Data gathering and analysis in gaming using Tobii Eye Tracking

Abstract: E-sports is growing and the price pools in e-sports tournaments are increasing, Valves video game DotA 2 is one of the bigger e-sports. As professional gamers train to increase their skill, new tools to help the training might become very important. Eye tracking can give an extra training dimension for the gamer. The aim of this master thesis is to develop a Visual Attention Index for DotA 2, that is, a number that reflects the player’s visual attention during a game. Interviews with gamers combined with data collection from gamers with eye trackers and statistical methods were used to find relevant metrics to use in the work. The results show that linear regression did not work very well on the data set, however, since there were a low number of test persons, further data collection and testing needs to be done before any statistically significant conclusions can be drawn. Support Vector Machines (SVM) was also used and turned out to be an effective way of separating better players from less good players. A new SVM method, based on linear programming, was also tested and found to be efficient and easy to apply on the given data set.