GOD'S EYE
Strategic risk convergence engine: nine systemic risks, one state vector
God's Eye is a strategic risk convergence engine: nine simultaneous systemic risks, from Gulf energy disruption to petrodollar strain to a potential yen carry unwind, tracked as one live state vector.
Most risk analysis watches one variable at a time. The premise here is that the danger is in the convergence: when five or more legs go active together, the system behaves in ways no single chart can show.
The engine tracks nine risk categories converging in one window: an energy war disrupting Gulf oil flows, dollar reserve fragmentation, private credit stress, food supply chain damage from LNG disruption, defense industrial strain, semiconductor supply risk, shipping route closures, a potential yen carry trade collapse, and AI-driven labor displacement.
Each leg is scored 0 to 100 from real observable data, z-score normalized against 2015-to-present baselines. Scores fall automatically when conditions normalize. The composite measures how many legs are active simultaneously.
Live data flows in from Supabase. A 400-run Monte Carlo simulation projects the state vector across a 29-week horizon, with sliders for counterfactuals: central bank hike probability, escalation multipliers, strategic reserve timing. Gaussian mixture models discover emergent regimes from the simulation cloud, with the number of regimes selected by BIC rather than assumed.
Every metric is tagged by epistemic status, observed, derived, inferred, or simulated, and traceable to its data source. A confirmed-signals panel keeps publicly verifiable events separate from model inference, so the dashboard always shows where a number came from and how much to trust it.
Recreated in-page from the live product. Scores shown are a captured state vector; the run button re-rolls with simulation jitter.
The risk isn't in any single variable. It's in the convergence.