The aim of the work is to determine how the absence of audience during the Covid-19 pandemic could have impacted on European soccer team performance and tactics with the employment of play by play data, or soccer logs. Three different seasons were considered, 18/19, 19/20 and 20/21 from the top five competitions in Europe and its first divisions: Spanish "La Liga", German "Bundesliga", French "La Ligue 1", Italian "Serie A" and English "Premier League". Two different approaches were used to work on this study, first the problem was considered as a classification task where two periods were establish: pre-covid (season 18/19 and season 19/20 until March 14th 2020) and covid (season 19/20 from March 15th and season 20/21. The classification models used were Logistic Regression, Decision Tree, Random Forest and AdaBoost. While the second approach used, a zone passing network analysis, was chosen to complement the results of the former, indicating most of the variables related to passing behavior as the most important to differentiate between the two periods, and also for capturing the teams game mode and tactics in a broader way.
L'obiettivo del lavoro è determinare in che modo l'assenza di spettatori durante la pandemia di Covid-19 potrebbe aver influito sulle performance e tattiche delle squadre di calcio europee attraverso l'utilizzo dei dati di eventi di calcio. Sono state considerate tre diverse stagioni, 18/19, 19/20 e 20/21 delle prime cinque competizioni europee e delle sue prime divisioni: "La Liga" spagnola, "Bundesliga" tedesca, "La Ligue 1" francese, "Serie A" italiana e "Premier League" inglese. Per lavorare a questo studio sono stati utilizzati due approcci diversi, in primo luogo il problema è stato considerato come un compito di classificazione in cui sono stati stabiliti due periodi: pre-covid (stagione 18/19 e stagione 19/20 fino al 14 marzo 2020) e covid (stagione 19/20 da marzo 15 in poi e stagione 20/21). I modelli di classificazione utilizzati sono stati Regressione Logistica, Alberi di Classificazione, Random Forest e AdaBoost. Mentre il secondo approccio utilizzato, un'analisi della rete di passaggio a zone, è stato scelto per integrare i risultati del primo, che ha dichiarato la maggior parte delle variabili relative al comportamento di passaggio come le più importanti per differenziare tra i due periodi, e anche per catturare la modalità di gioco e le tattiche delle squadre in modo più ampio.
Effects of crowd absence during the COVID-19 pandemic on the performance and tactics of European football teams
ABARCA JIMENEZ, NICOLE
2021/2022
Abstract
The aim of the work is to determine how the absence of audience during the Covid-19 pandemic could have impacted on European soccer team performance and tactics with the employment of play by play data, or soccer logs. Three different seasons were considered, 18/19, 19/20 and 20/21 from the top five competitions in Europe and its first divisions: Spanish "La Liga", German "Bundesliga", French "La Ligue 1", Italian "Serie A" and English "Premier League". Two different approaches were used to work on this study, first the problem was considered as a classification task where two periods were establish: pre-covid (season 18/19 and season 19/20 until March 14th 2020) and covid (season 19/20 from March 15th and season 20/21. The classification models used were Logistic Regression, Decision Tree, Random Forest and AdaBoost. While the second approach used, a zone passing network analysis, was chosen to complement the results of the former, indicating most of the variables related to passing behavior as the most important to differentiate between the two periods, and also for capturing the teams game mode and tactics in a broader way.File | Dimensione | Formato | |
---|---|---|---|
AbarcaJimenez_Nicole.pdf
accesso riservato
Dimensione
1.78 MB
Formato
Adobe PDF
|
1.78 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/38812