Directional Movement Index (DMI): A Practical Guide for Traders
Introduction
The Directional Movement Index (DMI), developed by J. Welles Wilder (1978), helps traders measure both the direction and the strength of a market trend. The system produces two directional lines—+DI (positive directional indicator) and −DI (negative directional indicator)—and is commonly used together with the Average Directional Index (ADX), which measures trend strength. Crossovers between +DI and −DI signal potential trend direction changes; ADX is then used to judge whether a trend is strong enough to trade.
Key takeaways
– +DI and −DI measure upward and downward pressure based on consecutive highs and lows.
– ADX (derived from +DI and −DI) measures trend strength—readings above 20–25 typically indicate a meaningful trend.
– Crossovers (+DI rising above −DI or vice versa) generate basic buy/sell signals, but they produce false signals in choppy markets.
– Combine DMI with ADX and other confirmation tools (trendlines, moving averages, price action, volume) to reduce whipsaws.
Core formulas (plain-language form)
– Directional Movement (per period)
– +DM = current high − previous high (if this is positive and greater than the corresponding down move; otherwise +DM = 0)
– −DM = previous low − current low (if this is positive and greater than the corresponding up move; otherwise −DM = 0)
– Wilder’s smoothing (N-period; commonly N = 14)
– Smoothed +DM_today = Smoothed +DM_yesterday − (Smoothed +DM_yesterday / N) + +DM_today
– Smoothed −DM_today = Smoothed −DM_yesterday − (Smoothed −DM_yesterday / N) + −DM_today
– Average True Range (ATR) — use Wilder’s ATR smoothing:
– ATR_today = ATR_yesterday − (ATR_yesterday / N) + TrueRange_today
– Directional Indicators (scaled to percentage)
– +DI = (Smoothed +DM / ATR) × 100
– −DI = (Smoothed −DM / ATR) × 100
– Directional Movement Index (DX) and ADX
– DX = (|+DI − −DI| / (+DI + −DI)) × 100
– ADX = Wilder’s smoothed average of DX (over N periods)
Step-by-step calculation (practical)
1. Choose a period N (default: 14).
2. For each period, compute:
a. True Range (TR) = max(high − low, |high − previous close|, |low − previous close|).
b. +DM and −DM as above (only one can be nonzero each period; if neither is positive, both are zero).
3. For the first smoothed values (period N), initialize:
– Smoothed +DM_initial = sum of first N +DM values
– Smoothed −DM_initial = sum of first N −DM values
– ATR_initial = sum of first N TR values
4. After initialization, apply Wilder’s smoothing each new period:
– Smoothed +DM_today = Smoothed +DM_yesterday − (Smoothed +DM_yesterday / N) + +DM_today
– Similarly for Smoothed −DM and ATR
5. Compute +DI and −DI: divide the smoothed directional movements by ATR, multiply by 100.
6. Compute DX each period and then smooth DX over N periods to produce ADX.
Interpreting DMI and ADX (practical rules)
– Directional bias:
– If +DI > −DI → upward directional pressure (bullish).
– If −DI > +DI → downward directional pressure (bearish).
– Crossovers:
– Buy signal: +DI crosses above −DI.
– Sell/short signal: +DI crosses below −DI.
– Use ADX to filter:
– ADX < 20 (or 20 (or 25) and ideally rising.
– Optional Condition 3: Price is above a chosen moving average (for longs).
– Enter on the close of the crossover candle or wait for a smaller pullback.
3. Stop and risk management:
– Place stop below recent swing low (for a long) or use ATR multiple (e.g., 1.5 × ATR) to set stop.
– Use position sizing to limit risk per trade (e.g., 1–2% of account).
4. Exit:
– Exit when −DI crosses above +DI AND ADX is still indicating trend reversal, or
– Use trailing stop based on ATR, or
– Exit if ADX falls below threshold (trend losing strength).
5. Avoid:
– Taking pure DMI crossovers when ADX is below threshold (range-bound markets).
– Overtrading on small intraday noise—observe multiple timeframes for confirmation.
Ways to make the DMI more reliable
– Add ADX threshold: require ADX > 20 or 25 and rising to take DMI signals.
– Timeframe confirmation: require the signal on your trade timeframe to be confirmed on a higher timeframe.
– Price confirmation: wait for a close beyond a recent swing high/low rather than reacting to a raw crossover.
– Add volatility filters: use ATR to size stops and filter signals when volatility is too low or too high.
– Combine with trend filters: moving averages, trendlines, or Ichimoku can help confirm the general market direction.
– Use volume or momentum confirmation: rising volume with a bullish DMI crossover or confirming momentum (RSI / MACD) reduces false signals.
– Adjust period N: increasing N reduces noise (fewer false signals but slower signals), decreasing N increases sensitivity (more signals but more noise). Stick with N = 14 as default and test alternatives.
What works well with DMI
– ADX: for measuring trend strength—essential companion.
– Moving averages: to determine market bias (long only above MA).
– Price action: swing highs/lows, support/resistance confirmation.
– Volume: rising volume on crossover is a stronger signal.
– Oscillators (RSI, Stochastics) for overbought/oversold or momentum divergence confirmation.
– ATR: for stop placement and volatility filtering.
Limitations and common pitfalls
– Lagging: DMI (and ADX) are based on historical highs/lows and smoothed averages—signals are lagged relative to price.
– False signals in choppy markets: +DI and −DI can crisscross frequently without a sustainable trend.
– Parameter sensitivity: choosing N improperly can either miss moves (N too large) or generate noise (N too small).
– Not predictive: indicators reflect past movement; a crossover doesn’t guarantee that price will follow.
DMI vs. Aroon indicator — main differences
– Purpose:
– DMI: measures directional movement and, with ADX, the strength of a trend.
– Aroon: measures how long it’s been since the most recent high and low over a lookback period; good for identifying trend formation and the start/end of trends.
– Calculation:
– DMI uses consecutive high/low comparisons, smoothed directional movement, and ATR scaling.
– Aroon uses the number of periods since last high/low (produces Aroon Up and Aroon Down lines scaled 0–100).
– Interpretation:
– DMI focuses on directional pressure and uses ADX to measure strength; crossovers between +DI and −DI drive signals.
– Aroon is more focused on timing: high Aroon Up indicates a recent high (bullish), high Aroon Down indicates a recent low (bearish).
– Which to use:
– Use DMI when you want direction and strength measures (trend-following).
– Use Aroon to spot the start/stop of trends and to detect consolidations.
Other indicators similar to or paired with DMI
– Average Directional Index (ADX) — part of the same system, measures trend strength.
– Parabolic SAR — trend-following stop and reversal indicator.
– MACD — momentum and trend-following crossover system.
– Moving averages / EMA crossovers — simple trend filters.
– Aroon — for trend start/end detection.
– RSI/Stochastics — momentum/overbought-oversold confirmation.
– On-balance volume (OBV) or volume indicators — add confirmation.
Backtesting and practical tips
– Backtest any DMI-based strategy across multiple instruments and market regimes (trending vs. ranging).
– Use walk-forward analysis and out-of-sample testing to avoid curve-fitting.
– Combine rules (ADX threshold, higher-timeframe confirmation) that improve real-world robustness.
– Evaluate commissions, slippage and trading hours (intraday liquidity) when testing DMI on short timeframes.
Example trade setup (concise practical checklist)
1. Chart: daily timeframe, DMI with N = 14, ADX plotted.
2. Entry long:
– +DI crosses above −DI.
– ADX > 25 and rising.
– Price above 50-day MA (optional filter).
– Place entry at close of crossover candle or wait for a pullback to a support level.
3. Stop: below recent swing low or 1.5 × ATR below entry.
4. Target: risk-reward 1:2 minimum, or trail with ATR as ADX stays elevated.
5. Exit: −DI crosses over +DI or ADX drops below 20 and is falling.
The bottom line
DMI is a useful trend-direction and trend-strength tool when used properly and combined with ADX and additional confirmation filters. By itself, DMI crossovers are noisy in sideways markets; using ADX thresholds, price confirmation, timeframe alignment, and risk management will significantly improve the quality of signals. Always backtest and paper-trade a DMI-based strategy before committing real capital.
Sources
– Investopedia — Directional Movement Index (DMI). (Investopedia / Michela Buttignol) https://www.investopedia.com/terms/d/dmi.asp
If you’d like, I can:
– Produce a worked numeric example with concrete price data and full calculations for +DI, −DI, DX, and ADX (using N = 14 or a shorter N to illustrate), or
– Create a sample trading rule set you can import to a backtesting platform (pseudocode or strategy outline). Which would you prefer?