Double Exponential Moving Average (DEMA)

The Double Exponential Moving Average is Patrick Mulloy's 1994 attempt to solve the eternal moving average problem: how do you reduce lag without amplifying noise? The solution: apply EMA smoothing once, then apply EMA smoothing again to that result, then subtract the double-smoothed line from twice the single-smoothed line. The mathematical effect is that the lag of the second EMA cancels much of the lag in the first EMA. The result is a curve that hugs price more closely than a standard EMA at the same period length. For traders who feel the standard EMA lags too far behind real price moves, DEMA is a faster alternative. The trade-off: DEMA reacts to noise as fast as it reacts to real moves. In choppy markets, you'll get more whipsaws than with a standard EMA. DEMA shines in clean trending markets and struggles in ranges. In crypto markets, DEMA is sometimes used on intraday charts (4-hour and lower) where reactivity matters and the noise can be filtered with confirmation indicators.

DEMA calculation in three steps: 1. Calculate a standard EMA over N periods. Call this EMA1. 2. Calculate an EMA of EMA1 — the EMA of the EMA. Call this EMA2. 3. DEMA = (2 × EMA1) − EMA2. The formula's logic: EMA2 represents the lag in EMA1. By subtracting EMA2 from twice EMA1, you cancel much of that lag while keeping the smoothing. The DEMA value will track price more tightly than a standalone EMA of the same period. A 20-period DEMA reacts faster than a 20-period EMA but slower than a 20-period WMA in most market conditions. Computational note: DEMA requires roughly 2N bars of history before it produces a stable value (because EMA2 needs EMA1 history to compute). Don't trust DEMA values from the first ~40 bars of a new chart load.

**Active trend-following on intraday timeframes.** DEMA(20) on the 1-hour chart can provide faster trend confirmation than EMA(20) without going full WMA-fast. **Stop placement in trending markets.** Some traders trail their stops at the DEMA line. Because it tracks price closely, you lock in more profit on continuation moves, but get stopped out earlier on minor pullbacks. **Building custom oscillators.** Quantitative traders sometimes use DEMA as the smoothing function inside larger indicator constructions, replacing the standard EMA to get faster oscillator response. **Crossover systems.** Two DEMAs (fast and slow) produce earlier crossover signals than equivalent EMAs. Fewer signals overall but each one is closer to the actual turning point. For most retail traders, the standard EMA is sufficient. DEMA matters most for active intraday traders who specifically need to reduce lag.

**Amplified whipsaws.** DEMA reacts faster, which means it reacts to noise faster. Backtesting will show more total signals than EMA — many of them false. **Computation requires substantial history.** Don't trust DEMA values until ~2N bars have passed. If you're loading a fresh chart, the first DEMA values shown may be misleading. **Overconfidence in "lag-free" claims.** Despite the marketing language, DEMA is still mathematically a lagging indicator. It lags less than EMA but it doesn't see the future. Don't expect prophetic signals. **Less universal than EMA.** Some platforms label DEMA inconsistently or compute the second EMA differently. Verify that backtest values match live values on whatever platform you trade.