Skip to content

Strategic Cost Management · Business Forecasting Models - Time Series and Regression Analysis

Trend Measurement Methods: Semi-Average, Moving Average, Least Squares

Updated 11 October 2026 · Fact-checked

Trend measurement finds the long-term direction of a time series by removing short-term ups and downs. You can use the freehand, semi-average, moving average or least squares method. Least squares gives the best-fit line Y = a + bX, which you extend to forecast future values by putting the next X value in the equation.

Understand Trend Measurement Methods

A time series is a set of values recorded over time, such as yearly sales or monthly output. The trend is its long-term movement: steady growth, steady decline, or no clear direction. Trend ignores seasonal swings, cycles and random noise.

You measure trend to see where the business is heading and to forecast. A cost manager uses it to project sales, demand or cost levels for budgets and pricing.

There are four methods. Freehand means plotting the data and drawing a smooth line by eye. It is quick but subjective. Semi-average splits the data into two equal halves, finds the average of each, and joins the two points. Moving average replaces each value with the average of a fixed number of nearby values, which smooths out fluctuations. Least squares fits the straight line that makes the sum of squared gaps between actual and trend values as small as possible.

Least squares is the most exact and is the one examiners prefer for forecasting. Semi-average and moving average are simple and are often asked as short numericals. Moving average does not give an equation, so it cannot forecast by itself. It also loses values at both ends of the series.

Once you have a trend line, forecasting is simple: put the future period into the equation. Treat the result as an estimate. It assumes the past pattern continues, and it becomes less reliable the further you go beyond the data.

Key rules to remember

Straight-line trend
Y = a + bX
a is the trend value at X = 0. b is the change in Y for a one-unit change in X.
Normal equations (least squares)
ΣY = na + bΣX ; ΣXY = aΣX + bΣX²
Use these when X is not coded from the middle. Solve the two equations for a and b.
Least squares with deviations from the middle (ΣX = 0)
a = ΣY ÷ n ; b = ΣXY ÷ ΣX²
Take X = 0 at the middle period. For odd n, X runs ..., -2, -1, 0, 1, 2, .... This is the fastest form.
Coding for an even number of years
X = 2 × (year − midpoint of the two middle years), where the midpoint is the half-year point between the two middle years
Find the midpoint by averaging the two middle years. For 2020 to 2025, the middle years are 2022 and 2023, so the midpoint is 2022.5 and X = 2 × (year − 2022.5). X values become ..., -3, -1, 1, 3, ... in half-year units. Then b is the change per half year, and the yearly change is 2b.
Semi-average slope
b = (Average of second half − Average of first half) ÷ (Years between the two mid-points)
Each average is placed at the middle of its own half. If n is odd, leave out the middle value.
Moving average of odd period m
MA = (sum of m consecutive values) ÷ m, placed against the middle period
For a 3-year average, the first value is placed against year 2.
Moving average of even period
Take the m-period totals, then average each pair of consecutive totals to centre it on a year
A 4-year moving average falls between years, so you centre it with a second 2-term average.

How to solve Trend Measurement Methods questions

Start by reading which method the question names. If it names none, use least squares for forecasting and moving average for smoothing.

  1. 1Write the data in a table with period and actual value, and note n.
  2. 2Identify the method asked: freehand, semi-average, moving average or least squares.
  3. 3For semi-average, split the data into two equal halves (drop the middle value if n is odd), find each average, and fix the mid-point year of each half.
  4. 4For moving average, add m values at a time, divide by m, and place each result at the middle period. For an even period, centre the result with a second 2-term average.
  5. 5For least squares, code X from the middle period, then find ΣY, ΣX, ΣXY and ΣX². Compute a = ΣY ÷ n and b = ΣXY ÷ ΣX² when ΣX = 0.
  6. 6Write the trend equation Y = a + bX and state clearly what X = 0 and the unit of X mean.
  7. 7To forecast, convert the future year into its X value, put it in the equation, and state the answer with units.
  8. 8Add one line of comment: the forecast assumes the past trend continues.

Quickest way: Code from the middle and use two sums

When to use it: Use this for any least squares question with equally spaced periods.

  1. Set X = 0 at the middle year (odd n) or use X = -3, -1, 1, 3 style codes (even n).
  2. Check that ΣX = 0 before you go on. This catches coding slips.
  3. Find ΣY and ΣXY only. Get ΣX² from the X column, or use the known total: 10 for X from -2 to 2, 28 for X from -3 to 3 (odd n). For even-n half-year codes, ΣX² = 20 for -3, -1, 1, 3 and 70 for -5, -3, -1, 1, 3, 5.
  4. Take a = ΣY ÷ n and b = ΣXY ÷ ΣX². No simultaneous equations are needed.
  5. For a forecast, count how many periods the target year is from the middle year, and put that X in the equation. For even n with half-year codes, count in half years.

Common mistakes in Trend Measurement Methods

  • Putting the moving average against the first or last period instead of the middle period.

    Students write the answer next to the last value they added.

    Fix: Always place the average at the centre of the group. A 3-year average of years 1 to 3 belongs against year 2.

  • Not centring a 4-year (even-period) moving average.

    The result falls between two years, and students ignore that.

    Fix: Average each pair of consecutive 4-year totals (or averages) so that the result sits on an actual year.

  • Using X = -2, -1, 0, 1 for an even number of years.

    Students copy the odd-n pattern, so ΣX is not zero and the shortcut formulas become wrong.

    Fix: For even n, use X = -3, -1, 1, 3 (half-year units) and remember that b then measures change per half year.

  • Forecasting with the wrong X value.

    Students count from the first year instead of from the year where X = 0.

    Fix: Write the X value of the last year first, then add one for each further year. For half-year codes, add two for each further year.

  • Taking the semi-average slope over the wrong gap.

    Students divide the difference of averages by the number of values in a half instead of the distance between the mid-points.

    Fix: Find the middle year of each half, subtract them, and divide the difference of the averages by that gap.

  • Forgetting to drop the middle value in semi-average when n is odd.

    Students try to make unequal halves.

    Fix: Leave out the middle value so both halves have the same number of years.

Worked examples

Example 1

The sales of a company (₹ in lakh) for 2020 to 2025 were: 2020: 40; 2021: 44; 2022: 48; 2023: 54; 2024: 59; 2025: 64. Fit a straight-line trend by the semi-average method and forecast sales for 2027.

Show the solution
  1. n = 6, which is even, so each half has 3 years. First half: 2020 to 2022. Second half: 2023 to 2025.
  2. Average of first half = (40 + 44 + 48) ÷ 3 = 132 ÷ 3 = 44. This is placed at 2021.
  3. Average of second half = (54 + 59 + 64) ÷ 3 = 177 ÷ 3 = 59. This is placed at 2024.
  4. Gap between the mid-points = 2024 − 2021 = 3 years.
  5. Annual slope = (59 − 44) ÷ 3 = ₹5 lakh per year.
  6. Trend equation: Y = 44 + 5X, where X = years from 2021 and Y is in ₹ lakh.
  7. For 2027, X = 2027 − 2021 = 6. Y = 44 + 5 × 6 = 74.

Answer: Trend line: Y = 44 + 5X (origin 2021, X in years, Y in ₹ lakh). The forecast sales for 2027 are ₹74 lakh, assuming the trend continues.

Example 2

Production of a plant (in '000 units) for 2021 to 2025 was: 20, 24, 29, 33, 39. (a) Find the 3-year moving averages. (b) Fit a straight-line trend by least squares and forecast production for 2028.

Show the solution
  1. (a) 3-year totals: 20 + 24 + 29 = 73; 24 + 29 + 33 = 86; 29 + 33 + 39 = 101.
  2. Divide by 3 and place at the middle year: 2022: 73 ÷ 3 = 24.33; 2023: 86 ÷ 3 = 28.67; 2024: 101 ÷ 3 = 33.67. No value exists for 2021 or 2025.
  3. (b) n = 5. Take X = 0 at 2023, so X = -2, -1, 0, 1, 2 and ΣX = 0.
  4. ΣY = 20 + 24 + 29 + 33 + 39 = 145.
  5. ΣXY = (-2 × 20) + (-1 × 24) + (0 × 29) + (1 × 33) + (2 × 39) = -40 − 24 + 0 + 33 + 78 = 47.
  6. ΣX² = 4 + 1 + 0 + 1 + 4 = 10.
  7. a = 145 ÷ 5 = 29. b = 47 ÷ 10 = 4.7.
  8. Trend equation: Y = 29 + 4.7X, with origin 2023 and X in years.
  9. For 2028, X = 5. Y = 29 + 4.7 × 5 = 29 + 23.5 = 52.5.

Answer: 3-year moving averages: 24.33 (2022), 28.67 (2023), 33.67 (2024). Least squares trend: Y = 29 + 4.7X (origin 2023). Forecast production for 2028 is 52.5 thousand units, that is 52,500 units.

Exam tips

  • Read the question for the method named. Marks usually go for the correct method, so do not switch to least squares when the question asks for semi-average.
  • Show the table with ΣX, ΣY, ΣXY and ΣX². Even if you slip on arithmetic, method marks are protected.
  • Always state the origin and unit of X (for example, origin 2023, X in years). Examiners look for this line.
  • In MCQs, check ΣX = 0 and test the answer. When X = 0 at the middle year, the trend value at X = 0 (a) must equal the mean of Y (ΣY ÷ n).
  • End forecast answers with one line that the projection assumes the past trend continues, and that cyclical or seasonal effects are not included.

Practice questions from Business Forecasting Models - Time Series and Regression Analysis

Trend Measurement Methods in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

Trend Measurement Methods: frequently asked questions

Which trend method is the most accurate?

Least squares is the most accurate of the four because it minimises the sum of squared deviations from the line. It also gives an equation you can use to forecast. Freehand is the least reliable because it depends on personal judgement.

How do I calculate a moving average trend?

Choose the period, such as 3 years. Add that many consecutive values, divide by the period, and place the result against the middle year. Then slide forward by one period and repeat. For an even period, centre the result by averaging two consecutive averages.

What is the semi-average method with an example?

You split the series into two equal halves and find the average of each. You place each average at the middle of its half and join the two points. In the example above, the averages are 44 and 59, placed at 2021 and 2024, which gives a slope of 5 per year.

Why do we take X = 0 at the middle year in least squares?

It makes ΣX equal to zero. The normal equations then become a = ΣY ÷ n and b = ΣXY ÷ ΣX², so you need no simultaneous equations. This saves time and reduces the chance of errors.

Can I use a trend line to forecast many years ahead?

You can, but the further you go from the data, the less reliable the figure is. The method assumes the same pattern continues. Use it for short-term forecasts and say so in your answer.