Predicting the metascore with a subjective user experience data

Abstract

The aim of this study is to test how well a subjective user experience (UX) data predicts the Metascore of a digital game. The Metascore calculated by the Metacritic.com is one of the most important indicators of a game's commercial success. Thus, game companies are interested in finding reliable in-house tools to estimate the Metascore before releasing their product. We utilized subjective survey data to test a preliminary regression model for Metascore. The model explained over 50% of the variance between the Metascores. Practically, this means that we can predict a correct Metascore class (e.g., universal acclaim) with 75% accuracy. These promising results provide good grounds for future research on the topic.