๐Ÿ“Š Portfolio-level dashboard. This is Walk-Forward Backtest โ€” it applies to HAL's whole book or universe-wide context, not to a single ticker. If you reached this page from a HAL 9000 memo expecting per-ticker drill-in, use one of the other dashboards (Factor Lens, Risk, Portfolio, News, Sector, Correlation, Attribution, Universe Research, Trade Idea) instead.
๐Ÿ“Š Reports โ€บ Walk-Forward Backtest

โช Walk-Forward Backtest

๐Ÿงช Practice Labs ๐Ÿ“š Manual Library ๐Ÿ“‚ Document Library ๐ŸŽฌ Video Library โ† Reports Hub ๐ŸŽฏ Portfolio ๐Ÿ›‘ Risk ๐Ÿ” Research ๐Ÿงฎ Factors โช Backtest ๐ŸŒ Regime ๐ŸŽฒ Monte Carlo ๐Ÿ“ˆ Attribution ๐Ÿ›๏ธ Sectors ๐Ÿ“ LP Letter ๐ŸŽ›๏ธ Mission Control ๐Ÿ’ก Trade Idea ๐Ÿ”ด HAL ๐Ÿ“– Manual
Pick params and click Run.
Walk-forward total
โ€”
annualized: โ€”
WF Sharpe
โ€”
Sortino: โ€”
WF Max drawdown
โ€”
peak-to-trough OOS
Win rate
โ€”
โ€” days in market
Folds
โ€”
โ€” OOS trading days
Annual vol
โ€”
annualized stdev

โš–๏ธ Honest vs Dishonest vs Dumb-Money

Three different ways to evaluate the same strategy on the same data โ€” each tells a wildly different story.
Run a backtest to populate.

๐Ÿ“ˆ Stitched Out-of-Sample Equity Curve

Strategy's stitched OOS equity (pure walk-forward, no curve-fit) vs SPY buy-and-hold over the same OOS window. Both normalized to start at $1.

๐ŸชŸ Per-Fold Walk-Forward Performance

Each fold = one 12-month train + 3-month blind test. The strategy's parameters are "fit" on the train window (deterministic in this build), then evaluated on the blind test window. Stitched OOS = concatenation of all test windows.
Fold Train range Test range Test bars Trades OOS Return OOS Sharpe OOS Max DD
Run a backtest to populate.