Date Published: July 10, 2017
Publisher: Public Library of Science
Author(s): Daniel Link, Martin Hoernig, Øyvind Sandbakk.
This paper describes models for detecting individual and team ball possession in soccer based on position data. The types of ball possession are classified as Individual Ball Possession (IBC), Individual Ball Action (IBA), Individual Ball Control (IBC), Team Ball Possession (TBP), Team Ball Control (TBC) und Team Playmaking (TPM) according to different starting points and endpoints and the type of ball control involved. The machine learning approach used is able to determine how long the ball spends in the sphere of influence of a player based on the distance between the players and the ball together with their direction of motion, speed and the acceleration of the ball. The degree of ball control exhibited during this phase is classified based on the spatio-temporal configuration of the player controlling the ball, the ball itself and opposing players using a Bayesian network. The evaluation and application of this approach uses data from 60 matches in the German Bundesliga season of 2013/14, including 69,667 IBA intervals. The identification rate was F = .88 for IBA and F = .83 for IBP, and the classification rate for IBC was κ = .67. Match analysis showed the following mean values per match: TBP 56:04 ± 5:12 min, TPM 50:01 ± 7:05 min and TBC 17:49 ± 8:13 min. There were 836 ± 424 IBC intervals per match and their number was significantly reduced by -5.1% from the 1st to 2nd half. The analysis of ball possession at the player level indicates shortest accumulated IBC times for the central forwards (0:49 ± 0:43 min) and the longest for goalkeepers (1:38 ± 0:58 min), central defenders (1:38 ± 1:09 min) and central midfielders (1:27 ± 1:08 min). The results could improve performance analysis in soccer, help to detect match events automatically, and allow discernment of higher value tactical structures, which is based on individual ball possession.
Technological innovations of recent years, particularly in the field of tracking systems, are leading to an increasing volume of data in sports. These enormous amounts of information present new challenges when it comes to analyzing and interpreting this data. What is required are answers to such questions as how sports clubs can best exploit the possibilities on offer to analyze game tactics, manage training processes and make better transfer decisions; how media companies can use this information to offer better and more innovative match coverage products; and how new scientific insights into the nature of sporting phenomena in general and the factors that influence performance can be gained.
In the following, a system of terms is introduced that defines the phenomenon of ball possession and differentiates its various types (Fig 2). In our approach, only time intervals during which the ball is in play are considered when determining ball possession. When the ball is in play, one of the two teams always has team ball possession and one of the players always has individual ball possession. When individual ball possession switches between two players, it is assumed that the pass reflects the tactical intent of the first player. This time interval is thus classified as ball possession for the first player. No individual player and neither team is assigned ball possession when the ball is not in play. This approach differs from that taken by some CIPs, who ascribe these time intervals to one or other of the teams (either the team that had possession of the ball up to that moment or the team that is in possession of the ball thereafter).
Automatic detection of ball possession involves a multi-step process (Fig 3). The first step involves pre-processing raw data provided by the CIPs in order to reduce the number of unrealistic position jumps and noise in the player and ball coordinates. The core of the procedure involves detecting the moments where the IBPs and IBAs start and the IBAs finish. As shown in the following sections, these time codes can be used to reconstruct all of the required time intervals. After that, the level of ball control within the IBAs is estimated using a Bayesian network. The following sections describe the individual steps of the process in detail.
The training and evaluation phase had two sub-goals: Firstly, training was performed to determine the model parameters. Secondly, the quality of IBP, IBA, IBC detection was examined. An evaluation of the team-specific constructs TBP and TBC was not necessary, since they are derived entirely from the IBPs and IBCs.
As a first application of the method, we evaluated new performance metrics based on the different types of individual ball possession. Our sample comprises 60 matches during the 2012/13 seasons of the German Soccer Bundesliga. The positional data were collected during the match using an optical tracking system (TRACAB). Kick detection was used to calculate the IBP and TBP, IBA, IBC and TBC intervals were determined using ball prediction without player prediction. Overall, we collected 69,667 IBA intervals, including 53,354 IBCs.
The evaluation suggests that the method is suited for determining individual ball possession in soccer. It can be used for a wide variety of potential applications. Firstly, it allows quantification of the amount of time a player in possession of the ball spends in different areas of the pitch or the distribution of ball possession times. Secondly, information about basic events, such as passes, tackles, or shots on goal, can be deduced directly from the individual ball possession data. These events are logged by CIPs as default, but due to the manual data collection process the event lists are not always complete and the events’ time stamps are sometimes imprecise. Automatic detection methods can thus help to ensure the quality of match data and potentially reduce the loggers’ workload. Thirdly, being able to detect ball possession is a fundamental prerequisite for discerning higher value tactical structures, such as the ability for players to receive a pass, pressing strategies or marking tactics. Finally, the ability to recognize ball possession types holds considerable potential for improving the quality of match analysis in professional soccer.