A smoothing constant is a variable used in time series analysis based on exponential smoothing. This constant determines how the historical time series values are weighted. The higher the smoothing constant, the greater weight assigned to the values from the latest period and as a consequence, the greater possibility for quick reaction to systematic changes in the time series. The smoothing constant must have a value between 0 and 1.

The three smoothing constants are: ( (alpha), ( (beta), and ( (gamma). These are used in exponential smoothing to calculate the following:

**( (alpha)**

- Base forecasts
- Seasonal index as each periodâ€™s share of annual demand
- Forecast MAD
- Mean forecast error
- Trend dampening.

**( (beta)**

- Trend quantities.

**( (gamma)**

- New seasonal indexes by weighting old ones.

When using forecast methods where the base forecast is calculated using adaptive exponential smoothing, a new smoothing constant is automatically calculated for each period.

This example considers using the smoothing constant in forecast calculation using exponential smoothing.

Assume the September forecast for an item was 122, and actual demand was 130. The ( smoothing constant for the item is 0.2. The October forecast is then calculated as follows:

0.2 x 130 + (1 - 0.2) x 122 = 26 + 98 = 124