Multi-dimensional player skill progression modelling for procedural content generation

Procedural Content Generation (PCG), i.e. how game content can be created algorithmically, is an increasingly important area and currently one of the most active topics within the software games industry and game research. One of the crucial aspects of PCG is the capacity to maintain the player engaged and in flow. In this work, we explore how player skill progression could be used by PCG to create more appropriate challenges for each player and propose a model for content adaptation that takes this concept as its core feature. Our approach introduces the player as an active element in the adaptation process and assumes both the player and the game should have an equal and active role in this process. Our adaptation explores how modelling the evolution of multiple dimensions of a same challenge while the game is played helps creating a better game experience for the player. To evaluate our approach, we present a novel validation process embedded in the game itself, with the purpose of providing a more direct and seamless way to analyse player preference. The results of the evaluation of our approach in the context of an endless running side-scrolling platformer game revealed that players have consistent and specific preferences regarding how difficulty should evolve over the course of a game, which should be taken into account when designing an engaging game progression.