EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

2880. Data science models for football scouting: the Racing de Santander case study

Invited abstract in session TD-16: Performance and scouting in football, stream OR in Sports.

Tuesday, 14:30-16:00
Room: 19 (building: 116)

Authors (first author is the speaker)

1. Manuel Duran
FIUBA, UBA
2. Diego Brunetti
IC/UBA
3. Sebastián Ceria
Racing de Santander
4. Guillermo Durán
Calculus Institute, University of Buenos Aires
5. Andrés Farall
Instituto de Cálculo, Universidad de Buenos Aires
6. Nicolás Marucho
IC/UBA
7. Pablo Mislej
Statistics, IC/UBA

Abstract

The use of data to improve the performance and operation of sports institutions is increasingly expanding.
Related to scouting, the most renowned case is that of Moneyball, which summarizes the successful use of data to improve the performance of a US professional baseball team that had a much smaller budget than that of most of its opponents.
In this work we apply different data science techniques to improve the scouting of a professional soccer team. In particular, we attack the problem of similarity between players, so that a sports manager can have information about who is the player most similar to someone he has to replace, or who he cannot access due to a budgetary issue.
Various similarity schemes were employed to compare players by considering their specific individual characteristics as covariates.
Some approaches adopted a non-supervised perspective, while others utilized supervised models that incorporated valuable information obtained from the players' known positions on the field as a target for supervision.
The disparity between these contrasting approaches yielded complementary results.
We take as a case study that of the Real Racing Club de Santander, from Spain Second Division,

Keywords

Status: accepted

Back to the list of papers