ADX Explained: How to Read Trend Strength in Futures
The ADX measures trend strength without direction. See how it is built from directional movement, how to read its levels, and how it differs from ATR and RSI.
By Imperial Analytics
The ADX is one of the few indicators built to answer a single narrow question: is the market trending, and how hard? It says nothing about which way. That restraint is the source of both its usefulness and the most common way traders misread it. This post defines what the ADX actually measures, walks through how it is built from directional movement, names the levels traders read from it, and marks the places where the reading quietly stops meaning what people think it means.
By Imperial Analytics
What the ADX actually measures
The ADX, short for average directional index, measures the strength of a trend on a scale from zero to one hundred, and nothing about its direction. A reading rises when price moves persistently one way and falls when price chops sideways. It was introduced by J. Welles Wilder in 1978 as the summary line of his Directional Movement System.1
Hold one idea from the start: the ADX is direction-blind by design. A market falling hard and a market rising hard can print the same ADX value, because the indicator rates only how orderly and one-sided the move is, not which side is winning. Direction is reported by two companion lines, the plus and minus directional indicators, and separating those two jobs is most of what it takes to read the tool correctly.
That separation is deliberate. Wilder built the ADX as the top layer of a system whose lower layers already carry the direction. The directional movement of each bar is split into an up-portion and a down-portion, those feed two directional lines, and the ADX is distilled from the gap between them. What survives at the top is a single question with a single answer: strong trend, or no trend.
How the ADX is built
The ADX is built in three stages. Directional movement splits each bar's range into an up-part and a down-part. Those are smoothed over fourteen periods and divided by true range to make the plus and minus directional indicators. The gap between those two lines becomes an index that is itself smoothed into the ADX. Fourteen periods is Wilder's default.1
The construction runs bottom to top. First, measure directional movement bar by bar. The up-move is the current high minus the prior high; the down-move is the prior low minus the current low. Only the larger of the two counts for a given bar, and only if it is positive, so an inside bar that stays within the prior bar's range contributes no directional movement at all. This is the raw signal that price is extending in one direction.
Second, smooth the up-moves and the down-moves over fourteen periods using Wilder's smoothing, the same recursive averaging used in ATR and RSI, then divide each by the smoothed true range over the same window. The results are the plus directional indicator and the minus directional indicator, each a percentage. The plus line answers how much of recent range was upward; the minus line answers how much was downward.
Third, take the absolute gap between those two lines, divide it by their sum, and scale to one hundred. That produces a directional index for each bar, and the ADX is a fourteen-period Wilder-smoothed average of that index. The two layers of smoothing are why the ADX is slow: it is an average of an average.
Data note
The arithmetic here is the methodology. Any numerical example in this post is illustrative. Imperial Analytics only surfaces pattern claims on a trader's own data when the sample meets the minimums defined in the AI Operating Charter: twenty trades in the matching condition for behavioral patterns, fifteen for time-of-day claims, and ten for day-of-week claims.
A concrete check makes the geometry clear. Suppose on a five-minute ES chart the smoothed plus directional indicator reads 28 and the minus reads 12. The current directional index is 100 × |28 − 12| ÷ (28 + 12) = 100 × 16 ÷ 40 = 40. If the fourteen-period average of that index currently sits at 32, the ADX reads 32. The plus line sitting above the minus line says buyers own the direction; the ADX at 32 says the move has real strength behind it. Neither number is an entry on its own.
How to read the ADX
Traders read the ADX two ways. The level rates trend strength: readings below twenty to twenty-five suggest no usable trend, while readings above twenty-five suggest a real one. The slope matters more than the level: a rising ADX says the current trend is gaining strength, a falling ADX says it is fading, regardless of how high the number is.
The level is the first read. Wilder treated a reading below about twenty as a market without a tradable trend, the regime where trend-following entries tend to get chopped up, and a reading climbing through twenty-five as confirmation that a directional move has organized itself. Those thresholds are conventions, not laws of the market, and they drift by instrument and timeframe, which is why they are a starting point rather than a trigger.
The slope is the read most beginners skip. A high ADX that has started to fall is describing a trend that is still strong in absolute terms but losing its grip, which is a very different situation from a low ADX turning up, where a new trend may be organizing. Because the ADX is twice-smoothed, its turns lag the price that caused them, so the slope is a lagging confirmation of a change already underway, not a forecast of one.
Direction still has to come from somewhere, and on this tool it comes from the two directional lines underneath. When the plus line is above the minus line, the strength the ADX reports is upside strength; when the minus line is on top, it is downside strength. The ADX itself never answers that question, which is exactly why traders pair it with the lines it was distilled from, or with the market structure they are already reading off the chart.
↳ Note
The ADX tells you the sea is rough. It will not tell you which way the current runs. Read the strength off the ADX and the direction off something else, or you will trade a strong number in the wrong direction.
What the ADX does not tell you
The ADX does not give direction, call tops or bottoms, or turn its threshold into a signal. A reading above twenty-five confirms a trend exists; it does not say whether that trend is young, old, or about to reverse. A high but falling ADX often marks the strongest part of a move, not its end.
The first limit is the one the whole post keeps returning to: no direction. A trader who buys because "the ADX is high" without checking which directional line is on top is acting on half the information, and half of those trades are into a strong downtrend. The number is agnostic, and treating it as bullish because it is large is the single most common ADX error.
The second limit is timing. Because the ADX is an average of an average, it confirms trends well after they begin and signals their weakening well after momentum has already turned. It is a state descriptor, not a leading edge. Using it to time entries bar by bar asks it to do a job its construction rules out, and the lag shows up as entries taken after the easy part of the move is gone.
The third limit is the threshold trap. The twenty-five level is a convention, and a market can trend cleanly at an ADX of twenty-two or churn at an ADX of thirty on a different instrument. Hard-coding a single cutoff across every contract and timeframe imports a false precision. The disciplined use is to treat the level as a rough filter and let the trader's own results, not the round number, decide where the useful cutoff sits. The same caution against leaning on any one reading is the point of the primer on why stacking more indicators does not strengthen an edge.
How the ADX differs from its near neighbors
The ADX is often confused with three indicators it sits beside. ATR measures volatility in price points, not trend strength. RSI measures momentum, the speed of price change, not whether price trends. MACD measures trend momentum but carries direction in its sign. The ADX alone rates trend strength on a fixed scale while deliberately discarding direction.
The ATR comparison is the closest, because the two share Wilder's authorship, the same fourteen-period default, and the same true-range input. But ATR reports the size of a bar's range in price points, a raw volatility figure that says nothing about whether the market is going anywhere. A quiet, orderly trend can have a low ATR and a high ADX at the same time. ATR sizes the stop; the ADX rates the trend. They answer different questions from overlapping inputs.
The RSI comparison trips people because both are bounded from zero to one hundred. RSI measures momentum, the recent balance of up-closes against down-closes, and a high RSI means price has risen quickly, not that a durable trend exists. A market can print an overbought RSI inside a choppy range where the ADX is flat and low. The scales look alike; what they count does not.
The MACD comparison is the subtle one. MACD also speaks to trend strength through the gap between two moving averages, but its value is signed and unbounded: a positive MACD means an uptrend, a negative one a downtrend, and the magnitude is a raw price difference with no fixed ceiling. The ADX strips the sign out on purpose and normalizes the magnitude to a fixed zero-to-one-hundred scale, which is what lets a reading of thirty mean roughly the same thing across instruments where a MACD of thirty would not.
How to read and log the ADX without leaning on one number
The ADX earns its place as a filter, not a trigger, with its useful cutoff found from your own results rather than a round number. Tag every trade with the ADX level, slope, and which directional line led, then check whether strong-trend states carry a different expectancy than flat ones. That turns it into a measurement.
The principle is the same one that governs every indicator. The indicator is not the edge. The edge is whatever pattern the trader's own results reveal when the indicator is used inside a defined process. If entries taken with a rising ADX above the trader's chosen cutoff and the correct directional line on top produce a positive expectancy on a given setup, and entries taken in a flat, low-ADX regime do not, that difference is worth keeping. If the two states produce roughly the same result after enough trades, that too is information, and the honest response is to stop treating the ADX as a filter on that setup.
The minimum journal fields to enable this check are the timestamp and instrument of the trade, the planned setup, the ADX period in use, the ADX level and slope at entry, which directional line was on top, and the realized P&L. Without those fields the question cannot be answered from the data. With them, it can be answered honestly the first time the sample reaches the threshold for a behavioral pattern claim, which is twenty trades in the matching condition. The reasoning behind that threshold is set out in the primer on what makes a behavioral pattern claim trustworthy, and the broader question of how many trades a signal needs before its win rate means anything is covered in the sample-size primer.
Frequently asked questions
Frequently asked questions
- q: Does the ADX tell you whether to buy or sell? a: No. The ADX rates trend strength on a zero-to-one-hundred scale and says nothing about direction. Direction comes from the two companion lines, the plus and minus directional indicators. When the plus line is above the minus line the strength is upside strength, and when the minus line is on top it is downside strength. Reading a high ADX as bullish is the most common error with the tool.
- q: What ADX level counts as a strong trend? a: By Wilder's convention, readings below about twenty suggest no tradable trend and readings above twenty-five suggest a real one. Those numbers are starting points, not laws. The useful cutoff drifts by instrument and timeframe, so the disciplined approach is to treat the level as a rough filter and let your own logged results, rather than the round number, decide where the cutoff belongs.
- q: What is the difference between the ADX and the plus and minus directional indicators? a: The plus and minus directional indicators measure how much of recent range was upward and how much was downward, so together they carry direction. The ADX is distilled from the gap between them and keeps only the magnitude, discarding the sign. The lines tell you which way; the ADX tells you how hard. They come from the same directional-movement math but answer different questions.
- q: Why does the ADX lag price so much? a: Because it is an average of an average. Directional movement is smoothed over fourteen periods to make the directional indicators, and the index built from those is smoothed again over fourteen periods to make the ADX. Each layer of averaging adds delay, so the ADX confirms a trend well after it starts and reports its weakening well after momentum has turned. It is a state descriptor, not a leading signal.
- q: Can the ADX be used on intraday futures charts? a: The ADX is defined on any series of highs, lows, and closes, so it can be computed on a one-minute, five-minute, or hourly futures chart. The construction does not change with the timeframe, but both the noise level and the lag scale with it, which is why intraday users test their own cutoff and period against logged results rather than assuming the fourteen-period default and the twenty-five level transfer unchanged.