David Hand. The Improbability Principle: Why Incredibly Unlikely Things Keep Happening. Bantam Press, 2014.

The other week something rather spooky happened. I was re-reading William Poundstone’s Big Secrets, a book published in the US in 1983 and the UK in 1985. One of the secrets discussed was how mentalists like (the then famous) Kreskin did their tricks. On page 206 there is this little excerpt:
Take the following hypothetical exchange:Kreskin: "Does the date September 11 mean anything special to anyone...?”.
September 11!? Do the shivers go down your spine at this? It’s amazing, spooky, that at random Poundstone chooses a date which now means so much to so many?
Well perhaps not according to mathematician David Hand, who argues that very improbable events happen all the time, not least because our understanding of statistics is weak and we forget the impact of very large numbers. Improbable events of the sort that fill the little spaces down the margins in Fortean Times: you someone loses a ring out on a boat and ten years later it is found in a fish caught by their father in law; you pick up a book left on a train by a stranger and it is the one your partner loved as a child and lost when they moved house twenty years ago. You get the sort of thing. [From my own experience: you are sent a tiny cutting from a local newspaper with an item that a relative thinks may be of some interest to you, and on the reverse is a report of the coroner’s court verdict on the suicide of someone you knew well twenty years earlier but hadn't been in contact with since! - JR]

Hand sees several factors at work; the law of large numbers, such an astronomical number of events are occurring in the world that some are bound to lead to very unlikely situations, such as being hit by lightning many times over, or getting three successive holes in one; misunderstanding the correct estimate of probabilities (this one, which involves the difference between ‘normal’ and ‘Cauchy’ distributions is very technical but it apparently means that something that is beyond astronomically unlikely on the former, has odds of less than 100 to 1!

Another error is to assume that things are unconnected and occur at random, when the reverse is the case. He points to the tragic case of a mother who was found guilty of murdering her two children, because a paediatrician (with no real knowledge of statistics) had estimated the chance of two children dying of cot death were 73 million or more to one against, but that assumed that one child dying of cot death did not increase the risk for a second, and that turned out to be wrong.

Another principle is that of ‘close enough’, the two events are not really all that close, and if one goes wide enough then all sorts of things are bound to occur. A case I remember from psychical research was of a woman who had a dream about a (real or imaginary) wartime film in which a British agent, played by Leslie Howard, was murdered by a German. This became a ‘precognition’ of the attempted assassination of Ronald Reagan (who, of course, was once an actor, and his assailant was rumoured to have neo-Nazi sympathies).

So Hand would argue that whatever date chosen, sooner or later something significant would have happened on that date. Furthermore on any given date things of great significance happen. Plug in any random date in Wikipedia and see what happened on that date.

The random date I had thought of was April 27, nothing much came to mind, but is filled with ‘significant’ events, and going down the list we see that on this date in 1878 was born the English athlete John Rimmer (died 1962). Someday, somewhere something truly momentous and earth shaking will occur on April 27 (or March 16, Oct 8, June 19 or any other date you care to choose) – Peter Rogerson.

1 comment:

  1. A good discussion of such things can be found in "Lady Luck" by Warren Weaver (Pelican books, 1977). Chapter 13 entitled 'Rare events, coincidences and surprising occurrences' gives some very good examples including a real world beater at the end.