Backtesting Bollinger Bands: Mean Reversion vs Breakout

2026-07-13

Bollinger Bands are one of the most recognizable indicators on any chart: a moving average with an upper and lower band set a number of standard deviations away. Their popularity hides a trap — the same bands can justify two completely opposite trades, and traders often pick the wrong one for their market without ever checking.

What the bands actually show

The middle band is a simple moving average (20 periods by default). The outer bands sit two standard deviations above and below it, so they widen when volatility rises and contract when it falls. Roughly 95% of price action stays inside the bands, which is exactly why people assume a touch of the outer band is "extreme." That assumption is the source of most Bollinger losses.

Two opposite strategies

Mean reversion (fade the edge): buy when price tags the lower band, sell when it tags the upper, betting that price returns to the middle. This works when the market is ranging.
Breakout (ride the move): buy when price closes above the upper band, betting that a band break signals the start of a real move. This works when the market is trending or expanding out of a squeeze.

The mistake is committing to one style by faith. In a strong crypto uptrend, fading the upper band shorts you into a rally over and over. In a chop, chasing breakouts buys every fake-out. The bands are neutral — your regime assumption is what wins or loses.

Backtesting both honestly

On this site, add a Bollinger Bands trigger and choose the mode: touch lower / touch upper for reversion, or break upper / break lower for breakout. Test the exact same market and period both ways. Then add a trend filter — for example, only take breakouts when a higher-timeframe trend agrees, or only take reversion trades when a longer moving average is flat. Compare the period table for each variant.

Watch the standard-deviation setting. Tightening it from 2.0 to 1.5 fires far more signals and usually more losses; loosening it to 2.5 trades rarely but on stronger extremes. If a strategy only works at one exact deviation and collapses on either side, that is overfitting, not an edge.

A band touch is not a signal by itself — it is a condition. Pair it with an exit rule and cost assumptions before you trust any equity curve. Try a Bollinger idea now on seven years of real Binance data and see which regime your chosen market actually rewards.