Deepening the Understanding of Mobile Game

Abstract

It is increasingly difficult for game developers to build a mobile game that achieves the top positions on app store charts. There is currently no clear strategy to build successful games. In this context, the main purpose of our work is to investigate the relationship between the mobile game features and their success in terms of the number of downloads and the gross revenue. This paper extends a previous work that analyzed the importance of 37 features of performing a linear regression on 34 games inside Top 100 games from both download and grossing charts. The current research analyses 60 games inside the Top 100 games and also 40 between the Top 400 and Top 500 games. Besides including more games that are more widespread in the chart, we also perform other analysis including data discrimination and classification techniques to compare successful games against unsuccessful ones. A decision tree model is trained to identify frequent patterns and discover useful associations and correlations within data. Besides that, a linear regression model that maps game features and charts performance is trained using a M5 prime classifier. Results show a different result from previous study. There is no correlation between features and game position on top download charts. Besides, it were identified 9 game features that influence the revenue performance of successful mobile games.