Bollinger Bands: What the Envelope Around Price Measures

Bollinger Bands consist of three lines. The middle is a simple moving average, conventionally over 20 periods. The outer two sit a fixed number of standard deviations above and below it, conventionally two. Because standard deviation is recalculated every period, the outer lines expand when price has been moving a lot and contract when it has been quiet.

That expansion and contraction is the entire point. A fixed percentage envelope would say the same thing about a calm stock and a violent one. A volatility scaled envelope adjusts to each.

What the position within the bands means

Price near the upper band means the current level is high relative to its own recent range. Price near the lower band means the opposite. The word "relative" is carrying the weight: the bands measure a stock against its own recent behaviour, not against any notion of value.

The most common misreading follows directly. Touching the upper band is routinely described as overbought and read as a sign that a fall is due. It is not. In a strong trend, price can ride the upper band for weeks, and each touch is the band confirming that the move is large by recent standards, which is what a strong trend is. The same misreading affects RSI and has the same answer: a measure of unusualness is not a measure of exhaustion.

The squeeze

When the bands narrow, recent movement has been small. Practitioners call this a squeeze and watch for it, on the reasoning that quiet periods in markets have historically tended to be followed by louder ones.

Two things are worth being precise about. First, the squeeze indicates that volatility has been low, which is a description of the past, and the observation that low volatility clusters and then breaks is a statistical tendency rather than a rule. Second, and more importantly, a squeeze contains no directional information at all. It says movement has been compressed. It does not say which way any subsequent move goes. Material that presents a squeeze as a directional setup is adding something the calculation does not contain.

The standard deviation assumption

Two standard deviations is chosen because, for a normal distribution, roughly 95 percent of observations fall inside that range. Financial returns are not normally distributed. They have fatter tails, meaning extreme moves occur considerably more often than the normal distribution predicts.

The practical consequence is that band touches happen more frequently than the statistics implies, and large breaches happen far more frequently. Treating a band touch as a rare event, at the frequency the underlying maths suggests, systematically overstates how unusual it is.

Settings, and why changing them proves little

The 20 period, two deviation default is convention rather than derivation. Shorter periods react faster and produce more touches. Wider deviation settings produce fewer. Every combination can be tested against history and some combination will always look best on any given stretch of it.

That last point is the important one and it applies to every configurable indicator. Optimising settings against past data finds the settings that fit that data, which is a different thing from finding settings that work. The same trap appears in MACD explained.

What it does well

Used descriptively rather than predictively, the bands answer a specific and genuinely useful question: is this move large or small by this stock's own recent standards? That normalisation is hard to do by eye, since a three percent day is dramatic for a utility and routine for a small cap biotech, and the bands handle it automatically.

They also make volatility regime changes visible. A stock whose bands have been widening for a month is behaving differently from one whose bands have been flat, and that is often the more informative observation than anything about where price sits inside them.

Volatility as one reading

How a stock is moving relative to its own recent range is one input to the price dimension EquityBias reads per covered stock. It sits alongside fundamentals, analyst activity and news sentiment, and no single one of those carries a conclusion on its own. Where they disagree, the gap is reported as divergence rather than averaged away. Related: moving averages for the line the bands are built around, and volume in trading for what participation adds. Current readings by sector are in the coverage directory.

EquityBias is a market data research tool. Nothing here is financial advice. Indicator readings describe past price movement and do not indicate future performance.