A moving average smooths out the ups and downs in time series data so you can spot the underlying trend. At GCSE you most often use a 3-point or 4-point moving average: add three or four consecutive values and divide by 3 or 4, then repeat, moving along by one.
Why use a moving average?
Raw time series data often shows seasonal variation — regular rises and falls that repeat on a cycle. For example, a shop's weekly sales might peak every Saturday and dip every Monday. This regular pattern makes it hard to see whether sales are genuinely increasing over time.
A moving average irons out those regular fluctuations so that the underlying trend — the long-term direction — becomes clearer. If the trend line slopes upward, sales are growing overall, even if individual days vary greatly.
How do you calculate a 3-point moving average?
A 3-point moving average averages three consecutive values. You start with values 1, 2, 3; then move along to values 2, 3, 4; then 3, 4, 5; and so on.
Worked example: A market stall records the following daily sales (£) over seven days:
| Day | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
|---|---|---|---|---|---|---|---|
| Sales (£) | 120 | 95 | 140 | 110 | 160 | 130 | 175 |
Calculate the 3-point moving averages.
| Average | Calculation | 3-point MA |
|---|---|---|
| Days 1–3 | (120 + 95 + 140) / 3 | 118.33 |
| Days 2–4 | (95 + 140 + 110) / 3 | 115.00 |
| Days 3–5 | (140 + 110 + 160) / 3 | 136.67 |
| Days 4–6 | (110 + 160 + 130) / 3 | 133.33 |
| Days 5–7 | (160 + 130 + 175) / 3 | 155.00 |
Notice: there are n − 2 = 5 moving averages for n = 7 data points.
Where do you plot each moving average?
Each moving average is plotted at the middle time point of the group it was calculated from.
- Days 1–3: plot at Day 2
- Days 2–4: plot at Day 3
- Days 3–5: plot at Day 4
- Days 4–6: plot at Day 5
- Days 5–7: plot at Day 6
Joining these plotted points gives the trend line (a smoother version of the original data).
How does a 4-point moving average work?
A 4-point moving average is used when data repeats over a cycle of 4 (for example, quarterly data: spring, summer, autumn, winter). Each average covers exactly one full cycle, so seasonal highs and lows cancel each other out perfectly.
Because 4 is even, the plotted position falls between two time points. To fix this, GCSE courses sometimes use a centred moving average: average pairs of consecutive 4-point moving averages, which lands on a whole time point.
Example: Quarterly sales data (£ thousands):
| Quarter | Q1 | Q2 | Q3 | Q4 | Q1 | Q2 | Q3 | Q4 |
|---|---|---|---|---|---|---|---|---|
| Year 1 | 30 | 45 | 55 | 25 | 35 | 50 | 60 | 30 |
First 4-point average: (30 + 45 + 55 + 25) / 4 = 155 / 4 = 38.75 (plots between Q2 and Q3 of Year 1)
Second 4-point average: (45 + 55 + 25 + 35) / 4 = 160 / 4 = 40.00 (plots between Q3 and Q4 of Year 1)
Centred moving average = (38.75 + 40.00) / 2 = 39.375 (plots at Q3 of Year 1)
In many GCSE courses, 4-point moving averages are plotted between time points and the centring step is not required — check your specification and the question's instructions.
What do you do with the trend line?
Once you have plotted the moving averages and drawn a trend line (usually a straight line by eye through the plotted points), you can:
- Describe the trend: "Sales show an overall upward trend of approximately £12 per day."
- Estimate a seasonal effect: subtract the trend value from the actual value at the same time point. Consistent positive or negative differences reveal the seasonal pattern.
- Predict future values: extend the trend line and read off a predicted value. Note that predictions are less reliable further into the future.
Common errors in exam questions
| Error | How to avoid |
|---|---|
| Plotting at the first or last time point instead of the middle | Always plot at the midpoint of the group |
| Dividing by the wrong number | Count how many values are in the average (3 or 4, not all of them) |
| Not enough moving averages | For n data points, a 3-point MA gives n−2 values; 4-point gives n−3 |
| Confusing trend with seasonal variation | Trend = long-term direction; seasonal variation = repeating short-term fluctuation |
Frequently asked questions
Why is the moving average plotted in the middle rather than at the end?
Plotting at the end would shift the trend line forward in time, making it look like changes happened later than they did. The middle time point is the natural balance point of the group, giving the most accurate representation of the trend at that moment.
How many moving averages can I calculate from a data set?
For a k-point moving average on n data values: you get n − k + 1 moving averages. So a 3-point MA from 7 values gives 7 − 3 + 1 = 5 averages; a 4-point MA from 8 values gives 8 − 4 + 1 = 5 averages.
Should the trend line be a straight line or a curve?
At GCSE, the trend line is almost always drawn as a straight line by eye through the plotted moving averages, using a ruler. Do not join all the individual moving average points — draw one best-fit straight line through them, similar to a line of best fit on a scatter graph.
Can I use a moving average to predict the next value?
Yes, but with caution. Extend your trend line to the next time period. If you also know the typical seasonal effect for that period, add it to the trend estimate. The further you extrapolate, the less reliable the prediction. Exam questions that ask for a prediction should note this limitation.
Work through moving average and time series questions with Professor Pi's step-by-step hints at aitutors.me.