Level 2 Patterns & indicators · Part 8/9
Measuring Volatility: Bollinger and ATR
Volatility is the one quantity in technical analysis that can be forecast reliably – not in direction, but in its own persistence: quiet phases follow quiet phases, active ones follow active ones. This clustering is among the most robust empirical findings in financial research.
Bollinger bands
A moving average, usually 20 periods, plus and minus two standard deviations. The width of the bands is the volatility.
The most common misuse is the same as with the RSI: price at the upper band is taken as expensive. It is not. In strong trends price rides the band; that is the definition of a strong trend, not a reason to sell.
The usable application is the squeeze: when the bands contract unusually tightly, expansion follows with elevated probability. What the squeeze does not say is the direction – and that is exactly where most descriptions stop.
Average true range
The true range of a period is the largest of three distances: high minus low, high minus previous close, previous close minus low. That last part matters – it accounts for gaps, which a plain high-low range would ignore. ATR is the average of those values over n periods, conventionally 14.
ATR is not a directional statement and not a signal. Its value lies in two practical applications:
Stops in ATR units. A stop “2 per cent below entry” means plenty of room when volatility is low and almost none when it is high. A stop “two ATR below entry” keeps the distance constant relative to actual movement – across instruments and market regimes.
Position size. With the stop expressed in ATR units, position size follows directly from the risk formula : risk in currency divided by (ATR × factor). More volatile instruments automatically get smaller positions – risk per position stays constant instead of fluctuating with the instrument’s temperament.
That second application is probably the single most practical building block on this level.
Keltner channels
Similar to Bollinger bands, but built around ATR instead of standard deviation. The difference is not cosmetic: standard deviation reacts strongly to individual outliers, ATR more evenly. For stop logic the ATR variant is usually the steadier one.
Why volatility is the more dependable quantity
Price direction is barely forecastable – that is the core of the critique of technical analysis. Volatility partly is, because it is strongly autocorrelated.
This is precisely why the practical value of this chapter lies in risk management rather than signals. You cannot know where price is going, but you can estimate reasonably well how far it will move over the next week. Translating that estimate into stops and position sizes uses the most reliable part of the entire toolbox.
Next
The final article of this level: how to test a pattern or indicator at all.