ATR (Average True Range)

ATR is an indicator that measures how wide the price moves within a bar, on average, over a chosen stretch of history. It answers a single question: what range should be considered normal for this security. ATR says nothing at all about the direction of the move — a rising reading only means that bars have become wider, and it falls just the same whether a rally or a sell-off is fading. In essence it is a numerical expression of volatility, tied to a specific instrument and a specific timeframe.

How the true range is built

The ordinary range of a bar is the distance from its high to its low, but it ignores any gap between bars. That is why the true range is defined as the largest of three values: the range of the bar itself, the distance from its high to the previous bar's close, and the distance from its low to that same close. This way an opening gap is captured in the calculation rather than lost.

The resulting series of true ranges is then averaged — the classic version uses exponential smoothing, so older bars carry less and less weight in the result but never drop out of it entirely. Hence its key property: ATR responds to a spike in range with a delay, and the longer the look-back period, the more sluggish the response. It is a lagging indicator: it describes what has already happened.

Example: one instrument, one timeframe

A bar's range can be read straight off the chart — from the length of the candle wicks and the distance between neighbouring bars:

Stretches where the candles lengthen and overlap one another with gaps produce a high ATR; stretches of a calm sideways market with short bodies produce a low one.

Where ATR is misread

The main mistake is comparing the ATR of two different securities. The indicator is expressed in the same units as the price, so an expensive stock almost always has a higher reading than a cheap one, and that says nothing about how "jumpy" it is. The measure only becomes comparable once it is divided by the price, that is, converted into relative terms.

The second mistake is carrying a value over between timeframes: ATR on an hourly chart and ATR on a daily chart measure different things and cannot be converted into each other by simple multiplication. The third is days with a scheduled gap: a record date, a stock split, the resumption of trading after a halt. The true range of such a bar includes the entire gap, and for a long time afterwards the indicator shows a range that is no longer present in trading. At such moments it is more useful to look not at ATR but at the market regime as a whole.