Football pass network based on the measurement of player position by using network theory and clustering

Research output: Contribution to journalArticle

Abstract

The present study proposed the new method to create a pass network based on the measurement of the pass positions. The pass positions were determined from the player positions measured by the automatic tracking system for soccer players. The pass positions were classified into clusters by clustering method. The pass network was created by the number of passes between different clusters. The present study analysed nine official games of Fagiano Okayama of Japan Professional Football League Division 2 in 2016 and 2017 years. The results showed that the pass network could abstractly represent the successful passes. Then, the network metrics such as the total links, degree centrality, scaled connectivity and cluster coefficient were evaluated. The total links and degree centrality were proportional to the number of passes. The scaled connectivity decreased with increasing the number of passes because the degree of the particular nodes increased. Moreover, the cluster coefficient of the node in the top 25% degree increased with increasing the number of passes. These results could provide useful information with respect to the team performance on the field.

Original languageEnglish
Pages (from-to)381-392
Number of pages12
JournalInternational Journal of Performance Analysis in Sport
Volume19
Issue number3
DOIs
Publication statusPublished - 2019 May 4

Keywords

  • cluster analysis
  • Football
  • network analysis
  • pass analysis
  • player position

ASJC Scopus subject areas

  • Orthopedics and Sports Medicine
  • Physical Therapy, Sports Therapy and Rehabilitation

Cite this

@article{c10fc05a378b414ca19a11dbf427418e,
title = "Football pass network based on the measurement of player position by using network theory and clustering",
abstract = "The present study proposed the new method to create a pass network based on the measurement of the pass positions. The pass positions were determined from the player positions measured by the automatic tracking system for soccer players. The pass positions were classified into clusters by clustering method. The pass network was created by the number of passes between different clusters. The present study analysed nine official games of Fagiano Okayama of Japan Professional Football League Division 2 in 2016 and 2017 years. The results showed that the pass network could abstractly represent the successful passes. Then, the network metrics such as the total links, degree centrality, scaled connectivity and cluster coefficient were evaluated. The total links and degree centrality were proportional to the number of passes. The scaled connectivity decreased with increasing the number of passes because the degree of the particular nodes increased. Moreover, the cluster coefficient of the node in the top 25{\%} degree increased with increasing the number of passes. These results could provide useful information with respect to the team performance on the field.",
keywords = "cluster analysis, Football, network analysis, pass analysis, player position",
author = "Takahiro Kawasaki and Kenichi Sakaue and Ryota Matsubara and Satoshi Ishizaki",
year = "2019",
month = "5",
day = "4",
doi = "10.1080/24748668.2019.1611292",
language = "English",
volume = "19",
pages = "381--392",
journal = "International Journal of Performance Analysis in Sport",
issn = "1474-8185",
publisher = "University of Wales Institute, Cardiff",
number = "3",

}

TY - JOUR

T1 - Football pass network based on the measurement of player position by using network theory and clustering

AU - Kawasaki, Takahiro

AU - Sakaue, Kenichi

AU - Matsubara, Ryota

AU - Ishizaki, Satoshi

PY - 2019/5/4

Y1 - 2019/5/4

N2 - The present study proposed the new method to create a pass network based on the measurement of the pass positions. The pass positions were determined from the player positions measured by the automatic tracking system for soccer players. The pass positions were classified into clusters by clustering method. The pass network was created by the number of passes between different clusters. The present study analysed nine official games of Fagiano Okayama of Japan Professional Football League Division 2 in 2016 and 2017 years. The results showed that the pass network could abstractly represent the successful passes. Then, the network metrics such as the total links, degree centrality, scaled connectivity and cluster coefficient were evaluated. The total links and degree centrality were proportional to the number of passes. The scaled connectivity decreased with increasing the number of passes because the degree of the particular nodes increased. Moreover, the cluster coefficient of the node in the top 25% degree increased with increasing the number of passes. These results could provide useful information with respect to the team performance on the field.

AB - The present study proposed the new method to create a pass network based on the measurement of the pass positions. The pass positions were determined from the player positions measured by the automatic tracking system for soccer players. The pass positions were classified into clusters by clustering method. The pass network was created by the number of passes between different clusters. The present study analysed nine official games of Fagiano Okayama of Japan Professional Football League Division 2 in 2016 and 2017 years. The results showed that the pass network could abstractly represent the successful passes. Then, the network metrics such as the total links, degree centrality, scaled connectivity and cluster coefficient were evaluated. The total links and degree centrality were proportional to the number of passes. The scaled connectivity decreased with increasing the number of passes because the degree of the particular nodes increased. Moreover, the cluster coefficient of the node in the top 25% degree increased with increasing the number of passes. These results could provide useful information with respect to the team performance on the field.

KW - cluster analysis

KW - Football

KW - network analysis

KW - pass analysis

KW - player position

UR - http://www.scopus.com/inward/record.url?scp=85066458512&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066458512&partnerID=8YFLogxK

U2 - 10.1080/24748668.2019.1611292

DO - 10.1080/24748668.2019.1611292

M3 - Article

AN - SCOPUS:85066458512

VL - 19

SP - 381

EP - 392

JO - International Journal of Performance Analysis in Sport

JF - International Journal of Performance Analysis in Sport

SN - 1474-8185

IS - 3

ER -