Simple and Double Exponential Smoothing (SES / Holt)
Also known as: SES, Holt's linear trend method, exponential smoothing forecasting, Basit ve Çift Üstel Düzleştirme (SES / Holt)
Exponential smoothing is a family of basic time-series forecasting models in which each new observation updates a smoothed estimate by a weighting parameter. Simple exponential smoothing (SES), introduced by Robert G. Brown in 1959, forecasts series with a stable level, while Holt's double exponential smoothing, introduced by Charles C. Holt in 1957, adds a trend term using the parameters alpha and beta.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use exponential smoothing for univariate continuous time series with a reasonable run of history (at least about 10 observations). Simple exponential smoothing fits series with a roughly constant level and no trend or seasonality, while Holt's double smoothing fits series that show a linear trend. Both assume the absence of seasonality (SES) and work best on stationary or only mildly trending data; they are not appropriate when strong seasonal cycles are present.
Strengths & limitations
- Very simple and fast, requiring only one or two smoothing parameters.
- Adapts continuously to new data because recent observations receive the most weight.
- Holt's extension captures a linear trend, letting forecasts rise or fall instead of staying flat.
- Simple exponential smoothing produces flat forecasts and cannot follow a trend.
- Neither SES nor Holt handles seasonality; a seasonal series needs the Holt-Winters extension.
- Performance depends on a sensible choice of the smoothing parameters alpha and beta.
Frequently asked
What is the difference between SES and Holt's method?
Simple exponential smoothing tracks only the level and produces flat forecasts, so it suits series with no trend. Holt's double exponential smoothing adds a second equation for the trend, so its forecasts can slope upward or downward over the forecast horizon.
What do alpha and beta control?
Alpha is the smoothing weight for the level: higher values react faster to recent observations. Beta is the smoothing weight for the trend in Holt's method: it governs how quickly the estimated slope adapts. Both lie between 0 and 1.
Can exponential smoothing handle seasonality?
No. SES and Holt assume there is no seasonal pattern. When the series has recurring seasonal cycles you need the Holt-Winters seasonal extension, which adds a third smoothing equation for the seasonal component.
How much data do I need?
A short history of about ten observations is enough to get started, though longer series give more reliable parameter estimates and forecasts.
Sources
How to cite this page
ScholarGate. (2026, June 1). Simple and Double Exponential Smoothing (SES / Holt). ScholarGate. https://scholargate.app/en/econometrics/simple-exponential-smoothing
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- ARIMAEconometrics↔ compare
- State Space ModelEconometrics↔ compare
- Structural Time Series ModelEconometrics↔ compare