What it measures

The average true range measures how much an asset moves in a bar, in price units, including any gap from the previous close. It is the standard yardstick for how far a stop should be and how large a position should be, and it is the third most used calculation in the strategies the desk screened.

The calculation

True range is the largest of three distances:

TR = max(high − low, |high − previous close|, |low − previous close|) ATR = Wilder's moving average of TR over length bars

On a 24-hour market gaps are rare, so TR is usually just the bar's range; the gap terms matter on anything that closes overnight.

Inputs and defaults

  • Length: 14. Shorter lengths track volatility changes faster and jump more.
  • Smoothing: Wilder's RMA by default; some scripts use SMA or EMA, which changes every stop built on it.

How it is read

As a distance, not a direction. A stop placed 2 ATR below entry sits outside the normal bar-to-bar noise; a position sized so that 2 ATR equals one percent of the account risks the same amount in calm and wild markets. Supertrend, Keltner Channels and the Chandelier exit are all ATR multiplied by something.

Where it fails

Wilder's smoothing makes it slow to come down: after a volatility spike the ATR stays high for many bars, so stops widen and position sizes shrink long after the market has calmed. In the other direction, a quiet stretch shrinks the ATR and the first big bar blows through every stop sized from it. ATR tells you how much the market has been moving, not how much it is about to.

How the desk reads it

The desk reads any ATR-sized strategy on what happened to it when volatility doubled in a week, because that is when the number is most wrong, and on how often the stop was hit by noise inside the bar rather than by a real reversal. Sizing in ATR units is a sensible default; it is not a defence against regime change.

In the desk's screen

Of the 614 screened strategies in the library, 295 (48.0%) use this indicator somewhere in their logic.

118 of those were gross-profitable on enough cells to reach the after-fee replay.

The count is read from the compiled code of each strategy in the library; a strategy that computes the indicator by hand under another name is not counted. The desk's own rules are not part of the count.