ForexAI Academy
Indicator Academy

Exponential Moving Average

EMA · module 02 of 33 · trend

A weighted average of closes that gives more weight to recent candles.

Exponential Moving Average on a chart
EUR/USD · 1h · demo data
Bear crossBull cross1.09051.0791
  • EMA 20
  • EMA 50
  • ▲/▼ where the rule triggers

EMAs weight recent candles more, so they turn sooner than SMAs. Pullbacks to the EMA 20 in an up-trend are the classic use.

How to apply it

  1. 1. Add it. Start with 9 and 21 on 1m/5m (Scalping).
  2. 2. Read the market first. A rising EMA(20) that price repeatedly returns to and bounces from is the classic pullback reference in an uptrend.
  3. 3. Adjust for ranges. Reacts faster than the SMA, which in a range means more false crossovers rather than fewer.
  4. 4. Manage the trade. Place the stop at 1.5 × ATR and the target at 3.5 × ATR, and only act when the chart above shows the trigger.

How it is calculated

EMA = (C × k) + (EMAprev × (1 − k)), where k = 2 ÷ (n + 1).

Inputs

  • · Period (n)
  • · Price source

Reading it correctly

In a trend

A rising EMA(20) that price repeatedly returns to and bounces from is the classic pullback reference in an uptrend.

In a range

Reacts faster than the SMA, which in a range means more false crossovers rather than fewer.

Settings by style

StyleSuggested values
Scalping9 and 21 on 1m/5m
Intraday20 and 50 on 5m/15m
Swing50 and 200 on 4h/1D

These are conventional starting points, not optimised values. Test any change on your own pair and timeframe before relying on it.

Worked example

With EMA(10) previously 1.0870 and a new close of 1.0890: k = 2/11 = 0.1818. EMA = (1.0890 × 0.1818) + (1.0870 × 0.8182) = 1.08736. The single new close moved the average about 0.4 pips — faster than an SMA(10) would, but still lagging.

When it failed

During the first minute after a rate decision, GBP/USD moved 70 pips. The EMA(9) turned up, price immediately retraced the whole move, and the crossover entry was stopped out within two candles. Fast weighting made the indicator react to a spike rather than a trend.

Limitations

  • Faster response means more sensitivity to single outlier candles.
  • Never fully forgets old data, so the first values after warm-up are approximate.
  • Still a lagging tool despite the weighting.

Works well alongside

  • · ADX / DMI
  • · ATR
  • · RSI

Redundant with

  • · SMA
  • · WMA
  • · KAMA
  • · SuperTrend

Stacking these produces agreement by construction, which is not confirmation.

Practice exercise

On 15m EUR/USD, mark every close that reaches within 0.25 ATR of EMA(20) while EMA(50) is rising. Record the next 20 candles' high and low in ATR units.

Check your understanding

1. Compared with an SMA of the same period, an EMA:
  • · Lags more
  • ✓ Responds faster to recent closes
  • · Ignores older data completely
  • · Predicts direction

The exponential weighting gives recent closes more influence, so the EMA turns sooner — at the cost of more sensitivity to outliers.

2. What is k for an EMA(19)?
  • · 0.05
  • ✓ 0.10
  • · 0.19
  • · 0.50

k = 2 ÷ (19 + 1) = 0.10.