[LINK] "Not so rare for rarities to occur in waves: Professor"

[rfmcdonald]

Recently, an above-average number of pedestrians have been killed in Toronto, fourteen so far, the worst number in a decade. The mass reaction to this involves a lot of soul-searching and concern about the worsening plight of the pedestrian. That may be useful, but as a University of Toronto statisticianpoints out chance also plays a role.

There is a tendency to grip these disturbing numbers and wring them for meaning. But for statisticians, maybe there is no meaning behind the numbers, just probability.

“Chances of a big clump are more than you would think,” said Jeffrey Rosenthal, a professor of statistics at the University of Toronto. “Yes, (the numbers) are rare and January was certainly much worse than usual, but it wasn’t something that is completely unexpected.”

Plane crashes, lightning strikes — for probability theorists, it’s no surprise that such rare happenings often occur in waves.

Rosenthal, who authored Struck by Lightning: The Curious World of Probabilities, explains that random events are all subject to a statistical phenomenon called Poisson bursts, named for the 19th-century mathematician Siméon Denis Poisson.

[. . .]

Seven isn’t that big a number when looked at through a statistician’s lens. Jeffrey Rosenthal calculates that between 2000 and 2009, Toronto witnessed an average of 31.9 pedestrian deaths per year and 2.7 deaths per month. Using Poisson distribution, this means there is about a 1.9 per cent chance of there being seven or more pedestrian deaths in a single month.

This principle applies elsewhere in life, of course.