WALK-FORWARD VALIDATION
Find out if a strategy actually held up out of sample.
BestFolio re-runs every strategy on rolling out-of-sample windows the model never saw at training time. The only honest way to ask whether a tactical strategy would have worked, not whether it was tuned to look like it would.
Already a Pro subscriber? Open the live tool →

BestFolio walk-forward optimization tool: select strategies, set lookback window and ranking criterion, run rolling out-of-sample portfolio analysis.
- Rolling out-of-sample
- 5-year in-sample / 1-year out-of-sample windows rolled forward across 30+ years.
- Per-strategy and per-portfolio
- Run walkforward on a single strategy, or on a blended portfolio of 2 to 20 strategies.
- Configurable splits
- Adjust window length and split ratio for sensitivity analysis. Defaults sit at conservative settings.
What it is
Walk-forward validation splits historical data into a training window and a separate testing window. The strategy parameters are fit on the training window, then evaluated only on the never-seen testing window. The window rolls forward and the process repeats. The result is a series of out-of-sample returns, drawdowns, and Sharpe ratios that show what would have happened if you had run the strategy in real time, not what would have happened if you had tuned it perfectly to the past.
Free vs Pro
Free
- ✓Walk-forward results for the 6 free strategies
- ✓Equity curve and CAGR / max drawdown / Sharpe per window
- ✓Default 5y / 1y split
Pro
- ✓Walk-forward across all 47+ Pro strategies
- ✓Walk-forward across blended portfolios (2 to 20 strategies)
- ✓Configurable window and split ratios
- ✓Per-strategy contribution analysis on portfolio walkforward
- ✓Worst-case window highlight (sequence-risk view)
Strategies that use this
HAA (Hybrid Asset Allocation)
Keller hybrid; recession-aware tactical.
View strategy →
GEM (Global Equities Momentum)
Antonacci dual momentum; absolute + relative.
View strategy →
BAA (Bold Asset Allocation)
Keller canary-based regime detection.
View strategy →
VAA (Vigilant Asset Allocation)
Keller offensive / defensive switching.
View strategy →
DAA (Defensive Asset Allocation)
Keller defensive variant of VAA.
View strategy →
ADM (Accelerated Dual Momentum)
Antonacci accelerated; faster signal.
View strategy →
Frequently asked questions
What is walk-forward validation in plain English?+
It's a way to check whether a strategy's backtest is real or a coincidence. You take historical data, hide a chunk of it, train the strategy on what's visible, then test it on the hidden chunk. If the strategy still works on data it never saw, the backtest is more believable. If it breaks on hidden data, the original backtest was probably a curve fit.
How does walk-forward differ from a regular backtest?+
A regular backtest runs the strategy across all available history and reports the result. A walk-forward test repeats that exercise on a rolling window where the testing portion was never seen by the strategy parameters. Regular backtests can be optimistic if the strategy was tuned to fit the data. Walk-forward removes most of that bias.
Does walk-forward eliminate overfitting?+
It doesn't eliminate it. Strategies can still be over-fit on a regime that lasts longer than the in-sample window, or the asset universe itself can be cherry-picked. Walk-forward is the strongest single defense available, but not a guarantee.
Which BestFolio strategies have walk-forward results?+
All published strategies have walk-forward results computed on the same engine. Free-tier strategies include HAA, GEM, Permanent Portfolio, Classic 60/40, Golden Butterfly, and RP Gold+SCV. Pro strategies cover the rest of the catalog.
Is walk-forward a free or Pro feature?+
Walk-forward results on the 6 free strategies are free. Walk-forward across the full catalog and across blended portfolios is part of Pro.
Start free, upgrade when you need it
6 strategies free forever. Upgrade to Pro for the full library, walk-forward, blending, and monthly signals.
30-day money-back guarantee on all paid plans.