Mining Controller Inputs to Understand Gameplay

Today's game analytics systems are powered by event logs, which reveal information about what players are doing but offer little insight about the types of gameplay that games foster. Moreover, the concept of gameplay itself is difficult to define and quantify. In this paper, we show that analyzing players' controller inputs using probabilistic topic models allows game developers to describe the types of gameplay -- or action -- in games in a quantitative way. More specifically, developers can discover the types of action that a game fosters and the extent that each game level fosters each type of action, all in an unsupervised manner. They can use this information to verify that their levels feature the appropriate style of gameplay and to recommend levels with gameplay that is similar to levels that players like. We begin with latent Dirichlet allocation (LDA), the simplest topic model, then develop the player-gameplay action (PGA) model to make the same types of discoveries about gameplay in a way that is independent of each player's play style. We train a player recognition system on the PGA model's output to verify that its discoveries about gameplay are in fact independent of each player's play style. The system recognizes players with over 90% accuracy in about 20 seconds of playtime.

[1]  Georgios N. Yannakakis,et al.  Player modeling using self-organization in Tomb Raider: Underworld , 2009, 2009 IEEE Symposium on Computational Intelligence and Games.

[2]  Ke Chen,et al.  Predicting skill from gameplay input to a first-person shooter , 2013, 2013 IEEE Conference on Computational Inteligence in Games (CIG).

[3]  Shrikanth S. Narayanan,et al.  Acoustic topic model for audio information retrieval , 2009, 2009 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics.

[4]  Alexei A. Efros,et al.  Unsupervised discovery of visual object class hierarchies , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[5]  Andrew Gelman,et al.  The No-U-turn sampler: adaptively setting path lengths in Hamiltonian Monte Carlo , 2011, J. Mach. Learn. Res..

[6]  George H. Mealy,et al.  A method for synthesizing sequential circuits , 1955 .

[7]  Jácint Szabó,et al.  Linked latent Dirichlet allocation in web spam filtering , 2009, AIRWeb '09.

[8]  Julian Togelius,et al.  Modeling player experience in Super Mario Bros , 2009, 2009 IEEE Symposium on Computational Intelligence and Games.

[9]  Letha H. Etzkorn,et al.  Source Code Retrieval for Bug Localization Using Latent Dirichlet Allocation , 2008, 2008 15th Working Conference on Reverse Engineering.

[10]  Kenneth O. Stanley,et al.  Galactic Arms Race: an experiment in evolving video game content , 2010, SEVO.

[11]  Juyong Park,et al.  Online game bot detection based on party-play log analysis , 2013, Comput. Math. Appl..

[12]  David M. Blei,et al.  Probabilistic topic models , 2012, Commun. ACM.

[13]  Michael I. Jordan,et al.  Latent Dirichlet Allocation , 2001, J. Mach. Learn. Res..

[14]  Simon Colton,et al.  Unsupervised Modeling of Player Style With LDA , 2012, IEEE Transactions on Computational Intelligence and AI in Games.

[15]  Andrew McCallum,et al.  Optimizing Semantic Coherence in Topic Models , 2011, EMNLP.

[16]  Shumin Zhai,et al.  Octopus: evaluating touchscreen keyboard correction and recognition algorithms via , 2013, CHI.

[17]  Julian Togelius,et al.  Predicting player behavior in Tomb Raider: Underworld , 2010, Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games.

[18]  Atsushi Fujita,et al.  Exploiting MMORPG log data toward efficient RMT player detection , 2010, Advances in Computer Entertainment Technology.

[19]  Michael I. Jordan,et al.  Modeling annotated data , 2003, SIGIR.

[20]  Santonu Sarkar,et al.  Mining business topics in source code using latent dirichlet allocation , 2008, ISEC '08.

[21]  Christian Bauckhage,et al.  Guns, swords and data: Clustering of player behavior in computer games in the wild , 2012, 2012 IEEE Conference on Computational Intelligence and Games (CIG).

[22]  Georgios N. Yannakakis,et al.  Real-Time Game Adaptation for Optimizing Player Satisfaction , 2009, IEEE Transactions on Computational Intelligence and AI in Games.

[23]  Bruce Phillips,et al.  Tracking real-time user experience (TRUE): a comprehensive instrumentation solution for complex systems , 2008, CHI.

[24]  Mark A. Girolami,et al.  Employing Latent Dirichlet Allocation for fraud detection in telecommunications , 2007, Pattern Recognit. Lett..

[25]  P. Donnelly,et al.  Inference of population structure using multilocus genotype data. , 2000, Genetics.

[26]  Julian Togelius,et al.  Experience-Driven Procedural Content Generation , 2011, IEEE Transactions on Affective Computing.

[27]  Jen-Tzung Chien,et al.  Latent dirichlet language model for speech recognition , 2008, 2008 IEEE Spoken Language Technology Workshop.

[28]  Scott E. Hudson,et al.  Dynamic detection of novice vs. skilled use without a task model , 2007, CHI.

[29]  Julian Togelius,et al.  Feature analysis for modeling game content quality , 2011, 2011 IEEE Conference on Computational Intelligence and Games (CIG'11).

[30]  Robert Zubek,et al.  MDA : A Formal Approach to Game Design and Game Research , 2004 .