Average True Range: Gaps, Smoothing and Units

Why the previous close matters

True range is the largest of the current high minus low, the absolute high minus previous close, and the absolute low minus previous close. Taking the largest distance allows a gap to contribute even when the current bar itself is narrow. Fidelity describes true range and ATR’s calculation.

The average depends on its smoothing convention. Under Wilder’s n-period update, new ATR equals (previous ATR × (n − 1) + current true range) / n. Other implementations may offer simple or exponential smoothing. The period count alone does not specify every arithmetic choice, especially near the beginning of a data series.

An example with a narrow bar after a gap

Assume the previous close was 100 and the next bar has a high of 105 and low of 103. Its visible high-low range is 2. The distances to the previous close are 5 and 3, so true range is 5. Using only high minus low would omit most of the change represented by the opening gap.

Suppose a previously initialized 14-period ATR equals 2.00. With this true range of 5, the next Wilder-style value is (13 × 2 + 5) / 14 = 2.2143. The calculation reacts to the gap while preserving earlier observations. It does not claim that the next bar will travel exactly 2.2143 units.

Now assume the instrument’s minimum price increment is 0.25. The new ATR corresponds to approximately 8.8572 ticks. An order cannot necessarily be entered at a fractional tick simply because the statistic has decimals. If a hypothetical rule multiplies ATR by two, its 4.4286-unit distance still needs a documented price-grid rule and a valid relationship to the entry price.

Volatility is separate from money at risk

ATR is in the same price units as its input. To estimate cash movement for a linear contract, multiply a price distance by the contract multiplier and quantity, with any required currency conversion. ATR alone does not know account size, commissions, gaps, spread or the broker’s execution policy.

A percentage version divides ATR by a selected price and multiplies by 100. That can aid scale comparisons, but it answers a different question from the raw price-unit series. Comparing a stock’s ATR in dollars with an FX pair’s ATR in quote-currency units without normalization is not meaningful.

Use the range estimate consistently

An unfinished bar can expand its high or low, so an intrabar ATR input may differ from the final value stored in a historical chart. Freeze the decision timestamp in a test. Futures rolls, stock adjustments and missing bars also deserve review because discontinuities in the input can appear as sudden volatility.

Check price-grid validity before constructing an order distance, and use the futures sizing calculator or its forex counterpart to translate a declared stop into a planned cash amount. Keltner Channels demonstrate another use of ATR: building a volatility envelope. None of these calculations guarantees a stop fill at its planned price.