Mean Reversion with Bollinger Bands: 10-Year Backtest Case Study

1. The Mechanics of Bollinger Bands

John Bollinger's volatility envelope model explained

Developed in the 1980s by John Bollinger, Bollinger Bands consist of a central moving average (typically a 20-day Simple Moving Average) and two outer bands placed at positive and negative standard deviations (typically ±2 standard deviations).

Under a Gaussian distribution assumption, approximately 95% of all price action should mathematically fall within two standard deviations of the mean. When prices breach the lower band, technical traders anticipate an imminent reversion back to the 20-day mean.

2. The 10-Year Backtest Setup (2014–2024)

Testing across 500 liquid US equities with realistic transaction friction

We evaluated three distinct rule sets across a decade of historical daily bars including the 2018 volatility shock, 2020 pandemic crash, and 2022 interest rate sell-off.

3. Python Code for Bollinger Mean Reversion

Vectorized signal generation with volatility filters

Here is the exact algorithmic implementation to test this strategy on your own dataset:

4. Execution Protocol & Stop-Loss Discipline

Managing catastrophic tail risk when mean reversion fails

1. Always set a hard catastrophic stop-loss at 3.0 times the 14-day Average True Range (ATR) below your entry price. If a stock drops 3 ATRs below the lower band, the statistical mean has broken.

2. Exit 50% of the position when price reaches the 20-period SMA, and trail the remainder with a 1.5 ATR trailing stop to capture occasional trend flips.