Modeling the gameplay actions of elite volleyball players and teams based on statistical match reports

Abstract

Background and Study Aim. In modern sports analysis statistical modeling of gameplay actions based on match data is becoming a key tool for optimizing training processes and tactical preparation. The aim of the research is to create models of volleyball players' actions based on statistical reports of the 2022 World Championship matches. Materials and methods. The study used statistical data on the World Volleyball Championship matches among men. The data was extracted from open internet sources and converted into tables in CSV format. These tables were processed in the PyCharm programming environment using Python code. The pandas library was used for data analysis and statistical operations, and 'scikit-learn' for machine learning. Results. Models are presented that best predict the results for teams and volleyball players. Important features for teams have been identified, indicating the successful execution of game elements for the team. The regression equations for the team represent a linear combination of various gameplay metrics that affect the total number of points the team scores in a match. They also emphasize the importance of action elements. Linear regression equations predict the total number of points a volleyball player scores based on various statistical indicators. Conclusions. It is recommended to use statistical modeling to optimize training and tactical strategies based on key gameplay metrics. Linear regression equations can assist in evaluating the effectiveness of a player and team. Regular data updates will ensure the relevance of models for better match preparation. Consideration should be given to the possibilities of implementing analytical tools based on the developed models into training programs to optimize the team's preparation for future matches.