Methodology

Unified Data Architecture for Orbital Credit Risk

Two integrated engines, the Saryn OR engine for multidimensional risk scoring and the Saryn CR engine for Bayesian PD estimation, operating over a unified telemetry and financial data layer. All outputs are audit-ready and mapped to IFRS 9 and Basel III/IV frameworks.

Five-Layer Stack
Layer 01
Data Ingestion

Continuous ingestion of orbital telemetry, TLE propagation data, operator financial filings, regulatory status feeds, and space situational awareness datasets. Unified schema across 80+ operators and 12,000+ assets.

Google Cloud Run · MongoDB Atlas
Layer 02
Saryn OR Engine

Seven-component risk scoring. Each component draws on a distinct data category: orbital mechanics, propulsion, power, communications, operator capacity, regulatory posture, and sovereign dependency. Components are combined and normalized against total credit exposure to produce a dimensionless Saryn OR Score. Aggregation methodology available under NDA.

Non-financial covariates
Layer 03
Saryn CR Engine

Bayesian probability-of-default engine: calibrated priors for sparse-history assets, operator-class pooling, failure-conditioned estimation on a labelled set of historical failure events, model ensembling, and sequential updating across the 1, 3, and 5-year horizons.

40-case SR 26-2 Backtest
Layer 04
Portfolio Analytics

Operator-level aggregation, concentration risk flagging, correlated failure modelling across constellation structures, and SPV tranche-level PD attribution. Supports structured finance due diligence and satellite-backed lending portfolio management.

SPV · Tranche · Concentration
Layer 05
Output & API

Audit-ready PDF report generation, structured JSON API for system integration, dashboard access for portfolio teams, and regulatory export formats mapped to IFRS 9 Stage classification and Basel III/IV Pillar II disclosure requirements.

REST API · PDF · Dashboard
Saryn OR Engine · Component Detail

Seven-Component Risk Scoring

OR Score produces a dimensionless risk score from seven independently scored components. Higher scores indicate elevated risk. Each component is bounded by an orbit-class-specific ceiling, and the combined score is normalised against total credit exposure so results are comparable across operators and orbits.

A failure in any single component propagates into the aggregate score, reflecting the physical reality that satellite asset failures are often single-point-of-failure events. Full component weighting and aggregation methodology is available under NDA.

Seven scored components
RFSRegulatory Friction ScoreFinancial impact of time-based delays from regulatory friction and policy opacity
TVOTime Value of OrbitLost revenue opportunity due to regulatory delay, scaled to total capital raised
FDRFinancial Debris ReserveUncollateralized financial liability for Active Debris Removal
DRSDecommissioning Reliability ScoreTechnical failure probability to de-orbit, relative to financial capacity
RDSRadiation Degradation ScorePremature failure risk from space weather and radiation environment
EPSEnd-of-Life Propellant ScoreDe-orbit failure risk from propellant reserve and telemetry uncertainty
CGFCislunar Governance FrictionGeopolitical and governance risk on Cislunar assets relative to credit exposure
Saryn CR Engine · Bayesian Stack

Five-component Bayesian probability of default

01
Sparse-History Priors

For assets with short or zero default histories, calibrated priors prevent the common failure of assigning near-zero PD to young assets.

02
Operator-Class Pooling

Assets are pooled by operator class (GEO comms, LEO SAR, MEO navigation, and similar) so information is shared across comparable assets, improving stability for thin-data assets while preserving asset-specific results.

03
Failure-Conditioned Estimation

Estimation is conditioned on a labelled set of historical on-orbit failure events, linking the risk components to realised default outcomes.

04
Model Ensemble

Independent model families are combined into a consensus probability of default across the 1, 3, and 5-year horizons, reducing single-model overfitting.

05
Sequential Updating

As new telemetry, financial filings, or operator events arrive, estimates update at the asset level without full retraining, supporting covenant monitoring and mark-to-market assessment.

Output Formats
PDF

Asset-Level Report

Full PDF output for a single satellite asset: OR Score component breakdown, posterior PD distribution at 1/3/5yr, data confidence score, comparables, and IFRS 9 Stage recommendation.

PD_1YR · PD_3YR · PD_5YROR_SCORE · COMPONENT_DETAILDATA_CONFIDENCE · IFRS9_STAGE
Web

Portfolio Dashboard

Web-based interface for lenders and fund managers monitoring multi-asset exposure. Concentration risk heatmap, risk distribution histogram, and operator-level aggregation with drill-down to asset level.

PORTFOLIO_PD · CONCENTRATIONRISK_DIST · OPERATOR_ROLLUPCOVENANT_TRIGGERS · ALERTS
REST JSON

API Integration

Structured JSON REST API for direct integration into lender credit systems, insurance pricing engines, or fund administration platforms. Supports batch queries across portfolios of up to 500 assets.

GET /asset/{id}/riskGET /portfolio/summaryPOST /batch · WEBHOOK_ALERTS
Model Validation
SR 26-2 · Phase 3

Operational Backtest

System-level validation across 40 historical default and near-default cases. Tests the full pipeline, Saryn OR scoring through Saryn CR probability-of-default output, against known outcomes using pre-event data. Zero missed defaults. Model risk framework aligned with SR 26-2 (US), PRA SS1/23 (UK), and the ECB Guide to Internal Models (EU).

1.00
Recall
0.9231
F1 Score
85.71%
Precision
92.50%
Accuracy