Inputs
Each assessment window submits oversight telemetry: escalation events, review timestamps, reviewer identities and roles, override actions, and the decisions they applied to.
Telemetry feeds the Judgment Independence Score. It is personal data, processed under GDPR Chapter II, and reduced to dimension scores before storage.
Per dimension scoring
Each of the seven dimensions is scored 0 to 100 from its inputs. Scoring functions are deterministic: the same telemetry produces the same dimension score, which is what makes a receipt verifiable.
| Code | Dimension | Weight |
|---|---|---|
| D1 | Escalation integrity | 18.0 |
| D2 | Review cadence | 16.0 |
| D3 | Decision authority | 15.0 |
| D4 | Reviewer competence | 14.0 |
| D5 | Independence | 13.0 |
| D6 | Auditability | 13.0 |
| D7 | Override capability | 11.0 |
Composite
The composite is the weighted sum of the seven dimension scores, scaled to the published 0 to 100 range. Weights are fixed within a major version so composites are comparable across assessments.
Calibration
Thresholds are calibrated against the case library: nineteen documented oversight incidents scored under the same functions. Calibration sets the meaningful-human-review floor on each dimension so that known failures fall below it.
Verdict logic
A pass requires the composite to clear the floor and every dimension to clear its own threshold. A composite above the floor with one dimension below returns a conditional verdict, not a pass.