Frame by frame completion probability of an American football pass

Detalhes bibliográficos
Autor(a) principal: Silva, Gustavo Pompeu da
Data de Publicação: 2022
Tipo de documento: Dissertação
Idioma: eng
Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
Texto Completo: https://www.teses.usp.br/teses/disponiveis/11/11134/tde-07042022-155700/
Resumo: American football is an increasingly popular sport, with a growing audience in many countries in the world. The most watched American football league in the world is the United States’ National Football League (NFL), where every offensive play can be either a run or a pass, and in this dissertation, it is the pass that matters. Not all pass plays in the NFL are created equal, many factors can affect the probability of a pass completion, such as receiver separation from the nearest defender, distance from receiver to passer, offensive formation, game score, among many others. When predicting the completion probability of a pass, it is essential to know who the target of the pass is. By using distance measures between players and the ball, it is possible to calculate empirical probabilities and predict very accurately who the target will be. The big question is: how likely is it for a pass to be completed in an NFL match while the ball is in the air? We develop a machine learning algorithm to answer this based on the aforementioned predictors. Using data from the 2018 NFL season, we obtained conditional and marginal predictions for pass completion probability based on a random forest model. This is based on a two-stage procedure: firstly, we calculate the probability of each offensive player being the pass target, then conditional on the target, we predict completion probability based on the random forest model. Finally, the general completion probability can be calculated using the law of total probability. We present animations for selected plays and show the pass completion probability frame by frame.
id USP_e663b4192edc359bbe117b4a6fcb8f16
oai_identifier_str oai:teses.usp.br:tde-07042022-155700
network_acronym_str USP
network_name_str Biblioteca Digital de Teses e Dissertações da USP
repository_id_str 2721
spelling Frame by frame completion probability of an American football passProbabilidade de completar um passe no futebol americano quadro por quadroAprendizado de máquinasMachine learningNational Football LeagueNational Football LeagueR SoftwareRandom forestRandom forestSoftware RAmerican football is an increasingly popular sport, with a growing audience in many countries in the world. The most watched American football league in the world is the United States’ National Football League (NFL), where every offensive play can be either a run or a pass, and in this dissertation, it is the pass that matters. Not all pass plays in the NFL are created equal, many factors can affect the probability of a pass completion, such as receiver separation from the nearest defender, distance from receiver to passer, offensive formation, game score, among many others. When predicting the completion probability of a pass, it is essential to know who the target of the pass is. By using distance measures between players and the ball, it is possible to calculate empirical probabilities and predict very accurately who the target will be. The big question is: how likely is it for a pass to be completed in an NFL match while the ball is in the air? We develop a machine learning algorithm to answer this based on the aforementioned predictors. Using data from the 2018 NFL season, we obtained conditional and marginal predictions for pass completion probability based on a random forest model. This is based on a two-stage procedure: firstly, we calculate the probability of each offensive player being the pass target, then conditional on the target, we predict completion probability based on the random forest model. Finally, the general completion probability can be calculated using the law of total probability. We present animations for selected plays and show the pass completion probability frame by frame.Futebol americano é um esporte cada vez mais popular, com uma audiência crescente em muitos países do mundo. Nos Estados Unidos existe a National Football League (NFL), onde toda jogada ofensiva pode ser uma corrida ou um passe, e nessa dissertação o interesse está nos passes. Nem todas as jogadas de passe são iguais, vários fatores podem influenciar a probabilidade de completar um passe, como separação do recebedor para o defensor mais próximo, distância entre o passador e o recebedor, formação ofensiva, placar do jogo e vários outros. Quando se tenta prever a probabilidade de completar um passe, é essencial saber quem é o alvo do passe. Usando medidas de distância entre os jogadores e a bola, é possível calcular probabilidades empíricas e prever com alta acurácia quem será o alvo. A grande questão é: quão provável é um passe ser completado em uma partida da NFL enquanto a bola está no ar? Foi desenvolvido um algoritmo de aprendizado de máquinas para responder a essa pergunta baseado nos fatores mencionados. Usando dados da temporada de 2018 da NFL, foram obtidas probabilidades condicionais e marginais de completar passes, baseadas em um modelo de floresta aleatoória. Foi feito um procedimento em dois estágios: primeiro, calcularam-se as probabilidades de cada jogador ofensivo ser o alvo do passe, depois, dado que o jogador é o alvo, é prevista a probabilidade do passe ser completado baseado no modelo de floresta aleatória. Por último, a probabilidade geral do passe ser completado pode ser calculada usando a lei da probabilidade total.Biblioteca Digitais de Teses e Dissertações da USPMoral, Rafael de AndradeSilva, Gustavo Pompeu da2022-03-08info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttps://www.teses.usp.br/teses/disponiveis/11/11134/tde-07042022-155700/reponame:Biblioteca Digital de Teses e Dissertações da USPinstname:Universidade de São Paulo (USP)instacron:USPLiberar o conteúdo para acesso público.info:eu-repo/semantics/openAccesseng2022-04-08T20:35:02Zoai:teses.usp.br:tde-07042022-155700Biblioteca Digital de Teses e Dissertaçõeshttp://www.teses.usp.br/PUBhttp://www.teses.usp.br/cgi-bin/mtd2br.plvirginia@if.usp.br|| atendimento@aguia.usp.br||virginia@if.usp.bropendoar:27212022-04-08T20:35:02Biblioteca Digital de Teses e Dissertações da USP - Universidade de São Paulo (USP)false
dc.title.none.fl_str_mv Frame by frame completion probability of an American football pass
Probabilidade de completar um passe no futebol americano quadro por quadro
title Frame by frame completion probability of an American football pass
spellingShingle Frame by frame completion probability of an American football pass
Silva, Gustavo Pompeu da
Aprendizado de máquinas
Machine learning
National Football League
National Football League
R Software
Random forest
Random forest
Software R
title_short Frame by frame completion probability of an American football pass
title_full Frame by frame completion probability of an American football pass
title_fullStr Frame by frame completion probability of an American football pass
title_full_unstemmed Frame by frame completion probability of an American football pass
title_sort Frame by frame completion probability of an American football pass
author Silva, Gustavo Pompeu da
author_facet Silva, Gustavo Pompeu da
author_role author
dc.contributor.none.fl_str_mv Moral, Rafael de Andrade
dc.contributor.author.fl_str_mv Silva, Gustavo Pompeu da
dc.subject.por.fl_str_mv Aprendizado de máquinas
Machine learning
National Football League
National Football League
R Software
Random forest
Random forest
Software R
topic Aprendizado de máquinas
Machine learning
National Football League
National Football League
R Software
Random forest
Random forest
Software R
description American football is an increasingly popular sport, with a growing audience in many countries in the world. The most watched American football league in the world is the United States’ National Football League (NFL), where every offensive play can be either a run or a pass, and in this dissertation, it is the pass that matters. Not all pass plays in the NFL are created equal, many factors can affect the probability of a pass completion, such as receiver separation from the nearest defender, distance from receiver to passer, offensive formation, game score, among many others. When predicting the completion probability of a pass, it is essential to know who the target of the pass is. By using distance measures between players and the ball, it is possible to calculate empirical probabilities and predict very accurately who the target will be. The big question is: how likely is it for a pass to be completed in an NFL match while the ball is in the air? We develop a machine learning algorithm to answer this based on the aforementioned predictors. Using data from the 2018 NFL season, we obtained conditional and marginal predictions for pass completion probability based on a random forest model. This is based on a two-stage procedure: firstly, we calculate the probability of each offensive player being the pass target, then conditional on the target, we predict completion probability based on the random forest model. Finally, the general completion probability can be calculated using the law of total probability. We present animations for selected plays and show the pass completion probability frame by frame.
publishDate 2022
dc.date.none.fl_str_mv 2022-03-08
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
status_str publishedVersion
dc.identifier.uri.fl_str_mv https://www.teses.usp.br/teses/disponiveis/11/11134/tde-07042022-155700/
url https://www.teses.usp.br/teses/disponiveis/11/11134/tde-07042022-155700/
dc.language.iso.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv
dc.rights.driver.fl_str_mv Liberar o conteúdo para acesso público.
info:eu-repo/semantics/openAccess
rights_invalid_str_mv Liberar o conteúdo para acesso público.
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.coverage.none.fl_str_mv
dc.publisher.none.fl_str_mv Biblioteca Digitais de Teses e Dissertações da USP
publisher.none.fl_str_mv Biblioteca Digitais de Teses e Dissertações da USP
dc.source.none.fl_str_mv
reponame:Biblioteca Digital de Teses e Dissertações da USP
instname:Universidade de São Paulo (USP)
instacron:USP
instname_str Universidade de São Paulo (USP)
instacron_str USP
institution USP
reponame_str Biblioteca Digital de Teses e Dissertações da USP
collection Biblioteca Digital de Teses e Dissertações da USP
repository.name.fl_str_mv Biblioteca Digital de Teses e Dissertações da USP - Universidade de São Paulo (USP)
repository.mail.fl_str_mv virginia@if.usp.br|| atendimento@aguia.usp.br||virginia@if.usp.br
_version_ 1815256781485506560