An analysis of affective state transitions in survival horror game with the aid of player self-reports and physiological signals

抄録

The current trend of indie survival horror games reveals some level of success in using randomly generated events or the randomness of AI behavior as main scaring technique and to extend game replayability. With more insight on how player’s affective states transitioned while playing survival horror games, the effectiveness of scary events could be increased. This paper analyzes the impact of different player affective states prior to (Neutral, Anxiety, Suspense) and after (Low-, Medium-, High-Fear) a scary event. Affect annotation tool was used to collect player’s pre and post scary event affect while playing a survival horror game. The player’s brainwave signals, heart rate and in-game actions were also collected for testing the potential of automatic measurement of horror game-related affect. Initial results indicated that players are more likely to get more fearful of a scary event when they are in suspense state, and the heart rate showed a good potential on differentiating players’ neutral, anxiety and suspense states. Game designers could utilize player’s affective state for maximizing player’s fear level by slowly building tension until the player got into the target pre-scary event emotion, and then play the scary event.