Most technical tools try to answer one of two questions: where is price headed, or how forcefully is it moving. The Average Directional Index, developed by J. Welles Wilder in his 1978 book New Concepts in Technical Trading Systems, was built specifically to answer the second question while staying silent on the first. That design choice makes it one of the more misunderstood tools in common use, because traders often expect it to do a job it was never built for.
What ADX actually measures
ADX quantifies the strength of a trend, not its direction. A rising ADX line tells you that whatever movement is happening — up or down — is gathering momentum in a directional sense. A falling ADX line tells you the opposite: price may still be moving, but the movement is becoming less orderly, more like drift than trend. The index itself is unsigned. It cannot distinguish an uptrend from a downtrend, which is why Wilder paired it with two supporting lines, the Plus Directional Indicator and the Minus Directional Indicator, usually written as +DI and -DI.
How the calculation works
The construction starts with directional movement: the difference between today's high and yesterday's high (+DM), and the difference between yesterday's low and today's low (-DM). Whichever is larger and positive becomes the directional movement for that period; the other is set to zero. These values are smoothed, typically over 14 periods, and expressed relative to the Average True Range to produce +DI and -DI. ADX itself is derived from the smoothed difference between +DI and -DI, expressed as a percentage of their sum. The result is a single line, usually plotted from 0 to 100, that rises when the gap between +DI and -DI widens and falls when the two converge.
Reading the levels
Wilder's original guidance treated readings below 20 as indicating an absence of trend, readings between 20 and 40 as a developing trend, and readings above 40 as a strong trend. These thresholds have been used loosely across decades of practitioner writing, but they were never derived from a rigorous statistical study of any particular market — they were heuristics from Wilder's own commodity trading experience in the 1970s. Applied to equities, currencies, or crypto markets with different volatility regimes, the same numeric threshold can mean something quite different. A reading of 25 on a low-volatility utility index may reflect a meaningfully more persistent trend than the same reading on a speculative small-cap name that swings wildly on small volume.
What ADX does not tell you
The most common misreading of ADX is treating a rising line as a bullish sign. It is not. ADX can rise sharply during a strong decline just as it does during a strong advance — the 2008 and 2020 drawdowns both produced elevated ADX readings well above 40 even as major indices fell more than 30% in a matter of weeks. The line describes the conviction behind a move, not its favorability to a long position. Direction has to come from +DI and -DI, or from some other framework entirely. ADX also says nothing about the sustainability of a trend going forward; it is calculated from past price ranges, so by construction it lags the shift it is describing. A trend can be in its late stages while ADX is still climbing, and can be starting to weaken well before the line turns down.
Common misapplications
A frequent error is using ADX as a standalone filter for whether to act on other signals, assuming that any reading above 25 confirms a tradable trend. This ignores the smoothing period embedded in the calculation. A 14-period average of directional movement responds slowly to sudden shifts, meaning ADX can remain elevated for several sessions after a trend has already begun losing steam, or stay low through the early days of a genuine breakout. Another common mistake is comparing ADX levels across instruments with very different volatility characteristics, as if 30 means the same thing for a government bond ETF and a high-beta technology stock. Because the underlying inputs are scaled by each instrument's own true range, the numeric output is only meaningfully comparable within the same security over time, not across different ones.
Where it fits in a broader framework
ADX is most useful as a contextual filter rather than a standalone decision tool. Many trend-following approaches described in academic and practitioner literature — including work summarized in Fischer and Fischer's writing on directional systems — treat ADX as a way of distinguishing between environments where directional tools like moving average crossovers tend to perform differently, versus range-bound conditions where they tend to produce more false signals. Used this way, ADX doesn't generate a view on the market; it describes the texture of the environment in which other tools are being applied. That distinction matters because it shifts ADX from being a would-be predictor to being a description of present conditions, which is closer to what the calculation actually supports.
The rule to internalise
ADX answers "how strong is the current directional movement" and nothing more. It does not say which direction is favorable, does not forecast persistence, and does not translate cleanly across different instruments or volatility regimes. Treating it as a measure of conviction in the recent past, rather than a forecast of what comes next, keeps its use grounded in what the mathematics actually supports.
Educational content only. Not investment advice.