Methodology / BRI 1.0
Craft gives it form. Behaviour proves it in use. Memory asks what survives.
One score / three independent lenses
The Brand Recall Index is a geometric score from 0 to 99. A weak lens pulls the whole read down. Every edition records its inputs, evidence, method and reason.
C.5 × B.3 × M.2
Try the index / illustrative controls
Change a lens. The index recomputes with the same versioned scorer used by the engine.
BRI / demonstration / not a published edition
When a signal is unavailable, its weight stays with the editor and is shown as carried. Public Read needs 100 eligible accounts; council reads remain independent and blind.
Forecast → observed
Behaviour begins as a Forecast. Each eligible observation replaces some prediction: w = 5 / (5 + n).
Forecast 76 / observed 82
100% Forecast / 0% observed
Observed evidence has an 18-month half-life. New evidence earns its place; old evidence remains in the record.
A brilliant drawing cannot erase weak behaviour or forgettable memory. Each lens affects the whole.
The current count-based width is a heuristic, not a calibrated empirical interval. Bootstrap confidence must name its evidence and method.
Evidence can arrive every day. The public score moves through a signed edition, normally capped at four points unless a material change is recorded.
Historical BMI editions stay in the Trail with their original method. They are never silently relabelled BRI.
A separate mean of individual geometric ratings from at least 100 distinct eligible accounts. Affiliated and held submissions do not count.
