Why is there no forecast error calculated for the first week in exponential smoothing?

Conditions

• The initial forecast is set to the first actual value ($F_1 = A_1$).
• You are calculating forecast errors to evaluate model accuracy.

Reasoning, step by step

1. Recall the initial forecast assumption: $F_1 = A_1$.
2. Calculate the error for the first period: $E_1 = A_1 - F_1 = A_1 - A_1 = 0$.
3. Recognize that a zero error does not provide useful information about the model's predictive accuracy.
4. Begin calculating errors from the second period onwards.

Example

The video explicitly states that when computing errors for exponential smoothing forecasts, we do not calculate an error for period one unless otherwise stated, and draws a red 'X' over the Error cell for Week 1.

Common misconceptions

• Calculating an error for the first period and including it in the Mean Squared Error (MSE) calculation.
• Believing that the first period's forecast is independent of the first period's actual value.

Watch the explanation

Explore next

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