Nearly every indicator you will meet is a moving average wearing a costume. Understand what an average of past prices can and cannot tell you and most of the rest of the toolbox becomes predictable.
What it computes
A simple moving average (SMA) adds the last n closes and divides by n. A 20-period SMA on a daily chart is the mean closing price of the last 20 days, recalculated each day as one price drops out of the window and a new one enters.
That is the whole mechanism, and it has one consequence worth sitting with: the line is a summary of the past. It contains no information that was not already on the chart. Its value is not prediction — it is that a single smooth line is easier to reason about than 200 individual closes.
Simple versus exponential
An exponential moving average (EMA) weights recent prices more heavily, with the weight decaying as you go back. It responds to new information faster.
The trade-off is exactly what you would expect, and it is unavoidable: faster response means more false signals, slower response means later ones. There is no setting that gives you both. An EMA will turn before an SMA at a genuine reversal — and it will also turn during noise that goes nowhere.
Short-term traders tend to prefer EMAs; position traders tend to prefer SMAs. Neither is superior. The choice is a statement about which error you would rather make.
The lag is structural
A moving average cannot lead price, because it is computed from price. The line turns after the move has begun, always.
Roughly, an n-period simple average lags by about half its length — a 50-day SMA reflects conditions centred about 25 days ago. This is not a flaw to be optimised away. It is arithmetic. Anyone selling you a "zero-lag" average has either added a projection (a guess) or reduced the smoothing (less smoothing, same trade-off).
Practically: moving averages confirm a trend that already exists and are useless for catching its start. Using one to time entries in a range is the most common way traders lose money with this indicator — price crosses the average constantly there, and each cross is a signal to do the wrong thing.
The common periods, and why
- 20 — roughly a month of trading days. A short-term bias line.
- 50 — a quarter. The standard medium-term reference.
- 100 and 200 — long-term. The 200-day is the most-watched line in markets.
These numbers have no mathematical privilege. They matter because enough participants watch them that reactions around them become partly self-fulfilling — institutional mandates, algorithmic triggers and financial-media narratives all cluster on the 200-day. That is a real effect, but it is a crowd effect, not a property of the average.
It also means obscure settings lose that advantage. A 137-period EMA may backtest better; nobody else is watching it.
How they are actually used
As a trend filter. The most defensible use. Price above a rising 200-day average means the long-term trend is up, so take long setups and skip short ones. This is a filter on other decisions, not a signal in itself.
As dynamic support and resistance. Pullbacks in a strong trend often stall near a rising average. This works while the trend works and fails the moment it ends — so it tells you nothing the trend was not already telling you.
Crossovers. A shorter average crossing a longer one — the "golden cross" and "death cross" at 50/200 — is the classic signal. Both lines lag, so their intersection lags doubly, and they whipsaw badly in ranges. Read them as a description of what has already happened, not a call to act.
Slope over position. A frequently overlooked point: the direction of the average is more informative than whether price is above or below it. A flat 200-day means no long-term trend exists, whatever side of it price happens to be on.
What to take from this
- A moving average summarises past prices — it adds no new information.
- EMA reacts faster and is wrong more often; SMA is slower and steadier. Pick your error.
- Lag is arithmetic, roughly half the period. "Zero-lag" claims are marketing.
- The 50 and 200 matter because of who watches them, not because of the maths.
- Use them to filter trend direction, not to time entries — and read the slope, not just the side.
Averages describe direction. The next lesson covers the other half of the indicator world: oscillators, which measure the speed of a move rather than its heading.