On this page
EMA — Averages & volatility, on the price chart.
What EMA measures
The series is seeded with the simple average of the first N samples, the convention used by TradingView and ATAS, then updated with ema = ema + alpha x (x - ema) where alpha = 2 / (N + 1). Every past sample keeps a non-zero weight that decays geometrically, so an EMA has no window in the strict sense: an old value is diluted, never dropped. With the close source the line is a smoothed price; with the delta source it smooths the per-bar difference between buy and sell aggression; with the volume source it smooths traded size.
The formula, as implemented
This is not how the indicator is usually defined elsewhere — it is what the shipped code computes, documented next to the implementation:
EMA_N(source) : graine = SMA des N premiers échantillons (convention standard, parité TradingView/ATAS), puis ema += α·(x − ema), α = 2/(N+1). Défaut N=20.
How to read it
- On the close source, slope comes first: the line turns only once enough closes have pulled it, so its direction confirms rather than anticipates.
- A price cross marks that the latest close broke away from its own decaying average. In a range that happens constantly and means very little.
- With the delta source, a line held above zero says buy-side aggression dominated over the recent bars. It does not say price rose, and the divergence between the two is the reason to look.
- With the volume source, the line gives a reference for what normal participation looks like right now, which helps in judging whether a breakout bar was actually traded.
- Two EMAs of different periods crossing compares two smoothings of the same data, so read it as a slow filter rather than an event.
Parameters and defaults
Period defaults to 20 and accepts 1 to 500; the smoothing factor follows directly as 2 / (N + 1), so period 20 gives each new sample about 9.5 percent of the result. The source selector is the parameter that changes the nature of the line rather than its speed: close, delta and volume are three different measurements.
| Parameter | Type | Default | Range |
|---|---|---|---|
| Period | number | 20 | 1 – 500 |
| Source | choice | Close | Close, Delta, Volume |
What it does not show
The indicator is drawn as an overlay on the price axis. With the delta or volume source the values have nothing to do with price levels, so the line sits far off-scale unless you read it knowing that. On the close source, an EMA is by construction late at turns and says nothing about how the move was traded. And because the seed is a simple average of the first N samples, two charts loaded with different history depth can show slightly different early values until the decay makes the difference irrelevant.
Related indicators
- SMA — Averages & volatility
- WMA — Averages & volatility
- Hull MA — Averages & volatility
- ATR (Wilder) — Averages & volatility
- Bollinger — Averages & volatility
- Keltner — Averages & volatility
This page in other languages
Frequently asked questions
- What is the difference between an EMA and an SMA?
- An SMA gives the same weight to each of the last N values and drops the oldest one entirely when a new bar arrives. An EMA weights recent values more and keeps a fading trace of everything before it, so it reacts sooner to a new move and never has a value fall out of the calculation abruptly.
- Can you apply a moving average to delta instead of price?
- Yes. This indicator exposes delta as a source alongside close and volume. The result smooths the per-bar balance between buy and sell aggression, which is a different measurement from a smoothed price and has to be read on its own scale.
- How is the EMA seeded on the first bars?
- The first N samples are averaged with a simple mean, and that value seeds the recursion. This is the convention used by TradingView and ATAS, so values match those platforms once past the warm-up. Nothing is plotted before the seed exists.