Raw daily numbers jitter; trends hide inside the noise. A moving average smooths a series by replacing each point with the average of its neighborhood, letting the underlying direction show through. This moving average calculator takes any list of values and computes both classic forms — the simple moving average (SMA) and the exponential moving average (EMA) — for a window you choose, listing the smoothed series point by point.
Simple Moving Average (SMA)
The SMA of window w at each point is the plain average of the last w values: every point in the window counts equally, and the oldest drops out as a new one enters. For the series 10, 12, 14, 16, 18 with a window of 3: (10+12+14)/3 = 12, then (12+14+16)/3 = 14, then (14+16+18)/3 = 16 — the SMA series 12, 14, 16. Two properties define it: it lags (the first value needs a full window, and the average always trails a trending series by about half the window) and it has memory edges — a spike echoes for exactly w points, then vanishes as if it never happened.
Exponential Moving Average (EMA)
The EMA weights recent values more: EMA today = α × value today + (1 − α) × EMA yesterday, with the standard smoothing factor α = 2 ÷ (w + 1) matched to the window. Same series, w = 3, so α = 0.5, seeded with the first value 10: the EMA runs 10, 11, 12.5, 14.25, 16.125. Compare the characters: the EMA starts immediately (no warm-up gap), reacts faster to new data, and never fully forgets — old values decay geometrically instead of dropping off a cliff. Traders prefer it for responsiveness; statisticians often prefer the SMA for its clean interpretation.
Choosing a Window
The window is a trade between smoothness and lag. A window of 3 barely smooths but tracks closely; a window of 50 irons a series flat and answers questions about last month. Finance conventions put the 50-day and 200-day SMAs on stock charts (their crossover is the famous golden cross / death cross signal), while short-term traders watch 9- or 12-period EMAs. For business data — weekly sales, monthly signups — a window matching the seasonal cycle (7 for daily data with weekly rhythm, 12 for monthly with yearly rhythm) removes the seasonality and exposes the trend underneath, which is usually the analytical point.
What Moving Averages Cannot Do
They describe the past; they do not forecast it, whatever chart lore suggests. A moving average of a random walk looks convincingly trendy — smoothing manufactures apparent patterns from noise, a fact every technical-analysis skeptic keeps in a drawer. They also smear genuine events: a one-day price shock becomes a w-day bump in the SMA and a decaying tail in the EMA, so event timing read off a smoothed chart is systematically late. Use them to see trend and to damp measurement noise, and keep the raw series nearby for the truth about timing.
Worked Example, Read Properly
The defaults rise by exactly 2 each step — a perfect ramp — so the SMA trails the latest value by exactly one step's rise times half the window's span... in numbers: the latest SMA (16) sits one full step (2) behind the latest value (18), the lag a window of 3 on a steady trend produces. The EMA's latest value, 16.125, sits closer: exponential weighting cuts the lag without cutting the window. On flat, noisy data the two converge; on trending data the gap between them is the lag made visible, and this calculator prints both precisely so the trade-off is a number, not a vibe.
Common Mistakes
Comparing an SMA and an EMA of "the same period" as if they were equivalent smoothers (the EMA of window w reacts like a shorter SMA) is the standard confusion. Filling the SMA's warm-up gap with partial averages without saying so quietly changes the statistic for early points. Centering confusion matters in statistics packages: some plot the average at the window's middle rather than its end, shifting apparent timing by half a window. And the eternal one: reading a crossover as a prophecy instead of a description of what already happened.
Related Tools
Find the rest of the market toolkit in the Finance Calculators hub — the EOQ Calculator for the operations side and the Churn Rate Calculator for subscription metrics worth smoothing.
Frequently Asked Questions
How do you calculate a simple moving average?
Add the last w values and divide by w, then slide the window forward one point at a time. For 10, 12, 14, 16, 18 with w = 3 the SMA values are 12, 14 and 16.
What is the difference between SMA and EMA?
The SMA weights every value in its window equally and forgets anything older. The EMA weights recent values exponentially more and never fully forgets, so it responds faster and starts producing values immediately.
How is the EMA smoothing factor chosen?
By convention alpha = 2 / (window + 1), which matches an EMA of "period w" to the responsiveness of a w-period SMA. A window of 3 gives alpha 0.5; a window of 9 gives 0.2.
What window size should I use?
Match the question. Match the seasonal cycle (7 or 12) to strip seasonality from business data; use classic chart windows (50, 200) for long stock trends; use short windows (3 to 9) when recency matters more than smoothness. Longer windows smooth more and lag more.
Why does the SMA start later than the EMA?
The SMA needs a full window of values before its first legitimate average exists — three values before a 3-period SMA can print. The EMA seeds from the first value and refines as data arrives.
Do moving averages predict the future?
No. They summarize the recent past with varying lag. Crossovers and trends in smoothed data describe what has already occurred; treating them as forecasts is the most persistent myth in chart analysis.
What does a golden cross mean?
A short moving average (usually the 50-day SMA) crossing above a long one (usually the 200-day). Traders read it as momentum turning positive — a description of the trend so far, whatever its predictive worth.