After the San Francisco earthquake of 1989, a woman announced to her seismologist friend that she was going to drive to Los Angeles to escape the aftershocks.
"I don't think that's a good idea," the seismologist said.
"Why not?" the woman asked, horrified.
"Well," said the seismologist, "the chances of being injured in a traffic accident are much higher than of being injured by aftershocks."
I like stats like this. (I say this as someone for whom statistics and the act of liking rarely happen in the same place.) And Tails You Win: The Science of Chance, a one-off BBC Four documentary, is full of just such facts.
Presenter David Spiegelhalter is professor of the public understanding of risk at the University of Cambridge. (Probability of documentary-makers choosing to show him on a punt: 100 per cent.) This means that people regularly approach him with questions that begin with bizarre coincidences and end with the phrase, "What are the chances?" (Probability that human nature incorporates a desire for control, and abject terror at our own powerlessness: 100 per cent.)
But we can control our own fate, to some extent. Drinking three pints takes half an hour off your life. So does smoking two cigarettes. Going for a half-hour run, by contrast, adds half an hour. "I hope you like running," says Spiegelhalter, "because that's how you've just spent your extra half-hour."
He then introduces us to the micromort, which is the unit used to measure your chances of death. A skydive is worth seven micromorts. Being 18 gives you a 500 micromort chance of death; by the age of 58, this has risen to 7,000. On the bright side, if you are a 58-year-old who goes skydiving (cue Spiegelhalter strapping on a parachute), your chance of death is 7,007 micromorts. So you may as well jump out of the plane.
Then we reach what the shampoo adverts refer to as the science part. Roughly: the world is just too complicated to understand fully. The best we can hope is to describe the subatomic world in terms of probability. (See? Science part.)
But probability is a handy way of measuring things. Take, for example, the weather. The most accurate form of prediction may involve simply reciting probabilities: 30 per cent chance of rain, or 15 per cent chance of strong winds. "Better a reliable probability than the wrong prediction," Spiegelhalter says.
And, he adds, "surprises will most certainly happen". For example, a programme about actuarial statistics might actually be enjoyable. And accessible. And genuinely engaging. What are the chances of that?