ATR (Average True Range): stops, filters and strategy ideas
ATR tells you how far a typical bar is travelling right now, and nothing about which way it is heading. Because it deals only in distance, futures traders lean on it for stop placement, exits and volatility filters rather than for entry signals. This page covers the common uses and how a strategy generator can compare different range-based rules on your own market.
What it measures
Average True Range looks at how much ground each bar actually covered and then averages that over a recent stretch. The word true matters: the range of a bar is not only its high minus its low, but also includes any gap from the previous close. Overnight gaps and opening jumps therefore show up in the reading instead of being quietly ignored. For futures contracts that trade separate day and night sessions, or that gap on news, this is a large part of why ATR is trusted.
ATR has size but no direction. A sharp rally and a sharp sell-off both push it higher. It is expressed in the same units as price, so on an equity index future it reads in index points and on crude oil it reads in price increments. That makes raw comparisons between different contracts meaningless, but comparisons over time on the same contract are immediately useful: you can see at once whether a typical bar today is larger or smaller than it was last month.
Because it is an average, one extreme bar is diluted, and the line changes at a measured, fairly smooth pace. It works more like a thermometer for the market than a signal generator. It describes whether the environment is hot or cold, not whether to buy or sell. That is why ATR is rarely a strategy on its own; it is usually the ruler that lets other rules size things correctly.
How it behaves in real markets
Volatility clusters. Wild sessions tend to be followed by more wild sessions, and calm stretches tend to stay calm. On a chart this shows up as plateaus and valleys rather than random noise, which is what makes ATR useful for telling one market environment apart from another.
In sharp declines and panic selling, ATR usually climbs faster than it does during a steady advance. That fits the way futures markets behave when bids disappear and spreads widen on the way down. Many traders also notice that ATR tends to hit a recent high during the most violent part of a move, not at the point where the move began.
Contract rolls, major data releases and session changes can make ATR jump in steps. Mixing day and night sessions in one calculation, or stitching contract months together without adjusting for the price difference between them, distorts the reading. How the data is prepared therefore has a direct effect on whether ATR can be relied on.
Common ways to use it (concepts)
Stops sized by volatility
The stop distance is expressed in units of the current average range instead of a fixed number of points. It widens when the market is volatile and tightens when it is calm, so the stop sits outside normal noise. Often paired with a trend entry based on a moving average.
Standing aside in quiet markets
When the average range falls below its recent norm, breakout rules are switched off so that entries are not taken in a market with no room to move. The reverse also works: allow a Bollinger Bands breakout only while the range is expanding.
An exit line that follows the range
The exit line trails the highest point reached during the trade by a distance measured in average ranges. It moves up as the trend extends, and the position is closed only when a pullback exceeds what normal fluctuation would explain.
Pairing it with other indicator families
ATR has no opinion on direction, so it almost always appears next to something that does. ADX describes how strong the trend is while ATR describes how far each bar travels. When both rise together, the market is usually in a large directional move. When ATR is high but ADX stays low, the market is often whipping back and forth violently, which is exactly where many trend rules struggle.
Channel indicators often borrow the ATR idea directly. The width of the Keltner Channel is built from the average true range, so running the two side by side duplicates a great deal of information. A more useful split is to let Keltner handle entry location while ATR handles stop distance and position size, so entries and exits are measured with the same volatility ruler.
Paired with VWAP, ATR converts the gap between price and the session's volume-weighted average into a distance measured in ranges. That makes deviations comparable across different days, so you can judge whether today's stretch away from the average is genuinely rare or simply looks large in points.
In AlphaPrime
In AlphaPrime, ATR seldom decides an entry by itself. It usually sits in the second half of a strategy: how far away the stop goes, how a trailing exit follows the trade, and which sessions are too quiet to bother with. The strategy generator takes these range-based exit and filter rules and pairs them with entry conditions from other indicator families.
The same entry idea can produce very different results depending on how the volatility stop is sized. The generator has the GPU backtest those variations one by one, then uses out-of-sample data to check which stop distances still make sense in conditions the strategy has never seen, rather than looking good only during the period used to pick them.
Combinations that pass can be exported as strategy code for your trading platform. The volatility stop and the quiet-market filter travel with the strategy, so on your platform they keep adjusting to the live range instead of turning into hard-coded point values.
FAQ
Does a rising ATR mean the market is going up?
No. ATR measures only the size of movement, and both rallies and sell-offs push it higher. For direction, pair it with the slope of a moving average or a trend tool such as ADX, and let ATR focus on distance.
How is ATR different from Bollinger Band width?
Both measure volatility. ATR looks at the actual high-low range of each bar and includes gaps, while Bollinger Bands take their width from how widely closing prices are scattered around their average. ATR moves more steadily; Bollinger width reacts more sharply to a run of one-directional closes.
If I export a strategy with ATR stops, do the stops still adapt?
Yes. The exported strategy code keeps the stop and exit logic expressed in average ranges, so when it runs on your trading platform the distances are recalculated from the current range rather than frozen at a fixed number of points.