評論
打印
This is Scientific American's 60-second Science, I'm Steve Mirsky.
With a full month of the regular season to play, the 2019 New York Yankees had set a Major League Baseball record for injuries. That medical crisis led to what the online sports publication The Athletic reported on January 3rd as "sweeping changes" to their training and strength-and-conditioning programs. When I read that news, I thought, of course, of traffic cameras, which sometimes, and reasonably, get placed at sites that have a disproportionate number of accidents in a given year.
"The fact that there's a higher rate of accidents will be partly due to chance because it will fluctuate over the course of time. Sometimes it will be less; sometimes it will be high."
David J. Hand, on the Scientific American Science Talk podcast in 2014. He's emeritus professor of mathematics and senior research investigator at Imperial College London, where he formerly held the chair in statistics. He was on the podcast to talk about then new book The Improbability Principle: Why Coincidences, Miracles, and Rare Events Happen Every Day.
So what does the Yankees' revamped training staff potentially have in common with traffic cameras?
"Now if we look back at last year and identify the places which have particularly high rates of accidents, the high rate of those places will be due to a sum of two things: the natural degree of dangerousness of those places, plus the fact that that particular year just happens to be a bad year—there were more accidents than normal at that year. But because it is a high rate of accidents, we're now going to put a camera there."
Again, a particular year just happens to be a bad year.
"Now what happens next year? The natural dangerousness of the place hasn't changed; it's still the same corner or intersection or whatever. But the chance bit of the number of accidents there, well, it could be low just as easy as high next year. On average, it will be lower than the high rate we saw. So next year, the rate will come down. It won't be because of the camera, it will just be because of natural fluctuation, removing that sort of chance part. But it will look as if putting the cameras there has improved things."
Hand continued: "The fact is, however, the speed cameras do work. There is no question that they do reduce the rate of accidents but not as much as a superficial analysis failing to take into account the Law of Selection and Regression to the Mean makes it look like."
As a big Yankees fans, I hope they have no injuries at all in 2020—and that the new training staff is outstanding. But when the team's injury rate falls—as it's almost certain to do after a record-setting bad year—I'll keep in mind that at least part of the drop may be due not to Tinkers, nor to Evers, but to chance.
For Scientific American's 60-second Science. I'm Steve Mirsky.
