Baselines · Free
Naive
Repeat the last observation. The baseline every other method must beat.
What it assumes
The next value equals the last observed value (random walk).
When to use
Short series, or as a baseline every method must beat (Forecast Value Added).
When to avoid
Strong trend or seasonality — it cannot look ahead of the last point.
Knobs
No smoothing parameter. Forecast = y_t for all horizons.
How it works
The naive forecast says the world stays where it just was: ŷ for every future period equals the last observed y. It has no trend, no season, and no smoothing. Forecast Value Added asks whether a more complicated method beats this last-value guess on a walk-forward test. If it does not, keep the naive.
Sample forecast question
Given these 8 monthly café covers, what are the naive forecasts for Sep, Oct, and Nov?
Step-by-step on these numbers
| Period | Actual | Fitted |
|---|---|---|
| Jan | 42 | 42 |
| Feb | 39 | 42 |
| Mar | 44 | 39 |
| Apr | 41 | 44 |
| May | 43 | 41 |
| Jun | 40 | 43 |
| Jul | 42 | 40 |
| Aug | 45 | 42 |
| Period | Forecast | 95% interval |
|---|---|---|
| Sep | 45 | 38.90 – 51.10 |
| Oct | 45 | 36.37 – 53.63 |
| Nov | 45 | 34.43 – 55.57 |
Parameters the engine found
last= 45
The question
Given these 8 monthly café covers, what are the naive forecasts for Sep, Oct, and Nov? Sample months: Jan=42, Feb=39, Mar=44, Apr=41, May=43, Jun=40, Jul=42, Aug=45 (covers).
Rule
Naive has no smoothing. Every future period equals the last observation: ŷ_{t+h} = y_t = 45.
Forecast
Sep = 45, Oct = 45, Nov = 45.
Interval
Residual σ = 3.114. For h = 1 the 95% band is [38.90, 51.10] around 45. The engine labels this interval native for the method.
Graph of this sample
The chart plots the canned table on this page (actuals, in-sample fitted, forecast, 95% interval). Not your Excel series. Free analysis never uploads raw data.
Educational only. Not investment, weather, or operational advice. In the add-in, rank is rolling-origin MASE — “best supported among candidates on this series.”