Classification of passes in football matches using spatiotemporal data

Sanjay Chawla, Joël Estephan, Joachim Gudmundsson, Michael Horton

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

A knowledgeable observer of a game of football (soccer) can make a subjective evaluation of the quality of passes made between players during the game, such as rating them as Good, OK, or Bad. In this article, we consider the problem of producing an automated system to make the same evaluation of passes and present a model to solve this problem. Recently, many professional football leagues have installed object tracking systems in their stadiums that generate high-resolution and high-frequency spatiotemporal trajectories of the players and the ball. Beginning with the thesis that much of the information required to make the pass ratings is available in the trajectory signal, we further postulated that using complex data structures derived from computational geometry would enable domain football knowledge to be included in the model by computing metric variables in a principled and efficient manner. We designed a model that computes a vector of predictor variables for each pass made and uses machine learning techniques to determine a classification function that can accurately rate passes based only on the predictor variable vector. Experimental results show that the learned classification functions can rate passes with 90.2% accuracy. The agreement between the classifier ratings and the ratings made by a human observer is comparable to the agreement between the ratings made by human observers, and suggests that significantly higher accuracy is unlikely to be achieved. Furthermore, we show that the predictor variables computed using methods from computational geometry are among the most important to the learned classifiers.

Original languageEnglish
Article number6
JournalACM Transactions on Spatial Algorithms and Systems
Volume3
Issue number2
DOIs
Publication statusPublished - 1 Jul 2017

Fingerprint

Spatio-temporal Data
Computational geometry
Observer
Predictors
Computational Geometry
Classifiers
Trajectories
Stadiums
Classifier
Trajectory
Game
Subjective Evaluation
Data structures
Learning systems
Object Tracking
Domain Knowledge
Tracking System
Complex Structure
Data Structures
Machine Learning

Keywords

  • Classification
  • Computational geometry
  • Feature engineering
  • Spatial algorithms
  • Supervised learning

ASJC Scopus subject areas

  • Computer Science Applications
  • Information Systems
  • Signal Processing
  • Discrete Mathematics and Combinatorics
  • Geometry and Topology
  • Modelling and Simulation

Cite this

Classification of passes in football matches using spatiotemporal data. / Chawla, Sanjay; Estephan, Joël; Gudmundsson, Joachim; Horton, Michael.

In: ACM Transactions on Spatial Algorithms and Systems, Vol. 3, No. 2, 6, 01.07.2017.

Research output: Contribution to journalArticle

Chawla, Sanjay ; Estephan, Joël ; Gudmundsson, Joachim ; Horton, Michael. / Classification of passes in football matches using spatiotemporal data. In: ACM Transactions on Spatial Algorithms and Systems. 2017 ; Vol. 3, No. 2.
@article{386ed043c8ca4062bba9d6b1329e8850,
title = "Classification of passes in football matches using spatiotemporal data",
abstract = "A knowledgeable observer of a game of football (soccer) can make a subjective evaluation of the quality of passes made between players during the game, such as rating them as Good, OK, or Bad. In this article, we consider the problem of producing an automated system to make the same evaluation of passes and present a model to solve this problem. Recently, many professional football leagues have installed object tracking systems in their stadiums that generate high-resolution and high-frequency spatiotemporal trajectories of the players and the ball. Beginning with the thesis that much of the information required to make the pass ratings is available in the trajectory signal, we further postulated that using complex data structures derived from computational geometry would enable domain football knowledge to be included in the model by computing metric variables in a principled and efficient manner. We designed a model that computes a vector of predictor variables for each pass made and uses machine learning techniques to determine a classification function that can accurately rate passes based only on the predictor variable vector. Experimental results show that the learned classification functions can rate passes with 90.2{\%} accuracy. The agreement between the classifier ratings and the ratings made by a human observer is comparable to the agreement between the ratings made by human observers, and suggests that significantly higher accuracy is unlikely to be achieved. Furthermore, we show that the predictor variables computed using methods from computational geometry are among the most important to the learned classifiers.",
keywords = "Classification, Computational geometry, Feature engineering, Spatial algorithms, Supervised learning",
author = "Sanjay Chawla and Jo{\"e}l Estephan and Joachim Gudmundsson and Michael Horton",
year = "2017",
month = "7",
day = "1",
doi = "10.1145/3105576",
language = "English",
volume = "3",
journal = "ACM Transactions on Spatial Algorithms and Systems",
issn = "2374-0353",
publisher = "Association for Computing Machinery (ACM)",
number = "2",

}

TY - JOUR

T1 - Classification of passes in football matches using spatiotemporal data

AU - Chawla, Sanjay

AU - Estephan, Joël

AU - Gudmundsson, Joachim

AU - Horton, Michael

PY - 2017/7/1

Y1 - 2017/7/1

N2 - A knowledgeable observer of a game of football (soccer) can make a subjective evaluation of the quality of passes made between players during the game, such as rating them as Good, OK, or Bad. In this article, we consider the problem of producing an automated system to make the same evaluation of passes and present a model to solve this problem. Recently, many professional football leagues have installed object tracking systems in their stadiums that generate high-resolution and high-frequency spatiotemporal trajectories of the players and the ball. Beginning with the thesis that much of the information required to make the pass ratings is available in the trajectory signal, we further postulated that using complex data structures derived from computational geometry would enable domain football knowledge to be included in the model by computing metric variables in a principled and efficient manner. We designed a model that computes a vector of predictor variables for each pass made and uses machine learning techniques to determine a classification function that can accurately rate passes based only on the predictor variable vector. Experimental results show that the learned classification functions can rate passes with 90.2% accuracy. The agreement between the classifier ratings and the ratings made by a human observer is comparable to the agreement between the ratings made by human observers, and suggests that significantly higher accuracy is unlikely to be achieved. Furthermore, we show that the predictor variables computed using methods from computational geometry are among the most important to the learned classifiers.

AB - A knowledgeable observer of a game of football (soccer) can make a subjective evaluation of the quality of passes made between players during the game, such as rating them as Good, OK, or Bad. In this article, we consider the problem of producing an automated system to make the same evaluation of passes and present a model to solve this problem. Recently, many professional football leagues have installed object tracking systems in their stadiums that generate high-resolution and high-frequency spatiotemporal trajectories of the players and the ball. Beginning with the thesis that much of the information required to make the pass ratings is available in the trajectory signal, we further postulated that using complex data structures derived from computational geometry would enable domain football knowledge to be included in the model by computing metric variables in a principled and efficient manner. We designed a model that computes a vector of predictor variables for each pass made and uses machine learning techniques to determine a classification function that can accurately rate passes based only on the predictor variable vector. Experimental results show that the learned classification functions can rate passes with 90.2% accuracy. The agreement between the classifier ratings and the ratings made by a human observer is comparable to the agreement between the ratings made by human observers, and suggests that significantly higher accuracy is unlikely to be achieved. Furthermore, we show that the predictor variables computed using methods from computational geometry are among the most important to the learned classifiers.

KW - Classification

KW - Computational geometry

KW - Feature engineering

KW - Spatial algorithms

KW - Supervised learning

UR - http://www.scopus.com/inward/record.url?scp=85045544550&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85045544550&partnerID=8YFLogxK

U2 - 10.1145/3105576

DO - 10.1145/3105576

M3 - Article

AN - SCOPUS:85045544550

VL - 3

JO - ACM Transactions on Spatial Algorithms and Systems

JF - ACM Transactions on Spatial Algorithms and Systems

SN - 2374-0353

IS - 2

M1 - 6

ER -