Technical analysis
Bollinger Bands: Standard Deviation, the Squeeze and Walking the Band
How Bollinger Bands are built from a 20-period average and two standard deviations, what %B and bandwidth measure, and how the squeeze and band walks are read.
Updated · 3 min read
How the bands are built
Bollinger Bands put a measure of volatility on the price chart itself. John Bollinger introduced them in the early 1980s. Three lines are drawn:
- The middle band is a 20-period simple moving average of the close.
- The upper band is the middle band plus two standard deviations of the same 20 closes.
- The lower band is the middle band minus two standard deviations.
The standard deviation is the one from the mathematics of spread: the square root of the average squared distance of each close from their mean, σ = √( Σ(x − μ)² / n ). Because it is recomputed every bar from the last 20 closes, the bands widen when price has been swinging and narrow when it has been quiet. That is the feature that distinguishes them from a fixed percentage envelope: the bands are a volatility gauge drawn in price units.
Two derived readings are often shown. %B locates price within the bands, %B = (P − Lower) / (Upper − Lower): 1 at the upper band, 0 at the lower, 0.5 at the middle, above 1 outside the top. Bandwidth measures how wide the bands are relative to the middle, Bandwidth = (Upper − Lower) / Middle, which makes the width comparable across time and across instruments.
What a band touch means
If prices were normally distributed around the moving average, about 95% of closes would fall within two standard deviations. They are not, and the average is estimated from the same twenty closes it is compared against, so in practice Bollinger reports that around 88 to 89% of closes fall inside bands set at 20 periods and two deviations. A close outside the bands is therefore unusual but not rare: it happens roughly one bar in nine.
The common misreading is that a touch of the upper band is a sell and a touch of the lower band a buy. In a trend, the opposite is closer to the truth. A market that is trending strongly walks the band: it closes at or outside the upper band for bar after bar, and each touch is a sign of strength, not exhaustion. The bands tell you that price is extended relative to the last twenty bars. Whether that is an extreme to fade or a trend to follow is decided elsewhere, by the trend on a higher timeframe, by RSI and its divergences, or by what price does on the bar after the touch.
| Reading | In a range | In a trend |
|---|---|---|
| Close above the upper band | Stretched; a return to the middle is likely | Strength; a band walk may be starting |
| Close below the lower band | Stretched; a return to the middle is likely | Weakness; a band walk down may be starting |
| Return to the middle band | A fade of the move succeeded | A pullback in the trend; often where the next leg starts |
The squeeze
The most reliable thing the bands say is about volatility itself. Volatility is cyclical: quiet periods are followed by active ones and vice versa. When bandwidth falls to its lowest level in roughly six months the bands are in a squeeze, and an expansion usually follows. The squeeze gives no direction; it says that a large move is becoming likely and that the breakout, when it comes, is worth taking seriously. A false start, where price pokes out of one band and then reverses through the other, is common enough that the first move out of a squeeze is often the wrong one.
Settings
The 20-period, two-deviation setting is the default for a reason: the moving average and the deviation are computed over the same window, and twenty bars gives a stable estimate without making the bands too slow. Shorter windows need a smaller multiplier (Bollinger suggests 1.9 at 10 periods) and longer ones a larger (2.1 at 50) to keep roughly the same share of closes inside. Changing the multiplier without changing the window simply moves the bands and changes how often they are touched.
On the desk
The symbol pages overlay Bollinger Bands on the candles beside the moving averages, and the same standard deviation, measured from daily closes and annualised, is what sizes the ranges and the risk meter in the analysis notebooks, so a range quoted there is the instrument's own volatility rather than a round number.
Questions
- What do Bollinger Bands measure?
- Volatility, drawn on the price chart. The bands sit two standard deviations above and below a 20-period moving average, so they widen when price has been swinging and narrow when it has been quiet. Price at a band is unusually far from its recent average.
- Is touching the upper Bollinger Band a sell signal?
- Not on its own. In a strong trend price walks the band, closing at or beyond it for many bars in a row, and each touch is a sign of strength. A band touch is a sell signal only when other evidence, such as a trend on a higher timeframe that is already down or a divergence on RSI, says the move is ending.
- What is a Bollinger Band squeeze?
- A squeeze is when the bands are at their narrowest in about six months, meaning volatility has fallen to a low. Because volatility is cyclical, an expansion usually follows, so a squeeze warns that a large move is becoming likely without saying which way.
- What percentage of prices stay inside Bollinger Bands?
- With the standard settings of 20 periods and two standard deviations, roughly 88 to 89% of closes fall inside the bands. The figure is below the 95% a normal distribution would give because prices are not normally distributed and the average is estimated from the same data.
Related guides
Technical analysis
RSI explained
The relative strength index step by step: Wilder's formula, what 70 and 30 mean, why RSI stays overbought in a trend, and how divergence is read.
4 min read
Technical analysis
EMA 50 and EMA 200
How the 50 and 200 exponential moving averages are calculated, what a golden cross and a death cross mean, and how an EMA touch is read across timeframes.
5 min read
Macro and rates
The VIX explained
How the VIX is derived from S&P 500 option prices, the rule of 16 for a daily move, what its levels and term structure mean, and how it tracks stocks.
4 min read
Technical analysis
Pivot points
The classic pivot point formula with R1 to R3 and S1 to S3, a worked example, the Fibonacci and Camarilla variants, and which session to compute them from.
4 min read