ATR Adaptive Stop Loss and Position Sizing

Fixed-dollar stops treat a 0.5% volatility stock and a 4% volatility stock identically — and get stopped on noise in one, under-risked in the other. A backtest of a 50/200-SMA crossover system shows ATR-based stops outperform fixed-percentage stops by 15-25% in net profit across liquid equities. The reason is simple: ATR (Average True Range) adapts stops and size to current volatility, keeping risk per trade constant regardless of regime. It is the foundation of every robust trend-following system.

Core Concept

ATR measures average true range over N periods, including gaps. The "true range" for a period is the largest of: (high − low), |high − previous close|, or |low − previous close|.

Formula:

True Range = max[(High − Low), |High − PrevClose|, |Low − PrevClose|]
ATR = exponential moving average of True Range over N periods

Default parameter: ATR(14). ATR is volatility in price units — directly translatable to stop distance.

Concrete example: a $50 stock with a daily ATR of $2.00 moves roughly $2 per day on average. A 2.5× ATR stop sits $5.00 from entry, placing the stop beyond the normal day's range so random noise does not reach it. If volatility spikes and ATR doubles to $4.00, the same 2.5× multiplier automatically widens the stop to $10.00 — no manual adjustment needed.

The standard multiplier is 2.0-3.0; 2.5 is a balanced default. Lower multipliers get hit more often (55-65% stopped on noise); higher multipliers risk more per trade for the same edge.

ATR also reflects regime. In a calm grind-up, ATR compresses and the stop tightens, letting you size up for the same dollar risk. In a crash, ATR explodes and the stop widens, automatically reducing position size so a single trade never blows through the risk budget. This self-regulation is why ATR-based systems survive regime shocks that destroy fixed-stop systems.

Practical Application

Rule 1: Match ATR Timeframe to Trade Timeframe

Stop = entry ± (ATR × multiplier). Adjust the multiplier by timeframe:

Timeframe	Strategy	ATR multiplier
Weekly	Long-term	3.0-4.0× weekly ATR
Daily	Swing	2.5-3.0× daily ATR
5M / 15M	Intraday	1.5-2.0× timeframe ATR

A daily ATR stop on a 5M chart entry creates mismatched noise tolerance — the stop is far too wide for intraday noise. Match ATR timeframe to trade timeframe. As a rule of thumb, the stop distance should sit outside roughly 90% of the timeframe's normal candle range; if 5-6 of the last 10 candles would have reached the stop, the multiplier is too low.

Rule 2: Size by Risk, Not by Shares

Risk-based sizing with ATR keeps the dollar risk constant:

Position size = Risk ÷ (ATR × multiplier).

A $10,000 account risking 1% ($100) on a stock with daily ATR $2.00 and 2.5× stop: size = $100 ÷ ($2.00 × 2.5) = 20 shares. On a higher-volatility stock with ATR $5.00: size = $100 ÷ $12.50 = 8 shares. Same dollar risk, different share count — volatility-normalized.

Rule 3: Trail With the Chandelier Exit

The Chandelier Exit trails the stop at 3× ATR from the highest high since entry. As price rises, the stop rises; as ATR expands during volatility spikes, the stop widens to avoid noise stops. Chandelier exits outperform fixed-percentage trailing stops on tested trend-following systems by 15-25% in net profit, at the cost of giving back more at the top.

Worked Trade Example

Daily chart, stock breaks above resistance at $30.00. ATR(14) = $0.80. Account = $20,000, risk = 1% ($200), multiplier = 2.5×.

Stop distance = $0.80 × 2.5 = $2.00
Stop = $30.00 − $2.00 = $28.00
Position size = $200 ÷ $2.00 = 100 shares
Entry: $30.00 on the breakout close
After first target at $34.00, trail with Chandelier (3× ATR from highest high)
Filters passed: ATR timeframe matches trade, risk fixed at 1%, trailing logic defined

Checklist

ATR(14) on the trade timeframe (not a mismatched one)
Multiplier set to strategy (2.0-3.0 swing, 1.5-2.0 intraday)
Position size = risk ÷ (ATR × multiplier)
Chandelier trail active after first target
Risk per trade ≤ 1-2% of account

Common Mistakes

Using ATR(14) on illiquid stocks. True range spikes on wide prints distort the average. Fix: filter to liquid names, or use ATR(20) to smooth the distortion.
Mixing ATR timeframes. A daily ATR stop on a 5M chart entry creates mismatched noise tolerance. Fix: match ATR timeframe to the trade timeframe exactly.
Forgetting that ATR includes gaps. A stock that gaps 3× ATR overnight stops you out at the open regardless of intraday stop placement. Fix: account for gap risk in sizing, or avoid holding through known gap events.

Advanced Tips

ATR adapts automatically to volatility regimes. During the 2020 COVID crash, equity ATR tripled in two weeks — fixed stops got hit on day-one gaps; ATR stops widened and survived. During low-volatility grind-ups, ATR shrinks, stops tighten, and position size rises for the same risk. The system never needs manual adjustment. Combine ATR stops with Bollinger Band squeeze breakouts (the ATR stop sits naturally beyond the opposite band) — see Bollinger Bands Squeeze Breakout Strategy. For trend-filtered entries to pair with ATR stops, see RSI Advanced Usage and MACD Histogram and Multi-Timeframe Resonance. For Fibonacci entries where stops anchor to the next level, see Fibonacci Retracement Selection.

Summary

Constant risk across volatility regimes is the foundation of robust trend-following. Use ATR(14) on the trade timeframe, set the stop at 2.5× ATR (adjusted by strategy), size by risk ÷ (ATR × multiplier), and trail with the Chandelier Exit after the first target. The system adapts to every regime without manual adjustment — calm or volatile, the dollar risk stays constant.