JFK Assassination: A Probability Analysis of Unnatural Witness Deaths


Richard Charnin (TruthIsAll)


Feb. 25, 2013  

Click this link to the most recent analysis update.




Original post: May 18, 2012


There has been much discussion and controversy regarding the large number unnatural witness deaths that occurred in the year following the 1963 JFK assassination and during the 1976-77 House Select Committee investigation of the JFK and MLK assassinations. The deaths were a combination of homicides, suicides, accidents and undetermined origin. The HSCA determined that both murders were probably due to conspiracies. The JFK Unnatural Witness Deaths Probability Spreadsheet  covers the period from 1963-1983. Most deaths occurred during the years in which the assassination was investigated: 1964 (Warren Commission) and 1977 (House Select Committee on Assassinations).





Suppose that on Nov. 22, 1963, 1400 individuals were selected at random from the entire U.S. population. Further suppose that within one year, 15 would die unnaturally under mysterious circumstances. Based on unnatural death mortality rates, only 1 in a group of 1400 would be expected to die.  But at least 15 did.


There are three possibilities. The 15 deaths were...

1) Unrelated. It was just a 1 in 167 trillion coincidence.

2) Unrelated. The individuals were selected in a scam to fool the public into believing that the assassination was a conspiracy.

3) Related. There was a common factor, a connection between them.


We can confidently rule out 1) and 2).

Then if the 15 were related, what was the connection?


Once you have eliminated the impossible, whatever remains, however improbable, is the truth – Arthur Conan Doyle


In any given year, only one unnatural death would be expected in a random group of 1,400. 

The probability that at least 15 of 1,400 randomly-selected individuals would die unnaturally is 1 in 167 TRILLION (mathematical proof below).


In the 3 years following the assassination, there were at least 33 unnatural deaths (only 2 would normally be expected).

The probability is 1.4E-33 or 1 714,705,498,316,173,300,000,000,000,000,000  

The odds of at least 33 UNNATURAL deaths of 1400 in a THREE year interval is 1 in 137 TRILLION TRILLION.


In the 14 years following the assassination, there were 70 unnatural deaths (11 would normally be expected).

The odds of at least 70 UNNATURAL deaths of 1400 in a FOURTEEN year interval is 1 in 714 MILLION TRILLION TRILLION.

That number is greater than all of the stars in the universe and grains of sand on earth.


The mathematical probabilities calculated in the JFK Witness Deaths Database are powerful evidence that the deaths could not have been a coincidence.

There had to be a connection between them.  Approximately 70 of the 100 deaths were unnatural (all were extremely suspicious).





The book Who’s Who in the JFK Assassination  presents vital information on each of more than 1,400 individuals (from suspects to witnesses to investigators) related in any way to the murders of President John F. Kennedy, Dallas Police Officer J. D. Tippit and alleged assassin Lee Harvey Oswald on November 22 and 24, 1963. It is based on years of research using a wealth of data sources and a detailed analysis of the Warren Commission's twenty-six volumes. This encyclopedic study includes entries on virtually all of the suspects, victims, witnesses, law enforcement officials and investigators involved in the assassination.


This is a summary of JFK-related deaths:



The original data source is:



Lee Harvey Oswald, the alleged assassin, must also be included in the list. Oswald was shot by Jack Ruby in front of millions of television viewers on Nov. 24, 1963 after claiming that he was "just a patsy". Transcripts of Oswald's interrogation were destroyed. He was conveniently disposed of before he could get a lawyer. This analysis indicates he was indeed a patsy. Ruby should also be included. He died in prison, claiming that he was injected with cancer cells because he wanted to  tell the truth.


In 1977, six top FBI officials who were scheduled to testify before the House Select Committee died within 6 months.


This graph displays the probabilities of 1-16 unnatural deaths among 1,000-10,000 randomly selected individuals.



This graph displays a table of probabilities that from 5 to 65 people in a random group of 2,000 would die UNNATURALLY in 1-15 year intervals.





An actuary engaged by the London Times calculated the probability that at least EIGHTEEN witnesses would die within 3 years of the JFK assassination as 1 in 100,000 trillion. But in a response to a letter from the 1977 House Select Committee on Assassinations, London Sunday Times Legal Manager Anthony Whitaker wrote:


Our piece about the odds against the deaths of the Kennedy witnesses was, I regret to say, based on a careless journalistic mistake and should not have been published. This was realized by The Sunday Times editorial staff after the first edition - the one which goes to the United States - had gone out, and later editions were amended. There was no question of our actuary having got his answer wrong: it was simply that we asked him the wrong question. He was asked ” what were the odds against 15 named people out of the population of the United States dying within a short period of time”  to which he replied -correctly - that they were very high. However, if one asks what are the odds against 15 of those included in the Warren Commission Index dying within a given period, the answer is, of course, that they are much lower. Our mistake was to treat the reply to the former question as if it dealt with the latter - hence the fundamental error in our first edition report, for which we apologize.


That settled the matter for the HSCA which did not bother to ask U.S. mathematicians to analyze the probabilities. One must ask: Why not?


Whitaker was only partially correct: True, the probability of 15 named individuals from the U.S. population dying is much lower than the probability that 15 out of 1400 listed in the Warren Commission report would die.  But he made two fundamental errors.


The first was misstating the problem definition. He assumed deaths of all types. He did not indicate that the probabilities are a function of the expected number of unnatural (not total) deaths within a given year. That was obfuscation based on a false premise.


The second was error by omission: avoidance of the mathematics. Whitaker did not include mortality statistics and show the probability calculations. Why not?

Because it would prove that the actuary's calculations were justified?




In fact, the answers to both questions show that in each case, the probabilities are at the vanishing point - assuming the deaths were independent events. The common factor in calculating the probability for both cases is the probability of death by unnatural causes in any given year. It is 0.000542.


1) The probability that 15 named individuals in the U.S. population would die unnaturally in any given year is p=0.000542^15 or 1.0e-49.

2) Yes, the probability that  least 15 out of 1400 persons in the Warren Commission index would die unnaturally in the year following the assassination is much higher: 1 in 167 trillion (6.0e-15).


The probability P of at least m unnatural deaths in a group of n persons during a time period t is

P(m)    = f (n,t,p), where p is the probability of an unnatural death in a given year. As t increases, the probability that at least m would die of unnatural causes also increases.


Probability of an unnatural death in a given year:

suicide.            0.000107                 

homicide         0.000062                 

accidental        0.000359                 

undetermined 0.000014


Total               0.000542 




The odds of dying (lifetime):

Accidental Injury                    1 in 36

Motor Vehicle Accident         1 in 100

Intentional (suicide)                1 in 121

Falling Down                         1 in 246

Assault by Firearm                  1 in 325






The Poisson distribution function is the perfect tool for calculating the probability of a rare event. It is derived from the Normal (Gaussian) probability distribution- the most important tool in statistical analysis. The function is used when the probability of an event (P) is very small but the number of trials (N) is so large that the expected number of events (P*N) is a moderate-sized quantity.

There are two parameters in the Poisson probability function: the expected number (a) of unlikely events and the actual number (m).  The probability is:

P (m) = a^m * exp (-a) / m!


We have determined that P =.000542 is the probability of an unnatural death in a group of 1400 in a given year.

The expected number (a) of unnatural deaths is:  a = 0.7588 = P*N = 000542*1400.


In other words, in a given year we would normally expect slightly lower than ONE (0.7588) unnatural death in a random group of 1400 people.

But there were 15 unnatural witness deaths within one year of the assassination.


In Excel, the probability P of an unlikely event is calculated by the function P = POISSON (x, a, type) , where

 x is the number of events; a is the expected numeric value; type is a logical value that determines the form of the probability distribution (discrete or cumulative)


Use the Poisson formula to compute the probability of exactly 15 unnatural deaths for N = 1400 witnesses in one year:

Once again, the actual calculation is:


a = 0.7588 = P*N = 000542*1400

Using the spreadsheet function,

P (15) = Poisson (15, 0.7588, false) = 5.70e-15

Or using the formula,

P (15) = 0.7588^15 * exp (-.7588) / 15! 


P (15) = 1 in 175,441,539,952,741

P = 1 in 175 TRILLION!


But we need the probability of AT LEAST 15 unnatural deaths - not exactly 15.  It’s virtually the same.

The probability is:

P = 1 – the sum of the probabilities for 0, 1 ... 14 deaths:

P = 1 – (prob (0) + prob (1) + prob (2) … + prob (14))

In mathematical notation:

P (m > 14) = 1 - ∑ P (i), i=0, 14


P (m > 14) = 5.98e-15

P (m > 14) = 1 in 167,145,910,421,722

P= 1 in 167 TRILLION!


This table displays the probability that at least m out of 1400 witnesses would die unnaturally in one year. The probability declines exponentially as the number of deaths increase.


m         1 in


0          1

1          2

2          6

3          24

4          132

5          892

6          7,195

7          67,346

8          718,040

9          8,593,044

10        114,073,493

11        1,663,713,384

12        26,445,366,889

13        455,051,758,699

14        8,427,523,639,942


15        167,145,910,421,722


16        3,534,913,873,810,260

17        79,526,916,217,848,800

18        1,966,037,843,894,810,000