Walk-Forward Results
Walk-forward analysis is the prediction checkpoint. Each row trains on an older window and evaluates on the later unseen window. A strategy must pass multiple unseen windows before its Day 30 forecast is treated as more than a lucky historical fit.
Forecast Lab
Select one or more models to compare their forward prediction paths from today. Actual rows are flat on Day
Forward estimate
Projected vs realized value by selected model
Dashed lines are estimated future paths. Solid actual points appear as the experiment advances and snapshots are recorded.
Capital sensitivity
Projected profit by initial investment
This tests the assumption that larger initial investment can reduce minimum-trade drag and improve implementation efficiency.