Using Recommendation Systems to Adapt Gameplay

Ben Medler

NOTE: This paper was selected by the program committee as a Meaningful Play 2008 Top Paper. It has been submitted to the Meaningful Play 2008 Special Issue of the International Journal of Gaming and Computer-Mediated Simulations (IJGCMS), which will be available in July-September 2009 issue. Due to the copyright requirements of the journal, only the abstract is available in the conference proceedings.


Recommendation systems are key components in many web applications (Amazon, Netflix, eHarmony). Each system gathers user input and searches for patterns that exist in order to determine user preferences and tastes. These preferences are then used to recommend other content that a user may enjoy. Games on the other hand are often designed with a one-size-fits-all approach not taking player preferences into account. This paper examines how current web application recommendation systems compare to current games that adapt their gameplay to specific users. The results from this comparison show that games have not to date used certain types of recommendation approaches. Design suggestions for how game developers may exploit the utility of these recommendation features are discussed and examined.