The triple exponential moving average (TEMA) is a technical indicator that smooths price data while reducing the lag of an ordinary exponential moving average. Despite its name, it is not simply an EMA applied three times. It combines a single EMA, an EMA of that EMA, and an EMA of the second in a way that cancels much of the delay introduced by smoothing, so the line follows prices more closely and turns sooner.
TEMA was introduced by Patrick Mulloy in an article in Technical Analysis of Stocks and Commodities magazine in 1994, alongside the closely related double exponential moving average (DEMA). It belongs to the family of lag-reducing averages that also includes the Hull moving average and adaptive measures such as Kaufman's adaptive moving average. It should not be confused with TRIX, an oscillator based on the rate of change of a triple-smoothed EMA.
Formula
For a chosen period , three exponential averages are calculated in sequence:
- is the exponential moving average of price over periods.
- is the exponential moving average of over periods.
- is the exponential moving average of over periods.
TEMA is then:
Each EMA uses the usual smoothing factor , so that the newest value is weighted by and the previous average by . The DEMA, for comparison, is .
Worked example
Suppose a share has been rising steadily and, on a given day with a 20-day setting, the three averages are , and . Each successive smoothing trails further behind the rising price. TEMA is:
The TEMA of 51.2 sits above the single EMA of 50.0, closer to where the price is trading in a rising market. The DEMA in the same case would be , between the two. In a falling market the corrections work the other way and TEMA sits below the EMA.
Why it reduces lag
Any moving average lags behind price because it includes older data. When an EMA is smoothed again, the lag roughly doubles, and a third smoothing adds more. The difference between successive EMAs is therefore a rough estimate of the lag itself. By adding back three times the gap between the first and second averages and correcting with the third, TEMA removes much of this delay while keeping most of the noise reduction of a long average. The weights 3, minus 3 and 1 sum to one, so the indicator stays on the same scale as price.
The result is that a 20-period TEMA responds to a turn in price considerably faster than a 20-period EMA or simple moving average, and on charts it hugs the price more closely. The improvement is largest in steady trends, where the gap between successive averages is a good guide to the lag. When prices reverse abruptly, the estimate of lag is out of date for a few bars, which is one source of the overshooting described below.
Uses
TEMA is used in the same ways as other moving averages. The most common are:
- Trend direction. A rising TEMA, or price holding above it, is read as an uptrend; a falling one as a downtrend.
- Crossovers. A price crossing TEMA, or a short TEMA crossing a longer one, is used as a signal in a moving average crossover system, with the aim of earlier entries and exits than standard averages give.
- Input to other indicators. Because it is a cleaner version of price, TEMA can replace an EMA inside oscillators such as the MACD to make them faster.
Limitations
Reducing lag has a cost. A TEMA reacts more to short-lived price swings than a conventional average of the same length, which produces more false crossover signals in sideways or choppy markets. Because the formula extrapolates from the gaps between averages, it can overshoot price after a sharp move and briefly sit beyond the most recent value. Settings also matter: shorter periods amplify both the speed and the noise, so many users pair a fast TEMA with a slower one or with a filter such as a trend-strength measure. Its readings also depend on enough history for all three EMAs to settle, so early values on a new chart are unreliable. Like any moving average, TEMA is built entirely from past prices; it describes trend rather than forecasting it, and its usefulness in a trading rule depends on the market, the period chosen and trading costs.