How do you calculate the next forecast in exponential smoothing with a smoothing constant of 0.2?

Conditions

• The smoothing constant is $\alpha = 0.2$.
• You know the actual sales value for the current period ($A_t$).
• You know the forecast value for the current period ($F_t$).

Reasoning, step by step

1. Identify the actual sales value for the current period ($A_t$) and the forecast value for the current period ($F_t$).
2. Multiply the actual sales value by the smoothing constant: $0.2 \times A_t$.
3. Multiply the forecast value by the complement of the smoothing constant: $0.8 \times F_t$.
4. Add the two products together to find the forecast for the next period ($F_{t+1}$).

Example

To calculate the forecast for week 5 ($F_5$), the video uses the actual sales for week 4 ($A_4 = 45$) and the forecast for week 4 ($F_4 = 40.00$). The calculation is $F_5 = 0.2(45) + 0.8(40.00) = 9 + 32 = 41.00$.

Common misconceptions

• Multiplying the actual sales by 0.8 and the forecast by 0.2 instead of the correct weights.
• Forgetting to multiply the forecast value by 0.8 and simply adding it to the weighted actual sales.

Watch the explanation

Explore next

• What is the initial forecast assumption if no prior forecast is given?
• How do you calculate the Mean Squared Error (MSE) for the forecasts?