Blog17 Haz 00:002 dk okuma
How to Read Football Statistics Like an AI Model
Sezi
AI Futbol Analiz Platformu
17 Haz 00:002 dk
# How to Read Football Statistics Like an AI Model
Modern football is swimming in statistics. xG, xA, PPDA, ELO — the acronyms pile up fast. But what do these numbers actually mean? And how do AI models use them?
The Core Metrics
Expected Goals (xG)
What it measures: The probability that a given shot results in a goal. A shot from 6 yards with a clear view might have 0.7 xG. A 30-yard volley through traffic might be 0.02. How models use it: xG is the foundation of modern football analytics. It tells you whether a team is creating quality chances — or just shooting a lot.Expected Assists (xA)
What it measures: The xG value of the shot that resulted from a pass. Measures creative contribution.ELO Ratings
What it measures: Team strength on a continuous scale. Originally from chess, adapted for football. Teams gain/lose points based on match results, weighted by opponent strength and competition importance. How models use it: ELO is Sezi's primary team strength signal. International ELO ratings (from eloratings.net) feed into the World Cup model.PPDA (Passes Per Defensive Action)
What it measures: Pressing intensity. Lower PPDA = more aggressive pressing.From Data to Prediction
Sezi's pipeline:
- Collect — fixture data, squad info, ELO ratings, recent form
- Model — Dixon-Coles Poisson for goal expectancy, XGBoost for ensemble
- Calibrate — isotonic regression ensures probabilities are honest
- Explain — AI generates a narrative for every prediction
The Bottom Line
Statistics don't predict football. But they do help you understand it better. That's the whole point of Sezi.
Dünya Kupası 2026 başladı. Maçı sayılarla anlayın.
Dünya Kupası sayfasına git