EBM: A Quantitative Usability Model for Video Games

Abstract

Usability is an important aspect of video games, because an unusable system can prevent a player from enjoying a game. This thesis presents a quantitative model to measure game control usability of platform video games, called EBM. It is based on the hypoth- esis that the e ort required to play a game is a good indicator of the game's usability. EBM produces a score called \EBM score", which estimates game control usability. It is calculated by combining the contributing factors of e ort, such as duration and intu- itiveness measures. The duration and intuitiveness measures of each game function are combined through a weighted average, weighted by each game function's relative function usage. Then the EBM score is calculated by a weighted sum of the overall duration and intuitiveness measures, using factor weights. EBM was validated by measuring the correlation between EBM scores and SUS scores for a total of 33 control schemes for 7 di erent platform games (5 schemes for 5 games, and 4 schemes for 2 games). Two di erent tests were conducted to validate di erent aspects of the model. The rst test investigated EBM's capability of estimating game control usability of di erent control schemes from the same game. The second test investigated EBM's capability of estimating game control usability of di erent control schemes from di erent games. For each control scheme 3 measurements were taken from 3 di erent participants. The correlations between the EBM scores and SUS scores of the 5 control schemes correlated signi cantly for each of the 7 games. Furthermore, the overall correlation between the EBM and SUS scores of all 33 control schemes across the 7 games was signi cant. Next, Controller Designer was developed to automate the calculation of EBM scores. It is based on EBM, and is used to support developers in designing control schemes. The tool was evaluated based on its usability and e ectiveness in helping the developers design a usable control scheme. The evaluation process involved using a set of interview questions to measure the e ectiveness of Controller Designer, and a SUS questionnaire to measure the usability of Controller Designer. The preliminary results showed that the tool contains problems that still need to be resolved before the tool can be useful.