Let's take an example.
I'm leaving home to go to the grocery store. My wife tells me, "you are going to run into Jane Doe in the parking lot of the grocery store."
I go to the store, and sure enough I see her.
Quite a coincidence. Indeed, so far-fetched a coincidence that I have trouble believing that my wife was just guessing. She must have had "foreknowledge." Maybe she just hung up the phone having talked to Jane.
But now let's suppose a slightly different scenario. My wife says nothing, and I drive to the grocery store and run into Jane Doe in the parking lot. That was "unlikely." It happens maybe once every 100 times I go to the grocery store. But do I think it's "suspicious?" No. I quite frequently run into somebody I know, and it's as likely to be Jane as anybody else.
Warren Commission critic John Hill made the following argument about this:
Please do your probability math for us. How many hotels in Chicago? A hundred? So, the odds of Ferrie (perhaps) just happening to call the one with Meyers lady friend in it would be 100 to 1. Terrible odds - for you.The problem with this argument is there was no reasonable a priori expectation that a plot to kill Kennedy would involve David Ferrie calling Jean Aase who would then contact Lawrence Meyers who then would go to Dallas and meet Jack Ruby.
So the real question is: what is the probability that one of the many people with a very tenuous and questionable "connection" with Lee Oswald would call a hotel in which lived a person who was one of the many people with a "connection" to another person who was one of the many people who had a "connection" to Jack Ruby?
It's not unlikely at all that something like that would happen. Indeed, probably many such chains of "coincidence" happened, and conspiracists have portrayed as uniquely sinister the few that they happen to have learned about.
It's considered an unlikely "coincidence" that, when Oswald put a phony "office address" on a pamphlet of his phony "Fair Play for Cuba Committee" it was the building (although not the address) where Guy Banister had an office.
Hill also believes that this was one of those sinister happenings:
There were tens, if not hundreds of thousands of addresses in N.O. The odds that Oswald just happened to accidentally pick one with Banister in the building? About 100,000 to one. And you'd have us believe that was most likely a coincidence?The problem again is one of bona fide prior expectations.
Let's suppose that we had, before the assassination, identified Banister as the single most sinister man in New Orleans. Identified him as the one person most likely to be involved in a JFK assassination conspiracy.
Further, suppose we had good reason to believe that the one bogus office address that Oswald put on his pamphlets and not his home address and not his work address and not the address of places where he demonstrated and not the address of WDSU Radio where he taped an interview would be the one that would reveal the nature of the conspiracy.
Then the odds of this connection happening "coincidentially" would be very poor. Not 100,000 to one, since there weren't 100,000 office buildings in New Orleans in 1963, but still poor.
The problem, however, is that nobody identified Banister as the single most sinister man in New Orleans before the JFK assassination. If he seems very uniquely sinister now, it's because a generation of conspiracists has portrayed him that way.
So the relevant question is: among the several addresses associated with Oswald in New Orleans in 1963, how many were also associated with at least one person who could be portrayed as uniquely sinister?
Remember, to get portrayed as sinister by conspiracists, it's sufficient to have strongly anti-communist political views, or be associated with the mafia, or be employed by government, or be associated with anti-Castro Cubans, or ever have had any connection to the CIA, or have any connection to the Trade Mart, or even (heaven help us!) be engaged in cancer research.
Thus it's not particularly unlikely that some address associated with Oswald would also be associated with somebody who, like Banister, could be seen as sinister.
On these issues, like many others, conspiracist "researchers" are rummaging through a massive amount of data looking for connections. It's not surprising that they can find several such connections. But none of these "connections" was one that, before the assassination, would have looked conspiratorial. Rather, conspiracists are, post hoc, interpreting very mundane events as sinister.
Indeed, the revised motorcade route was never published in the newspapers, which raises a fascinating question, namely: How did the alleged assassin even know that the President would pass by the Texas School Book Depository in order for him to shoot him? In an interesting study, "The Mathematical Improbability of the Kennedy Assassination," The Dealey Plaza Echo (November 1999), pp. 2-6, Ed Dorsch, Jr. has calculated that the probability of Oswald and JFK coming within 100 yards of each other at random during his Presidency is approximately 1 in 1 hundred billion! This suggests an encounter by the two was almost certainly no accident . . . . (p. 12)If one checks out the Dorsch article, one finds that it starts out reasonably enough:
Basically and logically stated, there are only three ways for an assassin to get within range of his target:So far, so good. The lone assassin theory really does posit that pure blind chance brought JFK into Oswald's sights.
Strangely, it seems that all recorded assassins have either stalked their victims, have been part of the victim's entourage, or have had help from the inside in setting up the target. (p. 3)
- by stalking
- by betrayal
- by accident
Now Dorsch entertains his readers for several paragraphs with an account of how no reputable academic would touch his analysis, and then proceeds as follows.
Firstly, we should examine our formula:Since Kennedy was in office for about 1,000 days, he calculates pt = .001. Thus, if he was going to be killed on a random day of his term, the probability of it being November 22, 1963 is 1/1000.
PROBABILITY = probability of time(pt) x probability of location pl
What the formula says is that the probability of the occurrence is equal to the time factor multiplied by the location factor.
To calculate the "location factor" Dorsch does the following:
The probability of location would be a circle with a radius of 100 meters, or .0314 square kilometers/area of the United States . . . in square kilometers. That figure is approximately 9,809,430 sq km.It would be trivial to ask how Fetzer turned three in one hundred billion into one in one hundred billion, or to point out that Dorsch really means ".0314/9809430" in the text above. There are more serious issues here.
So, our pl appears to be 9809430/.0314 = .000000003.
Thus, putting our formula together, we get:
PROBABILITY = .001 x .000000003 = .000000000003.
That figure represents 3 in ONE HUNDRED BILLION.
Dorsch is assuming that our assassin (Oswald, hypothetically) is dropped down into a random location in the U.S. He assumes that each day the President is dropped down into a random location in the U.S. What are the odds then that on November 22, 1963, Kennedy was dropped down within 100 yards of Oswald?
But of course, neither presidents nor would-be assassins are located randomly in the U.S. Both tend to be in population centers. Indeed, this is why such places are called "population centers." So dropping Kennedy down into a randomly selected location in a population center is much more likely to put him within 100 yards of Oswald. And of course, Kennedy didn't stay in one location when he made a political visit to a city. In Dallas, his motorcade traveled several miles through the heart of the metropolis.
But the more fundamental problem here is that there was no bona fide prior expectation that Oswald would be the shooter, or that the assassination would take place on November 22, 1963.
Suppose he had been shot in Miami in early November? Dorsch could run through his calculations and discover that there were only three chances in one hundred billion of that happening. Kennedy could have been shot anywhere on any date, and it would be equally unlikely.
The proper way of going about this would be to ask what the odds were of Kennedy passing any location that harbored a malcontent who might want to kill him and have the shooting skills to do so on any date during his presidency.
Indeed, we could take this a step further and point out that there was no bona fide prior expectation that Kennedy and not Eisenhower nor Johnson nor Nixon would be killed by an assassin. It seems obvious to us today that Kennedy was shot in Dallas on November 22, 1963, and that Oswald was charged with shooting him. But if Johnson had been shot in Bangor, Maine, on March 23, 1965 by a man named Joe Blow, that would seem equally "obvious" to us today. It would be equally "unlikely," and its unlikeliness would be equally meaningless.