On 12th January 2025, Keigo Kinoshita (M2) , Masayoshi Kurihara (M1), Hotaka Hayashi (B4), Shimada Lab., Department of Systems Innovation, receied the first prize at Sports Data Science Competition 2024 (Football).
From the left: Assoc. Prof. T. Shimada, K. Kinoshita, H. Hayashi, and M. Kurihara
2024 Sports Data Science Competition Best Award (Soccer Category)
The "Sports Data Science Competition" is a sports data analysis competition organized by the Sports Data Science Section of the Japan Statistical Society. This year's soccer category was competed by 42 teams. The team also won the Best Award last year, making this their second consecutive win.
About awarded research
"Evaluation of soccer player placement and free-running direction during attack by the path-cost model"
Keigo Kinoshita, Masayoshi Kurihara, Hotaka Hayashi, and Takashi Shimada
Football (soccer) is a sport in which the placement of 11 players on the pitch is extremely important. We have proposed a method of modeling the cost of attacking on a path ("difficulty of advancing") in the form of path integral based on the space dominance field on the pitch, which is determined by the player placement. The goodness of defensive player placement is evaluated by the cost of optimal paths.
In this study, we further developed this framework and proceeded with the analysis using the tracking data. We found that the importance of space dominance differs between the attacking and defensive directions (when a player loses the ball and is countered), and that the time horizon from the evaluation of placement to predicting the outcome of an attack is about 2~3 seconds. Furthermore, we defined the direction in which each attacking player should move as the direction that improves the evaluation of player placement, and showed that this can explain and classify the actual direction of the free run.
These are results that deepen our understanding of the complex dynamics of soccer, and at the same time, can easily be used to provide feedback to the soccer field.
Your impression & future plan
K. Kinoshita:
I am very happy to receive such a prestigious award two years in a row. I am also very grateful to Mr. Kurihara, Mr. Hayashi, and Dr. Shimada. This time, we have greatly developed the path cost model we proposed last year, which not only enables us to evaluate the placement of team players in various phases of soccer, but also shows the possibility of evaluating the positioning of individual players in the context of soccer. Our research has focused on producing results that are "ready-to-use on the field". We hope that you can expect to see our research applied in the soccer world in the future.
M. Kurihara:
We are very happy to receive such a prestigious award, as we did last year. We were able to obtain new and interesting suggestions by using last year's path cost model but developing it significantly. I am proud to have been involved in this research with Ms. Kinoshita, Mr. Hayashi, and Dr. Shimada. I will continue to work hard on my research, making the most of this experience and its results.
H. Hayashi:
I am truly delighted to receive this prestigious award. I am convinced that the experience of working together with my seniors on this wonderful research has been a deep learning experience and will be a source of great growth for me in the future. While the free-run directional evaluation, which was one of the subjects of this year's conference, was a great achievement, it also left us with a lot of room for further development. I have strong expectations for the future development of this research.