ABSTRACT
Chapter 25 discusses the process required to identify specific patterns in football using machine learning. It briefly introduces the way data are used for tactical analysis in football, and details what forms of data are typically available in professional environments (namely positional and event data). Then, machine learning, including the concept of supervised and unsupervised learning techniques, is broadly defined and the necessary steps to apply it on football data are shown. The chapter uses the supervised identification of the tactical instrument of counterpressing as an example, of how the process automatization can be implemented into practice. Additionally, an applied football example for unsupervised learning is presented. Finally, the advantages, ranging from time saving to objectifying the entire process, of using it in daily club processes are described in detail. Additionally, advice and important considerations when designing such a solution is outlined.
