Exponential Moving Average (EMA)

An exponential moving average (EMA) is a type of moving average that tracks the price of an asset over a chosen period while giving more weight to the most recent prices. Each new day's price pulls the average toward it by a fixed fraction, and the influence of older prices fades gradually rather than dropping out all at once. Because of this weighting, an EMA follows price changes more closely than a simple moving average (SMA) of the same length.

EMAs are among the most widely used indicators in technical analysis of shares, currencies and cryptocurrencies, and they appear by default on most trading platforms. Common settings include the 12-, 26-, 50- and 200-period EMA. They are building blocks for other tools, most notably the moving average convergence divergence (MACD) indicator, and they are used in crossover signals such as the golden cross and death cross. Like all moving averages, the EMA describes past prices and belongs among the lagging indicators.

Imagine keeping a running opinion of a restaurant. A simple average would treat every meal over the last month equally, so a terrible dinner a month ago counts as much as last night's excellent one. An exponential average pays most attention to the latest meals and lets older ones fade from memory, so the opinion updates faster when the kitchen improves or declines.

Traders use the EMA the same way, as a smoothed line that shows which way prices have been heading, updated quickly when the trend changes. The cost of reacting faster is that the line is also more easily fooled by a single unusual day.

The EMA for today is calculated from today's price and yesterday's EMA:

The weighting factor, often called the smoothing constant, is:

• $P_t$ is the latest closing price.
• $\text{EMA}_{t-1}$ is the previous period's EMA; the first value is often seeded with a simple average.
• $n$ is the number of periods chosen, such as 10 or 50.
• $k$ is the weight given to the newest price.

For a 10-day EMA, $k = 2 \div 11 \approx 0.1818$. Suppose yesterday's EMA was 50.00 and today's closing price is 52.00. Today's EMA is $52 \times 0.1818 + 50 \times 0.8182 \approx 9.45 + 40.91 = 50.36$. A 10-day simple average would move only if 52 replaced a lower price dropping out of the window, so the two averages can diverge. With a 50-day EMA, $k \approx 0.0392$, and the same price change would move the average by only about 0.08.

• Trend direction. Prices holding above a rising EMA are read as an uptrend, and below a falling EMA as a downtrend.
• Crossovers. A short EMA crossing above a longer one is taken as a bullish signal, and crossing below as bearish.
• Dynamic support and resistance. In trending markets, prices often pull back toward a popular EMA before resuming the trend, and some traders buy or sell near it.
• MACD. The MACD line is the 12-period EMA minus the 26-period EMA, with a 9-period EMA of that difference used as a signal line.

Short EMAs, such as 9 or 20 periods on intraday charts, are popular in day trading, while the 50- and 200-day EMAs are watched by longer-term investors, including on daily charts of Bitcoin. EMAs are often combined with other tools, such as Bollinger Bands, the relative strength index or the average directional index, to filter signals.

The simple moving average gives equal weight to each price in the window and ignores everything older, so it changes abruptly when a large price drops out. The EMA weights recent prices more and never fully forgets older ones, so it turns sooner at trend changes. Supporters prefer the EMA for faster signals; others prefer the SMA because it is steadier and generates fewer false signals in choppy markets. Neither is objectively better: the choice depends on the trader's time frame and tolerance for noise.

An EMA cannot predict prices; it summarizes what has already happened. In sideways markets with high volatility, prices cross back and forth through the average, producing many misleading signals. Results depend heavily on the period chosen, and traders who adjust settings until past signals look good risk fitting noise rather than finding a real pattern. Studies of moving-average rules have found that any advantage tends to be small and often disappears after trading costs, so the EMA is best treated as a way to describe trends rather than a stand-alone strategy.