Source-checked October 3, 2026 · Technical analysis education
SMA and EMA summarize the same price history differently. A faster response is a mathematical feature, not a promise of better trades.
This guide compares the two calculations and explains what a crossing can—and cannot—tell you. Neither is established here as the “best moving average” for day trading. If your question is which lookback to investigate, use our separate SMA selection and testing guide.
Each included observation contributes equally; the oldest leaves when a new one enters.
A recursive calculation carries forward older information with diminishing influence.
Compare like with like
Fidelity’s SMA explanation describes an arithmetic average over the selected period. Its EMA guide explains how heavier weighting of recent prices makes EMA generally more responsive than an SMA of the same length. Both react to observed prices and can lag a change in direction.
Scroll across the table to see all columns.
| Feature | Simple moving average | Exponential moving average |
|---|---|---|
| Calculation | Mean of the latest N selected prices | Current price combined with the previous EMA |
| Older observations | Drop out when they leave the window | Influence fades through the recursive calculation |
| Reaction to new prices | New price replaces the oldest included price | Recent price receives greater weight under the usual convention |
| Key limitation | Can remain slow after a sharp price move | Can react to short-lived moves that quickly reverse |
| What it does not provide | A guaranteed price floor, ceiling or return | A guaranteed earlier profitable entry or exit |
Keep the same symbol, bar interval, price input and session when comparing the lines. A 20-bar average on a five-minute chart and a 20-day average summarize different observations. A comparison of a short EMA against a long SMA changes two variables at once.
Also check the indicator’s timeframe override and display offset. A line shifted on the screen is not earlier knowledge of a price. For a transparent comparison, use no visual offset and document the calculation settings.
A worked SMA-versus-EMA example
The following is an original, invented six-bar price sequence. We use a three-bar lookback to keep the arithmetic small; it is not a suggested trading setting. Closing prices are $100, $102, $101, $105, $99 and $100.
For SMA, average the latest three closes. For EMA, use the common weighting factor α = 2 ÷ (N + 1), so α = 0.5 when N = 3. Start the EMA at bar 3 with the first three closes’ SMA, $101. Thereafter:
The update rule
New EMA = α × latest close + (1 − α) × previous EMA. With α = 0.5, each new value here is halfway between the latest close and the previous EMA. This is an illustration of the standard EMA calculation documented by TradingView.
Scroll across the table to see all columns.
| Completed bar | Closing price | Three-bar SMA | Three-period EMA |
|---|---|---|---|
| 1 | $100.00 | Not yet available | Not yet seeded |
| 2 | $102.00 | Not yet available | Not yet seeded |
| 3 | $101.00 | $101.00 | $101.00 (initial seed) |
| 4 | $105.00 | $102.67 | $103.00 |
| 5 | $99.00 | $101.67 | $101.00 |
| 6 | $100.00 | $101.33 | $100.50 |
At bar 4, the SMA is ($102 + $101 + $105) ÷ 3 = $102.666…, displayed as $102.67. The EMA is 0.5 × $105 + 0.5 × $101 = $103. At bar 5, the EMA becomes 0.5 × $99 + 0.5 × $103 = $101. Calculations retain precision; the table rounds only for display.
The EMA follows this sequence’s sharp rise and reversal more closely. The table contains no order executions and supports no conclusion about profitability. Platform values can differ near the start of a chart if their initialization or available history differs; compare those conventions before treating a mismatch as an error.
What a crossing actually says
Price crosses an average
The selected price moves from one side of the line to the other. That describes its relationship to a summary of recent history. It does not show what the next price will be.
A fast average crosses a slow one
The shorter and longer summaries change their relative order. The crossing occurs after enough of the underlying price data has changed; it does not establish an executable trade by itself.
In the worked example, the close is above both averages at bar 4 and below both at bar 5. Acting on every crossing can therefore reverse a decision quickly. Repeated changes of this kind are often called whipsaws. A series of such trades may accumulate execution costs without capturing a sustained move.
Specify whether a rule waits for a completed candle. A crossing displayed during a forming bar may disappear before the close. If the rule depends on that final close, a backtest should not quietly assume the decision was available earlier in the bar.
A moving average is not a barrier
Calling a line “support” or “resistance” does not oblige buyers or sellers to defend it. Prices can cross it, gap through it or stay far from it. A nearby line cannot guarantee an exit price or cap a loss.
Adding several indicators also does not automatically provide independent confirmation. If they reuse the same underlying prices, agreement may reflect related calculations. Any claimed improvement needs its own evidence.
A fair comparison workflow
- State the purpose. Decide whether you are describing a trend, defining a trigger or evaluating a filter. Avoid judging one line as a trigger and the other as background context.
- Match the inputs. Use the same prices, bars, session and lookback. Record the EMA initialization and allow adequate prior data before comparing signals.
- Keep the trading rules fixed. If testing a strategy, hold entry timing, exit logic, position sizing and cost assumptions constant while changing the average type.
- Examine reversals as well as trends. Inspect where the averages disagreed, where both were late and where a signal reversed. Keep those cases in the record.
- Evaluate on unused data. Retain a separate evaluation period and report net results and drawdowns. A paper-trading exercise can check a workflow, but cannot establish future returns or guarantee live fills.
Research-bias guidance from QuantConnect explains why repeatedly tuning on the same history can overfit a strategy. A visually appealing pair of lines is not a substitute for that evaluation. This guide makes no claim that either average wins a performance comparison.
Keep execution and risk in the decision
FINRA’s day-trading disclosure cautions that losses can be substantial, including all trading funds; margin and short selling can create larger obligations. A change from SMA to EMA does not change those basic risks.
As FINRA explains for stop orders, a stop-market trigger does not guarantee its execution price, and a stop-limit order can remain unfilled. Check the specific account’s rules and order behavior before committing money.
Move from the line to the question
Choose and test an SMA lookback · Explore position-size assumptions · Compare brokerage requirements
Sources and method
Substantively rewritten October 3, 2026 using the linked Fidelity and TradingView indicator documentation, QuantConnect research guidance and FINRA risk disclosures. The six-bar calculation is an original synthetic example with an explicitly stated EMA seed and weighting factor. InvestPips checked its arithmetic; we did not conduct a live trading test, a historical market backtest or a broker-execution study.
This article covers the distinction between SMA and EMA, principally for U.S. stock-chart education. Lookback selection is covered separately. Advertising is separate from this analysis. See our editorial and advertising standards.
What changed in this update
Removed the unsupported “Sarah” trading-success narrative and claims that EMA is inherently better for day trading. Corrected statements that crossovers avoid sideways markets or that longer averages supply more reliable signals. Removed inaccurate descriptions of alternative average types and replaced them with a focused, reproducible SMA-versus-EMA comparison. The original URL and publication date remain intact.
Educational information, not personalized investment advice or a recommendation to trade. No indicator, historical test or simulation guarantees a future result.