June 2026
Predictor Mundial
Statistical model that predicts World Cup match outcomes using ELO ratings, Poisson distribution, and Monte Carlo simulation.
Python
The Problem
Predictor Mundial estimates the outcome of a specific World Cup match by combining three statistical techniques instead of relying on a trained machine learning model.
Demo
Architecture
Built as a Python script/notebook, using a historical match dataset (sourced from a public GitHub repository) as the foundation for the model. The prediction pipeline combines:
- ELO ratings to rank team strength based on historical performance
- Poisson distribution to model the expected number of goals for each team
- Monte Carlo simulation to run the match 10,000 times and estimate the probability of each possible outcome
Key Features
- Team strength ranking via ELO
- Goal-scoring probability modeling with Poisson distribution
- Match outcome estimation based on 10,000 simulated runs
Technical Challenges
- Translating historical match data into reliable ELO ratings
- Modeling realistic goal distributions per team using Poisson
- Running and aggregating results from thousands of Monte Carlo simulations efficiently