ATR Trailing Stop: A Choice About Time
An ATR trailing stop sets the trail distance as a multiple of the average true range, so it widens automatically in volatile conditions and tightens in quiet ones. The multiple is a decision about holding period rather than about safety.
Covered on this page: TradingView.
How it works
The distance is a multiple of volatility rather than a fixed amount. Instead of trailing by fifty points, you trail by two average true ranges, and the actual number of points changes as the market does.
That solves the problem a fixed stop has. Fifty points is a wide stop in a quiet session and a tight one in a fast session, so a fixed distance is wrong in both directions at different times.
The average true range is a smoothed measure of how far a market moves per bar. Because true range adds any gap from the previous close to the bar’s own high-to-low span, the average never reads below the plain average of those spans over the same bars.
And it varies a great deal. Say the reading is 1.00 in a calm month and 2.50 in a fast one: a two-range trail is then 2.00 points in the first and 5.00 in the second. A fixed stop distance is correct at one point in that spread and wrong across the rest.
What the multiple buys
The multiple maps mostly onto holding period. It changes how long you stay in far more than it changes whether you eventually get stopped.
A wider trail lets more ordinary noise pass without an exit. The back-and-forth has to exceed the distance to trigger the stop, so each step up in the multiple keeps the position open through swings that would have closed a tighter one — and at a fixed position size, each step also adds risk per trade in direct proportion.
The trade-off worsens as the width grows. Large counter-moves are rarer than small ones, so each extra range of room tends to buy less useful time than the step before while costing the same added risk, which is what makes two to three the interesting region rather than an arbitrary convention. And a trail of any width is still reached eventually by an ordinary reversal, so the width is not buying survival.
In practice
True range is a price measure only. A wide bar on no volume and a wide bar on heavy volume produce the same reading, and only one of them means the market is genuinely more volatile.
The reading is per bar, so it inherits the timeframe. A two-range trail on a daily chart and on a five-minute chart are entirely different risk decisions with the same label.
Worth stating plainly: true range differs from the bar’s own high minus low only when a gap occurs. On a market that trades continuously, each bar opens where the last closed, so true range and bar range match and the average is simply a smoothed bar range; on a stock that gaps overnight, the gaps are folded into the reading.
The tool gives a distance and nothing else. Where to start, which way to trade and when to enter all come from somewhere else entirely.
A one-range trail means constant turnover. A stop that close is reached by ordinary noise within a few bars, and each cycle pays a round trip in spread and commission, which is the hidden cost of a tight volatility stop.
Nobody is defending the level. Unlike a swing low, an ATR stop sits at a price nothing happened at, which is a genuine argument for placing it beyond structure rather than instead of it.
Setting the period as well as the multiple
The lookback is the second parameter and it is usually left at 14 without thought. A shorter period reacts faster and produces a stop that tightens quickly after a calm stretch; a longer one is steadier and slower to widen when conditions change. Both effects show up as changes in holding time, in the same way the multiple does.
The practical approach is to fix the period and vary only the multiple. Two parameters that both control the same thing make results impossible to attribute, and there is no evidence that tuning the lookback adds anything the multiple cannot do more legibly. Write both numbers into the plan and change one at a time.
What an ATR trailing stop is not
It is not safer than a fixed stop. It is better scaled.
It is not a market view. The distance says nothing about direction.
It is not gap protection. Nothing that rests at a price is.
And it is not a fixed risk. The distance moves, so position size must too.
How this connects
This is a stop-loss whose distance is set by volatility rather than by preference. ATR supplies the distance; a multiple of it supplies the room.
That is the argument for it. A fixed-point stop is too tight in a volatile market and too loose in a quiet one, and average true range is a direct measure of which you are in.
It is nearly the same tool as a chandelier exit, which anchors the same ATR multiple to the highest high since entry rather than to the current bar.
It only ratchets one way. The stop follows price in your favor and never retreats, which converts an open position into a defined worst case that improves over time.
And it still fills where the market is. A gap through the level skips it, so the stop promises an order rather than a price.
When it fails
In a range a tight multiple stops out repeatedly with nothing to show for it, and because the readings are low in quiet conditions, the stop is at its tightest exactly where the noise is proportionally largest.
The second failure is holding position size constant. A wider stop with the same size is more money at risk, and the whole point of scaling the distance is lost if the size does not scale inversely.
A third is measuring the range at entry and never updating it. A trail that keeps a distance set in a calm session is a fixed stop wearing a volatility label.
A fourth is treating 2× as standard. It is a common default, and the right multiple comes from how long the method needs to hold a position, which a survival test on your own instrument answers.
And a fifth is expecting the width to prevent stop-outs. Even a four-range trail is reached by a large enough reversal, and a gap can jump straight past it to a worse fill.
The original data
The S&P 500’s worst fall shows what the multiple is really choosing. From the 9 October 2007 close of 1,565.15 the index fell 56.8% to 676.53 on 9 March 2009, and it did not close above the 2007 peak again until 28 March 2013, about 5.5 years later. A trail of any width would have exited somewhere in that decline; the multiple decided only how much was given back first. A trail sitting 10% under the peak close, at about 1,409, would have gone early in a fall that ran more than five times deeper.
The swing in the volatility reading itself is the argument for the whole approach. Daily moves in the calm of 2007 and the panic before the March 2009 low were nothing like the same size, so a stop distance that does not move with them is materially too tight in one condition or too wide in the other. Run the survival table on your own instrument and pick the multiple whose median holding period matches your method’s — it converts the most argued-about number in stop placement into an arithmetic question with an answer.
Related
ATR is the underlying measurement and how it is calculated. Trailing stop is the general mechanism and the full survival table. And chandelier exit is a named version anchored to the highest high.
Switching from a fixed point stop to a volatility-scaled one removes an entire category of mistake. You stop being knocked out constantly in fast conditions and stop sitting through pointlessly wide risk in quiet ones. The multiple can change over time; the fact that it scales should not.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money. Some links on this page earn a commission if you buy through them. It costs you nothing and it does not decide what appears here or in what order — how these pages are made is set out in our methodology.