Tune the governed actuarial + ML blend, decision thresholds, and rate-table factors. Every saved change is audit-logged; the preview scorer re-scores a hypothetical quote with your unsaved values so you can see the impact before applying.
Decision thresholds and score mapping. Actuarial weight is always the complement of ML weight.
Multipliers applied to the base default rate (PD). Values >1 raise risk; <1 lower it.
Active blend version —.
Scores a hypothetical quote with your current unsaved edits.
Saved changes write the engine_knobs / factor rate tables and record an audit entry. They take effect on the next underwriting run.