Statistics & Probability

Verified calculator with a transparent formula

Statistics & Probability

Exponential Smoothing Calculator.

Calculates the smoothed value for the next period using simple exponential smoothing, given the most recent actual value, the previous smoothed value, and a smoothing constant alpha.

Results update live as you type.
ƒ

Your inputs

How it works

  1. 1

    Enter the most recent actual observation.

  2. 2

    Enter the previous smoothed value (or initial forecast).

  3. 3

    Choose the smoothing constant alpha between 0 and 1.

  4. 4

    The calculator computes the new smoothed value as alpha times actual plus (1 minus alpha) times previous smoothed.

alpha * actual + (1 - alpha) * previous_smoothed

Frequently asked questions

What does alpha represent?

Alpha is the smoothing constant. Higher alpha gives more weight to the most recent actual value, making the forecast more responsive to changes. Lower alpha makes the forecast smoother and less reactive.

How do I choose alpha?

Common practice is to try values between 0.1 and 0.3 for stable series, and higher values (0.5-0.9) for series with sudden changes. You can also optimize alpha by minimizing forecast errors on historical data.

What if I don't have a previous smoothed value?

If no previous smoothed value exists, you can use the first actual observation as the initial smoothed value, or the average of the first few observations.

Explore this calculator category

Statistics & ProbabilityBrowse all

Results

Formula checked

Next smoothed value

0

Ask Anything

Estimate for general guidance only — verify important decisions with an appropriate professional.

How it works

Calculates the smoothed value for the next period using simple exponential smoothing, given the most recent actual value, the previous smoothed value, and a smoothing constant alpha.

  1. Enter the most recent actual observation.
  2. Enter the previous smoothed value (or initial forecast).
  3. Choose the smoothing constant alpha between 0 and 1.
  4. The calculator computes the new smoothed value as alpha times actual plus (1 minus alpha) times previous smoothed.

Formulas

The math behind this calculator, written out so you can verify the result.

Simple Exponential Smoothing

S_t = α * X_t + (1 - α) * S_{t-1}

The new smoothed value S_t is a weighted average of the current actual value X_t and the previous smoothed value S_{t-1}.

Example:

Input: X_t = 120, S_{t-1} = 110, α = 0.3

Calculation: 0.3*120 + 0.7*110 = 36 + 77 = 113

Result: S_t = 113

Real-world use cases

Where this calculation shows up in everyday life.

Demand Forecasting

Smooth sales or demand data to generate short-term forecasts for inventory management.

Example: Forecast next month's sales based on this month's actual and previous forecast.

Trend Analysis

Remove random fluctuations from time series data to identify underlying trends.

Example: Smooth weekly website traffic to see the overall direction.

Quality Control

Monitor process measurements to detect shifts while filtering out noise.

Example: Track the smoothed average of product weights to ensure consistency.

Tips and common mistakes

Tips

  • Start with alpha = 0.2 and adjust based on how quickly you want the forecast to react.
  • Use a holdout sample to test different alpha values and pick the one with lowest error.
  • For very noisy data, use a smaller alpha to avoid overreacting to random fluctuations.
  • If the series has a trend, consider double exponential smoothing instead.

Common Mistakes to Avoid

  • Using alpha > 1 or < 0; alpha must be between 0 and 1.
  • Forgetting to update the smoothed value each period; always use the latest smoothed value as the previous one.
  • Assuming exponential smoothing works well for data with strong seasonality without adjustments.

Assumptions and limitations

  • Use the stated inputs and units.
  • Results are estimates for planning and education.
  • Check measurements and source data before making an important decision.