Pendragon
03-27-2009, 07:10 PM
Why Are Valve's Games So Polished?
Mike Ambinder, PhD, experimental psychologist for Valve, was at the Game Developer Conference 2009 . For Valve and Ambinder, the goal is to create a fun game, and everything else is secondary.
Traditional methods of playtesting include numerous aspects. Ambinder first went into direct observation. This involves simply watching others play the game, how they act, all the while attempting to simulate an at-home experience. In Left 4 Dead, Valve noticed in playtesting that people weren't going back to save teammates who were getting pounced on by hunters. Initially there were only auditory cues to indicate a hunter had attacked, but through testing Valve later on put in bright outlines that could be seen through walls which proved more effective. Direct observation is critical because it can show actual behavior, not what they just say. The artificial environment will bias the experience to some degree, but the importance of what people actually do in game aside from what they say is critical because, as Ambinder pointed out, people tend to lie.
Verbal reports were the next method of playtesting outlined, where testers think aloud and describe their actions as the play session proceeds. It's unprompted and uncorrected, and works well alongside direct observation. In Left 4 Dead, for instance, players initially didn't recognize when they were healing their teammates, calling out questions like 'Am I healing?' So added in were a number of visual cues as well as a shift in perspective. This is helpful because it allows for a real-time perspective into players' thoughts and provided immediate feedback for motivations behind the actions. However, because the testers are talking to someone who made the product along with being distracted actually playing the game, what they say isn't always trustworthy.
Mike Ambinder, PhD, experimental psychologist for Valve, was at the Game Developer Conference 2009 . For Valve and Ambinder, the goal is to create a fun game, and everything else is secondary.
Traditional methods of playtesting include numerous aspects. Ambinder first went into direct observation. This involves simply watching others play the game, how they act, all the while attempting to simulate an at-home experience. In Left 4 Dead, Valve noticed in playtesting that people weren't going back to save teammates who were getting pounced on by hunters. Initially there were only auditory cues to indicate a hunter had attacked, but through testing Valve later on put in bright outlines that could be seen through walls which proved more effective. Direct observation is critical because it can show actual behavior, not what they just say. The artificial environment will bias the experience to some degree, but the importance of what people actually do in game aside from what they say is critical because, as Ambinder pointed out, people tend to lie.
Verbal reports were the next method of playtesting outlined, where testers think aloud and describe their actions as the play session proceeds. It's unprompted and uncorrected, and works well alongside direct observation. In Left 4 Dead, for instance, players initially didn't recognize when they were healing their teammates, calling out questions like 'Am I healing?' So added in were a number of visual cues as well as a shift in perspective. This is helpful because it allows for a real-time perspective into players' thoughts and provided immediate feedback for motivations behind the actions. However, because the testers are talking to someone who made the product along with being distracted actually playing the game, what they say isn't always trustworthy.