Predicting Users' Difficulty Perception in a VR Platformer Game
A poster presented at ACM MIG 2024 on predicting how users perceive difficulty in a VR platformer game.
Publication details
Abstract
Level design challenges posed by novel virtual reality (VR) gaming experiences call for new techniques. Reports of issues in VR games have pointed to drastic differences in difficulty perception between users. To address this issue, we propose a novel approach that tracks user difficulty perception on the manipulation of various game parameters. The collected user data is used to train a recurrent neural network (RNN) to predict the perceived difficulty of game levels. When predicting difficulty perception on a scale of 0-10, our model predicts on average 1.19 points away from the user's actual rating across 16 levels. Our approach presents an effective proof of concept that can open up a plethora of future research avenues for manipulating game difficulty in virtual reality.