Forecasting is the process of using historical data to estimate future values. Common methods include moving averages and exponential smoothing.

When should I use moving average vs exponential smoothing?

Moving averages are simple and good for stable data without strong trend or seasonality. Exponential smoothing reacts faster to recent changes by giving more weight to recent observations.

Can I forecast more than one period ahead?

This tool produces the next-period forecast for both methods. Then it propagates the same level forward for additional horizons.

Forecasting Calculator – Moving Average & Exponential Smoothing

Forecasting Calculator (Moving Average & Exponential Smoothing)

Paste your historical time series (one value per line, oldest to newest) and compute forecasts using simple moving average or single exponential smoothing. The tool also shows error metrics (MAD, MAPE) for the in-sample fit.

Time series (one value per line)

Window (m)

number of periods to average

Forecast periods

Blue: actual, Orange: forecast (from first available point). Only first 24 points shown.

Next forecast

—

MAD

—

MAPE (%)

—

Method

—

t	Actual	Forecast	Error	\|Error\|	\|PE\| %

How the calculator works

1. Simple Moving Average

\( \hat{Y}_{t+1} = \dfrac{Y_t + Y_{t-1} + \dots + Y_{t-m+1}}{m} \)

where m is the window size.

Use this when your data is fairly stable. A larger window means smoother forecasts but slower reaction to changes.

2. Single Exponential Smoothing

\( F_t = \alpha Y_{t-1} + (1-\alpha) F_{t-1} \)

with starting \( F_1 = Y_1 \) (simple initialization).

Use this when you want the most recent data to matter more. α close to 1 makes the forecast very reactive.

Error metrics

MAD (Mean Absolute Deviation): average of |error|.
MAPE: average of |error|/actual, as percentage.