Tell us how you trade and we'll point you to the right research segment.
Talk to Our Team →Start with our beginner-friendly guides on market basics, order types, and risk management before you place your first trade.
Start Learning → Browse All Articles →Stop loss ATR sizing is the practice of setting how far a stop-loss sits from entry using the Average True Range indicator, rather than a fixed number of points or a round distance chosen without reference to how the instrument is actually behaving. Average True Range measures the typical range a price has covered per period over a lookback window, folding gaps, intraday swings and volatility spikes into a single figure that rises and falls as conditions change. This piece works through what the indicator is actually measuring, how a stop distance is derived from it, why a volatility-based stop behaves differently across calm and turbulent regimes, and the situations where even a well-sized ATR stop still fails to protect a position.
Complete Guide
Every Section
Takeaways
Average True Range is built from a simpler concept called true range, which for any single period is the largest of three measurements: the distance between that period’s high and low, the distance between the previous close and the current high, and the distance between the previous close and the current low. Using the largest of the three rather than just the high-low range is what lets the indicator capture gaps, since a period that opens sharply away from the prior close would otherwise be under-measured by a plain high-low calculation.
Average True Range then smooths this true-range figure over a chosen lookback window, producing a single rolling number that represents the typical range the instrument has covered per period recently. It is expressed in the same price units as the instrument itself, which is what makes it directly usable for sizing a stop rather than remaining an abstract volatility score.
Using one period’s range alone would make the measurement noisy and reactive to a single unusual session. Smoothing it over a window produces a more stable read of typical movement, one that still adjusts as conditions genuinely shift but does not lurch on the back of one outlier candle. The length of that window is itself a judgement call: a shorter window reacts faster to a change in regime but is noisier, a longer window is steadier but slower to reflect a real shift in how the instrument is currently trading.
The mechanical step is straightforward: take the current Average True Range reading and multiply it by a chosen factor to get the distance the stop sits from the entry price. A stop set at entry minus a multiple of ATR on a long position, or entry plus a multiple of ATR on a short position, is directly anchored to how much the instrument has actually been moving rather than to a distance picked from habit.
The multiplier is the part that carries the real judgement. A small multiplier keeps the stop close to entry, which limits the loss on any single trade but increases how often ordinary volatility alone triggers an exit before the underlying thesis has had a chance to play out. A larger multiplier gives the position more room to breathe through normal noise, at the cost of a wider loss if the stop is eventually hit. Neither choice is inherently correct; it depends on the holding period, the instrument, and how much drawdown per trade is actually tolerable.
A fixed-point stop applies the same distance regardless of how the instrument is currently trading. In a quiet stretch that distance can be needlessly wide relative to actual movement, giving up more room than the setup requires. In a volatile stretch, the same fixed distance can be far too tight, getting hit repeatedly by ordinary swings that have nothing to do with the trade thesis being wrong.
An ATR-based stop adjusts automatically as the indicator’s reading changes. When the market is calm, the calculated distance narrows on its own; when volatility picks up, the distance widens without requiring a manual recalculation. This is the core argument for using it: the stop tracks current conditions rather than staying frozen at whatever distance felt reasonable when the trade was first planned.
The same adaptiveness that makes ATR sizing responsive also means the stop distance can change meaningfully between one trade and the next on the same instrument, purely because volatility has shifted. A trader used to thinking in terms of a fixed distance has to get comfortable with position size itself varying trade to trade, since the risk per unit is no longer constant even when the amount of capital risked per trade is kept constant.
An ATR-based stop only completes the risk picture when it is paired with a position-sizing rule, because the stop distance alone does not say how many units to trade. The usual approach is to decide the amount of capital willing to be risked on a single trade first, then divide that amount by the ATR-based stop distance to arrive at the position size that keeps the risk on that trade consistent with the plan.
This is where volatility-based stops earn their keep relative to fixed-point stops: because the stop distance itself reflects current volatility, the resulting position size automatically shrinks in more volatile conditions and grows in calmer ones, for the same amount of capital risked. A fixed-point stop paired with a fixed position size does not make this adjustment, which is part of why traders who rely on it tend to be over-exposed precisely when volatility is highest.
Recalculating position size only at the moment a trade is opened is usually sufficient, since the position size decision is made once and then held for the life of that trade. What does need periodic attention is checking whether the ATR reading used at entry still reflects current conditions if a position is held for an extended stretch, since a volatility regime that has shifted materially since entry can leave the original stop distance either unnecessarily wide or dangerously tight relative to what the instrument is doing now.
There is no single correct lookback period or multiplier that applies across every instrument and holding period; both are choices that should reflect how the position is actually being traded. A shorter lookback period reacts quickly to a genuine change in volatility regime but can also whipsaw around a single unusually wide or narrow session. A longer lookback period is steadier but slower to reflect a real shift in conditions, which matters if volatility has just stepped up sharply and the stop needs to widen without much delay.
A short-holding-period trade generally uses a tighter multiplier, since the position is not expected to absorb much adverse movement before the thesis is either confirmed or invalidated. A position intended to be held through several sessions or longer typically uses a wider multiplier, giving the trade room to breathe through ordinary daily noise without being stopped out on a swing that says nothing about whether the broader view is still correct. Matching the two deliberately, rather than defaulting to one multiplier for every trade regardless of intended holding period, is one of the more common adjustments experienced users of this method make over time.
In a clearly trending market, an ATR-based stop can be trailed behind price as the trend develops, recalculating the distance at each new period so the stop tightens as the position moves favourably while still respecting current volatility rather than an arbitrary breakeven-plus rule. This lets a trending position capture more of the move while still protecting against a genuine reversal.
In a range-bound market, the same logic applies differently: an ATR-based stop set too tight relative to the width of the range gets triggered by ordinary oscillation within that range, well before the range itself has actually broken. Recognising which regime the instrument is currently in — trending or range-bound — before deciding how tightly to apply the stop is part of using the method correctly rather than mechanically.
A few habits undermine the method more often than the calculation itself does:
No volatility-based calculation removes the risk of a gap that jumps straight past a stop level, since ATR is derived from recent past behaviour and cannot anticipate an overnight or pre-market move that exceeds anything in the lookback window. A stop that would have comfortably absorbed ordinary volatility can still be executed well beyond its intended level when the market opens with a large gap through it.
ATR-based sizing also assumes recent volatility is a reasonable guide to near-term volatility, which holds most of the time but breaks down around scheduled events — earnings, policy announcements, major data releases — where the market can reasonably be expected to move well outside its recent range. Being aware of upcoming events that could invalidate the assumption behind the calculation is part of using the method responsibly rather than applying it mechanically in every situation.
Liquidity also matters in a way the calculation itself cannot capture. A wide bid-ask spread or thin order depth on a particular instrument means a stop order can be filled meaningfully worse than its trigger price even without a gap, simply because there is not enough resting interest at the intended level when the order needs to execute. An ATR-derived distance calculated from price data alone says nothing about how cleanly that stop will actually be filled once it is triggered, which is a separate consideration worth checking before relying on the method for an illiquid instrument.
Neither approach is universally better; they solve different problems. An ATR-based stop adapts to current volatility automatically, while a fixed-point stop is simpler but can be badly mismatched to conditions that have changed since it was set.
There is no single correct multiplier; it depends on holding period and how much room the trade thesis genuinely needs. Shorter-term trades tend to use tighter multiples, longer-term positions tend to use wider ones.
No. ATR is calculated from recent past price behaviour, and a stop order can still be executed well beyond its intended level if the market gaps through it before the order can be filled at the set price.
Not necessarily. Different instruments and holding periods can call for different lookback windows, since a period that suits a fast-moving instrument may be poorly matched to a steadier one.
Explore our All Services service or get in touch with our research team.