Gameplay metrics is a method and approach that is growing in popularity amongst the game studies research community for its capacity to assess players’ engagement with game systems. Yet, little has been done, to date, to quantify players’ responses to feedback employed by games that conveys information to players, i.e., their audio-visual streams. The present thesis introduces a novel approach to player experience assessment - termed feedback-based gameplay metrics - which seeks to gather gameplay metrics from the audio-visual feedback streams presented to the player during play. So far, gameplay metrics - quantitative data about a game state and the player's interaction with the game system - are directly logged via the game's source code. The need to utilise source code restricts the range of games that researchers can analyse. By using computer science algorithms for audio-visual processing, yet to be employed for processing gameplay footage, the present thesis seeks to extract similar metrics through the audio-visual streams, thus circumventing the need for access to, whilst also proposing a method that focuses on describing the way gameplay information is broadcast to the player during play.In order to operationalise feedback-based gameplay metrics, the present thesis introduces the concept of gameplay performance segmentation which describes how coherent segments of play can be identified and extracted from lengthy game play sessions. Moreover, in order to both contextualise the method for processing metrics and provide a conceptual framework for analysing the results of a feedback-based gameplay metric segmentation, a multi-layered architecture based on five gameplay concepts (system, game world instance, spatial-temporal, degree of freedom and interaction) is also introduced.Finally, based on data gathered from game play sessions with participants, the present thesis discusses the validity of feedback-based gameplay metrics, gameplay performance segmentation and the multi-layered architecture. A software system has also been specifically developed to produce gameplay summaries based on feedback-based gameplay metrics, and examples of summaries (based on several games) are presented and analysed. The present thesis also demonstrates that feedback-based gameplay metrics can be conjointly analysed with other forms of data (such as biometry) in order to build a more complete picture of game play experience. Feedback based game-play metrics constitutes a post-processing approach that allows the researcher or analyst to explore the data however they wish and as many times as they wish. The method is also able to process any audio-visual file, and can therefore process material from a range of audio-visual sources.This novel methodology brings together game studies and computer sciences by extending the range of games that can now be researched but also to provide a viable solution accounting for the exact way players experience games.
展开▼
