Meaning and weighting
An exponential moving average combines a new observation with the previous smoothed value. If the weight of the new observation is alpha, the weight of the previous estimate is one minus alpha. With repeated application, older data receive declining weight. Fidelity distinguishes this method from a simple average with equal weights.[1]
Author’s calculation
The initial estimate is 100, the new price is 110 and the coefficient is 0.5. The new average is 105. With a different starting point, a short calculation produces a different value, so initialization must be disclosed.
Limitation
The EMA uses past and current observations, not new independent information about the future. Giving more weight to the new price changes sensitivity and memory. A crossing of two averages is a relationship between transformations of one series, not independent proof of future direction.
Further reading and sources
A comparison of EMA, the simple moving average and the linearly weighted moving average on one artificial series appears in the [separate guide](/guides/moving-averages/). Checking an indicator’s arithmetic and checking a trading claim should be done separately.
[1] Fidelity Investments, Exponential Moving Average / Simple Moving Average; definitions of smoothing methods.
Primary material: [Fidelity Exponential Moving Average](https://www.fidelity.com/learning-center/trading-investing/technical-analysis/technical-indicator-guide/ema)
Additional primary material: [Fidelity Simple Moving Average](https://www.fidelity.com/learning-center/trading-investing/technical-analysis/technical-indicator-guide/sma). Definition only, without validation of trading signals.