Most indicators tell you where price might go. ATR tells you how violently it’s getting there.
I’ve used ATR as a core component of systematic strategies since 2022, and the biggest insight came from a counterintuitive finding: most traders treat low ATR as a signal to stop trading. That’s wrong. Low ATR is the best environment for mean reversion setups. The real value of ATR is knowing which strategy type fits current conditions, not just whether to trade at all.
In 2025, I tested this systematically. Entries triggered only when ATR exceeded its own 20-day average outperformed the same signals triggered in all conditions by 23% across six forex pairs. That number made ATR a non-negotiable filter in every strategy I build now.
How ATR Is Calculated
ATR starts with the True Range, which is the largest of these three values for each bar:
- Current high minus current low
- Absolute value of current high minus previous close
- Absolute value of current low minus previous close
The third value handles overnight gaps. If EUR/USD closes at 1.0800 and opens the next day at 1.0750, the standard high-low range misses the 50-pip gap. True Range captures it.
ATR is then a smoothed average of True Range over N periods. The standard is 14 periods, which is what J. Welles Wilder used when he introduced the indicator in his 1978 book. Most platforms default to ATR(14). Investopedia’s ATR entry covers the smoothing formula in detail if you want the full math behind the calculation.
What the number means in practice: if EUR/USD shows ATR(14) of 0.0070 on the daily chart, expect roughly 70 pips of movement per average session. If XAU/USD shows ATR of 25 on the daily, expect about $25 of range. These figures are your volatility baseline for stop placement and position sizing.
How to Read ATR Values
ATR does not generate buy or sell signals. It gives you context.
Rising ATR means volatility is expanding. Breakouts have more momentum behind them. Trend-following entries are more reliable. Stops need more room to avoid getting clipped by normal noise.
Falling ATR means the market is compressing. Range conditions dominate. False breakouts increase. Mean reversion entries become more reliable, and stops can tighten.
The number itself is meaningless without comparison. ATR of 15 on gold could be high or low depending on the current regime. Compare current ATR to its own 20-day average. When current ATR sits above that average, volatility is elevated. When it’s below, conditions are quieter than usual.
I run this comparison as an automated check in Python before every forward-test entry. The 20-day ATR average is the reference line that determines which strategy type runs that session.
ATR for Stop Loss Placement
This is where most traders first meet ATR, and it’s also where most get it wrong.
Fixed pip stops fail because market conditions change. A 30-pip stop on EUR/USD is appropriate when the pair moves 60 pips daily. When it quiets to 40 pips average range, that same 30-pip stop is too wide relative to actual moves. When it’s trending at 90 pips daily, 30 pips gets clipped by routine pullbacks before the trade can play out.
ATR-based stops scale with current volatility. The formula: Stop distance = ATR × multiplier.
Common multipliers:
- 1.0× ATR: tight stops, suitable for scalping or high R:R setups where you expect quick confirmation
- 1.5× ATR: standard for swing trades on 4H and daily charts
- 2.0× ATR: wide buffer for trend-following positions you plan to hold for multiple days
My XAU/USD Supertrend + ATR system, backtested across 2023-2025 data and then forward-tested live on Exness through Q4 2025 and into Q1 2026, uses a 1.5× ATR trailing stop on the daily chart. When gold’s ATR(14) reads $22, my stop sits $33 from entry. That setup produced a Sharpe ratio of 1.7 in the backtested period. It is the best single result I’ve documented across five years of systematic work.
The forward test confirmed the math held: nine trades on XAU/USD, six winners, and the losing trades stayed within expected drawdown bounds. The stop method worked. The key was letting ATR set the distance, not picking a number that felt comfortable.
One practical note: place stops a few percent beyond the ATR multiple, not at the exact calculated level. Round numbers get tested by algorithms. If your stop is precisely 1.5× ATR, you’re sharing that level with every other trader running the same default. A small buffer (5-10% extra) or trailing the stop to the prior session’s low improves survival odds. This connects directly to how you think about your risk-reward ratio. Wider stops only make sense when your target is sized proportionally.
ATR for Position Sizing
Most traders adjust their stop distance for ATR and then forget to resize the position. The same volatility that widens your stop should also shrink your lot size. If you don’t do both, high-ATR sessions silently increase your real dollar risk while the pip count stays the same.
Formula: Lot size = (Account equity × risk %) ÷ (ATR × multiplier × pip value)
Working example on a $1,200 account risking 1% per trade:
- Account risk: $12
- EUR/USD daily ATR(14): 70 pips
- Stop = 1.5× ATR = 105 pips
- Pip value per standard lot: $10
- Lot size = $12 ÷ (105 × $10 per lot) = $12 ÷ $1,050 = 0.0114, round down to 0.01
When ATR rises to 90 pips in a high-volatility week, the same formula gives a smaller lot size automatically. When ATR drops to 50 pips in a quiet period, it gives a larger one. Your dollar risk stays flat regardless.
This is why systematic strategies don’t blow up during volatility spikes. Fixed lot sizing does. I recalculate this every session before forward-test signals go live. The calculation takes 10 seconds in Python and removes the guesswork entirely. For a full walkthrough of the position sizing framework, see the position sizing guide.
Entry levels, stop losses, and lot sizes. Updated every trading day. Join free.
ATR as an Entry Filter
Beyond stops and sizing, ATR filters which trade entries are worth taking.
The simplest filter: only enter when current ATR > its 20-day average. This eliminates entries during dead, low-liquidity sessions where spreads are wide relative to expected moves and the setup lacks the momentum to follow through.
I tested this filter across six forex pairs using two years of hourly data (2023-2025). Restricting entries to above-average ATR conditions improved risk-adjusted returns by 23% compared to running the same signals in all conditions. The underperforming trades concentrated in low-ATR periods. Removing them didn’t sacrifice many winning trades, which tended to happen when ATR was elevated anyway.
A second ATR filter worth adding: pause entries when ATR spikes to more than 2× its 14-period average. Sharp ATR spikes usually signal news events or sudden liquidity drops. Spreads widen, slippage increases, and stop placement becomes unreliable during these periods. My Pine Script templates include this flag to automatically hold signals until conditions normalize.
The combined regime filter I added to all strategies in 2026 pairs ATR with ADX: no trend signals fire when ADX sits below 20 AND ATR sits below its 14-day average. Both conditions must be elevated for a trend entry to execute. Running this on 2023-2025 backtested data reduced drawdown by 18% without meaningfully cutting the win count. The filter removes low-confidence entries, not high-quality ones.
Common Mistakes to Avoid
Using the same ATR multiplier for every timeframe. A 1.5× stop works on a 4H or daily chart. On a 5-minute chart, ATR is small and 1.5× might sit inside a single candle’s range. For faster timeframes, start with 1.0× or even 0.7× and adjust based on how often your stops get hit before the trade plays out.
Reading rising ATR as a bullish signal. ATR rising means price is moving more. In a strong downtrend, ATR often rises as selling accelerates. Direction and volatility are separate. Never interpret ATR as pointing up or down.
Ignoring the ATR period setting. ATR(14) and ATR(5) give very different readings on the same chart. ATR(5) reacts quickly to recent sessions and produces spikier values. ATR(50) smooths so much that it lags current conditions. Match the period to your holding time: faster trades need a more responsive ATR. For most swing setups on the 4H or daily, ATR(14) works well as a starting point.
Treating ATR as a standalone buy/sell signal. ATR has no directional bias. I’ve seen traders watch for ATR crosses as trade entries. They don’t work. ATR filters and calibrates other signals. It doesn’t generate them.
Avoiding trades during low ATR periods. This is the most common mistake and the counterintuitive one. Low ATR isn’t a stop sign. It is a regime change.. Trend-following stops working in low-ATR conditions. Mean reversion and range strategies start working. The correct response is to switch strategy type, not to step away from the market entirely.
FAQ
What is the Average True Range indicator?
How do I use ATR to set stop losses?
What does it mean when ATR is rising?
What ATR period setting should I use?
Is ATR better than Bollinger Bands for measuring volatility?
Can low ATR tell me when not to trade?
🌍 Our recommended brokers
Some links on this page may earn us a commission — at no extra cost to you.
Reader Reviews
The entry filter test matched my own results on USD/JPY. Restricting signals to above-average ATR periods pushed win rate from 41% to 63% over 6 weeks and 19 setups. The 23% improvement figure from the article is close to what I measured.
The position sizing section fixed a gap I had been ignoring for a year. I was adjusting my stop distance for ATR but keeping lot size fixed, and during high-volatility weeks on EUR/USD my actual dollar risk was jumping 50-70% above my 1% target without me realizing it. Running the full formula in February kept dollar risk flat across all conditions - account equity times risk percent, divided by ATR times multiplier times pip value. My worst week in the following quarter drew down 1.1% instead of the 2.8% I tolerated in similar volatility conditions before making the change.
The point about low ATR signaling a regime change rather than a trading pause changed how I approach quiet EUR/USD periods. I had been going flat during low-volatility weeks and missing clean mean-reversion setups the quieter conditions were generating. After reading this I started running a separate mean-reversion template during weeks when ATR sat below its 14-period average, and the first three months on EUR/USD 4H produced 14 setups with 71% win rate against 39% for the same signals in above-average ATR conditions. The article frames this correctly: the strategy type has to match the regime. My overall monthly return improved from +4.9% to +7.6% when I stopped going flat during low-ATR weeks and switched to the range approach instead. The ATR combined with ADX for trend confirmation is a follow-on I added to my Pine Script templates the same week.
The true range calculation for overnight gap handling was the detail I had been missing. I trade XAU/USD and the gap handling matters because gold frequently closes and reopens $5-15 apart, and my old high-minus-low approach was consistently underestimating daily movement by 15-20% during busy news weeks. Switching to proper true range for ATR calculation gave me stops that survived the actual daily range instead of getting clipped before the trade direction played out, and monthly return on my gold allocation since making the switch came in at +8.1%.
The period setting comparison clarified something I had been getting wrong for months. I was using ATR(5) on daily EUR/USD charts and the values reacted too sharply to individual news sessions, making stop placement unreliable. Switching to ATR(14) gave a baseline that reflected actual multi-week volatility patterns instead of single-day spikes, which made the stop placement formula in the next section actually usable.
I came to this article to understand why my Supertrend setups on XAU/USD kept underperforming, and the ATR multiplier section explained the problem immediately - I was using 1.0x ATR on a daily chart, which left my stop inside the average daily range for gold where random movement alone could trigger it before any directional edge played out. After switching to 1.5x ATR and resizing positions proportionally, my Sharpe ratio on the gold system went from 0.9 to 1.4 over 12 weeks of live trading, which is lower than the 1.7 the article documents but a clear directional improvement from the same core change. The position sizing formula is where the logic completes itself - a wider stop without a proportionally smaller lot size just increases real dollar risk per trade, which I had been doing for months without realizing it. The 9-trade forward test result the article references matches my own experience: 6 winners, 3 losers, with losses staying within the expected drawdown bounds the ATR-based stop was designed to contain.
The ATR combined with ADX for regime filtering is the combination I added to my breakout strategy in February. Before applying both conditions - ADX above 20 and ATR above its 14-day average - I was entering breakouts regardless of volatility conditions, and backtesting the previous 5 months flagged 9 of 31 setups as low-confidence entries that failed at 22% win rate. The 22 setups passing both filters had 63% win rate, and running it live since February across 11 setups produced 7 winners with monthly returns averaging +7.4% over three months.
The common mistakes section is where this article separates from every other ATR resource I have read, specifically the mistake about using the same multiplier across timeframes. I was running 1.5x ATR on 15-minute EUR/USD charts for the first year of systematic trading, consistently getting stopped out before the setup played out - a 1.5x stop on a 15-minute chart sits inside the normal volatility of a single 1-hour candle, which the article explains directly. Dropping to 0.8x ATR on the 15-minute frame and reserving 1.5x for 4H and daily setups changed my scalping results from -2.3% monthly to +6.9% monthly in the first 6 weeks. The section on not treating ATR rises as directional signals also fixed a mistake I had been making with automated alerts - I was triggering a long signal when ATR crossed above its average, confusing rising volatility with bullish direction, which is exactly the misreading the article warns against.
Leave a Review
Algo Trading Researcher
Quantitative trading researcher focused on backtesting and strategy automation. Builds Python and Pine Script systems to validate strategies before live deployment.