SCORE
Scoring Engines
1. Overview
Scoring Engines (SCORE) are a capability class of Micro Engines (MICE) responsible for producing composite indicators, ratings, indices, or maturity scores from one or more input signals.
SCORE engines:
- combine multiple metrics into a single evaluative output,
- apply explicit weighting, thresholds, or models,
- express interpretation, not raw measurement.
SCORE engines answer: “How does this perform, relative to a model?”
They do not claim factual truth — they express structured judgment.
SCORE is a type, not an identity.
Concrete implementations are identified by MEID, versioned via ZAR, and leave CMI lineage stamps at runtime.
2. Position in the MICE Model
Capability, not identity
A Scoring Engine:
- does not compute base metrics (CALC),
- does not aggregate raw structures (AGGR),
- does synthesize meaning across signals.
Multiple MEIDs may implement SCORE capabilities, across:
- different domains (climate, governance, operations),
- different scoring models (internal, regulatory, partner),
- different disclosure or decision contexts.
Typical placement
| Dimension | Value |
|---|---|
| Capability type | SCORE |
| Common tiers | Tier-2, Tier-3 |
| Typical classification | Contract-Engine |
| Versioning | ZAR (CMI-level) |
| Lineage impact | Appends scoring CMI stamp to USO |
3. Design Principles
-
Explicit Scoring Model
All weights, thresholds, and aggregation logic must be declared and versioned. -
Interpretive Transparency
A score must be explainable in terms of its contributing inputs. -
Deterministic Evaluation
Given the same inputs and model version, outputs must be reproducible. -
Non-Authoritative by Default
Scores are evaluations, not certified facts unless explicitly assured. -
Auditability
Every score references the exact scoring model and input set used.
4. Scope of Responsibility
What SCORE engines do
- Combine multiple signals into a composite indicator
- Apply weights, bands, or scoring functions
- Produce ratings, indices, or maturity levels
- Express relative performance or alignment
Typical examples:
- climate transition readiness score
- ESG maturity index
- governance compliance rating
- supplier risk score
- internal benchmark or percentile ranking
5. What SCORE Engines Do Not Do
SCORE engines intentionally do not:
- ❌ compute primary values (CALC)
- ❌ project future scenarios (SCEN)
- ❌ validate data correctness (VALI)
- ❌ certify trust or compliance (ASSURANCE / AAE)
- ❌ enforce policy decisions (ZSSR)
A SCORE engine evaluates —
it does not decide or certify.
6. Inputs
Typical SCORE inputs include:
- validated metrics or aggregates,
- optional scenario-adjusted values (from SCEN),
- scoring model definitions (weights, thresholds),
- contextual parameters (peer group, industry, region).
All scoring models must be:
- explicit,
- versioned,
- referenceable.
7. Outputs
SCORE engines emit:
- a composite score or rating,
- optional sub-scores or component breakdowns,
- explicit references to the scoring model,
- lineage metadata linking all contributing inputs.
Outputs are commonly consumed by:
- dashboards and analytics,
- reporting and disclosure layers,
- RMAP engines (risk interpretation),
- assurance workflows (optional certification).
8. Interaction with ZSSR
ZSSR does not route to “SCORE” generically.
Instead, it:
- selects specific MEIDs implementing scoring logic,
- binds them to an approved scoring model,
- enforces policy constraints (e.g. “no external ratings allowed”).
Tags such as score, rating, or maturity
may be used as fallback selectors, but explicit rules always dominate.
9. Example Scoring Engines
10. Summary
Scoring Engines:
- synthesize multiple signals into interpretable indicators,
- make evaluative logic explicit and auditable,
- support benchmarking, ratings, and decision support,
- remain separate from assurance and governance by design.
They are the interpretive synthesis layer of the ZAYAZ platform.
Stable