TL;DR. (upper band − lower band) / middle band. A BandWidth of 0.04 means the Bollinger Bands are 4% of price apart. A low reading is John Bollinger's Squeeze, a quiet market that often comes before a large move; a high reading is a market already moving hard.
What it measures
- How wide the bands are, next to price. The Bollinger Bands node gives the bands themselves and their width in points. BandWidth divides that width by the middle band, so it compares across prices and instruments.
- Volatility through the standard deviation. The bands sit a number of standard deviations from a simple average, so BandWidth moves with how spread out recent closes are, where the ATR follows the bars' ranges.
- Compression and expansion. It says when the market is coiled and when it has broken out, never in which direction.
Formula
Inputs in AlgoLift
| Setting | Default | Range | Notes |
|---|---|---|---|
| Period | 20 | 1 to 500 | Bars for the middle average and the deviation. Bollinger's value. |
| StdDev multiplier | 2.0 | 0.1 to 10 | How many deviations the bands sit from the middle. BandWidth scales with it in a straight line. |
| Source | Close | Close, Open, High, Low, HL2, HLC3, OHLC4 | Which value of each bar the bands are built on. |
Recommended periods
- 10 to 14 (fast): Short squeezes inside a session.
- 20 (default): Bollinger's own setting, the one his Squeeze work uses.
- 50 to 100 (slow): Regime-length compression over days or weeks.
Outputs in AlgoLift
| Handle | Type | Plotted | Notes |
|---|---|---|---|
| BandWidth | Numeric | Always | The bands' width as a fraction of the middle band, drawn in its own panel. |
How to read it
- At a multi-month low: the Squeeze. Volatility has contracted as far as it has in a long while, and a large move is more likely than usual. The reading does not say which way.
- Rising sharply from a low: the expansion has started. Breakout rules want this moment.
- At a high and turning down: a big move is losing energy; trends often pause or reverse into a range here.
- Compared across instruments: because it divides by price, 0.03 on two markets is the same relative width.
Best in, worst in
Best in breakout strategies, as the filter that says a range has been quiet long enough to be worth breaking.
Worst in steady trends, where the bands stay wide for weeks and BandWidth offers little beyond "the market is moving".
Three setups
1. The Squeeze breakout
Compare BandWidth with its own lowest value of the last 120 bars (a Rolling Min/Max Modifier) and AND that "at a low" condition with price crossing above the upper Bollinger Band through a Crossover node. Long on the break up, short on the mirror.
2. Expansion confirmation
Require BandWidth to be above its value 5 bars ago (a Bars Ago Modifier) before a breakout entry fires, so the strategy only trades breaks that widen the bands.
3. Range-only mean reversion
Fade touches of the bands only while BandWidth is below its 100-bar average. When the bands start to widen, the filter switches the fade off before the breakout runs over it.
Common mistakes
- Comparing BandWidth with %B. %B is where price sits inside the bands; BandWidth is how far apart they are. Bollinger Bands outputs %B as
percent_b. - A fixed Squeeze threshold on every market. What counts as narrow differs by instrument and timeframe. Compare BandWidth with its own history instead of a constant.
- Changing the multiplier to tune the signal. BandWidth is proportional to the multiplier, so a threshold and a multiplier tuned together are one parameter counted twice (overfitting in trading).
BandWidth is the Bollinger Bands' width next to price: low is a coiled market waiting to move, high is a market already moving. It never picks the direction.