Simple Moving Average (SMA): Formula, Signals, and What Tests Reveal

A simple moving average (SMA) shows the average price over a chosen number of periods. It turns a busy price chart into a smoother line, helping traders read trends and define trading signals.

The calculation is simple. Its trading behavior is less obvious: making an average smoother can sometimes make a crossover rule trade more often. Below, we’ll cover the basics, then explore that behavior with interactive charts and results from 267,280 historical configuration tests.

What does a simple moving average show?

An SMA gives every price in its window the same weight. A 20-period SMA usually uses the last 20 closing prices. A rising line means that average is increasing; a falling line means it is decreasing.

The period follows the chart. Twenty periods on a daily chart means 20 daily bars. On a two-hour chart, it means 20 two-hour bars. Changing the chart interval changes how much time the average covers.

Traders use shorter averages to follow recent moves and longer ones to see a broader trend. The familiar 50- and 200-period averages are examples, rather than special values in the formula. Fidelity’s guide gives a concise overview of these uses.

Because an SMA includes older prices, it responds to a turn after price has already moved. Longer averages generally smooth more of the short-term movement, with more delay. The line describes past prices; it cannot tell us what comes next.

How to calculate an SMA

Add the prices in the window and divide by the number of prices:

SMA = sum of the last N prices ÷ N

Five closes: 10, 11, 12, 13, 14
SMA = 60 ÷ 5 = 12

When the next close arrives, replace the oldest one:

Next close: 13
New window: 11, 12, 13, 14, 13
SMA = 63 ÷ 5 = 12.6

Price fell, but the average rose. The new 13 replaced an older 10. An SMA’s next step depends on the price leaving its window as well as the price entering it. A rising average does not necessarily mean the latest candle rose.

How traders use SMA signals

There are three common ways to read the indicator:

• Trend direction: use the slope of the SMA to describe the direction of the average price.
• Price crossings: compare price with one SMA. A basic long-only rule buys above the line and exits below it.
• Moving-average crossovers: compare a shorter, fast SMA with a longer, slow SMA. A basic rule buys when the fast line is above the slow line and exits when it is below.

Here is the price-crossing rule on an invented, repeating pattern:

The crossings repeat because the prices repeat. If we extend this exact pattern, we can anticipate its signals. Real prices give us no such promise. A crossing tells us that two values have changed order; it does not establish that a lasting trend has begun.

Why price can keep crossing the same SMA

A market can move back and forth around an average, producing repeated entries and exits. This is called whipsaw.

The chart below uses historical Bitcoin as a worked example. The three interactive charts share the same prices, so we can change the rule and compare its decisions. Triangles show recorded TrainBard executions. We count each buy or sell as one execution.

Select Back and forth. Four consecutive closes alternate above, below, above, and below SMA 110. The rule buys, sells, buys, and sells over six hours. Each crossing is real, but those crossings do not develop into four distinct trends.

One response is to replace price with a short SMA. A brief move in one candle then has less influence on the comparison. Across our nine historical datasets, replacing price with a 10-bar average reduced executions in more than 99% of matched settings; the rest were unchanged.

It is tempting to extend that idea: if a little smoothing helps, surely more smoothing will quiet the rule further.

Why smoother SMAs can produce more signals

The smoothness of each line and the number of times they cross are different things. Two calm-looking lines can exchange places repeatedly when they stay close together.

Select Less than a dollar apart to see how small the gap can become. Then select Smooth, but still switching. The rule reacts to which line is higher, even when their difference is tiny.

There is a useful clue in the calculation. SMA 100 and SMA 110 share the same 100 recent closes. The longer average also includes ten older closes. The fast line is above the slow line precisely when the recent 100 prices average more than those older ten. Both groups change as the window moves. Their averages can trade places without either plotted line looking jagged.

In all nine tested datasets, a 100/110 crossover rule produced more executions than a 50/110 rule. These datasets cover Bitcoin and Ether at several candle sizes. A denser sweep then showed the increase occurring frequently as the two lengths approached each other, with exceptions.

Line smoothness alone cannot tell us how often a crossover rule will trade. The tests measure this behavior in the datasets we examined; they do not establish a rule for every market or every pair of lengths.

How a buffer changes crossover signals

Instead of adding more smoothing, we can require more separation. A buffer gives the rule a band in which to keep its existing position.

With a 0.25% buffer, the fast SMA must rise more than 0.25% above the slow SMA to trigger a buy. Once invested, the rule waits until the fast SMA falls more than 0.25% below the slow one to sell. Small crossings inside the band cause no action.

Select Ignore the small crossing. The buffered rule is already invested and stays invested as the lines change order. Finally outside the band shows its later exit. The averages have not changed; the decision boundary has.

Across the nine datasets, this small buffer reduced executions in roughly three quarters of matched settings and left the others unchanged. That measures activity, not whether the resulting trades were better. The chart also shows the cost of waiting: a wider gap can postpone an exit or leave a move untraded.

An SMA supplies a line. The trading rule decides whether a close crossing, a smoothed comparison, or a wider gap is enough reason to act. Testing those choices reveals behavior that a single attractive chart can hide.

How we checked the behavior

We started with nine historical datasets on Bitcoin and Ether, then examined crossover activity more closely with a denser Bitcoin sweep. Together, the studies cover 267,280 valid configuration evaluations, including deliberate repeats for verification. They describe related historical tests, rather than independent trials or predictions.

The calculations apply to any price series. The empirical coverage here is narrower: two crypto assets, four candle sizes, and two historical windows. The charts show selected moments; the comparisons below use full test windows. We have not established how frequently these behaviors occur in equities, foreign exchange, or every other market.

To try the same progression, open TrainBard with the supplied script: begin with a price crossing, add a fast average, then change the buffer. Keep the other settings fixed and look at the decisions each change adds or removes.

For a different weighting of past prices, see our exponential moving average guide. Our related RSI, MACD and Bollinger Bands guides examine how thresholds, smoothing and rolling windows change indicator behavior.