I Built an AI to Measure Football Miracles. The Greatest One Is Moroccan.

2026-06-14 · Source: Machine Learning on Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Intermediate, medium

Summary

An AI model, "The Miracle Index," quantifies the improbability of football upsets using Elo ratings trained on 49,363 international matches dating back to 1872. It calculates a "Miracle" score based on the inverse of the winning probability. The model identified Cameroon's 1-0 win over Brazil in 2022 (4.5% chance, Miracle: 95.5) as the biggest single-match World Cup upset. By simulating the 2022 World Cup 40,000 times, it determined Morocco's semi-final run was a 2.5% event, yielding a Miracle Index of 97.5, making it the most improbable tournament journey in the dataset. The author is also developing "The Hope Index" to measure fan sentiment against statistical probabilities, creating a "Belief Gap," set to go live on June 11 for the 2026 World Cup.

Key takeaway

For data scientists building predictive models for sports or analyzing complex event probabilities, this work demonstrates the power of combining historical data with simulation. You can quantify "miracles" by measuring the distance between expectation and reality using Elo ratings and Monte-Carlo simulations. Consider integrating sentiment analysis to explore the "Belief Gap" between statistical likelihood and public perception, potentially revealing overlooked factors in event outcomes.

Key insights

An AI model quantifies football upsets using Elo ratings and 40,000 tournament simulations, revealing Morocco's 2022 World Cup run as the greatest miracle.

Principles

• Elo ratings learn team strength from raw results.
• Miracle Index = (1 − winner's probability) × 100.
• Tournament simulations reveal sustained improbability.

Method

An Elo engine, trained on 49,363 matches since 1872, calculates win probabilities. A Monte-Carlo simulator runs 40,000 tournament scenarios. A multilingual transformer (`XLM-R oBERTa`) measures fan sentiment.

In practice

• Quantify historical sports upsets.
• Predict tournament outcomes.
• Analyze fan sentiment vs. statistics.

Topics

• Elo Ratings
• Sports Analytics
• Monte-Carlo Simulation
• Sentiment Analysis
• Football World Cup
• Predictive Modeling

Code references

• achrafi04/hope-index

Best for: Data Scientist, Machine Learning Engineer, AI Engineer

Related on AIssential

• Who Will Win the 2026 Soccer World Cup? — Building transparent, auditable forecasting models, where every assumption is explicit, offers a defensible alternative to black-box ML.
• Can Machine Learning Predict the World Cup? — Predicting football outcomes requires robust feature engineering, as complex models offer minimal gains over simpler regression approaches.
• Using FC26 to simulate the world cup ? [D] — Game simulations are unreliable for complex real-world predictions due to inherent randomness and unmodeled variables.
• Inside soccer’s data renaissance — Machine learning reveals counter-intuitive soccer strategies, like "Boot It," by analyzing complex game data.
• Goal Analytics: A Fully Built Data Pipeline, Six Models, and a Real Backtest — A robust sports forecasting system integrates diverse models, rigorous point-in-time backtesting, and a meticulously cleaned data pipeline.
• A Universal Dense Football Event Representation Based on TabTransformer — A TabTransformer learns dense, semantic representations of heterogeneous football event data, improving downstream task performance.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning on Medium.