Indicators / WORKSHOP
ATR: is there room for a normal price fluctuation?
After a gap, True Range is 8 but ATR14 is about 2.43. Why is that consistent, and why does an average range not promise a safe stop?
How it works
A candle’s True Range is the maximum of High−Low, |High−previous Close| and |Low−previous Close|. ATR smooths a series of those ranges. This exercise uses Wilder smoothing (RMA): after initialization, new ATR14 = (13 × previous ATR14 + new True Range) / 14. It measures range in price units, without a direction.
Match instrument, timeframe, trading session, candle history and smoothing method. TradingView defaults to RMA but offers other methods. SMA14 requires the outgoing candle’s range; the previous average alone is insufficient. A historical decision before the new candle must use ATR already known then; do not insert the value calculated after its close. Price distance, quantity and monetary value per price increment separately determine monetary risk. ATR does not include all costs or specify a stop’s execution probability.
What to check
- Record ATR14 settings and the observation time: before the candle or after its close.
- Calculate all three True Range components, including distances from the previous close.
- Update RMA separately. For SMA, identify the outgoing range.
- Compare this scale with price-based invalidation and monetary risk; do not move protection solely to increase size.
A hypothetical example
All numbers are invented and expressed in price units. Before the new candle, the close is 100 and an already initialized ATR14 using RMA is 2. The new completed candle has High 108 and Low 105. True Range = max(3, 8, 5) = 8. New ATR14 = (13 × 2 + 8) / 14 = 34 / 14 ≈ 2.43. Counterexample: calling ATR 8 confuses one candle with a smoothed series. A separate SMA14 model also starts with an average of 2, so its 14 ranges sum to 28. If the outgoing range is 1, new SMA14 = (28 − 1 + 8) / 14 = 2.50; if it is 4, the result is (28 − 4 + 8) / 14 ≈ 2.29. Equal previous averages do not make different calculation methods interchangeable.
When to skip
Do not use the calculation when smoothing, candle composition, observation time or monetary value per price increment is unknown. If the minimum size exceeds the risk budget, no position size fits.
The claim “ATR is about 2.43, so a move above 2.43 is impossible” is disproved by the hypothetical candle’s True Range of 8. ATR is not a ceiling, confidence interval or guaranteed exit price. Using the updated ATR in a decision made before that candle introduces future information.
Try it yourself
In a historical replay, record ATR from the last closed candle before revealing the next one. Then reveal the new candle and manually check its True Range and updated ATR with the chosen smoothing method. Preserve both observation times. State the price units and separately convert a predefined risk distance into money using quantity and instrument specifications. This checks a calculation, rather than searches for a signal.
With previous close 100, High 108, Low 105 and previous RMA14 = 2, find True Range and updated ATR14. For a separate SMA14 previously averaging 2, calculate the new average when the outgoing range is 1 and when it is 4. Which value was known before the new candle?
Show explanation
True Range = 8; updated ATR14 with RMA = 34 / 14 ≈ 2.43. SMA14 gives 2.50 and about 2.29. Before the new candle, the known previous value was 2; the new values use the completed new candle. None determines price direction or maximum possible loss.
Separate a single candle’s range, smoothed ATR and monetary risk. Use only information available when the decision is made.
Sources and definitions
TradingView: Average True Range ↗TradingView · RMA and SMA definitions ↗Scenarios and exercises are educational FINBORS examples, not a verified trading strategy.