Bollinger Bands: what they are, and whether they work

2,400 of 12,035 out-of-sample tests beat simply buying and holding the same asset — 19.9%. On the other 9,635 it did not. That is indistinguishable from the 20.1% rate across all 382 indicators we test (one pooled rate over all 660,005 tests we have run, not a mean of the per-indicator rates) — the difference is inside the margin this many tests can resolve, so read it as ordinary, not better or worse.

One test is one variant on one asset on one timeframe. We tested 7 variants of Bollinger Bands, so an asset that carries several of them contributes several tests — this is the family's record across everything we ran, not a count of assets. Those variants are not alike: individually they beat buy-and-hold between 17.4% and 24.0% of the time, and 19.9% is the pooled rate over all of them, not a property every version of this indicator has. The table below gives each one separately. "Beat" means a higher return than holding that same asset over that same window. Measured out-of-sample, on data the setup was not chosen on.

The middle band is a simple moving average of price, conventionally 20 periods. The upper and lower bands sit a multiple of the standard deviation of price over that same window above and below it — conventionally two. Because standard deviation is recomputed every bar, the width of the channel is itself a volatility reading: a 'squeeze' is the band width contracting, and an expansion is it widening.

How it's read. Two incompatible readings are both conventional, which is the main thing to understand about them. The mean-reversion reading treats a touch of the upper band as stretched and expects a return toward the middle. The breakout reading treats the same touch as strength and expects continuation. They cannot both be right on the same bar, and which one works is an empirical question rather than a matter of interpretation — which is what the numbers below are for.

Where it struggles by design. In a strong trend, price can ride the upper band for a long stretch, and the mean-reversion reading will fight that move the entire way. Conversely, in a quiet range the breakout reading produces a stream of failed signals as price crosses the band and immediately returns.

Origin: Devised by John Bollinger in the early 1980s.