Intelligence Sources

What Goes Into the Score

Cade Market aggregates signals from structured data repositories, open-source intelligence, and expert forecasting networks. Each source category carries calibration weights that reflect historical accuracy per event type.

Source Categories

Four Layers of Intelligence Input

The scoring engine draws on four source categories. Not all categories carry the same weight on every event type. Calibration history determines contribution per category.

Structured Data

Government and Institutional Data

Legislative voting records, budget publications, trade statistics, central bank communications, and electoral commission data. Structured sources are parsed and scored against event definitions with high precision.

OSINT

Open-Source Intelligence

News wire services, regional press, social listening signals, and public research publications. OSINT signals are weighted by publication type, regional relevance, and historical reliability on similar event categories.

Economic Indicators

Quantitative Economic Signals

Currency volatility, credit default swap spreads, equity market reactions to political events, and macroeconomic release data. Economic signals calibrate well against political stability and policy event categories.

Expert Networks

Structured Expert Forecasting

Probability estimates from regional specialists and structured forecasting panel members. Expert signals are scored against historical calibration to determine contribution weight per event type and region.

Data Pipeline

From Source Signal to Probability Score

Sources are ingested, classified, weighted, and fed into the Bayesian scoring engine. The pipeline runs continuously, updating scores as new signals arrive.

SOURCES Structured Data OSINT Feeds Economic Signals Expert Forecasts Calibration Weight by accuracy per event type Bayesian Update Prior + weighted likelihood ratio Probability Score 68% CI: 61 - 74%

Quality Controls

Source Integrity and Calibration

Sources are not treated as equally reliable. The calibration layer applies historical accuracy weights so that sources with poor predictive records on a given event type contribute proportionally less.

Historical Accuracy Scoring

Each source category's weight is derived from its historical Brier score on similar resolved events. Accuracy degrades weight; reliability increases it.

Category-Specific Weights

A source that calibrates well on economic policy events may not calibrate well on civil unrest. Weights are assigned per source category and per event type, not globally.

Source Attribution Transparency

Analysts see which source categories contributed to a score and by how much. This attribution is visible in the dashboard and in API response payloads.

Divergence Flagging

When source categories diverge significantly, the score carries a wider confidence interval and a divergence flag. Analysts see when sources disagree, not just the aggregate.

See Source Attribution in the Dashboard

Request access to review source contribution breakdowns for live tracked events in your region or sector of interest.