Double Exponential Moving Average (DEMA)
The double exponential moving average (DEMA) is a technical indicator designed to track price more closely and with less lag than a standard exponential moving average (EMA) of the same length. It does this by combining two EMAs mathematically: DEMA = 2 × EMA(n) - EMA(EMA(n)), where EMA(EMA(n)) is an EMA calculated on top of the first EMA rather than on raw price.
The effect of that formula is to strip out a large portion of the lag that is inherent to any moving average, since a plain EMA always trails price to some degree by construction - it is, after all, an average of recent values. By subtracting a smoothed version of the smoothing, DEMA reacts faster to genuine changes in direction while still filtering out a meaningful amount of short-term noise, sitting somewhere between a fast, choppy raw price line and a slow, heavily lagging simple moving average.
As an example, if a stock gaps up sharply and holds the new higher level, a standard 20-period EMA might take five or six sessions to fully "catch up" to the new price level because it is still weighting the pre-gap prices in its average; a 20-period DEMA calculated on the same data typically closes that gap in two or three sessions instead, making crossovers and trend signals based on it noticeably more timely.
The trade-off is that faster reaction also means more sensitivity to short-term noise and a higher chance of whipsaws (false signals) in choppy, range-bound markets - DEMA tends to shine specifically in trending conditions and can generate more false signals than a slower moving average when a market is going sideways.
DEMA is a purely mathematical construct applied to a price series, so it works identically whether that series is a forex pair, a stock, a cryptocurrency, or a commodity future - the underlying calculation does not change based on what is being charted.
Worth being clear about what the lag reduction can and cannot do, because the marketing around these averages tends to blur it. DEMA removes lag relative to an EMA of the same period; it does not remove lag. Every moving average is built from prices that have already happened, so none of them can turn before price does. What DEMA changes is how quickly the line acknowledges a turn that has already begun - which is valuable, and is a different claim from being early.
The practical consequence is that the choice between DEMA and a plain EMA is not really about which is better but about what the strategy needs. A trend-following system that suffers most from entering late benefits from the faster line. A system that suffers most from being shaken out of good positions by noise is made worse by it, and would do better with a slower average than the one it already has. Traders who switch to a faster average without changing anything else usually find they have traded one problem for the other, at which point the honest test is whether the whole approach performs better, not whether the line hugs price more closely on the chart.