Bollinger Band Squeeze: Trading the Volatility Breakout

A Bollinger Band squeeze is a period when the bands contract unusually close to the middle line, meaning price has been moving less than it usually does. Contracted volatility tends to be followed by expanded volatility, so a squeeze is a timing signal: something is more likely to happen soon.

It is not a direction signal. The bands are built from standard deviation, which measures the size of moves and is blind to their sign. Every workable squeeze strategy gets its direction from somewhere other than the squeeze itself.

What the bands are

Three lines. The middle is a moving average, conventionally twenty periods. The outer two sit a set number of standard deviations above and below it, conventionally two.

Standard deviation is a measure of dispersion: how far recent closes have strayed from their own average. When price has been moving in small steps, the number is small and the bands draw in tight. When price has been swinging, the number is large and the bands widen.

That construction has a consequence worth stating plainly, because a great deal of bad advice depends on not noticing it. The bands adapt to whatever price has just done. They are not fixed levels and they are not support and resistance. Price touching the upper band does not mean expensive; it means price is roughly two standard deviations above its recent average, which in a strong trend is an ordinary place to be for weeks at a time.

Measuring the squeeze properly

Eyeballing “the bands look narrow” is unreliable, because narrow compared with last month and narrow compared with last year are different claims. Bandwidth turns it into a number.

Bandwidth is the distance between the bands expressed as a fraction of the middle line:

bandwidth = (upper band − lower band) ÷ middle band

Chart that value underneath price. A squeeze is bandwidth sitting near the bottom of its own range over whatever lookback you care about — six months is a reasonable default.

Two properties matter. Dividing by the middle line makes the figure comparable across time for the same instrument, because it adjusts for price level: bands four points wide mean something different on a 40 stock than on a 400 one. And it is not comparable between instruments, since each is being measured against its own volatility history. A bandwidth of 0.04 is tight for one stock and ordinary for another.

Why the direction has to come from elsewhere

The squeeze tells you a quiet period is likely to end. It has no view on whether the resolution is up or down, and no amount of staring at the bands will produce one.

Three places to get direction instead.

The higher timeframe trend. A squeeze on the daily chart inside a rising weekly trend resolves upward more often than one inside a falling weekly trend. This is the ordinary argument in trade with the higher timeframe trend, applied to a specific setup.

Structure. A squeeze forming just under a level that has capped price three times is a different proposition from one forming in open space. The level, not the bands, is the thing you are trading.

The resolution itself. The most honest answer is to let the market choose, and take the trade after it has. That costs the first part of the move and removes most of the guessing.

Trading it

Wait for a close, not a touch. Price tags the band intraday constantly, including during the squeeze itself. A close beyond the band is a weaker filter than people assume but still far better than a touch.

Prefer confirmation on the following session. The classic failure is a push out of the range that triggers the stops resting just beyond it, then reverses through the other side. A quiet market accumulates orders close to the range on both sides, which is exactly what makes that whipsaw so common.

Place the stop on the other side of the range, not at the middle band. The range that formed during the squeeze is the structure. If price re-enters it and crosses to the far side, the breakout failed, and that is the level that says so.

Size from the stop distance. This is the part that catches people out. A squeeze means a narrow range, which means a close stop, which means the position size the arithmetic allows is unusually large. That is correct only if the stop is genuinely where the idea fails. The method is in how to set a stop loss using ATR and the position size calculator.

What a squeeze does not tell you

How long the quiet lasts. Contracted volatility can persist far longer than seems reasonable. A squeeze is not a countdown, and entering early on the grounds that it has gone on long enough is a position taken on impatience.

How far the expansion runs. Volatility expanding does not imply a large trend. Many squeezes resolve into a brief burst and then a new quiet range at a slightly different level.

Whether the first move is the real one. Frequently it is not, which is the single most expensive thing about this pattern.

The band-walk, and why it contradicts the usual advice

The most common misuse of Bollinger Bands is treating the outer band as a reversal signal: price reached the upper band, so sell. In a genuine trend, price frequently “walks the band”, closing near or outside it session after session while the trend continues.

That behaviour is not an anomaly. Two standard deviations contains roughly the bulk of recent moves given the recent distribution, and a stock entering a sustained trend has changed its distribution. The indicator is measuring the old one, and lagging.

So a close outside the band in a strong trend is better read as evidence of strength than as a signal to fade. It is the opposite conclusion from the one most introductions to the indicator offer, and the difference between the two is whether there is a trend, which the bands themselves cannot tell you. Reading that from the chart’s structure is covered in support and resistance.

Why two standard deviations is a convention, not a law

The default places the bands two standard deviations from the middle line, and that figure gets repeated as though it carried a guarantee about how often price stays inside.

It does not, for a reason worth understanding. The statistical rule people have in mind assumes returns follow a normal distribution. Financial returns do not: large moves occur considerably more often than that assumption predicts, which is the well-documented fat-tails problem.

The practical consequence is that price trades outside the bands more frequently than the textbook figure implies, and that excess is concentrated in exactly the sessions that matter most — the sharp ones. An indicator calibrated on the assumption that such moves are rare will be surprised by them regularly.

So treat the bands as a description of recent dispersion rather than as a boundary with known odds attached. The question “how unusual is this move” has a useful answer from the bands; the question “how often will price exceed this level” does not.

A workable routine

1. Chart bandwidth beneath price, with six months of history visible.
2. Flag instruments where bandwidth is near the low end of that range.
3. Establish direction from the higher timeframe and from structure, before anything moves.
4. Wait for a close beyond the band, then for the next session to hold it.
5. Stop on the far side of the squeeze range; size from that distance.
6. If price re-enters the range, the setup is finished. Take the loss rather than waiting for the next attempt, because a failed squeeze frequently produces several.

Step three is the one that separates this from gambling on a coin flip with extra steps. The squeeze supplies the timing. Everything else has to supply the direction.