http://richardcharnin.wordpress.com/2013/05/15/jfk-witness-deaths-calculating-the-probabilities/
We must distinguish between unnatural and natural deaths. The failed to do
so. In its attempt to dismiss the London Sunday Times actuary’s 100,000
TRILLION to 1 odds of 18 witness deaths in three years after the
assassination, the HSCA made the following fundamental errors:
1) considered just 21 of 46 suspicious deaths from Nov 1963-Feb 1967,
avoiding Oswald, Ruby, Ferrie, etc.
2) did not specify that the deaths were UNNATURAL
3) did not consider witness CONNECTIONS to the case and MOTIVATION for
their elimination
4) ignored the need to utilize UNNATURAL MORTALITY RATES
5) falsely stated that the UNIVERSE of witnesses was UNKNOWABLE
6) ignored the KNOWN UNIVERSE of 552 Warren Commission witnesses:14 unnatural,
10 suspicious deaths
7) ignored the SUSPICIOUS deaths in 1977 of 7 top FBI officials called to
testify at HSCA
8) did not question why no one at the London Times knew the actuary's name
9) failed to show the actuary's ASSUMPTIONS and calculation METHOD
10) did not consider using the POISSON distribution function to calculate
probabilities
The HSCA succeeded in misinforming the public by inferring that the
actuary’s odds were not proof of a conspiracy.
The key to calculating the probability of a given number (n) of unnatural
deaths occurring in a given group (N) in a given time period (T) is to
recognize that it is based on the difference between <em>expected and
actual unnatural deaths</em>. The larger the discrepancy between the
actual observed and expected number of deaths, the lower the probability.
These are the relevant probability input parameters:
N= total number of witnesses
n= number of observed unnatural deaths
T= time period in years
R= unnatural mortality rate
The tables in the JFK Witness Spreadsheet Database display unnatural death
probabilities for 552 Warren Commission and 1400 material witnesses based
on unnatural, weighted and homicide mortality rates over 3, 7 and 14 year
time intervals.
In 1963, the annual national homicide rate was 6 per 100,000 (0.00006). At
least 4 of the 552 Warren Commission witnesses were murdered in the first
3 years after the assassination. The equivalent Warren Commission rate was
0.0024 = 4/(552*3) or 240 per 100,000 (40 times the national rate).
If the 5 accidental deaths and 1 suicide in the three year period were actually
homicides, then the equivalent Warren Commission rate was 0.006038 = 10/(552*3) or 603 per 100,000 (100 times the national average).
The Poisson Distribution calculates the probabilities of rare events over
time.
We must first calculate E, the expected number of unnatural deaths: E= N*R*T
The probability of EXACTLY n unnatural deaths is: P = POISSON (n, E,
false).
The probability of AT LEAST n deaths is P = 1- POISSON (n-1, E, true)
EXAMPLE:
The probability of AT LEAST 14 Warren Commission witness homicides in the
14 years from 1964-77:
E = 0.479 = N*T*R = 552*14*0.000062
P = 1 - Poisson (13, 0.479, true)
P = 4.44E-16 = 1 in 2,251,799,813,685,248
P = 1 in 2251 trillion</strong>
View the witness database and the probability sensitivity tables...
https://docs.google.com/spreadsheet/ccc?key=0AjAk1JUWDMyRdDFSU3NVd29xWWNyekd2X1ZJYllKTnc#gid=1