Statistical
Football Match Prediction

Match outcome probabilities from a Poisson distribution model, with a 10,000-trial Monte Carlo simulator for custom scenarios.

🏆 Today’s fixtures

Live and upcoming matches from Europe’s top four leagues, with model-based outcome probabilities.

⚙️ Custom match simulator

Adjust each side’s attack, defence and form, then run 10,000 simulated matches.

HOME TEAM
Attack coefficient 65
Defensive rating 75
Recent form 70
VS
AWAY TEAM
Attack coefficient 70
Defensive rating 60
Recent form 65

💡 An away-side adjustment is applied automatically.

📊 Team ratings by league

Attack and defence coefficients and Elo ratings for each club in the model.

Club Elo rating Attack (xG) Defence (xGA) Last 5

📖 How the model works

The statistical methods behind the predictions.

📈

Poisson distribution

Goals are rare, roughly independent events occurring over a fixed 90 minutes, which makes the Poisson distribution a natural fit. Each side’s attack and defence coefficients give an expected goal rate, and the model computes the probability of every scoreline from 0-0 to 5-5.

⚡

Elo rating system

Originally developed for chess and now widely used in football, Elo rates a club relative to its opponents rather than in isolation. The rating gap between two sides adjusts the expected goal rates, capped at ±40%.

🎲

Monte Carlo simulation

Rather than solving the distribution analytically, the simulator draws random goal counts 10,000 times using a Knuth Poisson sampler. Counting the outcomes gives empirical win, draw and loss frequencies that reflect real match variance.