What Is a Moving Average? SMA vs EMA, Explained

A moving average is the average price of an instrument over a fixed number of recent periods, recalculated on every new bar. It converts a jagged price series into a single line that follows the market with a delay. That delay is not a defect to be tuned away; it is what the calculation is.

How it works

The two common forms differ only in how they weight the past.

SMA(n) = (P1 + P2 + ... + Pn) / n

EMA(n): k = 2 / (n + 1)
        EMA_today = (Price_today × k) + (EMA_yesterday × (1 − k))

The simple moving average weights every period equally and drops the oldest bar completely when a new one arrives. This has a consequence people rarely account for: an SMA can turn because of a bar that left the window, not because of anything that just happened.

The exponential moving average weights recent prices more heavily and never fully discards old data — each past bar's influence decays geometrically. It responds sooner to a change in direction, and it produces more false turns for the same reason.

Lag is measurable. For a price series moving in a straight line, an SMA of n periods sits roughly (n−1)/2 periods behind. A 200-period average of daily bars is, by construction, describing where price was about a hundred days ago. Shortening the period reduces the lag and increases the number of times the line reverses on noise. There is no setting that removes both.

Why it matters

Every rule built on a moving average inherits the lag. You enter after the move has started and exit after it has ended, and you pay the spread, the commission and any slippage on each of those transactions. In a trending market that cost is small against the move captured. In a range it is the whole result: the line crosses back and forth, each cross is a paid transaction, and the account bleeds without a single dramatic loss.

This is why a moving average is a component, not a strategy. It says something about the recent average price. It says nothing about position size, stop placement, or whether the current market is the kind the rule works in.

What the data shows

Across the public accounts on ShowMyTrades with trading history (August 2026), the median account has closed 171 trades, the median trade lasts 2.4 hours, and the median profit factor is 1.28. A crossing rule read on an intraday timeframe fits that shape closely: frequent entries, short holds, and a margin over break-even thin enough for transaction cost to decide the outcome.

Turnover is therefore the figure to hold against any crossover system. Those accounts have traded 1,724,575 lots, and the commission billed on that volume comes to $4,782,670. A signal rule that fires more often does not merely need to be right more often — it needs to be right by enough to cover a fee levied on every crossing.

One honest limit: ShowMyTrades does not infer which indicator or expert advisor produced a trade, so none of these figures can be attributed to moving-average systems specifically. They describe the whole published population, which is the only thing they can honestly describe.

Where you see it on ShowMyTrades

The charts viewer plots equity, growth, profit and drawdown — the account, not the price chart. There are no indicator overlays, because the platform's job is to show what a strategy did, not to help design one.

What you can do is measure an MA-based system's real behaviour. In Custom Analysis, filter by Magic Numbers to isolate a single expert advisor on a shared account, then narrow with Trade Duration (seconds), Days of Week and Trading Hours (Server Time) to see when the rule actually fires. In Advanced Statistics, Total Trades, Avg. Trade Length and Total Commissions together tell you whether the turnover is being paid for. The Breakdown Statistics panel, on its By Symbol tab, shows whether the result came from one instrument that happened to trend — which for any trend following rule is the question that matters.

Common misunderstandings

• "The crossover is the system." A crossover is an entry trigger. Without sizing, an exit and a cost assumption it is not a strategy, and the backtest of it is measuring the market, not the method.
• "The 200-day average is significant because everyone watches it." Popularity produces attention, not edge. Nothing in our data supports or refutes any specific period, and no honest source can offer you one.
• "EMA is better because it reacts faster." Faster reaction means earlier entries and more false ones. The choice is a trade-off, not an upgrade.
• "I optimised the period and 34 worked best." Selecting a period because it performed best on history you have already seen is curve fitting, whatever the number turns out to be.

For the metrics that tell you whether any rule-based system is actually earning its costs, see tracking your trading performance.

Related terms