This post is by Leo.
After a long and exciting journey, the book I co-authored with Dimitris Karlis, and Ioannis Ntzoufras Predictive Modelling for Football Analytics edited by CRC Press is available!

The book discusses the most well-known classical and Bayesian models, along with the main computational tools used in the football analytics domain. It also introduces the footBayes R package (built on Stan and CmdStan), which accompanies the reader through all the examples proposed in the book. It aims to be both a practical guide and a theoretical foundation for students, data scientists, sports analysts, and football professionals who wish to understand and apply predictive modelling in a football context.

This text is primarily for senior undergraduates, graduate students, and academic researchers in mathematics, statistics, and computer science who are interested in learning about football analytics. For sure, we really enjoyed writing this book.

Here’s the table of contents:

Chapter 1 – A short introduction to football analytics

Chapter 2 – Methods, algorithms and computational tools

Chapter 3 – Tournament and game prediction via simulation

Chapter 4 – Implementation of basic models in R via footBayes

Chapter 5 – Additional statistical models for the scores

Chapter 6 – Modelling international matches: the Euro and World cup experience

Chapter 7 – Compare statistical models’ performance with the bookmakers

You can order the book here.

And you can download all the data and reproducible R code of the book here.

P.S. I still remember fitting my first Stan model on football data when I was a visiting scholar at Columbia, in the Department of Statistics, back in 2016. I was sitting in Andrew’s office, hoping for a decent fit, and discussing with Jonah how to improve it. Almost ten years later, we now have some proof that we can always improve our models;)

P.S.2 As the great coach José Mourinho once said, “He who knows only about football, knows nothing about football”. For me, football is simply a tool to make statistics more accessible and engaging especially for those outside the field.