A Brief Look at Statistics and Year Dates in the Bible Codes
Since much of the prophetic value of the Bible Codes may be related to providing a time at which a future event might occur, it’s important to know at least roughly the probability of randomness affecting our interpretation of year dates showing up in Bible Code matrices. I’ve run four experiments to estimate the impact on the expectation of finding Bible Codes by year from HC 5759 through HC 5779 and found significant differences in expectation for the different dates, differences that cannot be ignored as we attempt to extract meaning from simple Codes.
The experiment is very simple. I used each HC year date (including the “heh”) from 5759 (AD 1999) through 5779 (AD 2019) as the main search term in the Torah to develop Table 1 (see below).
An N/A in the column indicates that the number was too large for the software to determine (over 10,000). The normalized expectation was determined by multiplying the expectation over the Torah by 0.00265…, or the ratio of the number of characters in a typical matrix divided by the total number of characters in the Torah. It is not in any way a real measure of the expectation of seeing a year on any given page at any given magnification or skip number, nor have I seen a good way to calculate it.
The expectation estimate for decade years, e.g. 5760 and 5770 must be corrected by subtracting the expectation for all of the following years in the decade since the software can’t really distinguish between a unique 5770 and the 5770 embedded in 5771, the 5770 embedded the 5772, …, and the 5770 embedded in 5779.
The statistical data is straight from Computronics Bible Codes 2001 edition.
Also note that for every year where I could do “found” instances to compare with the expectation, more were found than expected and in several years (colored) there were enough extra finds to put the odds of random chance at about one in a million or better. Those may be very important years, or a natural variation, but most consider odds like one in a million suspiciously intentional.
Q: I’ve found a prophecy and a year date associated with it. Is that the year it’s going to happen?
A: Maybe. Maybe not.
Q: How can I tell?
A: Well, that’s the $64,000 question, isn’t it?
(1) First you have to decide the basic philosophy of G-d and the history of mankind. Do you believe there are really codes in the Torah and Tanach? Do you believe in free will for yourself? Do you believe G-d can see the future without influencing it? Do you believe in predestination, and on what scale? Do you believe that history is more or less than the sum of all individuals exercising free will? Your answers to these questions make a difference.
(2) It’s clear to me that if the year designator falls within a very long code and is self-consistent within the code, then the validity of that date has the same order of believability as the overall code in which the date is embedded.
See http://biblecodes.us/pages/endofdays.pdf for an example of a very long code with year dates in it, if the site is still available. In summary, there are 7 long codes (>20 characters), 3 of which have an unambiguous 4-character designator for 5770, and no other year to compete. Further, the discoverer, Roy Reinhold, states that he looked across a 30-year span and found no other unambiguous dates in the general area of the matrix to compete. The probability that these codes exist together is beyond accident. The year is good. Hebrew year 5770 is likely a milestone in the end of this age and the transition to the next.
(3) In a simple code as published by Drosnin and also by many others, where the central term is on the order of 10-characters, and all of the other terms deemed valuable are on the order of 5-characters, then you have to decide based on the number of competing dates (there’s often more than one), the skip number of the date compared to the central term, the probability of a random date being found, and the distance from the central term to the date. A fourth test is consistency between the matrix under consideration and all of the other matrices with similar central terms. Remember, depending on the question, there may be more than one valid answer, there may be no valid answers, or there may be one-and-only-one valid answer. In general, I prefer year dates with a strong vertical component as near as possible to the central term, preferably sharing a letter with the central term or crossing it, with no visibly credible conflicting years. You develop a “feel” for these things after looking at a thousand matrices or so. Sometimes there’s just no good year available.
(4) Even if you have an unquestionably valid date, do you know when that occurs in the common calendar of the era, the Gregorian calendar? That depends on whether Hebrew calendar dates are ceremonial dates counted from 1 Nissan (also 1 Aviv … usually in March) or whether the dates are civil calendar dates counted from 1 Tishri (usually in September). For half the year, 1-Tishri through the end of Adar(s), it doesn’t matter much because the ceremonial calendar and the civil calendar will have the same year number. From 1-Nissan through the end of Elul the calendars will be different by a year. Based on my own birth date and Bible Code matrices, my prejudice would be for using the ceremonial calendar. My birth date is 23 March 1949 Gregorian, 22 Adar 5709 HC (civil), and 22 Adar 5710 HC (ceremonial). Working backwards, I find evidence to support the 22 Adar 5709 date, and also a 22 Adar 5710 date. The 5710 is found roughly 20 times more often than the 5709 date, and it is expected more often. An example of one is hardly proof. Most Bible Code researchers use civil calendar dates based on the same kind of argument. Dates that are found which correlate to real events are in civil Hebrew calendar dates, and the software is based on civil dates. It’s hard to search for ceremonial dates. Will dates regarding the advent of the Messiah or other religious and legal events be in the common civil calendar or in the ceremonial calendar as commanded by G-d? We don’t know right now. We can only speculate. It appears to me that the answer just may be “both” if you know how to search, and there may be a preference for the ceremonial calendar which hardly anyone is using.
(a) Many of the year dates in our recent past and near future are present in significantly larger numbers than one would predict by probability theory. We live in interesting times. Did anyone doubt it?
(b) The year 5770 (AD 2010) – see Table 2 below - is potentially present in such large numbers that it is almost impossible to estimate its probability of significance statistically. If the alphabet and numbering system were actually designed with coding in mind, then 5770 is a significant year just by its shear numbers. Otherwise, 5770 is so common “by accident” that its appearance in a simple code might be completely disregarded based on a high probability of random occurrence, and the same might be said for many other years, e.g. 5760.
(c) The year 5776 (AD 2006) is present in numbers so much larger than expected (sigma > 8) that 5766 is very likely to be a significant year. A. McCracken’s interpretation is that there may be 400-500 extra dates over expectation that are probably encoded for information based on a private email, and L. Piperov of Bulgaria has found other evidence for significance based on his soon to be published manuscript. The year 5776 is also present in small enough numbers that if you see it in your matrix, the odds are pretty high that it’s meaningful.
(d) One important date we know fairly well, still in our future, is the potential destruction of Jerusalem during (at the start of?) Word War 3. The date is 9 Av 5766. (See these Exodus2006 web pages - page 1 page 2 page 3). 9 Av 5766 translates into 3 August 2006 no matter what assumptions you make about calendars. It’s in that part of the year where the Hebrew civil calendar and the Hebrew ceremonial calendar have the same year date. If it matters, the Gregorian calendar and the US government fiscal calendar also bear the same year date.
(e) Other dates in our future may not be so clear-cut, depending on which side of 1 Nissan (1 Aviv) they fall. We should be neither disappointed nor relieved if an event “scheduled” for 5770 (nominally 2010) falls as early as Sep 2009 or as late as Mar 2011.
Respectfully,
Al Sutton, California, USA, 1330 local time, 28 June 2004, 10 Tammuz 5764
Table 1
Full Year Dates in the Torah (includes the “heh” for 5th Millennium)
|
HC |
Nom |
Gregorian |
Gregorian |
Unique |
Normalized |
FOUND |
Significance |
|
Year |
From HC Civil |
From HC Ceremonial |
Expectation |
Expectation |
in |
in sigma |
|
|
Torah |
|||||||
|
5759 |
1999 |
Sep 1998 - Sep 1999 |
Mar 1999 - Mar 2000 |
1,767 |
5 |
1,953 |
4.43 |
|
5760 |
2000 |
Sep 1999 - Sep 2000 |
Mar 2000 - Mar 2001 |
36,106 |
96 |
N/A |
N/A |
|
5761 |
2001 |
Sep 2000 - Sep 2001 |
Mar 2001 - Mar 2002 |
3,439 |
9 |
3,845 |
6.93 |
|
5762 |
2002 |
Sep 2001 - Sep 2002 |
Mar 2002 - Mar 2003 |
2,077 |
6 |
2,419 |
7.50 |
|
5763 |
2003 |
Sep 2002 - Sep 2003 |
Mar 2003 - Mar 2004 |
268 |
1 |
321 |
3.24 |
|
5764 |
2004 |
Sep 2003 - Sep 2004 |
Mar 2004 - Mar 2005 |
894 |
2 |
1,013 |
3.99 |
|
5765 |
2005 |
Sep 2004 - Sep 2005 |
Mar 2005 - Mar 2006 |
3,566 |
9 |
3,959 |
6.59 |
|
5766 |
2006 |
Sep 2005 - Sep 2006 |
Mar 2006 - Mar 2007 |
3,878 |
10 |
4,385 |
8.14 |
|
5767 |
2007 |
Sep 2006 - Sep 2007 |
Mar 2007 - Mar 2008 |
279 |
1 |
317 |
2.25 |
|
5768 |
2008 |
Sep 2007 - Sep 2008 |
Mar 2008 - Mar 2009 |
914 |
2 |
981 |
2.23 |
|
5769 |
2009 |
Sep 2008 - Sep 2009 |
Mar 2009 - Mar 2010 |
229 |
1 |
254 |
1.63 |
|
5770 |
2010 |
Sep 2009 - Sep 2010 |
Mar 2010 - Mar 2011 |
222,503 |
592 |
Table2 |
Table 2 |
|
5771 |
2011 |
Sep 2010 - Sep 2011 |
Mar 2011 - Mar 2012 |
21,106 |
56 |
N/A |
N/A |
|
5772 |
2012 |
Sep 2011 - Sep 2012 |
Mar 2012 - Mar 2013 |
12,749 |
34 |
N/A |
N/A |
|
5773 |
2013 |
Sep 2012 - Sep 2013 |
Mar 2013 - Mar 2014 |
1,645 |
4 |
1,705 |
1.48 |
|
5774 |
2014 |
Sep 2013 - Sep 2014 |
Mar 2014 - Mar 2015 |
4,585 |
12 |
5,972 |
6.58 |
|
5775 |
2015 |
Sep 2014 - Sep 2015 |
Mar 2015 - Mar 2016 |
21,884 |
58 |
N/A |
N/A |
|
5776 |
2016 |
Sep 2015 - Sep 2016 |
Mar 2016 - Mar 2017 |
23,800 |
63 |
N/A |
N/A |
|
5777 |
2017 |
Sep 2016 - Sep 2017 |
Mar 2017 - Mar 2018 |
1,714 |
5 |
1,909 |
4.70 |
|
5778 |
2018 |
Sep 2017 - Sep 2018 |
Mar 2018 - Mar 2019 |
5,607 |
15 |
6,121 |
6.86 |
|
5779 |
2019 |
Sep 2018 - Sep 2019 |
Mar 2019 - Mar 2020 |
1,407 |
4 |
1,513 |
2.82 |
ADDENDUM 1
Table 2
Statistics for the First 18,000 Skip Codes for 5770
|
5770; 2010 |
||||
|
skip min |
skip max |
expectation |
found |
sigma |
|
1 |
1,000 |
6,209 |
6,294 |
1.07 |
|
1,001 |
2,000 |
6,148 |
6,426 |
3.55 |
|
2,001 |
3,000 |
6,087 |
6,368 |
3.61 |
|
3,001 |
4,000 |
6,025 |
6,384 |
4.62 |
|
4,001 |
5,000 |
5,964 |
6,189 |
2.92 |
|
5,001 |
6,000 |
5,902 |
6,101 |
2.59 |
|
6,001 |
7,000 |
5,841 |
6,020 |
2.34 |
|
7,001 |
8,000 |
5,779 |
5,957 |
2.34 |
|
8,001 |
9,000 |
5,718 |
5,820 |
1.35 |
|
9,001 |
10,000 |
5,657 |
5,831 |
2.32 |
|
10,001 |
11,000 |
5,595 |
5,886 |
3.89 |
|
11,001 |
12,000 |
5,539 |
5,765 |
3.03 |
|
12,001 |
13,000 |
5,478 |
5,734 |
3.46 |
|
13,001 |
14,000 |
5,411 |
5,752 |
4.64 |
|
14,001 |
15,000 |
5,350 |
5,662 |
4.27 |
|
15,001 |
16,000 |
5,288 |
5,494 |
2.83 |
|
16,001 |
17,000 |
5,227 |
5,535 |
4.26 |
|
17,001 |
18,000 |
5,165 |
5,597 |
6.01 |
|
1 |
18,000 |
102,383 |
106,815 |
13.85 |
The potential importance of the statistical significance of the year 5770 is too nagging to ignore it just because the software is unable to handle more than 10,000 codes at a time. I ran the first 18,000 skips in the Torah in increments of 1,000 as shown in Table 2. Frankly, I got tired after running 18 trials, hand recording the results, and then looking forward to more than 100 additional runs. A cursory look at Table 2 has some interesting features worth mentioning.
(1) For the first 18,000 skips, every interval of 1,000 has a monotonically decreasing expectation, as expected.
(2) For the first 18,000 skips, every interval of 1,000 has more found codes than expected.
(3) Although each interval is significant beyond even odds, the totals reflect an outstanding suspicion that the number 5770 is significant beyond a doubt at 13+ sigma.
Even the lazy can do a least-squares best-fit extrapolation to estimate the aggregate for a full experiment. I ran a few trials at the highest skip numbers numbers to anchor the line at the high end, and then ran a full table of “expectations” and “founds” similar to Table 2. Without boring you with a 130 column table, the results suggest approximately 3,000 more unique codes for “5770” exist than expected for a estimated significance of 6-7 sigma.
Conclusion: 5770 is a statistically significant code in its own right at better than 1 in a million odds as well as the most common date code in the Torah of the 20 tested.
