ADX Indicator: How to Measure Trend Strength (Not Direction)
A full walkthrough of the ADX indicator: what it measures and what it structurally cannot, the DX formula behind the line, the 25 threshold and the sources that disagree with it, how +DI and -DI supply the direction ADX leaves out, and concrete period settings by chart timeframe.
Almost every indicator on a chart has an opinion about direction. ADX does not, and that is the entire point of it. Of the thirteen reference guides I went through to write this, all thirteen state some version of the same sentence: ADX measures the strength of a trend and says nothing about which way that trend is running. It is also the claim that gets ignored most often in practice, because a rising line on a chart looks bullish and the human eye does not care what the axis means. A stock falling 9% on the day with ADX climbing from 18 to 41 is not a stock turning around. It is a stock whose downtrend is picking up force. Here is the formula behind the line, the threshold everyone quotes and the sources that quietly disagree with it, and the settings that actually change what you see intraday.
Quick Answer
The ADX indicator measures how strong a trend is on a 0 to 100 scale, and says nothing about which way it runs. Above 25 is the conventional line for a trend worth trading. Below 20 reads as a range. A rising ADX while price falls means the downtrend is getting stronger, not that price is about to turn up. Direction comes from the +DI and -DI lines plotted alongside it.
What the ADX Indicator Actually Measures
ADX stands for Average Directional Index, and it came out of J. Welles Wilder Jr.'s New Concepts in Technical Trading Systems in June 1978. Wilder had an unusually productive year. The same book gave us RSI, ATR and Parabolic SAR, which means a large share of any modern chart traces back to one volume written for hand-calculated daily futures data. If you have read the walkthrough of trading RSI, you have already met the other half of the same trader's toolkit, and the two were designed to answer completely different questions.
The output is a single line bounded between 0 and 100. Every source agrees on the range and every source agrees on what the line means: low values mean there is no trend worth the name, high values mean there is one. What it does not encode, at all, is direction. Trading Setups Review has the cleanest framing for the split anyone has published. ADX is a speedometer, not a compass. It will tell you that the car is doing 70. It has no idea whether you are driving toward the office or away from it.
This is where most of the misreading happens, and five of the thirteen guides call it out explicitly as the single most common error attached to the indicator. Somebody sees ADX rising and reads it as bullish. CMC Markets states the correction flatly: a high ADX reading does not mean prices are rising. The Forex Geek and Trading Strategy Guides both give the same concrete illustration, which is worth internalising because it is the whole lesson in one sentence. Price falling with ADX rising means the sellers are winning harder. It is a bearish confirmation, not a bullish one.
ADX rises in an uptrend and in a downtrend. It only falls when the trend does.
Why the line cannot carry a sign, by construction
Most guides state the strength-not-direction rule and move on. The reason is more useful than the rule, and it sits in the arithmetic. ADX is built from the absolute difference between the two directional lines. Whether +DI is 40 and -DI is 10, or +DI is 10 and -DI is 40, the absolute gap is 30 either way, and the calculation produces the identical value. The sign is discarded before the ADX line is ever computed. There is no scenario, no setting, and no timeframe in which ADX becomes directional, because the information required to make it directional has already been thrown away. That is not a limitation somebody could patch. It is the design.
How Is the ADX Indicator Calculated?
The headline formula is one line, and it is the last step of a longer chain.
DX= ( | +DI − −DI | / ( +DI + −DI ) ) × 100
ADX= a 14-period smoothed moving average of DX. Some implementations use an exponential average instead of Wilder's original smoothing, which produces slightly different values from the same data.
Reading it from the inside out: the numerator is how far apart the two directional lines are, stripped of sign. The denominator is how much directional movement there was in total. So DX is the share of all movement that went one way rather than both ways, expressed as a percentage. A market pushing hard in one direction has a big gap and a big DX. A market chopping back and forth has +DI and -DI sitting almost on top of each other, a tiny gap, and a DX near zero. That single ratio is the whole idea, and everything before and after it is smoothing.
Getting to +DI and -DI takes a few steps of its own. Each bar, the calculation measures how much the high extended above the previous high and how much the low extended below the previous low. The larger of the two becomes that bar's directional movement, up or down, and the smaller is set to zero, so a single bar only ever contributes in one direction. Those values are smoothed over 14 periods, divided by a smoothed true range, and turned into the +DI and -DI percentages you actually see. Then DX, then another 14-period smoothing to get ADX. StockCharts' ChartSchool entry on the Average Directional Index works through all six steps with the arithmetic if you want to rebuild it in a spreadsheet.
Two consequences fall directly out of that chain. The first is that ADX is a lagging indicator, and unusually so. There are two separate rounds of averaging between raw price and the final line, which is why it confirms trends rather than predicting them. Nobody should be surprised by this and plenty of people still are. The second is a detail only StockCharts bothers to mention: because the smoothing is recursive and compounds, ADX values do not fully stabilise until roughly 150 periods of price history have gone into them. Load a fresh chart with a short history and the early readings are arithmetic artefacts. It converges, and it converges slowly.
How Do You Read the ADX Indicator on a Chart?
Two things to read, and they are independent of each other. The level tells you whether a trend exists. The slope tells you what is happening to it right now.
The slope half is simple and it is the half people underuse. Rising ADX means the current trend is strengthening. Falling ADX means it is weakening. Neither statement says anything about direction, which is the trap all over again. A falling ADX in a downtrend does not mean price is going up. It means the selling is losing conviction, which might resolve into a bounce, a range, or a pause before the next leg down. Ten of the thirteen sources state the rising-and-falling rule in those neutral terms, and it is worth keeping them neutral in your own head.
The level half is where the published guidance genuinely disagrees, and flattening that into one confident number would be dishonest.
| ADX reading | Standard interpretation | The caveat worth knowing |
|---|---|---|
| 0 to 20 | No trend, or one too weak to be worth trading. Range conditions | Agreed across almost the whole corpus. Wikipedia uses 20 as its weak cutoff as well |
| 20 to 25 | Grey zone. A trend may be forming but has not proved anything yet | Some traders pull the line down to 20 and trade this band. StockCharts notes 20 is increasingly common |
| 25 to 40 | The conventional strong-trend band. Most sources call this tradeable | Where the disagreement lives. LiteFinance reads this as a trend beginning or confirming, not a strong one |
| 40 to 50 | Strong by every convention in the set, including the strict ones | This is where LiteFinance and Wikipedia start using the word strong |
| 50 to 75 | Very strong. A trend this established is usually well advanced | Late-stage territory. Strength and remaining upside are not the same thing |
| 75 to 100 | Extreme, and rare enough that you should check your settings first | Fidelity notes readings above 60 do not occur frequently |
Eleven of thirteen sources put the strong-trend line at 25 and the dead zone below 20, so that is the convention to work from and the one this site uses elsewhere. The breakdown of Parabolic SAR leans on the same range when it argues that a trailing-stop system is only reliable once ADX is rising through roughly 20 to 25, and the two posts should agree, because it is the same indicator doing the same job. What you should also know is that LiteFinance treats 25 to 40 as a trend that is merely beginning or confirming and reserves strong for 40 and up, and Wikipedia's entry on the average directional movement index uses 40 and 20 as its two cutoffs rather than 25. The honest summary is that everything between 20 and 40 is read very differently depending on who is doing the reading. Pick a line, write it down, and stop moving it after losing trades.
A strong trend and a good entry are two different questions, and ADX only answers one.
Upload the chart screenshot and SnapPChart reads that one image against a fixed rubric, then returns a setup grade, an entry, a stop with the reasoning behind the level, targets, and the reward-to-risk those levels imply. The grade is the part worth arguing with before you put money on it.
Grade this chart+DI, -DI, and Why ADX Is Not the Same Thing as DMI
The direction ADX leaves out comes from the two lines it was built from. +DI above -DI is a bullish bias. -DI above +DI is a bearish one. Twelve of the thirteen sources state it in roughly those words, and the pairing is what makes the whole system usable: ADX for whether, the DI lines for which way.
The naming trips people up on a live platform. DMI, the Directional Movement Index, is Wilder's complete system including the ADX line. ADX on its own is one component of it. TradingView's help page for the Average Directional Index ships them as two separate indicators, where ADX plots the single line alone and DMI plots all three together. Add ADX expecting the coloured lines and you will not get them. That is a naming convention, not a broken indicator.
The crossover signal, and the counter-argument to it
Seven sources describe the standard trigger. Buy when +DI crosses above -DI, sell when -DI crosses above +DI, and filter both with an ADX threshold so you only take crossovers that happen while a trend actually exists. The filter is the important half. An unfiltered DI crossover in a flat market is exactly the kind of signal that shows up constantly and means nothing, for the same structural reason a MACD crossover fires repeatedly in a range: two lines derived from the same price series will cross whenever that price series wiggles.
There is a real counter-current here worth knowing about. Corporate Finance Institute notes that many analysts now watch the ADX line by itself and skip the DI crossovers entirely, on the grounds that those crossovers frequently generate false signals. That is one source against seven, so treat it as a minority position rather than settled practice, but it is a coherent one. If you are using ADX as a regime filter on top of a setup you found some other way, the DI lines may be adding screen clutter and nothing else. The same source flags a second use that gets very little coverage anywhere: price and ADX pulling apart. Price grinding to new highs while ADX turns down says the move is running on less force than it was, which is an early warning rather than a reversal signal, and it is the kind of thing worth noticing before you add to a winner.
What Are the Best ADX Settings for Day Trading?
The default is 14 and it is not a magic number. It is the number Wilder chose in 1978 for daily futures charts, and ten of the thirteen sources list it as the standard because every platform ships it. Seven of them also make the point that the period is a genuine tuning parameter: shorter reacts faster and produces more false readings, longer smooths the noise and lags further. That much is uncontroversial and also vague, so here are actual numbers.
The 1-minute and 5-minute rows below come from two independent sources that arrived at the same pair, The Forex Geek and Trading Strategy Guides, which is about as close to corroboration as indicator-settings advice gets.
| Chart / use | Period | What it does | What it costs you |
|---|---|---|---|
| 1-minute chart, intraday | 3 | Registers a push within a few bars, so an opening-range burst shows up while it is still running | Fires on noise. Three strong bars inside a range will read as a trend |
| 5-minute chart, intraday | 7 | Fast enough for a session-length move, meaningfully smoother than the 1-minute setting | Still noticeably noisier than the default. Expect false strong readings in the first 30 minutes |
| 15-minute and hourly | 14 | Wilder's default and the number every platform ships with | On intraday charts it often confirms a move that is already half spent |
| Daily and swing | 14 to 20 | The standard read, or a slightly slower one for multi-day holds | Several bars late to every turn, which on a daily chart is several days |
| Regime filter only | 20 to 30 | Very smooth. Only a sustained trend pushes it above 25 | Late to everything. Usable as a background filter, not as a trigger |
| The contrarian intraday setting | 2 | Trading Setups Review proposes a 2-period ADX specifically to defeat the standard lag | At this length the output is nearly as jumpy as the price it is measuring |
One thing that table does not show, and that matters more than any row in it: shortening the period changes what a given reading means. A 3-period ADX reaches 30 on moves that would leave a 14-period ADX sitting at 19. If you run a fast setting and keep the 25 threshold you inherited from the default, you have quietly loosened your filter without deciding to. Either raise the line to match, or watch where ADX typically tops out on your instrument at your setting and calibrate from that.
There is also a timeframe argument that is not resolved. LiteFinance advises against M5 and M15 charts for this indicator and recommends H1 and above, on the grounds that the lower timeframes are too noisy for the smoothing to mean anything. That is one source's house view rather than a corpus-wide rule, and the same source takes the same position on other indicators, so weight it accordingly. It is also not wrong about the mechanism. Less data per bar means a lower signal-to-noise ratio, and no amount of parameter tuning creates information that is not in the bars. If you are trading a 1-minute chart, you are accepting that trade, and choosing a 3-period ADX is accepting it twice.
What Should You Pair ADX With?
Eight of the thirteen sources say do not trade it standalone. The reason is the same one that runs through this whole post. ADX answers one question, answers it well, and cannot be made to answer any of the others. It has no direction, no entry price, and no invalidation level. Anything you build on it has to supply those three things from somewhere else.
The most natural partner is price structure, because levels are exactly what ADX lacks. Knowing the trend is strong tells you nothing about where to get in, and the method for marking levels that actually hold covers the half of the problem ADX cannot touch. Wilder's own Parabolic SAR is the other classic pairing, and it works because the two indicators are strictly complementary: SAR has a direction and an exit level but no view on strength, ADX has strength and no view on anything else. Where each fits in a full intraday workflow is laid out in the ranked list of day trading indicators, and the broader question of how indicators sit on top of structure rather than replacing it is the argument of the technical analysis overview.
Where the tool I build fits here is narrower than I would like it to be, and it is worth saying plainly. SnapPChart does not calculate ADX, does not compute directional movement, and has no ADX field anywhere in its analysis output. What it does is read a chart screenshot you upload, which means if you plot ADX or the full DMI on your own chart before taking that screenshot, the visible lines are part of what the analysis sees, in the same way a visible moving average or MACD panel is. That is a real but much weaker claim than the tool deriving an ADX value on its own, and the two should not be confused. A neutral description of what a single chart read covers is on the AI chart analysis page. The stronger connection to this post is simpler: knowing what trend strength looks like is how you sanity-check any written read of a setup, mine included. If an analysis describes a strong trend and you cannot see one on the chart, that disagreement is information worth having before you size the trade.
ADX is a filter, not a signal. Use the level to decide whether to be trading trend setups at all right now, use the slope to tell whether the one you are in is gaining or losing force, and get direction from the DI lines or from price. Default period 14, strong-trend line at 25, and if you change either one, change them together and write down what you changed.
Frequently Asked Questions
What is considered a strong ADX reading?
The majority convention is 25. Eleven of the thirteen reference guides I worked through to write this treat a reading above 25 as a trend strong enough to trade and a reading below 20 as a range. A minority hold out for more: LiteFinance reads 25 to 40 as a trend that is only beginning or confirming and reserves the word strong for 40 and above, and Wikipedia uses 40 and 20 as its two cutoffs rather than 25. So 25 is the standard answer and it is not a settled one. The part almost nobody mentions is that the number is not comparable across settings. A 3-period ADX crossing 25 on a 1-minute chart is a far lower bar than a 14-period ADX crossing 25 on a daily, because the shorter setting reaches higher values off much smaller moves. The more useful habit is to look at where ADX sits relative to its own range on that specific symbol over the last few weeks, and to treat the fixed number as a rough sanity check rather than a trigger. Fidelity adds one useful bound at the top: readings above 60 do not occur often, so if you are waiting for one you will mostly be waiting.
Can you trade off the ADX indicator alone?
No, and it is worth being precise about why rather than repeating the rule. A trading decision needs three things: a direction, a price to act at, and a price where you are wrong. ADX supplies none of them. It supplies one number describing how much force is behind whatever is already happening. Add the +DI and -DI lines and you get a direction, which is two out of three, and you still have no level. Every ADX system you will find published is really a system built on something else with ADX bolted on as a filter, which is the honest description of what the indicator is for. Used that way it is genuinely good, because the question it answers is one that price alone answers slowly and one that most indicators cannot answer at all.
What is the difference between ADX and DMI?
DMI, the Directional Movement Index, is the whole system Wilder published: +DI, -DI, and the ADX line derived from the gap between them. ADX is one component of that system, the smoothed average of the directional movement index value. In practice the distinction shows up as a platform quirk. TradingView ships them as two separate indicators, where ADX plots the single line on its own and DMI plots all three together, so a trader who adds ADX and then wonders where the coloured lines went has hit a naming convention rather than a bug. You will also see the full name Average Directional Movement Index used interchangeably with Average Directional Index, which is the same thing again.
Is ADX a leading or a lagging indicator?
Lagging, and more so than most. Price feeds into directional movement, directional movement gets smoothed, the smoothed values produce DX, and DX gets smoothed again to produce ADX. Every one of those stages is an average of the past, so by the time ADX has climbed through 25 the trend it is describing has usually been running for a while. That is a real cost and it is also the job. ADX exists to confirm that a move has force behind it, not to call the move before it happens, and a tool that confirmed instantly would be measuring noise. What the lag means practically is that ADX is close to useless for timing an entry and quite good at telling you whether to keep taking entries at all this session. Read it as a regime switch you check, rather than a signal you act on.
Who invented the ADX indicator and when?
J. Welles Wilder Jr., published in June 1978 in New Concepts in Technical Trading Systems. That single book is responsible for an unusual share of what sits on a modern chart: ADX and the wider directional movement system, RSI, ATR, and Parabolic SAR all came out of it. Wilder was a mechanical engineer who moved into commodities, and he was designing for hand calculation on daily futures data, which shows in the structure of the indicator. The smoothing is recursive so each new bar only needs the previous value and the current one, which is exactly what you want when you are updating figures with a pencil at the end of the session and exactly why the early values on a freshly loaded chart are not yet reliable.
This article is for educational and informational purposes only and is not investment, financial or trading advice. The ADX and directional movement formulas, the 14-period default, and the threshold bands described here are the conventional formulations published by J. Welles Wilder Jr. and reproduced by charting platforms and reference sources; the period settings quoted by timeframe are ranges commonly cited in that literature rather than tested or recommended parameters, and no combination of them is claimed to be profitable. Where sources disagree, for example on whether a strong trend begins at 25 or at 40, that disagreement is reported rather than resolved. Nothing here is backtested performance and no results are claimed or implied. Indicator readings describe what price has already done and do not predict what it will do next. Day trading and active trading carry a substantial risk of loss and are not suitable for every investor. SnapPChart grades a static chart screenshot you upload and returns a setup grade, entry, stop, targets and reasoning for that single image; it does not calculate ADX or directional movement, does not track indicator values, does not scan the market, and does not track your account, positions or P&L. It can only account for indicators that are visibly drawn on the image you upload. Do your own analysis, size positions so that being wrong is survivable, and consider speaking to a licensed financial professional about your own circumstances before trading.
Writes about AI-assisted day trading, technical analysis, and the systems traders actually use to stay disciplined.
ADX tells you there is a trend. It never tells you whether this setup is any good.
A reading of 34 is the same reading on a clean pullback and on a chart you should not be touching. Upload the screenshot and SnapPChart reads that single image against a fixed rubric, then returns a setup grade, an entry, a stop with the reasoning behind the level, targets, and the reward-to-risk those levels imply. It does not calculate ADX. It does give you a written second opinion before you commit size.