Bollinger Bands

A moving average tells you the typical price lately. But it can't tell you whether today's price is a normal little wander away from that average or an unusually big stretch. On a calm stock a 3% gap from the average is huge; on a volatile one it's nothing. So "how far from average is far?" depends entirely on how jumpy the stock has been.

Bollinger Bands, invented by John Bollinger, answer exactly that. Take a moving average, then draw two rails above and below it — but make the rails elastic, set by how volatile the stock has recently been. When price is wild the rails spread wide; when price goes quiet they squeeze in tight. The result is a self-adjusting envelope that shows, at a glance, how stretched price is right now relative to its own recent behaviour. This module explains how the bands are built, what the famous "squeeze" means, and why a band touch is emphatically not a buy or sell signal on its own.

The bands exist to add a missing dimension to the moving average: a sense of scale. On its own, a moving average is a single line, and price wandering 2 above it looks the same whether the stock normally swings by 1 or by 10. Bollinger's idea was to measure the recent swinginess and use it to size the rails.

The measure he used is the standard deviation — a statistic for how spread out recent prices have been around their average. A calm stretch has a small standard deviation; a wild one has a large one. The standard recipe is:

• The middle band: a 20-period simple moving average — the "typical price lately."
• The upper band: the middle band plus 2 standard deviations of the last 20 prices.
• The lower band: the middle band minus 2 standard deviations.

Because the rails are set by standard deviation, they breathe with the market. In a violent stretch the bands balloon outward; in a sleepy one they pinch together. That breathing is the whole value of the tool: it turns "how far is far?" into a moving, self-calibrating answer instead of a fixed guess.

One honest caveat lives right inside the maths. The "2 standard deviations" language borrows from the bell curve, where about 95% of values fall within two standard deviations. Real prices are not neatly bell-shaped — they have fat tails, meaning extreme moves happen far more often than a bell curve predicts. Price escapes the bands more often, and more violently, than the tidy statistics suggest. So treat "95% of the time price stays inside" as a rough rule of thumb, never a law.

Three behaviours are worth knowing, and none of them is a mechanical trade signal.

The squeeze. When the bands pinch tightly together, volatility has collapsed — the stock has gone unusually quiet. This matters because quiet periods tend to be followed by loud ones: a squeeze often precedes a larger move. But — and this is the crucial part — the squeeze tells you nothing about direction. Price can burst up or down out of a squeeze; the bands only say "a move is coming", not "here is which way." Traders who add a direction to a squeeze are inventing it.

Walking the band. In a strong trend, price does not bounce off the far band — it rides it. In a powerful uptrend price can hug the upper band for many bars while the trend runs; in a hard downtrend it walks the lower band down. This is why "touch equals reversal" is so dangerous: in exactly the strongest moves, the band touch means continuation, not a turn.

The snap-back. In a range-bound, going-nowhere market, price does tend to stretch to one band and then revert toward the middle. This is the behaviour the "mean reversion" crowd trades — buy near the lower band, sell near the upper — and it works only while there is no trend. The moment a trend starts, the same tactic has you selling strength and buying weakness.

Put those three together and the pattern is familiar from the whole of this part of the book: the bands behave one way in a trend and the opposite way in a range, and the tool cannot tell you which you are in.

Walk the composite in Figure 1 from left to right. illustrative

On the left, price barely moves. The bands pinch into a narrow ribbon — the squeeze. A beginner sees "nothing happening" and looks away. A careful reader sees the opposite: a coiled spring. The squeeze doesn't say when or which way, only that the current calm is unusual and rarely lasts. So the reader notes it and waits for price to declare a direction, rather than guessing one.

Then price breaks up out of the quiet range. As it accelerates, the standard deviation jumps and the bands balloon outward. Now watch the trap unfold: price is pressed right against the upper band. The "touch equals sell" crowd shorts here — and gets run over, because in a strong trend price walks the upper band, hugging it bar after bar as the move continues. The upper band is not a ceiling; it is a measure of how far price has stretched, and in a real trend it keeps stretching.

Eventually the move tires, the bands stop widening, price falls back toward the middle band, and — if the stock now chops sideways — the range behaviour returns, where stretching to a band tends to snap back. Nothing about the bands changed. The market state changed, and with it the correct way to read the same picture. That switch, from trend behaviour to range behaviour, is the entire skill.

Bollinger Bands cannot give you a direction from a squeeze. The squeeze is one of the most useful things the bands show — but it is strictly direction-neutral. Anyone who reads "a squeeze means up" (or "means down") is adding information the tool does not contain.

They cannot make a band touch a signal. Touching, or even briefly closing outside, a band means only that price is stretched relative to its recent average. In a range that stretch tends to revert; in a trend it tends to continue. A band touch nudges the odds depending on context and settles nothing on its own — which is why "touch the band and trade" is a losing rule half the time by construction.

They cannot honour the tidy statistics. The "2 standard deviations = 95%" framing assumes a bell curve, and prices are not bell-shaped; they have fat tails. Price breaks outside the bands harder and more often than the normal-distribution story implies, so a big move outside the bands is not the rare freak the maths suggests — it is a regular event you must expect.

And they cannot tell you what to own. The bands are drawn on price and know nothing of the business. The bands describe price volatility, not company value. They can help you gauge how stretched and how quiet a price is; they can never tell you whether the underlying company deserves a place in your portfolio.

1. Fading every band touch. The single most common Bollinger mistake: selling the upper band and buying the lower one regardless of trend. It works in a range and is ruinous in a trend, where price walks the band. The tactic and the market state must match.

2. Reading a direction into the squeeze. A squeeze forecasts volatility, not direction. Deciding in advance which way price will break out of a squeeze is guessing dressed as analysis.

3. Trusting the 95% statistic. Because prices have fat tails, price spends more time outside the bands, and lunges further past them, than the normal-curve figure implies. Sizing a position as if a band break were a once-in-twenty rarity underestimates the real risk.

4. Optimising the settings to old data. "22-period, 2.1 standard deviations backtests best on this stock." It fits this stock's past noise. Bollinger settings tuned to look perfect on history are usually fitted to that history's quirks, and the default 20-and-2 is popular mainly because everyone uses it — which, as ever, makes it a little self-fulfilling and no more magic than any other choice.