The Directional Movement Index (DMI): Reading Trend Strength and Its Blind Spots

A moving-average crossover tells you a trend may have turned. It says nothing about whether that trend has any force behind it. That gap is exactly what the Directional Movement Index was built to close.

J. Welles Wilder introduced the DMI in 1978, in the same book that gave traders the RSI and ATR. The idea is simple to state and harder to use well: separate the direction of a move from its strength, and put a number on each. Get that separation right and you can tell a real breakout from a market drifting sideways while pretending to go somewhere.

What the DMI Actually Measures

The DMI compares each bar's high and low against the previous bar's to work out whether buyers or sellers are gaining ground. From that comparison it produces three lines.

Positive Directional Indicator (+DI) tracks the strength of upward price movement. It measures how far the current high extends above the prior high, then smooths the result over a lookback period. A rising +DI means buyers are pushing the market up with conviction.

Negative Directional Indicator (-DI) does the mirror image for the downside, measuring how far the current low sits below the prior low. A higher -DI points to sellers in control.

Average Directional Index (ADX) is the one that sets DMI apart from most trend tools. It ignores direction entirely and measures only how strong the trend is. A market can be falling hard or rising hard and the ADX will read high in both cases. As a rough guide, an ADX above 40 signals a powerful trend, while a reading below 20 says the market is weak or ranging.

Here is how the pieces are calculated:

You do not need to compute these by hand — every charting platform does it for you. But knowing that +DI and -DI are normalized by ATR explains why the indicator behaves oddly when volatility spikes: the denominator is moving too.

Reading the Signals

The headline signal is the crossover between +DI and -DI.

When +DI crosses above -DI, upward momentum has overtaken downward momentum — a potential long setup. When +DI crosses below -DI, the balance has tipped the other way, and it reads as a potential short or an exit from a long. That is the whole mechanic. It is clean, which is part of the appeal and part of the trap.

The crossover on its own is only half the picture. A +DI/-DI cross in a listless market flips back and forth and means very little. This is where the ADX earns its place: it tells you whether the crossover is worth acting on.

• ADX rising through 20, then 25: a trend is taking hold. Crossover signals in this environment carry more weight because there is genuine momentum behind the move.
• ADX above 25 and climbing: the trend is strengthening. Wilder's original use case was to let winners run here — and, if you are managing an open position, this is the zone where you might trail a stop rather than tighten it prematurely. See stop-loss placement for how to move a stop without choking a good trade.
• ADX below 20: no meaningful trend. Trend-following signals from the DMI are effectively void, and a crossover here should be treated as noise, not an invitation.
• ADX turning down from a high level: momentum is fading. The trend may be rolling into a consolidation. That is not automatically a reversal — it is a reason to stop adding and to reassess.

The honest way to frame the ADX is as a filter, not a signal generator. It never tells you which way to trade. It only tells you whether the market is trending strongly enough that a directional read is worth taking.

Where the Trade Is Invalidated

A signal without an exit plan is a wish. The DMI is unusually good at defining invalidation because its own logic gives you the level.

If you take a long on a +DI/-DI bullish cross, the setup is invalidated when the lines cross back — -DI reclaiming +DI is the market telling you the premise was wrong. A second, softer form of invalidation: the crossover fires but the ADX refuses to rise above 20. That is a trend signal with no trend behind it, and it is a common way DMI traders get chopped up.

Because the crossover level is objective, it pairs naturally with a hard price stop rather than replacing one. Plenty of traders set the structural stop at a swing point and use the DMI cross-back as an earlier warning to reduce or close. Whichever you choose, decide the size before the entry, not after the position moves against you. If sizing is the part you skip, start with position sizing — the indicator does not protect the account, the size of the loss does.

What the DMI Is Good At

In a trending market, the DMI does real work. It filters out the small, directionless price wiggles that fool simpler tools and keeps you focused on the moves with force behind them. When the ADX confirms a strong trend, the +DI/-DI relationship gives a clear, unambiguous read on which side is winning.

The signals are also explicit. There is no interpreting a shape or drawing a line by feel — either +DI is above -DI or it is not, either the ADX is rising or it is not. For traders who want rules they can write down and check the same way every time, that clarity is the main draw.

And the ADX component fills a gap most indicators leave open. Momentum oscillators tell you how fast; the ADX tells you how committed. A trade taken with a rising ADX is a different proposition from the same crossover taken with the ADX flat at 15.

Where It Fails

The DMI has a specific and well-known weakness: it is poor in choppy, sideways, or highly volatile markets. When price is ranging, +DI and -DI cross each other repeatedly, and each cross looks like a signal. Most of them are false. Acting on that stream of whipsaws is a reliable way to bleed the account through a series of small losing trades — the losses are individually minor and collectively serious.

This is not a flaw you can tune away with a different setting. It is inherent to how the indicator works, which is why the ADX filter matters so much: refusing to trade DMI crossovers when the ADX is below 20 removes most of the bad signals, at the cost of sitting out some early trend entries.

The second limitation is that the DMI should not stand alone. Leaning on it as your only input is risky. It pairs well with tools that measure something it does not — a MACD for momentum, an RSI for overbought and oversold conditions, or plain price structure for support and resistance. The point of stacking indicators is not to pile on confirmation until you feel certain; it is to catch the cases where one tool is lying to you. When the DMI says trend and the RSI says exhausted, that disagreement is information.

A Note on Comparing Indicators

The DMI is often set against tools like the Aroon indicator, which also tries to identify trend changes by looking at the timing of recent highs and lows. The broad point holds across all of them: no single indicator sees the whole market, and each measures the same price action from a slightly different angle. Pick the one whose logic you actually understand, then test how it behaves on the instruments you trade before you size up.

The DMI's real contribution is the discipline it forces. It makes you ask two questions in order — is there a trend, and how strong is it — before which way do I trade it. Answer them in that sequence and the crossover signals become far more useful. Skip the strength question, trade every cross, and the indicator will happily hand you a losing streak in a flat market.