This is a review of the soccer analytics research applying a generalized linear spatial regression model, conducted by Iavor Bojinov and Luke Bornn.
Soccer is the most popular sport in the world and the Barclays English Premier League has the largest fan base of all soccer leagues. Soccer is a dynamic game dependent on team strategy and individual player skills. Offensive and defensive strategies are extremely important in bringing teams to ultimate success.
At the end of every season analysts, coaches and fans analyze the teams and how the season played out, typically using summary statistics like who scored the most goals or who saved the most goals. However, in all of the analysis it is important to remember that soccer is basically a spatial game. To deal with this aspect summary statistics that quantify a team’s ability to retain possession of the ball and to disrupt the opposing team when they have possession are created. After this, a map is created that lays out the strengths and weaknesses of a team’s offense and defense.
A generalized linear spatial regression model is used to determine the average disruption surface over a season, mapping areas having a high probability of a disruption and areas having a low probability. A disruption of the attacking team is an action taken by the defensive team that leads to an interruption of the flow in play, including such plays as interceptions and tackles. From there the conditional probability of a disruption at a given location is determined.
Typically, teams that finish in the top half of the league have the highest controlling coefficient and an average disruption coefficient. The number of shots taken by a team and the value of both the control and disruption coefficients are clearly positively related. Intuitively, this makes sense, as a team that retains possession of the ball longer will ultimately have more shots. Next, maps are generated for a team’s defensive disruption surface and one for their offensive control surface.
Coaches can use the maps to gain an understanding of their team’s weaknesses and employ methods to correct them. They can also use the maps to determine their opponent’s weaknesses and determine strategy to take advantage of those weaknesses. Teams with a weak defense will have a disruption surface below average. Disruption surfaces map where a team presses and where they are more passive. This provides coaches with information regarding the vulnerabilities of their opponent and how to strengthen their own weaknesses. Looking at the control and disruption coefficients of their team allows coaches to determine their team’s weakness and strengths as well as their opponents and use this in strategizing for the game.
Analyzing team disruption and control surfaces provides analysts with the opportunity to study what impact a coach has on a team’s playing style. When a team changes coaches, the team typically does not instantly change their playing style, so any major changes made over time are due to strategic decisions made by the new coach. Consequently, the tactics of coaches can be analyzed from year to year and team to team.
Find out how Sports Analytics Expert Victor Holman can give your team the competitive advantage.
How mature is your team’s analytics program? Take the Sports Analytics Maturity Assessment.
Discover the Groundbreaking Sports Analytics Application and Framework coaches and sports analysts are talking about!
Learn all about sports analytics in Victor Holman’s Sports Analytics Blog.