Mean Reversion: Profiting from Overshoots
The Idea: Extremes Don't Last
When a stock moves way up (far above its trend), it's likely to revert down. When it moves way down, it's likely to revert up.
We call this mean reversion — prices bounce back to their average. It's the opposite of momentum: momentum says "buy winners," mean reversion says "short winners" (they've moved too far).
Mean reversion works in choppy, sideways markets where momentum fails. It exploits overshooting — prices moving too far before correcting. Best combined with [[volatility]] estimation so you know what "far" means.
Measures how many standard deviations away the price is from its moving average. Z > 1.5 → overbought (short). Z < -1.5 → oversold (long).
Price is 2 standard deviations above its MA. What's the signal?
When does mean reversion work better than momentum?
Mean reversion: prices deviate from MA, then snap back
Z-score: how many standard deviations away from MA (0 = at MA)
Signal: Long when Z < -1.5 (oversold), Short when Z > 1.5 (overbought)
Key insight: threshold choice matters — tighter (Z > 1.0) = more signals; looser (Z > 2) = fewer, higher-conviction trades
Opposite of momentum: works in choppy markets, fails in trends
Pairs trading: apply mean reversion to correlated assets (exploit relative mispricing)