Bollinger Bands: What They Actually Show

Bollinger Bands are a moving average with two boundaries set at a distance of standard deviations. The main misunderstanding around them is that the boundaries get treated as support and resistance, when they show something else entirely.

Bollinger Bands: what they are made of

The centre line is a simple average, usually over 20 bars. The upper and lower boundaries sit two standard deviations of price away from it over the same period. From this follows the key property: the width of the channel is volatility. The bands contract in a quiet market and flare on impulse, and that movement of the boundaries carries more information than price touching them.

How often it signals on real data

We ran the rule "close outside the Bollinger band (20, 2σ)" over daily bars for 12 currency pairs — 518 bars each, about two years. We counted frequency rather than profit: how often the state changes and what share of days the indicator spends in signal.

• State changes over the period: 62 (median across pairs)
• Annualised: 31.1
• Share of days in signal: 10.8%
• Sample: 12 pairs, 518 daily bars

The measurement is reproducible: the script and data are in the site's repository and the conditions are described in the methodology. Frequency is not return: it says only how often you will have to make a decision.

Why touching a band is not a signal

The textbook says roughly 5% of observations should fall beyond two sigma. Our measurement gave 10.8% — twice that, consistent with our fat-tails measurement. The practical conclusion is blunt: a move outside the band is not a rare event. In a trend price rides the upper boundary for weeks, and reading a touch as 'overbought' produces a run of losing trades against the move.

What the bands do tell you

A squeeze as a precursor to expansion. Volatility is cyclical, and a period of narrow range is statistically more likely to be followed by a wide one — confirmed in our volatility clustering measurement. A squeeze says nothing about the direction of the breakout, but it does say the range ahead is likely to be wider.

Bandwidth as a separate figure

A derived measure is often taken from the bands: the ratio of channel width to the centre line, expressed as a percentage. It removes the main inconvenience — absolute width depends on the price level, so comparing EUR/USD with USD/JPY by it is meaningless. Normalised bandwidth is comparable across instruments and against its own history. This is what people mean by a squeeze: a reading in the lower part of its own annual range.

How to use them

As a volatility gauge for stop distance and position size, not as a source of entries. Channel width suggests what stop distance is appropriate to the current market. Parameter choice is covered separately in how deviation and period change signal frequency.

This material is for educational purposes and is not individual investment advice.