.  There are an additional two cases where one term of a pair simply consists of another term with two added Hebrew letters.  In these eight pairs, each term is counted separately and 0.763 is subtracted from R(sum) for each.]   Of these 38, there are 33 horizontal terms and 5 terms with vertical components.  R0 is the Torah R-value of the central term '13 Shevat' and is equal to -0.516.  R(sum) is the sum of the matrix R-values of all the other terms and is equal to 50.912.  H = 33, the number of horizontal terms.  Finally, the row split R is 2.  Thus,

Antilog [R(sum) + R0 - 0.763H]/R = Antilog [50.912 - 0.516 - 0.763(33)]/2 = Antilog(25.217)/2 = 8.24 X 10^24.

It is important to remember that this is NOT a probability calculation but rather a threshold calculation.  A probability calculation only applies to the results from a controlled experiment containing an a priori search list.  My threshold calculation, on the other hand, is a means that I have developed to determine whether a candidate array which also meets other stated criteria can plausibly be called a valid Bible code array.  In this case, the minimum threshold for an array with 33 horizontal terms is definitely surpassed and I judge that this matrix can very plausibly be called a valid Bible code array.