Ever noticed how after an incredible season, a young talent in football often slips back to ordinary performance? Superstar breakout player this year. Kind of average the next year.
Doesn’t this happen surprisingly often?
It does. And there’s a reason for it: Regression to the mean.
In many areas of life—sports, business, personal relationships—the outcomes we observe hinge on two main factors:
- The underlying trend: The true level of skill, performance, or ability.
- Random fluctuations around that trend: Luck—both good and bad—that causes temporary deviations.
When someone or something performs exceptionally well or poorly, chance usually plays a part. And since luck isn't consistent over time, these extreme outcomes tend to move back toward the average—they regress to the mean.
In other words: if something goes exceptionally well (or poorly), chances are that things will get worse (or better) over time.
Some of these might sound familiar:
- After a streak of unusually good luck in a game, your next plays tend to return to your typical performance.
- A student scores exceptionally high on a test, but future scores revert closer to their average grades.
- Following an extremely bad hair day, your hairstyle tends to look more like it usually does the next day.
- A viral social media post brings a surge of followers, but your daily follower count soon goes back to normal.
- After experiencing an intense bout of creativity, an artist's work returns to their usual level of output.
- Your favorite team wins by a huge margin one game, but the next game's score is more typical.
- A restaurant has an unusually busy night, but customer traffic returns to average levels afterward.
- You sleep exceptionally well one night, but the following nights your sleep quality reverts to normal.
- An investment yields unusually high returns one year, but the next year it aligns with market averages.
- After a period of extreme stress, your mood and stress levels gradually return to your personal baseline.
Randomness influences everything. Extremes tend to even out over time. And predicting the future from the past must account for mean regression.
Here are 3 other concepts you might benefit from: