Predicting players social behaviour in online multiplayer videogames
Online multiplayer videogames are popular because most of their content is generated by interactions among players and thus extremely dynamic. Predicting players’ behaviour in these games is necessary to better design and maintain the game environment and to keep the players engaged in the game. These predictions are difficult because a player’s behaviour varies as a function of the behaviour of other players and with past experience in the game environment. Current approaches used to predict player behaviour are based on machine learning, which can optimize a game environment already in place but does not allow for robust predictions of players’ actions in the future. This project will improve our ability to predict players' behaviour and test hypotheses developed in the fields of ecology and evolution using a scientific and analytical approach. I will use the videogame ‘Dead by daylight’ produced by Behaviour Interactive to predict how a focal player’s in-game behaviour is shaped by behaviours of other players. Results of this project will help managers and designers to make informed decisions over the game's life cycle and maintenance. Ultimately, this will increase players' engagement and revenue for the company.