Knowledge Base

Football AI Glossary

Decoding the language of modern football analytics. From Expected Goals to Neural Networks, understand the tech stack reshaping the beautiful game.

The Modern Football Tech Stack

Data Collection

Optical Tracking: Cameras installed in stadiums (e.g., ChyronHego, Hawk-Eye) capturing player coordinates 25x/second.

Wearables: GPS vests (e.g., Catapult) measuring physical load, distance, and sprints.

Processing & AI

Computer Vision: Algorithms that identify players, ball, and events from video feeds automatically.

Machine Learning: Models trained on historical data to predict outcomes (xG) or tactical patterns.

Application

Recruitment: Identifying undervalued players matching specific profiles.

Performance Analysis: Real-time feedback to coaches on tablets during matches.

Metric Dasar

Expected Goals (xG)

Shooting

Ukuran kualitas peluang tembakan. Nilai xG berkisar antara 0 hingga 1, yang menunjukkan probabilitas tembakan tersebut menjadi gol berdasarkan data historis dari ribuan tembakan serupa.

Expected Assists (xA)

Passing

Probabilitas operan menjadi assist gol. Mengukur kualitas operan akhir, terlepas dari apakah penerima bola berhasil mencetak gol atau tidak.

Advanced AI

Possession Value (PV / xT)

Tactical

Model yang mengukur seberapa besar sebuah aksi (operan, dribel) meningkatkan peluang tim untuk mencetak gol. Ini menghargai pemain yang memajukan bola ke area berbahaya, bukan hanya yang melakukan aksi akhir.

PPDA (Passes Allowed Per Defensive Action)

Pressing

Metrik untuk mengukur intensitas pressing. Semakin rendah angka PPDA, semakin intens tim tersebut melakukan pressing (membiarkan lawan melakukan sedikit operan sebelum mencoba merebut bola).