Bayes' Law can be expressed in many ways, but if the evidence for just one link in someone's theory is practically zero, then the same goes for the entire theory.
Nevertheless, as the subtitle says, incredibly unlikely things do happen. We see it all the time, as when I contacted an internet stranger in Texas to buy a book, only to discover she lived right next door to friends I had once stayed with (who sent best wishes by return!).
The person who turns out to have been uncannily correct may convince themselves, and perhaps others, they had some superior clarity of mind, reasoning ability or access to privileged information. Generally they are deluded, and the simple explanation is the reverse.
If a large number of people guess the answer to a mystery, then obviously, after it's solved, somebody's guess will have ranked closest.
Of all those who prophesied the history of last year before it began, somebody, somewhere, can inevitably claim with hindsight to have been the most clairvoyant. Vagueness is self-evidently an ingredient of success, and precision notoriously elusive.
David Hand presents a rich mix of true stories, hypothetical examples and the reasons why our intuitive solutions are so often incorrect.
Nothing seems to lodge more indelibly in the human mind than coincidence, but our collective failing is to look for a cause when there isn't one. Tossing a coin 20 times and getting 20 heads is calculably unlikely by a million to one, but it cannot cause the 21st coin to land heads, or tails, it's always 50/50. Forget the notion of a winning streak.
Unfortunately, exceedingly rare happenings can induce depressingly ridiculous explanations, but common sense is the foundation of probability theory. By avoiding sales-killing equations, the author has cleverly projected the basic maths of statistics almost unnoticed.
For a sensible answer there has to be a precise question. How many people must gather for two to have the same birthday? To be certain, there need to be 367. But with as few as 23 people, it's 50/50 that two will have a common birthday - unless you specify the day of the year, then the number is a much larger 253.
If you want 100% certainty of matching one specific birthday, the number becomes astronomical.
Sadly it's more than a truism that probability theory can be employed to prove or disprove practically anything. A so-called pattern of aircraft problems may be nothing more than a statistical blip. Innocent people have tragically gone to prison after self-proclaimed experts have failed to understand their own shonky statistics.
Anyone tempted to rush to judgement, in whatever sphere, would do well to stop and read this book first. Being so packed with anecdotes, it's easy to read in short unpredictable bursts.
Clive Trotman is a Dunedin arbitrator and science presenter.
