Trend & directionAverage Directional Index · ADX
Wilder's 0–100 gauge of trend strength — how strong the move is, not which way it points.
Works best in trending marketsEngine-computed on a fixed sample series
What it is
The Average Directional Index, developed by J. Welles Wilder, is a 0-to-100 gauge of how strongly a market is trending, regardless of which direction it is heading. Its single most important feature is that it is direction-agnostic: a powerful uptrend and a powerful downtrend can both produce a high ADX, because the tool measures the strength of the move, not its sign. The question it answers is the one that decides which other tools to use — is there a real trend here worth following, or is the market ranging and likely to chop up trend-following signals? ADX is part of Wilder's Directional Movement system, which also produces the +DI and −DI lines that supply the direction ADX deliberately omits. It is used above all as a regime filter.
How it's calculated
The system starts by measuring directional movement each bar: the positive directional movement (+DM) is how much the current high exceeds the previous high, and the negative directional movement (−DM) is how much the previous low exceeds the current low, with only the larger of the two counting on any given bar. These are smoothed with Wilder's method over 14 periods and divided by the Average True Range to produce +DI and −DI, the positive and negative directional indicators. The Directional Index (DX) is then 100 times the absolute difference between +DI and −DI divided by their sum, which captures how one-sided the directional pressure is. Finally, ADX is Wilder's smoothed average of DX over 14 periods, which is why it is often described as doubly smoothed. That double smoothing makes ADX steady but lagging.
Reading it, step by step
The headline read is the level: above roughly 25 the market is considered to be trending and trend-following tools are in their element, while below 20 it is ranging and those tools will misfire. Just as important is the slope — a rising ADX means the trend is strengthening whether price is climbing or falling, and a falling ADX means the current trend is losing momentum even if price is still moving. Because ADX ignores direction, you must read it together with the +DI and −DI lines or with price to know which way to trade: high ADX with +DI above −DI is a strong uptrend, high ADX with −DI on top a strong downtrend. A very high ADX (above 40 or 50) that then turns down often flags an exhausting climax rather than continued strength. The tool is thus read as a two-part message: trend-or-range from the level, strengthening-or-weakening from the slope.
Best timeframes and settings
The standard is a 14-period ADX, which Wilder designed for daily charts and which serves swing trading well; it applies to any timeframe, with intraday traders using it to filter shorter trends and position traders on weekly charts. A shorter period makes ADX more responsive and quicker to cross the 25 threshold but noisier, while a longer period smooths it further and lags more. Because ADX is already doubly smoothed, it is inherently a lagging, deliberate indicator, so shortening the period only partly offsets that. The 25/20 thresholds are conventions, not laws, and some traders adjust them per instrument — a market that trends weakly may warrant a lower trending threshold. Most keep the 14-period default and change timeframe to suit the trade horizon.
When and where to use it
ADX is the archetypal regime filter, so its best use is deciding when to deploy trend-following tactics and when to switch to mean-reversion. In a trending market with ADX above 25, breakout and moving-average systems are favored; in a low-ADX range, oscillator fades are favored instead. It works on any liquid instrument and timeframe. Do not use ADX to pick direction — that is precisely what it refuses to tell you — and do not act on it alone; it is a companion to a directional method. Avoid reading a low ADX as bearish or a high ADX as bullish, a common beginner error, since both are agnostic to which way price is going.
Strategies that use it
The regime-filter strategy is the classic: only take trend-following entries — moving-average crossovers, breakouts — when ADX is above 25 and rising, and stand aside or fade when ADX is below 20. The DI-crossover-with-ADX strategy buys when +DI crosses above −DI while ADX is rising above 25 and sells on the mirror, using ADX to confirm the crossover occurs in a real trend rather than a chop. An exhaustion strategy watches for an extremely high ADX (above 40 to 50) that rolls over, tightening stops or taking profit on the trend it had been riding as the reading signals the move is spent. Across these, ADX never generates the entry direction by itself; it gates and confirms signals that come from price or the DI lines.
Combining it with other indicators
ADX is built to be combined with its own +DI and −DI lines, which supply the direction, and this Directional Movement package is the most natural pairing. Beyond that, it filters moving-average and breakout systems, permitting their signals only when a trend is present, and it complements Parabolic SAR, which provides trailing stops in the trends ADX confirms. Momentum oscillators like RSI pair with it in a division of labor: ADX decides whether to trust trend signals or oscillator fades, so a low ADX greenlights RSI mean-reversion while a high ADX greenlights trend trades. Aroon is a cousin that times the trend's start while ADX measures its strength, so the two together give onset and magnitude. Volume can corroborate that a rising ADX reflects genuine participation.
Where it fails
ADX lags because it is doubly smoothed, so it confirms a trend well after the trend has begun — by the time ADX clears 25, a good part of the move may be over. It can also still read high just as a trend is exhausting, since it reflects recent strength that may be about to evaporate, so a high ADX is not a promise of continuation. Its cardinal trap for beginners is expecting it to tell direction; on its own it never says whether to buy or sell, and treating a rising ADX as bullish is a real mistake. In very choppy markets ADX hugs the low teens and gives no actionable trend read, which is correct but frustrating. The defenses are to pair it always with a directional tool, to treat it as a filter rather than a trigger, and to remember that its level measures strength while its slope measures change, neither of which is a direction.
A worked example
Imagine a stock emerging from a quiet range where ADX has been sitting at 15, telling trend traders to wait. Price breaks out to the upside, and over the next several sessions ADX rises through 20 and then above 25 while +DI climbs above −DI, together signaling that a genuine uptrend with upward direction has taken hold. A trader using the regime filter now takes the moving-average crossover long that fired around the same time — a signal they would have ignored while ADX was low — entering near $60. ADX continues rising to 38 as the stock trends to $70, confirming strengthening momentum and keeping the trader in. When ADX finally peaks near 45 and rolls over while price stalls, the trader reads the exhaustion, tightens the stop, and exits near $69, having used ADX to confirm the trend's presence and strength without ever asking it for direction.