Secrets of soccer: Neural network flows and game performance. (January 2020)

• Record Type: Journal Article
• Authors: Marchiori, Massimo; de Vecchi, Marco
• Is Part Of: Computers & electrical engineering. Volume 81(2020)
• Publication Date: 2020-01
• Subjects: Soccer -- Football -- Games -- Sport -- Networks -- Flows -- Efficiency; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2019.106505
• Languages: English
• ISSNs: 0045-7906

Abstract

Abstract: Soccer is the most popular sport in the world, with currently over three billion fans. There are various reasons for this success, but a unique feature of soccer stands out: every match has a high level of unpredictability. For instance, it is not uncommon for a less skilled team to sometimes defeat a better team. Moreover, a match that is apparently decided in favor of a team can suddenly change course, even with few minutes left, ending with a completely opposite result. These highly dramatic effects have brought popularity to this sport, making every match a complex event where the outcome is far from granted. This same complexity has however challenged all attempts of data analysis: the "secrets of soccer", that is to say the recipes for success, are still an unknown realm, defying all common statistical approaches. In this study we try to shed some light on these secrets by introducing a novel approach that uses neural network flows. We transform a team play into a corresponding brain-like structure, an abstraction that we analyze using measures of efficiency, assessing the "quality of thinking" of the brain. This way, we can view any soccer match as an alternate battle of minds and explore how far this parallelism can help to solve some fundamental open problems, like finding an effective recipe for success, and establishing the best field control strategies.