MODELING AND ANALYSIS OF FEATURES OF TEAM PLAY STRATEGIES IN ESPORTS APPLICATIONS

Современные информационные технологии и IT-образование

Alexander M. Kadan¹  Tsiango Chen¹  Le Li¹ 

[1] Yanka Kupala State University of Grodno, Grodno, Belarus;

关键词: Cybersport;    eSports;    eSports game;    first-person shooter game;    Counter Strike;    Global offensive;    Global offensive;    CS: GO;    CS: GO;    strategy games;    team play;    team play strategy;    data analysis;    machine learning;   

DOI  :  10.25559/SITITO.14.201802.397-407

来源: DOAJ

【 摘 要 】

The perfect combination of sports and information technologies, implemented in video and computer games, has made eSports unusually popular, rapidly developing and meeting the challenges of the modern world. The scale of the competitions and the amount of financial investments in eSports have increased rapidly. In this regard, special attention is given to activities related to the preparation of sports teams for competitions. Of particular interest for coaches is the study of the best competitive practices, strategies and tactics of both teams of rivals and their individual players. The article proposes and confirms the hypothesis that the methods of training eSports sportsmen based on mathematical approaches to data analysis are much more substantive and justified, which formally confirm the advantages of various team tactics and allow drawing conclusions about the prospects for using various methods of conducting the game. As the main source of data for analyzing the strategies used by the teams, the demo files of the saved games of the Counter Strike tournaments were used. Based on them, a dataset was created, including games’ metadata, data on game situations related to the behavior of individual players - movement, use of equipment, results of interaction with the enemy, as well as additional computed features of team play related to the level of game activity, the use of open space, traps, group character and speed of movement. Based on the methods of machine learning, recommendations are made on important features of priority team strategies that assume high / low values of these features in case of victory or defeat of the respective team. Also, was made a study of the features that affect the victory of individual teams with antagonistic gaming interests.

【 授权许可】

Unknown