RABBI GLAZERSON'S CHEMICAL ALI MATRIX & METHODOLOGY
A matrix that brings statistical problems into sharp focus...
In January, 2010 I published a matrix about Chemical Ali after his execution in Iraq. Rabbi Matiyahu Glazerson then informed me that in the Israeli press the murderer’s name was the equivalent of “Ali the Chemical.” He supplied the Israeli transliteration of the Arabic nickname (his actual name was Ali Hassan al-Majeed). The rabbi mentioned that he had a matrix about the man, but did not have a version to transmit. He supplied me with the key terms and skips involved. I reproduced it with CodeFinder software. It is shown below.
The matrix, at first glance, looked to be significant. But one of my initial instructors in Torah Codes, former Pentagon Code breaker Harold Gans, taught me early on back in 1997 to only do probability calculations based on a priori material. Ditto with my other mentor, Dr. Robert Haralick. I have been faithful to their teachings over the past 13 years. And so, when I saw the matrix done by Rabbi Glazerson, and when he asked me to calculate its significance, I had to ask which terms were specified before his search (a priori) and which were found only after the matrix was produced (a posteriori).
With an axis term of CHEMICAL ALI at its 9th lowest skip (13,153), only three terms were a priori. They were HANGED at skip -2, 10 SHEVAT (the day he was hung) at skip +1 and a short version of 5770 (2010), the year he was hung. 5770 was at skip -6,575.
The a posteriori terms were SURELY DIE at skip +1, NAZI at skip +1, MURDER at skip -1 and VILLAIN at skip +1.
In determining the frequency of key words by my normal protocol, when terms appear on a matrix in the open text or at skip +1, I only search the Torah for such words at skip +1. I consider it a best case scenario when an a priori term appears in the open text on a matrix. If the key word is at skip -1, or at the same skip as the axis term, I search the Torah at skips +1, -1, N (the skip of the axis term), and -N. These four skips I consider “special case.” N and –N results in a term being parallel to the axis term.
If the key term is not at a special case skip, I use a combination of the Roffman Skip Formula and skip tables to get a sense of frequency. Finally, when an axis term is not at its minimum skip, I divide the value of the matrix by the rank order of the axis term.
If all the terms on Rabbi Glazerson’s matrix were a priori, the matrix would truly be spectacular from a statistical point of view. The p value would be 6.627 X 10-9 which means that we would only expect it to occur by chance about 1 time in 150,896,055 searches. But, Rabbi Glazerson was kind enough to tell me what he found only after the matrix was produced. And that data dropped its p value to about 0.481. That means that we would expect to find such results about 1 in every 21 times. In fact, most of that significance comes from the date 10 SHEVAT. No other a priori term was significant by itself. The 3-letter form of the year 5770 is not one that I would normally use. It really amounts to just 770, not 5770. There was about an 85% chance to have the 4-letter word HANGED on the matrix.
As for the a posteriori term NAZI, certainly Chemical Ali acted like a Nazi to the Kurds. He gassed 5,000 of them in the Kurdish town of Halabja. But Ali was not an actual member of the Nazi party, so this term is somewhat questionable. Had it been a priori, it would have been the best term in the matrix. NAZI at skip +1 only occurs once in the 304,805 letters of Torah.
So what do we take away from the rabbi’s matrix? If we hold to the normal a priori rules, it is not significant. But the matrix is good enough that it raises the question of whether a new system is needed to rate the relevance of a posteriori terms. Certainly SURELY DIE, MURDER, and VILLAIN would have been appropriate a priori terms had the rabbi chosen to name them first. The reality is that most Codes researchers have an idea of what they are looking for, but don’t start out with a thesaurus in hand to be certain that they name every appropriate term on the first pass. If we assume that the matrix was deliberately intended by the Torah Author, it seems unfair to penalize Him for our failure to think of all terms that He had in mind. However, if we go down this path, we must also adopt something like the methodology of Dr. Robert Haralick. He has a program that names multiple synonyms and spellings for key terms. His overall p values then are based on this larger search group. The problem is that his software is not commercially available. In the past he has also not required a conditional axis term that must be on a matrix, but he has shown flexibility on this issue.
As mentioned earlier, I have worked with Dr. Haralick on Torah Codes for 13 years. He has had many terrific ideas about new software, but taking an idea from a concept or even a working model to one that is both commercially available and as user-friendly as CodeFinder is not an easy matter. There must be a market for such a product. Torah Codes software sales do not support this now, but one effective television documentary (or breakthrough) might change sales potential.