Averages — Arithmetic Mean, Weighted & Moving Average

An average is a single number that summarizes a set of values. In school, "average" usually means the arithmetic mean, but the weighted average and the moving average each answer a different question, so the right choice depends on what you are trying to capture.

Use the arithmetic mean when every value should count equally, a weighted average when some values should count more, and a moving average when ordered, time-based data needs smoothing.

The arithmetic mean: every value counts equally

The arithmetic mean adds all values and divides by how many there are. Why divide by $n$? Because giving each of the $n$ observations the same share of the total is exactly what "equal influence" means.

The mean uses every value, which makes it easy to compare across groups, but it is sensitive to outliers — one unusually large or small number can pull it away from what feels typical.

The weighted average: some values matter more

When importance differs, weight each value:

Here $x_i$ is a value and $w_i$ is its weight. Larger weights give more influence. The formula divides by $\sum w_i$ because the weights, not the count, now define how the total should be shared. Course grades, investment returns by portfolio share, and average prices by quantity all follow this pattern. One condition matters: the total weight $\sum w_i$ must not be $0$.

The moving average: smoothing data over time

For data in time order, average a rolling window of recent values instead of the whole set. With window length $k$:

This smooths noise so a short-term trend is easier to see. It does not remove variation and does not predict the future; it summarizes recent data using the window you chose. A longer window looks smoother because it reacts more slowly.

One worked example that shows all three

A student's practice scores over five weeks are $70$, $75$, $80$, $85$, and $100$.

For one overall average across all five weeks, use the arithmetic mean:

Now suppose recent work should count more, with weights $1, 1, 1, 2, 2$:

The newer scores carry more weight, so the result rises to $85$. To smooth the recent trend instead, take a $3$-week moving average of the last three weeks:

The same five numbers produced three different answers because the goal changed. That is the core idea behind choosing the right average.

Try it yourself, then check the answer

Take five numbers from your own study data. Find the arithmetic mean, then a weighted average where the last two values count double, then a $3$-value moving average for the last window. If your weighted average lands between the plain mean and the largest weighted values, your setup is consistent.

Calculation traps to watch for

• Using the plain mean when the data already has weights. Equal weighting only makes sense when each value should contribute equally; otherwise it misleads.
• Averaging averages without the original weights. If one class has $10$ students and another has $30$, you cannot average the two class averages as if the groups were equal. Use the underlying counts.
• Forgetting to divide by the total weight. Multiplying values by weights is only half the job; you still divide by $\sum w_i$.
• Naming a moving average without its window length. A $3$-day and a $30$-day moving average are not interchangeable.

Where each average is used

Use the arithmetic mean for test scores or measurements where every observation counts equally. Use a weighted average when importance is already assigned, such as grade categories or quantities sold. Use a moving average for time-based data — temperatures, sales, traffic, study progress — when raw values jump around from period to period.