Skip to content

Strategies – Mean Reversion

Overview

Mean reversion is the statistical tendency for an asset's price to return to its historical average after deviating from it. While Momentum strategies bet on continuation, Mean Reversion strategies bet on reversal — buying when something is "too cheap" and selling when it is "too expensive" relative to recent history.

This utility implements a Bollinger Band + RSI Mean Reversion Strategy using a synthetic Ornstein-Uhlenbeck price process.

Key Concepts

Mean Reversion

  • Rooted in the statistical concept of regression to the mean.
  • Works best in ranging, sideways markets where there is no dominant trend.
  • Commonly applied to: pairs of correlated assets, interest rates, commodity spreads, volatility indices (e.g., VIX).

Bollinger Bands

  • Middle Band: Simple Moving Average (SMA) over window periods — our estimate of fair value.
  • Upper/Lower Band: SMA ± num_std × rolling standard deviation.
  • When price touches the lower band, it is statistically cheap (a potential buy).
  • When price touches the upper band, it is statistically expensive (a potential sell/short).

RSI Confirmation Filter

  • RSI < 35 confirms oversold conditions — strengthens the lower band buy signal.
  • RSI > 65 confirms overbought conditions — strengthens the upper band sell signal.
  • Using dual confirmation dramatically reduces false entries in trending markets.

Ornstein-Uhlenbeck (OU) Process

  • The continuous-time stochastic model of mean reversion.
  • Used in interest rate models (Vasicek), commodity pricing, and pairs trading.
  • Governed by: dx = kappa * (theta - x) * dt + sigma * dW

Logic Implemented

Entry signals (both conditions required): - Long (+1): Price < Lower Bollinger Band AND RSI < 35 - Short (–1): Price > Upper Bollinger Band AND RSI > 65

Exit signals: - Exit Long when price reverts above the SMA (fair value reached). - Exit Short when price reverts below the SMA.

Files

  • mean_reversion_strategy.py: OU data generation, Bollinger Bands, RSI, signal logic, backtest engine, and performance metrics.

How to Run

python mean_reversion_strategy.py

Financial Applications

1. Statistical Arbitrage

  • Trade the spread between two cointegrated assets (e.g., pairs trading).
  • The spread is the OU process; enter when spread deviates too far.

2. Fixed Income / Interest Rates

  • Short-term interest rates exhibit strong mean reversion.
  • The Vasicek model (based on OU process) is used for bond pricing and interest rate derivatives.

3. Volatility Trading

  • Implied volatility (VIX) is mean-reverting by nature.
  • Selling options when VIX is high, buying when low, exploits this reversion.

4. Market-Making

  • Market makers implicitly bet on short-term mean reversion.
  • They buy at the bid and sell at the ask, profiting when prices revert within the spread.

Best Practices

  • Market Regime Detection: Mean reversion strategies fail badly in strong trends. Use a regime filter (e.g., ADX indicator) to disable the strategy when the market is trending.
  • Transaction Costs: Mean reversion often requires many trades; commissions and slippage must be accounted for.
  • Parameter Sensitivity: Test multiple window lengths (10, 20, 30 days) and band widths (1.5, 2.0, 2.5 std) — results can vary significantly.
  • Combining with Momentum: Many professional quant funds run both momentum and mean reversion strategies simultaneously to diversify across regimes.