Evaluation of creating scoring opportunities for teammates in soccer via trajectory prediction

Evaluating the individual movements for teammates in soccer players is crucial for assessing teamwork, scouting, and fan engagement. It has been said that players in a 90-min game do not have the ball for about 87 minutes on average. However, it has remained difficult to evaluate an attacking player without receiving the ball, and to reveal how movement contributes to the creation of scoring opportunities for teammates. In this paper, we evaluate players who create off-ball scoring opportunities by comparing actual movements with the reference movements generated via trajectory prediction. First, we predict the trajectories of players using a graph variational recurrent neural network that can accurately model the relationship between players and predict the long-term trajectory. Next, based on the difference in the modified off-ball evaluation index between the actual and the predicted trajectory as a reference, we evaluate how the actual movement contributes to scoring opportunity compared to the predicted movement. For verification, we examined the relationship with the annual salary, the goals, and the rating in the game by experts for all games of a team in a professional soccer league in a year. The results show that the annual salary and the proposed indicator correlated significantly, which could not be explained by the existing indicators and goals. Our results suggest the effectiveness of the proposed method as an indicator for a player without the ball to create a scoring chance for teammates.

[1]  Yoshinobu Kawahara,et al.  Estimating counterfactual treatment outcomes over time in complex multi-agent scenarios , 2022, ArXiv.

[2]  K. Fukui,et al.  SoccerTrack: A Dataset and Tracking Algorithm for Soccer with Fish-eye and Drone Videos , 2022, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[3]  Ulf Brefeld,et al.  Who can receive the pass? A computational model for quantifying availability in soccer , 2022, Data Mining and Knowledge Discovery.

[4]  Keisuke Fujii,et al.  How does AI play football? An analysis of RL and real-world football strategies , 2021, ICAART.

[5]  Ulf Brefeld,et al.  Rating Player Actions in Soccer , 2021, Frontiers in Sports and Active Living.

[6]  Pascal Bauer,et al.  A Goal Scoring Probability Model for Shots Based on Synchronized Positional and Event Data in Football (Soccer) , 2021, Frontiers in Sports and Active Living.

[7]  K. Kushiro,et al.  Evaluation of soccer team defense based on prediction models of ball recovery and being attacked: A pilot study , 2021, PloS one.

[8]  Keisuke Fujii,et al.  Data-driven Analysis for Understanding Team Sports Behaviors , 2021, J. Robotics Mechatronics.

[9]  Bernard Ghanem,et al.  SoccerNet-v2: A Dataset and Benchmarks for Holistic Understanding of Broadcast Soccer Videos , 2020, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[10]  Y. Kawahara,et al.  Cognition and interpersonal coordination of patients with schizophrenia who have sports habits , 2020, PloS one.

[11]  Masakiyo TERANISHI,et al.  Trajectory prediction with imitation learning reflecting defensive evaluation in team sports , 2020, 2020 IEEE 9th Global Conference on Consumer Electronics (GCCE).

[12]  Oliver Schulte,et al.  Deep soccer analytics: learning an action-value function for evaluating soccer players , 2020, Data Mining and Knowledge Discovery.

[13]  Yoshinobu Kawahara,et al.  Policy learning with partial observation and mechanical constraints for multi-person modeling , 2020, ArXiv.

[14]  P. Poupart,et al.  Inverse Reinforcement Learning for Team Sports: Valuing Actions and Players , 2020, IJCAI.

[15]  O. Bousquet,et al.  Google Research Football: A Novel Reinforcement Learning Environment , 2019, AAAI.

[16]  Lotte Bransen,et al.  Measuring Football Players' On-the-ball Contributions from Passes during Games , 2018, MLSA@PKDD/ECML.

[17]  Stefan Becker,et al.  RED: A Simple but Effective Baseline Predictor for the TrajNet Benchmark , 2018, ECCV Workshops.

[18]  Oliver Schulte,et al.  Deep Reinforcement Learning in Ice Hockey for Context-Aware Player Evaluation , 2018, IJCAI.

[19]  Yisong Yue,et al.  Generating Multi-Agent Trajectories using Programmatic Weak Supervision , 2018, ICLR.

[20]  Jenna Wiens,et al.  The Advantage of Doubling: A Deep Reinforcement Learning Approach to Studying the Double Team in the NBA , 2018, ArXiv.

[21]  Jesse Davis,et al.  Actions Speak Louder than Goals: Valuing Player Actions in Soccer , 2018, KDD.

[22]  R. Zemel,et al.  Neural Relational Inference for Interacting Systems , 2018, ICML.

[23]  Yoshua Bengio,et al.  Z-Forcing: Training Stochastic Recurrent Networks , 2017, NIPS.

[24]  Oliver Schulte,et al.  A Markov Game model for valuing actions, locations, and team performance in ice hockey , 2017, Data Mining and Knowledge Discovery.

[25]  Xinyu Wei,et al.  Not All Passes Are Created Equal: Objectively Measuring the Risk and Reward of Passes in Soccer from Tracking Data , 2017, KDD.

[26]  Alexander J. Smola,et al.  Deep Sets , 2017, 1703.06114.

[27]  Yisong Yue,et al.  Coordinated Multi-Agent Imitation Learning , 2017, ICML.

[28]  Jesse Davis,et al.  Predicting Soccer Highlights from Spatio-Temporal Match Event Streams , 2017, AAAI.

[29]  Daniel Link,et al.  Real Time Quantification of Dangerousity in Football Using Spatiotemporal Tracking Data , 2016, PloS one.

[30]  Matthew Kerr,et al.  Developing a Data-Driven Player Ranking in Soccer Using Predictive Model Weights , 2016, KDD.

[31]  Søren Kaae Sønderby,et al.  Sequential Neural Models with Stochastic Layers , 2016, NIPS.

[32]  Oliver Schulte,et al.  A Markov Game Model for Valuing Player Actions in Ice Hockey , 2015, UAI.

[33]  Yoshua Bengio,et al.  A Recurrent Latent Variable Model for Sequential Data , 2015, NIPS.

[34]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[35]  Yoshua Bengio,et al.  On the Properties of Neural Machine Translation: Encoder–Decoder Approaches , 2014, SSST@EMNLP.

[36]  Kirk Goldsberry,et al.  A Multiresolution Stochastic Process Model for Predicting Basketball Possession Outcomes , 2014, 1408.0777.

[37]  Max Welling,et al.  Auto-Encoding Variational Bayes , 2013, ICLR.

[38]  Tianjun Xiao,et al.  GRIN: Generative Relation and Intention Network for Multi-agent Trajectory Prediction , 2021, NeurIPS.

[39]  Pegah Rahimian,et al.  Inferring the Strategy of Offensive and Defensive Play in Soccer with Inverse Reinforcement Learning , 2021, MLSA@PKDD/ECML.

[40]  Javier Fernández,et al.  Decomposing the Immeasurable Sport: A deep learning expected possession value framework for soccer , 2019 .

[41]  L. Bornn,et al.  Wide Open Spaces: A statistical technique for measuring space creation in professional soccer , 2018 .

[42]  Austin Basye,et al.  Physics-Based Modeling of Pass Probabilities in Soccer , 2017 .

[43]  Yisong Yue,et al.  Data-Driven Ghosting using Deep Imitation Learning , 2017 .

[44]  Kirk Goldsberry,et al.  NBA Court Realty , 2016 .

[45]  Iain Matthews,et al.  "Quality vs Quantity": Improved Shot Prediction in Soccer using Strategic Features from Spatiotemporal Data , 2015 .

[46]  Kirk Goldsberry,et al.  POINTWISE: Predicting Points and Valuing Decisions in Real Time with NBA Optical Tracking Data , 2014 .