Bollinger Bands Strategy Guide

At a glance

Construction: A moving average plus and minus k standard deviations
Default: 20 periods, 2 standard deviations
What a touch means: Price is statistically extended, nothing more
Two opposite uses: Reversion at the bands, or breakout after a squeeze

Key takeaways

The bands adapt to volatility automatically, which makes them more useful than fixed percentage channels across different instruments.
Price touching a band is normal: with 2 standard deviations, roughly 5 percent of closes fall outside the bands even in a random series.
In trends, price can ride the upper band for weeks. Fading band touches without a trend filter is a classic losing pattern.
The squeeze, a period of unusually narrow bandwidth, identifies volatility compression that frequently precedes expansion, though not in a predictable direction.
Percent B and bandwidth turn the bands into numerical values that can be tested, which is more useful than visual interpretation.

How the bands are built

Middle band = SMA(close, 20)
Upper band  = Middle + (2 x standard deviation of close over 20)
Lower band  = Middle - (2 x standard deviation of close over 20)

Percent B   = (close - lower) / (upper - lower)
   1.0 = at the upper band
   0.5 = at the middle band
   0.0 = at the lower band
   Values outside 0 to 1 mean price is beyond the bands

Bandwidth   = (upper - lower) / middle
   Low bandwidth  = compressed volatility (a "squeeze")
   High bandwidth = expanded volatility

The formula and the two derived measures.

Because the width is derived from standard deviation, the bands widen automatically in turbulent markets and narrow in calm ones. That adaptivity is the main advantage over fixed-percentage envelopes, and it means the same settings can be applied across instruments with very different volatility.

What a band touch actually means

With two standard deviations, a normal distribution would place about 5 percent of observations outside the bands. Financial returns have fatter tails, so the real figure is higher. Touching a band is therefore a routine event, not an unusual one.

Observation	Valid reading	Invalid reading
Close beyond the upper band	Price is extended relative to recent volatility	"It must reverse"
Repeated closes on the upper band	A strong trend is in progress	"Extremely overbought, sell"
Bands narrowing sharply	Volatility is compressing	"A big move is coming up"
Bands expanding rapidly	Volatility has expanded; risk is higher	"The trend is confirmed"
Price crossing the middle band	The short-term average has been crossed	A meaningful trend change

Mean reversion rules using the bands

Instrument: Liquid index ETFs or large caps. Not single volatile stocks, where the tail risk is severe.
Trend filter: Price above the 200-day moving average. Without this, you are buying declines in downtrends.
Entry: Close below the lower band, that is Percent B below 0, with the trend filter satisfied.
Exit: Close above the middle band (the 20-period average).
Time stop: Exit after 8 sessions regardless of outcome.
Disaster stop: 3 x ATR(10) below entry.
Size: Risk 0.5 percent of equity; maximum three concurrent positions, because band signals cluster.

This is a mean reversion strategy with the standard profile: a high win rate, small average wins, and occasional large losses when a decline continues. The trend filter and the disaster stop are what make it survivable.

The squeeze and breakout use

A squeeze occurs when bandwidth falls to the low end of its own historical range, indicating unusually compressed volatility. Volatility is mean reverting, so compression tends to be followed by expansion. What the squeeze does not tell you is the direction of that expansion.

Squeeze condition: Bandwidth in the lowest 20 percent of its trailing 6-month range, sustained for at least 5 bars.
Entry: Close beyond either band after the squeeze condition, in the direction of the break. Some versions require a trend filter to permit only one direction.
Stop: The opposite band, or the middle band, whichever is closer.
Exit: Trail using the middle band, or use an ATR trailing stop once the move extends.
Failure rule: If price closes back inside the bands within two bars, exit immediately. Failed expansions reverse quickly.
Caution: Squeeze breakouts have a substantial false-signal rate. The structure works because the occasional large expansion pays for them.

Practical notes

Use Percent B and bandwidth rather than visual band touches. Numbers can be backtested; eyeballing cannot.
The default 20 and 2 settings are reasonable. Widening to 2.5 standard deviations reduces signal frequency and improves reliability for reversion use.
Standard deviation assumes a distribution that financial returns do not follow, so band excursions are more frequent than the normal-distribution intuition suggests.
Bands based on a short lookback adapt quickly but become unstable; 20 periods is a reasonable balance.
Combining bands with Keltner Channels is a common squeeze definition, since Keltner uses ATR rather than standard deviation and the two respond differently to compression.

Frequently asked questions

Should I sell when price touches the upper Bollinger Band?

Not on its own. In trending markets price can remain at or beyond the upper band for weeks, so fading band touches means fighting trends. Band touches are useful as reversion signals only within a range-bound regime or with a trend filter that permits trades in one direction only.

What is the Bollinger squeeze?

A period when the bands narrow to an unusually small width, indicating compressed volatility. Because volatility mean reverts, compression often precedes expansion. The squeeze identifies the timing of a likely expansion but says nothing about its direction, which must come from another rule.

What are the best Bollinger Band settings?

The default 20 periods and 2 standard deviations is a sound starting point and has the advantage of not being fitted to your data. For reversion strategies, 2.5 standard deviations produces fewer and more meaningful signals. As always, prefer settings that work across a range rather than one that works uniquely.

Do Bollinger Bands work in all markets?

The construction adapts to volatility, which is why the same settings can be applied across instruments. What differs is the regime: bands work well for reversion in range-bound markets such as many currency pairs and poorly in strongly trending markets such as crypto during a bull phase. Match the strategy to the regime rather than assuming the indicator handles it.

What is Percent B used for?

It converts the band position into a single number, which makes it testable and comparable across instruments. A value above 1 means price closed above the upper band, below 0 means below the lower band. Using Percent B rather than visual inspection is what turns Bollinger Bands from a chart overlay into a strategy component.

Test this idea before you trade it

Describe the rules in plain language and AlgoTrader AI turns them into a structured strategy blueprint with a configurable historical backtest, cost assumptions, and exportable code.

Build a backtest

Keep reading

Referenced by

Educational use only. This guide explains how a strategy works. It is not investment advice, not a recommendation, and no result described here is a forecast. Test any approach on historical and out-of-sample data, size positions conservatively, and never risk money you cannot afford to lose.