Bollinger Bands Explained: Bands, Squeezes and Limits
Bollinger Bands are a price-chart indicator built around a moving average, with upper and lower bands set a chosen number of standard deviations away. Band width describes a form of recent price dispersion, while a squeeze refers to relatively narrow bands. Neither touching a band nor observing a squeeze guarantees a directional move.
The three lines in a standard configuration
A common configuration uses a 20-period simple moving average of closing prices as the centre line. The upper band is SMA plus two times the standard deviation of prices across the same 20 observations; the lower band is SMA minus two standard deviations. Different charting packages may use population or sample standard deviation conventions, leading to slightly different values. The settings are conventions, not universal statistical requirements.
Worked example: three hypothetical bands
Suppose a fictional 20-period SMA is 100 and the corresponding standard deviation is 3 price units. With a multiplier of two, the upper band is 106 and the lower band 94. If the next price is 107, it is above the currently calculated upper band. That observation does not by itself establish a 95% probability of reversal or a profitable short sale; prices are serially dependent and their distribution need not be normal.
What a squeeze actually means
A squeeze is a period of relatively low band width, often measured as (upper−lower)/middle band. If bands are 94 and 106 around a centre of 100, relative width is 12%. If the bands later contract to 98 and 102 around 100, width is 4%. Narrow bands reflect reduced dispersion of recent prices under the chosen method. A breakout might follow in either direction, or the sideways range can persist.
Why touching an upper band can coincide with strength
A strongly trending asset may repeatedly close near or outside its upper band. The indicator is not a fixed ceiling: bands expand and move when price changes. Selling solely because price touches the upper band can be dangerous in a sustained rally, and buying a lower-band touch can fail during a steep decline. Trend context and volatility conditions must be assessed independently.
Band width versus return volatility
Bollinger Bands commonly calculate standard deviation of price levels, not log returns. This is different from annualised historical volatility derived from daily return variation. A trending market can produce wider price-level bands even when day-to-day return variability is stable. Comparing band width across assets with very different prices requires normalisation and consistent windows.
Using band location as descriptive context
The percentage location within the bands can be calculated as (price−lower band)/(upper band−lower band) when band width is nonzero. Values above one indicate price above the upper band; below zero means beneath the lower band. This describes the price’s position relative to a recent envelope and does not reveal fundamental value or future expected return.
Common methodological problems
Historical charts can make squeezes look obvious after an expansion; in live conditions the narrowest point is only known in hindsight. A backtest that marks a ‘squeeze’ by comparing with future widths leaks information. Define a threshold using information available at each historical date and account for transaction costs, overnight gaps and stop execution.
Research checklist
Identify input price, lookback, standard-deviation convention and multiplier. Compute a few values independently, document the squeeze threshold and define a directional rule separately. Evaluate trend and range-bound conditions with out-of-sample data. Do not infer a probabilistic price range from a two-standard-deviation construction without testing the price distribution and time dependence.
For further learning on this method and its limitations, see mean reversion and range conditions.
Bollinger Bands: how changing volatility widens bands
This educational comparison uses hypothetical figures and is not a prediction or investment recommendation.
| Input | Low-volatility case | Higher-volatility case |
|---|---|---|
| 20-period moving average | 100 | 100 |
| Rolling standard deviation | 2 | 5 |
| Band multiplier | 2 | 2 |
| Upper band (mean + 2σ) | 104 | 110 |
| Lower band (mean − 2σ) | 96 | 90 |
| Total band width | 8 | 20 |
This simplified calculation illustrates that the same moving-average level can coexist with much wider bands when dispersion rises. The standard-deviation convention, data source, sample length and use of adjusted prices can differ across implementations. Price touching a band does not establish that a reversal is imminent: during persistent trends, prices can remain near an outer band, and during sudden events they can move well outside it. For comparisons, specify the input window, multiplier and asset rather than assuming all Bollinger Band indicators are directly interchangeable.
Frequently asked questions
Does an upper-band touch mean overbought?
Not necessarily. The asset may be trending strongly.
Does a squeeze predict breakout direction?
No. It describes narrow recent dispersion, not direction.
Are Bollinger Bands a 95% prediction interval?
No. Price levels and financial returns do not automatically meet the assumptions behind that interpretation.