Project Summary
Most World Cup predictions rely on outdated signals like FIFA rankings or long-term national team performance. I asked a different question:
Does current club-level player form predict international match outcomes better than national team reputation?
To answer this, I built the World Cup 2026 Form Intelligence Engine! It is an end-to-end system that aggregates real player performance data and translates it into team-level insights and predictions.
I ingested recent club statistics (minutes, goals, xG, assists, defensive actions) and mapped players to their national teams. Using this, I computed a Team Form Score (0–100), weighted by expected starting XI contributions. These scores are broken down into attack, midfield, and defense to reflect how teams actually perform on the pitch today (not historically!).
WIthen trained a simple, interpretable model to predict match outcomes based on these form-based features. Crucially, I added an explainability layer that highlights why a team is favored - surfacing key factors like midfield dominance, defensive consistency, or over-reliance on a few players.
🔍 What I found
- Current player form is a stronger and more responsive signal than national team reputation
- Several high-ranked teams appear overrated due to poor recent club form
- Lower-ranked teams with strong player performance emerge as underrated contenders
🎯 Why it matters
Football analysis (and fan intuition) often lags behind reality. Decisions are based on reputation, not current performance. This leads to biased predictions and missed insights.
Our system reframes match prediction as a data-driven, explainable decision problem:
- For analysts: better evaluation tools
- For fans: deeper understanding
- For bettors: more informed decisions
Ultimately, this project shows that teams don’t win matches -> players in form do.
Log in or sign up for Devpost to join the conversation.