Download PDFOpen PDF in browser

Move it or lose it: Exploring the relation of defensive disruptiveness and team success

EasyChair Preprint 989

8 pagesDate: May 12, 2019

Abstract

Introduction: Due to the increasing number of tracking data available for official matches in different leagues there are new ways to capture the performance of teams. To not rely on notational data, we previously introduced the D-Def (Goes et. al, 2018), an aggregated variable to quantify passing solely based on tracking data. This value captures the change of organisation by a pass (defensive disruptiveness). In this study, we updated the D-Def by including an automated classifier for subunits, instead of using starting formations, and investigated the relation of the D-Def on team success. Methods: Position tracking data of all players and the ball collected during 89 Dutch Premier League matches was used. Alignment of subunits was automatically identified, using a K-Means classifier, for every pass. D-Def was calculated for every pass (N= 63601) as an aggregate in the change in movement as a result of the pass-based team- and line centroids of subunits and surface and spread of the defending team. Team success was evaluated via wins and losses. We excluded 21 matches because they resulted in a draw. The predictive value of the D-Def for success was calculated using logistic regression analysis. Results & Discussion: The regression model achieved a R² of 0.69 which is high in comparison to other key performance indicators in the literature. This shows that the approach previously introduced as a proof of concept is related to match outcome. Therefore D-Def can be a useful tool to evaluate team performance. Conclusions: This study highlights that performance can be predicted through spatio-temporal aggregates based on player tracking data and we do not need to rely on notational data anymore. Reference: Goes FR, Kempe M, Meerhoff LA, Lemmink KAPM (2018). Not every pass can be an assist: a data-driven model to measure pass effectiveness in professional soccer matches. Big Data, DOI: 10.1089/big.2018.0067

Keyphrases: Analytics, Big Data, Soccer, performance analysis

BibTeX entry
BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:
@booklet{EasyChair:989,
  author    = {Matthias Kempe and Floris Goes},
  title     = {Move it or lose it: Exploring the relation of defensive disruptiveness and team success},
  doi       = {10.29007/gwn6},
  howpublished = {EasyChair Preprint 989},
  year      = {EasyChair, 2019}}
Download PDFOpen PDF in browser