top of page
Writer's pictureCris

He's not getting worse, he's regressing to the mean!



How often did you see a top player performing poorly after a while?


Very often, in any sport or discipline, top-notch players cannot keep high standards. I always thought it was because of the pressure, or maybe because they couldn't handle being famous and so on.


This phenomenon has a very solid statistical rule called regression to the mean.


People who excel in a field will eventually regress to the average

This means that the soccer top player who had an extraordinarily good season will very likely perform worse during the next year.

In fact, top soccer players represent a small minority of all the soccer players in the world and, because of that, they are meant to get closer to the average of the other players. In the book Thinking, Fast and Slow Daniel Kahneman explains this with a very nice anecdote.


During a training session at a pilot school, he explained why punishing a pilot for a bad performance won't make the next flight better, in the same vein as praising a good performance won't make the next one worse.


That's when a veteran stood up and explained that, on the contrary, it MAKES a difference. In fact, he explained that every time he punished a bad flight, the next flight was usually better; and every time he happened to praise a good pilot, he would have flown poorly.


Daniel proposed then an experiment. For two rounds, all the pilot teachers had to throw a coin behind them, trying to get as close as possible to a pawn. During the first round, some pilots were very good and threw the coin very close to the goal, while others performed very badly. During the second round, things changed. The ones who performed well during the first attempt now performed worse, and vice versa. What they testified was the perfect essence of "Regression to the mean".


Pilots that (by chance) were very good during the first attempt during the second regressed to the mean (by getting worse). Similarly, pilots who performed below the average regressed also to the mean (by getting better).


With this experiment, Daniel showed why pilots with a very bad performance would have performed better anyway (with or without punishment) and why pilots with a good performance would have performed worse the next time.


That was something I never thought of before, but it makes extreme sense!


What do you think?


Comments


Subscribe to the newsletter

Join to receive an email every Friday with the best citations I came across during the week!

Thanks for submitting!

bottom of page