What is the exponential smoothing forecast formula used in this video?

Conditions

• The formula is used to compute the forecast for period $t+1$.
• The same smoothing constant $\alpha$ is used throughout the worked example.
• The initial forecast $F_1$ must be specified before recursion can begin.

Reasoning, step by step

1. Identify the current period's actual demand, denoted as $A_t$.
2. Identify the current period's forecast, denoted as $F_t$.
3. Multiply the actual demand $A_t$ by the smoothing constant $\alpha$.
4. Multiply the forecast $F_t$ by the complementary weight $(1-\alpha)$.
5. Sum the two products to obtain the forecast for the next period, $F_{t+1}$.

Example

In the worked example, the formula is applied as $F_{t+1} = 0.4 A_t + 0.6 F_t$. For Year 3, the instructor substitutes the Year 2 actual demand ($A_2=365$) and the Year 2 forecast ($F_2=310$) into the formula: $F_3 = 0.4(365) + 0.6(310) = 332$.

Common misconceptions

• Assuming the formula calculates the forecast for the current period rather than the next period.
• Believing that the weights $\alpha$ and $(1-\alpha)$ change from period to period in simple exponential smoothing.

Watch the explanation

Explore next

• How do you initialize the first forecast before applying exponential smoothing?
• What do alpha and 1 minus alpha represent in this forecasting method?