← ML Lab  ·  Dashboard

OmmAlpha · ML Phase 1B

Historical RS reconstruction and market-context coverage audit. Original backtest signals are not modified.

Production dataset scope: BACKTEST Run #1. Readiness and coverage figures on this page are scoped to that official run. The historical RS cache itself is reusable by stock/date, so its background processing queue may contain rows useful to other experiments too.

Signal window

2024-10-09First BACKTEST signal
2026-09-25Last BACKTEST signal
4,662Historical signals
480Distinct signal dates

1 · Historical momentum calculation

498 / 498Stocks completed
227,037Historical momentum rows
0Stock errors
100.0%

2 · Cross-sectional RS ranking

484 / 484Signal dates ranked
227,037Ranked stock-date rows
4,662Signals with reconstructed RS
100.0%Signal RS coverage
100.0%
Membership provenance: historical RS is calculated from prices as they existed on each date, but the cross-sectional membership uses the current OmmAlpha NIFTY 500 survivor universe. The exporter labels these rows BACKFILLED_CURRENT_NIFTY500 rather than pretending historical constituent membership is exact.

3 · Breadth coverage

2024-08-23First NIFTY500 breadth date
2026-09-30Last NIFTY500 breadth date
523Breadth sessions stored
YESCovers ML signal window
Breadth covers the complete BACKTEST signal window. Exported breadth rows retain their membership_approximate quality flag.

Processing controls

Current stage: COMPLETE