Quick Answer: Regression to the Mean is a statistical phenomenon where an extreme outcome (unusually good or bad) is statistically likely to be followed by an outcome that is closer to the average. It happens simply because extreme outcomes require extreme luck, and extreme luck doesn't last.
The Sports Illustrated Cover Jinx
There is an old superstition that appearing on the cover of Sports Illustrated curses athletes, causing their performance to crash immediately after. There is no curse—only regression to the mean. To get on the cover, an athlete must perform exceptionally well over weeks. That extreme performance was a mix of immense skill and temporary good luck. In the weeks that follow, their skill remains, but their luck returns to normal. Their performance drops. They regressed to the mean.
The Illusion of Punishment vs Reward
Daniel Kahneman, a Nobel laureate, famously dealt with flight instructors who swore that yelling at pilots for bad landings worked (the next landing was better), and praising them for great landings failed (the next landing was worse). Kahneman realized the instructors were being fooled by regression. An exceptionally bad landing is likely to be followed by a closer-to-average (better) landing, regardless of yelling. An exceptionally great landing is likely to be followed by a closer-to-average (worse) landing, regardless of praise. The instructors falsely attributed statistical regression to their discipline tactics.
The Math Behind It
Performance = Skill + Random Variance (Luck). Skill is stable. Variance is a coin flip. If someone scores a 99% on an exam, they have high skill AND they got lucky guesses on the edge questions. On the next test, their skill is still high, but their luck will probably be average. So they score a 94%. We see a "drop" in performance, but it's just the absence of extraordinary luck.