BRIDGEPOINT INTEGRITY INDEX · BII · SHADOW CALIBRATION

A score that is allowed to say “not enough evidence.”

The BridgePoint Integrity Index is being built as a transparent 0–100 property integrity/resilience context index. Higher is intended to mean stronger evidence-supported integrity context. It is not a structural-survival probability, insurance rate, lending score, investment rating, coverage determination or claim outcome.

Model statusConnectingnumeric publication is gated
Health assessments—current evidence assessments
Publication-grade BII—must pass strict evidence gates
Dynamic intelligence scores—separate opportunity/intelligence model
Why BridgePoint is not publishing fake 84/100 badges today: the current property-health layer has real assessments, but the strict BII gate requires multiple observed component families and materially higher confidence. Missing evidence is not silently converted into a “healthy” score.

What BII is intended to combine

  • Physical property/component health evidence
  • Hazard exposure and event history
  • Repair, restoration and mitigation evidence
  • Data freshness and source provenance
  • Coverage/completeness and uncertainty

Weights will be versioned and calibrated against verified outcomes before BII is presented as a decision-grade quantitative index.

What BII does not claim

  • No “exact survival probability.”
  • No promise that a roof will fail or survive a specific storm.
  • No direct insurance pricing, lending or investment recommendation.
  • No inference that missing data means low risk.
  • No use of protected-class or behavioral tracking inputs.
Current publication gate
Minimum confidence0.50
Observed component families3
Health statusMust not be INSUFFICIENT_DATA

Transparent architecture

BIIv = calibrated f(physical health, hazard exposure, recovery/mitigation, freshness, evidence coverage, uncertainty)

The important part is the subscript v: every released model is versioned. A future change in weights, data sources or validation status must be auditable rather than silently changing what “75” means.

Separate the score from confidence

A 78 with 0.92 confidence and a 78 with 0.18 confidence are not equivalent. BridgePoint keeps the index value, confidence, observed component count, source freshness and reason codes separate. If confidence is too low, the public value is withheld.

Inspect a BridgePoint property ID

Enter an exact street address or a BridgePoint canonical property UUID. The preview resolves the property and only returns a numeric BII if the publication gate passes.

Distribution strategy

The score can become infrastructure through voluntary, documented integrations: BridgePoint’s own terminal, a public read API, partner APIs, QGIS/ArcGIS plugins, enterprise dashboards, and browser tools that users explicitly install. Third-party property portals should be integrated by permission or user-controlled extension behavior—not deceptive injection.

Financial / insurance use

BridgePoint can support research, triage and portfolio context now. Any underwriting, pricing, lending or trading use that materially affects decisions needs calibration, validation, jurisdiction review, reason codes, drift monitoring and governance appropriate to that use.

OPEN INSTITUTIONAL TERMINALCLAIMS LIFECYCLE TERMINALMACHINE CONTRACTEMBED BII BADGE