Analysis and Prediction of the Cybersports Match Outcome in DOTA 2 Discipline Based on Multi-factor Regression on the Example of the Team Virtus Pro
This article is devoted to the question of constructing a regression model, which will make it possible to analyze and predict the outcome of e-sports meetings in the DOTA 2 discipline using the Virtus Pro team as an example. In addition, throughout the text, a brief analysis of the e-sports industry these days, as well as a brief explanation of the essence of the chosen e-sports discipline are taken into consideration. Convincing arguments are presented that make it possible for the reader to understand that e-sports today can equal traditional sports in the following indicators: the number of followers, peak views as well as the amount of money circulating in the industry. The law introduced during the research is intended for the following external users: e-sports organizations, potential investors, breeders as well as for the players themselves. The resulting model was analyzed and described using various statistical indicators. A correlation grid was obtained, indicators of regression statistics were described, analysis of variance was made with a detailed description of its results. Factor signs considered during the research are described in detail for reader’s understanding. It is important to note that the resulting law makes it possible to apply it not only to the object of study, but also to other e-sports organizations having a team in this discipline, provided that their styles of playing the game have similarities. The tool for deriving the regression law was Microsoft Excel. The source of data is the team’s statistics, which were collected during real e-sports meetings by the leading site of statistics and the gaming community DOTA 2 — Dotabuff.