Global Conflict Predictor

Multi-horizon probabilistic forecasting · Game theory + ensemble · Backtesting report card.

Data: — View source

Conflict & historical data

Add historical data points below. The model uses duration, fatality level, recent trend, and key events to produce a predictive outcome.

Prediction brief

Select a conflict and click “Predict outcome” to generate an actionable brief. The API is configured at build time; if predictions fail, the backend may be temporarily unavailable.

Backtesting report card

Rolling-origin backtest: Brier score (lower is better), ECE calibration (lower is better).

Correlation model & predictability

Tracks cross-signal relationships, driver stability, and time-split forecast predictability.

Load analytics to compute correlation insights.

Country deep dive

Identify risk for a specific country using current conflict-linked signals.

Enter a country to run country-level analysis.

Risk by region

Risk by conflict type

Top 15 risks

Trip scenario

Travel risk outlook

Select a region and travel date to generate advisory.

Top conflict-driven risks

    Global fuel and inflation datapoints

    Europe and China datapoints

    Global flight and travel datapoints (FAA, FlightAware, overseas)

    A research project by Brad. This platform combines multiple data sources (ACLED, UCDP, GDELT, World Bank, UNHCR, CFR and others) with scientific principles from game theory, power transition and conflict diffusion theory, early-warning indicators, and ensemble forecasting to produce multi-horizon probabilistic conflict forecasts. For research and educational use.
    Connect on LinkedIn: bradolton