Published 2026-10-14
Walk-forward analysis
Walk-forward analysis re-fits a strategy on a rolling window and tests it on the next unseen period, repeating forward in time. It is closer to how a strategy is actually used than a single in-sample / out-of-sample split.
What it is
Walk-forward analysis fits the strategy on a window, tests it on the next unseen period, then rolls the window forward and repeats.
Instead of one split, you get a sequence of out-of-sample tests — each one made with only what was known at the time.
Why it beats a single split
| Approach | What it tells you |
|---|---|
| One out-of-sample split | One result, on one period |
| Walk-forward | Many results, across regimes and time |
A single split can be lucky. A walk-forward sequence is much harder to get lucky across. It also shows how the fitted parameters drift — which is often the real story.
How to check
Ask how many folds, how long each window is, and whether the parameters were re-fit at every step. A walk-forward that re-fits on all the data first is not a walk-forward. See overfitting.