All trust-facing pages read from this canonical layer to avoid narrative drift.

Readiness: Tracking started

Tracking started. A larger rolling sample window is needed before full trust metrics are stable.

Pending metrics: 7-day accuracy, MAE, False GO rate, False GO count, Independent actual coverage, Evaluated rows, Coverage-ok days (7d)

Freshness

Dataset date: 2026-04-04

Freshness status: fresh

Trust generated: 2026-04-04T22:46:24.141Z

7-day reliability

Accuracy (7d): Unavailable

False GO rate (7d): Unavailable

False GO count (7d): Unavailable

MAE: Unavailable

Coverage + calibration

Independent actual coverage: Unavailable

Evaluated rows: Unavailable

Coverage-ok days (7d): Unavailable

Truth conflict share (7d): Unavailable

Some values are intentionally withheld until the rolling sample window is large enough.