Averages — Arithmetic Mean, Weighted & Moving Average

An average is a single number that summarizes a set of values. In school, "average" often means the arithmetic mean, but a weighted average or moving average may be the better choice because each one answers a different question.

Use the arithmetic mean when every value should count equally. Use a weighted average when some values should count more than others. Use a moving average when the data is ordered over time and you want to smooth short-term ups and downs.

Arithmetic mean: use it when every value should count equally

The arithmetic mean is the usual average:

$$\text{mean} = \frac{x_1 + x_2 + \cdots + x_n}{n}$$

This works when each observation deserves the same influence. If one value should matter more than another, the arithmetic mean is not the right summary.

The mean uses every value in the set, which makes it useful and easy to compare across groups. It is also sensitive to outliers, so one unusually large or small number can pull it away from what feels typical.

Weighted average: use it when some values have more importance

A weighted average gives different importance to different values:

$$\text{weighted average} = \frac{\sum w_i x_i}{\sum w_i}$$

Here, $x_i$ is a value and $w_i$ is its weight. Larger weights give a value more influence on the result.

This is the right tool when the problem already tells you that some parts matter more. 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 result is only meaningful if the weights actually match the situation you are modeling.

Moving average: use it to smooth data over time

A moving average is used for data listed in time order. Instead of averaging the whole set at once, you average a rolling window of recent values.

For a simple moving average with window length $k$:

$$\text{MA}_t = \frac{x_t + x_{t-1} + \cdots + x_{t-k+1}}{k}$$

This helps smooth noisy data so the short-term trend is easier to see. It does not remove variation, and it does not predict the future. It only summarizes recent data using the window you chose.

The window length matters. Change the window, and the moving average changes too. A longer window usually looks smoother because it reacts more slowly.

One worked example that shows the difference

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

If you want one overall average across all five weeks, use the arithmetic mean:

$$\frac{70 + 75 + 80 + 85 + 100}{5} = \frac{410}{5} = 82$$

So the arithmetic mean is $82$.

Now suppose the teacher wants recent work to count more, using weights $1, 1, 1, 2, 2$. Then the weighted average is

$$\frac{1(70) + 1(75) + 1(80) + 2(85) + 2(100)}{1 + 1 + 1 + 2 + 2} = \frac{595}{7} = 85$$

So the weighted average is $85$. The newer scores matter more, so the result rises.

If you want to smooth the recent trend instead, use a $3$-week moving average for the last three weeks:

$$\frac{80 + 85 + 100}{3} = \frac{265}{3} \approx 88.3$$

This does not replace the full-course average. It answers a different question: what has recent performance looked like?

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

Common mistakes when working with averages

Using the arithmetic mean when the data already has weights

If test categories, quantities, or percentages have different importance, a plain mean can mislead. Equal weighting only makes sense when each value should contribute equally.

Averaging averages without keeping the original weights

If one class has $10$ students and another has $30$, you usually cannot average the two class averages as if they were equally large groups. You need the underlying counts or weights.

Forgetting to divide by the total weight

For a weighted average, multiplying values by weights is only part of the job. You still have to divide by $\sum w_i$.

Naming a moving average without the window length

A moving average is incomplete unless you say what window you used. A $3$-day moving average and a $30$-day moving average are not interchangeable.

When each type of average is used

Use the arithmetic mean for test scores, measurements, or other data where every observation should count equally.

Use a weighted average when the problem already assigns importance, such as grade categories or quantities sold.

Use a moving average for time-based data such as temperatures, sales, traffic, or study progress when the raw values jump around from one period to the next.

Try a similar problem

Take five numbers from your own work or 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. That quick comparison usually shows which kind of average your problem actually needs.