Two charts, same pair, same week. One of them ground sideways in a forty pip box and stopped out three breakout attempts. The other ran two hundred pips in a straight line. On both, the moving averages crossed, and on both the crossover looked identical at the moment you had to press the button. The difference was trend strength, and that is the only thing ADX measures.
J. Welles Wilder published it in 1978 alongside RSI and the parabolic stop. It has aged better than most indicators from that period, mainly because it does not pretend to be a signal.
How the number is built
Start with directional movement. For each bar, compare today's high to yesterday's high and today's low to yesterday's low. If the up move is larger than the down move and positive, that bar contributes to positive directional movement. If the down move is larger, it contributes to negative directional movement. Ties and inside bars contribute nothing.
Those raw values are smoothed over the lookback period, normally 14, and divided by the average true range to make them comparable across instruments. That gives you two lines, DI+ and DI-, each a percentage. DI+ measures how much of recent range came from upward extension, DI- how much came from downward extension.
ADX is then the smoothed absolute difference between DI+ and DI-, divided by their sum. Because it uses the absolute difference, direction is discarded. A market falling hard and a market rising hard produce the same high ADX reading. That is not a flaw, it is the design.
Reading the levels
The conventional bands are widely quoted and worth treating as rough zones rather than thresholds.
| ADX reading | Usual interpretation | What it argues for |
|---|---|---|
| Below 20 | No effective trend | Range tactics, fade the edges, expect breakout failure |
| 20 to 25 | Trend forming or fading | Wait, or size smaller until the direction resolves |
| 25 to 40 | Established trend | Trend continuation entries, trail rather than target |
| Above 40 | Strong, often extended | Late for entries, tighten stops, expect a pause |
The direction of ADX matters more than its level. A reading of 22 that has risen from 12 over six bars describes a market waking up. A reading of 34 that has fallen from 50 describes a trend running out of participants, even though 34 sits in the strong band. Rising ADX means the gap between DI+ and DI- is widening. Falling ADX means the two sides are converging, which happens both in a genuine reversal and in a plain consolidation.
ADX is direction-blind, so pair it
Because ADX cannot tell you which way, every practical use has a second component supplying direction. The common pairings are the DI lines themselves, where DI+ above DI- means the upside is dominant, or an external filter such as price relative to a longer moving average, or simply the higher timeframe structure.
The classic Wilder setup takes a DI crossover only when ADX is above 25 and rising, and uses the extreme of the crossover bar as the stop reference. It is a slow system that catches large moves and gives back a lot at the end of them, which is the profile of every trend following approach.
The use I prefer is negative. ADX below 20 is a reason not to take a breakout. If the market has not been extending in either direction, the odds of a clean continuation after a break of the box are worse, and range logic is the better fit. Using an indicator to veto trades rather than generate them removes a lot of arguing with yourself.
The lag, and what it costs you
ADX is a double-smoothed derivative of a smoothed value. With the default 14 period it takes roughly a full cycle of that length before a new trend registers above 25, which means the reading confirms after the first leg is already gone. That is the price of the filter, and shortening the period to chase it just gives you a noisier line that whipsaws through the threshold.
Two consequences. First, ADX is a poor entry trigger and a reasonable environment filter, so run it on a timeframe above the one you execute on. Second, a high ADX is evidence the move has been going on for a while, which cuts both ways: momentum persists more often than not, and the risk-to-reward on a fresh entry at ADX 45 is usually worse than it looks. That is a sizing question, and it belongs alongside your risk rules rather than in the indicator.
No indicator reading changes the arithmetic of a losing trade. ADX describes conditions, not outcomes, and leveraged trading carries a high risk of loss regardless of what any filter shows.
Settings and honest expectations
Fourteen is the default and there is no strong argument for anything else on daily and four hour charts. Traders on lower timeframes sometimes drop to ten to get a faster read, at the cost of more crossings through 20 and 25. If you change it, change it once and test it, do not tune it per instrument until it fits the last three months. That is overfitting and it will not survive contact with next quarter.
What ADX will not do: predict reversals, mark a top, or tell you a trend is safe. It compresses one dimension of price behaviour into a number, and used as a yes-or-no gate on strategy selection it earns its space on the chart. Used as a signal it will disappoint you slowly.
"I use ADX to stop myself trading breakouts in a dead market. That is most of the value, and it is worth more than another entry signal."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- ADX is built from the absolute difference between DI+ and DI-, so it measures strength and ignores direction entirely.
- The slope matters more than the level: 22 and rising is a better trend argument than 34 and falling.
- Below 20 works best as a veto on breakout trades rather than as a signal to fade anything.
- Double smoothing makes ADX lag by roughly its lookback period, so run it above your execution timeframe.
Frequently Asked Questions
What ADX value confirms a trend?
Readings above 25 are conventionally treated as an established trend and readings below 20 as no effective trend, with the band between them ambiguous. The direction of travel matters more than the exact number, since a rising line from a low base often describes a trend earlier than the level alone suggests.
Does ADX tell you whether the market is going up or down?
No. It is calculated from the absolute difference between the two directional indicators, so a strong downtrend and a strong uptrend produce the same reading. Direction has to come from DI+ versus DI-, from price against a moving average, or from higher timeframe structure.
What is the best ADX setting?
Fourteen periods is Wilder's original and remains the standard on daily and four hour charts. Shorter settings respond faster and cross the thresholds far more often, so any change should be tested once across a wide sample rather than tuned separately for each instrument.
About the Author
Alex Onta is an Executive Director at SINGUARD. He built eTrader, the terminal, the mobile apps, eTrader Broker, Copytrading, Business and Community, along with the worldwide clustered-server infrastructure it all runs on, with his brother Roman Onta helping on the design, and he leads that division today. Together with Roman he builds the Prop Firm CRM, the Broker CRM, Scalegram and CopySignals, and the two of them carry worldwide compliance, payment processing and international business structuring side by side. He lives and works in Dubai for most of the year. Meet the executive duo leading Singuard's five divisions.