Exponential Smoothing Calculator
Forecast the next value in a chronological series of up to 8 data points using simple exponential smoothing with an adjustable smoothing factor.
Only the first this-many chronological periods below are used.
Higher values react faster to recent changes; lower values smooth more.
Your results
Simple exponential smoothing with α = 0.3
| Period | Actual | Smoothed |
|---|---|---|
| 1 | 105 | 105 |
| 2 | 110 | 106.5 |
| 3 | 108 | 106.95 |
| 4 | 115 | 109.36 |
| 5 | 120 | 112.56 |
| 6 | 118 | 114.19 |
| 7 | 125 | 117.43 |
| 8 | 130 | 121.2 |
Calculation breakdown
- Smoothing
- Sₜ = α × Aₜ + (1 − α) × Sₜ₋₁, with S₁ = A₁
- Forecast
- next forecast = α × latest actual + (1 − α) × latest smoothed value
Worked example
With sales of 105, 110, 108, 115, 120, 118, 125 and 130 and a smoothing factor of 0.3, the smoothed series gradually tracks the upward trend, forecasting the next period at roughly 125–126 units.
Assumptions
- Values must be entered in chronological order (oldest first).
- Simple exponential smoothing works best for data without a strong trend or seasonality; more advanced methods (Holt-Winters) handle those better.
How this calculator works
Enter up to 8 chronological values (oldest first) and a smoothing factor (alpha). The calculator applies simple exponential smoothing, weighting recent observations more heavily, and forecasts the next period's value.
Frequently asked questions
What does the smoothing factor (alpha) control?
A higher alpha makes the forecast react faster to recent changes, while a lower alpha smooths out noise but responds more slowly to trends.
Is this suitable for seasonal data?
No, simple exponential smoothing doesn't account for trend or seasonality; more advanced methods like Holt-Winters are needed for those patterns.
How is this different to a moving average?
Exponential smoothing weights all past values, giving more importance to recent ones, whereas a simple moving average weights only the most recent window equally.