What is Mean Reversion? Definition, Formula, and Example
Mean reversion is the statistical tendency of an asset's price, spread, or volatility to return toward its historical average after an extreme deviation — the foundational principle behind pairs trading, volatility arbitrage, and statistical arbitrage strategies.
What is Mean Reversion?
Mean reversion is the tendency of a measurable quantity — price, volatility, spread, valuation ratio — to return to its long-run average after deviating significantly from it. The concept underpins a broad class of quantitative trading strategies: identify a quantity that has moved beyond its normal range, enter a position betting on its return toward the mean, and exit when it normalizes. Mean reversion is the mathematical opposite of momentum — it assumes extremes are temporary, not self-reinforcing. Most quantitative hedge fund strategies combine both: momentum in trending regimes, mean reversion in ranging ones.
How Mean Reversion Is Quantified
Ornstein-Uhlenbeck process — the standard continuous-time model for mean reversion:
dX_t = θ(μ − X_t)dt + σdW_t
Where θ = speed of mean reversion, μ = long-run mean, σ = volatility, dW_t = Wiener process.
Z-score — the practical implementation:
Z = (Current Value − Rolling Mean) / Rolling Standard Deviation
Half-life — time for a deviation to revert 50% toward the mean, estimated via AR(1) regression:
Half-life = −ln(2) / ln(β₁), where β₁ is the lag-1 autocorrelation coefficient.
Hurst Exponent — H < 0.5 indicates mean reversion; H = 0.5 is random walk; H > 0.5 indicates trending.
Worked Example
SPY has a 20-day rolling mean of $520 and a 20-day standard deviation of $8. During a sharp intraday selloff, SPY drops to $496.
Z-score = ($496 − $520) / $8 = −3.0
A z-score of −3.0 is three standard deviations below the mean. Historically, SPY has closed at ≤ −3 sigma fewer than 0.3% of trading days. A mean-reversion trader enters long at $496 with a target of $512 (−1 sigma) and a stop at $492 (−3.5 sigma). Separately, a pairs trader noting QQQ at −1.1 sigma goes long SPY / short QQQ, targeting normalization of the spread between them.
When Traders Use Mean Reversion
Mean reversion strategies operate across timeframes. Intraday traders fade gap-opens beyond 2 standard deviations from prior close or VWAP, targeting a return to the VWAP midpoint. Swing traders use Bollinger Band extremes — price touching or breaching the outer bands at high RSI readings — as mean-reversion signals. Statistical arbitrageurs systematically trade pairs of cointegrated equities, going long the underperformer and short the outperformer when their spread reaches a z-score threshold. Volatility traders apply mean reversion to implied volatility itself: IV rank and IV percentile measure how far current IV has deviated from its own history.
Limitations and Misconceptions
Mean reversion requires correctly defining the mean — a 20-day mean and a 200-day mean produce opposing signals in a trending market. Not all deviations revert: fundamentally impaired companies, regime changes, and structural breaks can drive permanent mean shifts (General Electric, Peloton). The Hurst exponent and half-life are both sensitive to the lookback period chosen. Mean reversion positions can move substantially further against you before reversing — position sizing must account for this path risk. A drift in the mean itself (a stock in secular decline) makes any static mean definition misleading. Always define maximum acceptable loss before entering a mean-reversion trade; the strategy has asymmetric loss exposure without it.