Regression to the mean
Regression to the mean is the tendency for extreme results to be followed by more average ones, simply because luck evens out. An exceptional performance is usually part skill, part chance — and the chance part rarely repeats.
How it works
Most outcomes mix skill and luck. When you observe an extreme, an unusual dose of luck is often involved; next time luck reverts to typical, so the result drifts back toward the average — with no other cause needed.
How to use it
- Don’t over-react to a single extreme result, good or bad — expect it to moderate.
- Beware crediting an intervention for improvement that was just regression.
- Judge ability over many trials, not one outlier.
Worked example
A rookie’s record-breaking first season is often followed by a weaker second — the “sophomore slump” — not because praise spoiled them, but because their peak mixed skill with luck that reverted. The same logic explains why punishing a bad result often “seems to work.”
Where it fails
It’s easy to mistake regression for cause — to credit a coach, a drug, or a scolding for improvement that would have happened anyway. Always ask whether an extreme baseline made reversion inevitable.
The deeper point
Half of what looks like cause is just luck reverting. Most management "wins" — punishing the slump, praising the streak — are quietly taking credit for regression, which is why so much advice "works" and so little replicates.
Frequently asked
- What is regression to the mean?
- The tendency for extreme outcomes to be followed by more average ones, because the luck component of an extreme result rarely repeats.
- Why does regression to the mean fool people?
- It makes useless interventions look effective — after a terrible result things improve on their own, and whatever you did gets the credit.
- Is regression to the mean a real cause?
- No — it’s a statistical artifact of mixing skill and luck. No mechanism is needed; extremes simply tend to be followed by less-extreme values.
Related
Editorial synthesis © ReadGlobe 2026, drawing on the mental-models tradition (Charlie Munger, Farnam Street) and the primary sources for each model. · Last reviewed 2026-05-29.