World Cup 2026 Predictor
World Cup 2026 Predictor
End-to-end match simulator that runs 50,000 Monte Carlo tournaments to forecast the World Cup
Problem
Naive win-probability takes ignore team form trends and don't quantify uncertainty across a full tournament bracket.
Solution
Built a prediction pipeline using Poisson regression and histogram gradient boosting on time-decayed form features, then simulates the full 2026 bracket 50,000 times via seeded Monte Carlo to produce calibrated advancement probabilities.
- Stack
- Pythonscikit-learnpandasNumPy
- Last updated
- February 2026