Bollinger Bands Explained

What Bollinger Bands actually measure

Three lines. The middle one is a 20-day simple moving average. The outer two sit two standard deviations of price above and below it. John Bollinger built them in the early 1980s, and the innovation wasn't the average: it was making the envelope around it adaptive.

Standard deviation is a measure of how spread out recent prices have been. When a stock starts moving violently, standard deviation rises and the bands widen. When it goes quiet, they contract. The bands breathe with the stock.

Which means Bollinger Bands are, first and foremost, a volatility indicator. They answer: is this move large relative to how much this stock has been moving lately? Not is this stock cheap or expensive? Almost every mistake people make with the bands comes from reading a volatility statement as a direction signal.

The mistake almost everyone makes

Touching the upper band is not a sell signal. This is the single most common misuse of the indicator, and Bollinger himself has said so directly and repeatedly.

Here's why it seems like one, and why that intuition fails. Price at the upper band is unusual: it's two standard deviations above the 20-day average, which by construction doesn't happen often. The mistake is concluding that unusual means unsustainable.

In a strong uptrend, price will walk the band: ride along the upper edge for days or weeks, tagging it repeatedly while the stock keeps climbing. That behaviour isn't exhaustion, it's what genuine strength looks like on this indicator. A trader who sells every upper-band tag systematically sells the best-performing stocks earliest, over and over.

The same applies inverted at the lower band during a decline. A band tag tells you the move is large relative to recent volatility. It tells you nothing about whether it will continue. To get direction you need something else: a trend reference like a longer moving average, or a momentum read like MACD.

The squeeze, and what it doesn't tell you

The squeeze is the most genuinely interesting thing the bands do. When volatility collapses, the bands narrow toward the moving average, sometimes dramatically.

The reasoning behind trading it is sound and is one of the better-supported observations in market behaviour: volatility clusters. Quiet periods tend to be followed by active ones, and violent periods tend to be followed by more violence. A squeeze says the stock has gone unusually quiet, and quiet does not last indefinitely.

But there are two limits, and both are load-bearing:

There's a further trap worth naming: the false breakout out of a squeeze is common enough to be its own pattern. Price breaks one way, triggers the obvious entries, then reverses through the whole range. This is why squeeze strategies usually pair the breakout with a confirmation rule rather than firing on the first bar outside the band.

What "20, 2" means, and how often price really stays inside

The 20 is the moving average length. The 2 is how many standard deviations out the bands sit. Both are adjustable, and the defaults are a reasonable place to stay.

Widening to 2.5 standard deviations means fewer, more genuinely extreme touches. Narrowing to 1.5 means price is at a band constantly and the signal loses meaning. Changing the 20 changes the middle band and both outer bands at once, since the standard deviation is computed over the same window, so it alters every reading simultaneously.

Now the part usually stated wrongly. You'll often read that two standard deviations contains "about 95% of price action". That figure comes from the normal distribution, and stock returns are not normally distributed. Real returns have fat tails: extreme moves happen considerably more often than the bell curve predicts.

In practice, something closer to 85-90% of closes land inside the bands. And the excursions aren't randomly scattered; they cluster during exactly the volatile, fast-moving periods when you most need the indicator to behave. Treating the bands as a reliable 95% container is how people end up short a stock that keeps going.

How traders actually use them

Where Bollinger Bands break down

They're derived entirely from recent price. The bands only know the last 20 days. A stock that has been quiet for a month has tight bands going into an earnings report that will move it 15%, and the bands offer no warning, because nothing in their input knows the event is coming.

They adapt after the fact. Volatility rises, then the bands widen. In the first days of a genuine regime change the bands are still calibrated to the old, calmer market, which is precisely when a band tag is most misleading.

And like every indicator here, they describe rather than predict. Whether "buy the lower band in a stock above its 200-day average" makes money on the stock you trade isn't answerable from a chart or an article. It's answerable by running the rule and reading the numbers.

Frequently asked questions