Bollinger Bands
Also called Bollinger band
Bands set a number of standard deviations, by default two, above and below a moving average, by default 20 periods. Developed by John Bollinger, they widen when volatility rises and narrow when it falls.

- Level
- Intermediate
- Reading
- 2 min
- Sources
- 5Original authorBookReference
- Updated
On this page· 5 sections
A fixed band around price, say 3% above and below an average, fits a quiet market and fails in a wild one. Traders who used such percentage bands had to keep redrawing them, and John Bollinger’s answer was to let volatility set the width[1].
The result is three lines: a moving average in the middle and a band on each side, placed by the standard deviation of recent prices[1][4]. The bands widen when volatility rises and narrow when it falls[3]. Bollinger set out the method in the 2001 book Bollinger on Bollinger Bands[2].
How it works
The defaults are 20 periods and two standard deviations[1][5]:
- middle band = 20-period simple moving average
- upper band = middle band + 2 × 20-period standard deviation
- lower band = middle band − 2 × 20-period standard deviation[3]
Bollinger uses the population form of the standard deviation[1]. The middle band is a simple average because the standard deviation is calculated from a simple average too, and Bollinger wants the two to be consistent[5].
The defaults can change. To contain price in a similar way, Bollinger suggests 1.9 standard deviations at 10 periods and 2.1 at 50 periods[5].
The bands give a relative definition of high and low: by definition, price is high at the upper band and low at the lower one[5]. Two tools follow from them[1][5]:
- %b shows where price sits relative to the bands, using an adaptation of the stochastic formula: (price − lower band) ÷ (upper band − lower band).
- BandWidth shows how wide the bands are, normalized by the middle band: (upper band − lower band) ÷ middle band. Its most popular use is to spot “the Squeeze”, a stretch of unusually narrow bands.
Example
Say a currency pair’s 20-period SMA is 1.1000 and its 20-period standard deviation is 0.0040.
- Upper band = 1.1000 + 2 × 0.0040 = 1.1080. Lower band = 1.1000 − 0.0080 = 1.0920.
- BandWidth = (1.1080 − 1.0920) ÷ 1.1000 = 1.45%.
- With price at 1.1060, %b = (1.1060 − 1.0920) ÷ (1.1080 − 1.0920) = 0.875: high in the range, still inside it.
Now say the market calms and the standard deviation halves to 0.0020. The bands narrow to 1.0960 and 1.1040, and BandWidth falls to 0.73%. The same price of 1.1060 now sits above the upper band, with %b at 1.25.
Watch out
A touch is not a signal. Bollinger’s rules say a tag of the upper band is not in itself a sell signal, and a tag of the lower band is not in itself a buy signal[5]. In a strong trend, price can walk up the upper band or down the lower one[3][5].
Closes outside the bands are, in Bollinger’s reading, at first continuation signals rather than reversal signals[5].
The bands are not a probability fence. Bollinger warns that prices are not normally distributed and that 20 periods is too small a sample for statistical significance. With the default settings, Bollinger typically finds about 90% of the data inside the bands, not 95%[5].
Sources
- Original authorBollingerBands.com (John Bollinger)Bollinger BandsChecked November 9, 2026
- BookOpen Library (Internet Archive)Bollinger on Bollinger Bands (Bollinger, McGraw-Hill, 2001)Checked November 9, 2026
- ReferenceStockCharts ChartSchoolBollinger BandsChecked November 9, 2026
- ReferenceStockCharts ChartSchoolGlossary - BChecked November 9, 2026
- Original authorBollingerBands.com (John Bollinger)Bollinger Band RulesChecked November 9, 2026
