How heavily used is psychology in the making of video games? 'Cause I got to thinking, if a video game was able to sense the player's enjoyment level, then couldn't the game also be programmed to isolate what sorts of elements in levels elicit a positive response, then alter the level design to include more elements similar to the ones that had previously elicited a positive response? Because a game that learned how to be more fun would be pretty cool. Not sure how one would go about sensing the player's enjoyment level.
The best would be is if there was an identifiable pattern of input that could predict enjoyment level. This would be much more difficult to generalize to action segments, but for example, it seems doubtful that a player who's bashing the A button during a cutscene is enjoying it. If such relationships exist, they would likely be the most difficult to discover, but they would be the easiest to implement. Beyond that, more advanced controllers would have access to more relationships; the gyroscopic sensors found in some handhelds, for example might be used for some other sort of relationship. If sufficiently advanced face recognition exists, then the use of a camera and such technology could also be of use.
And even if programming levels that alter themselves isn't feasible, I think such data collection would be valuable to developers, as they could see what segments had the most positive responses, and use this data in future titles.
But, my question is, are psychologists and developers teaming up to try out stuff like this, or intentionality using psychology in creating games? I would think there would be some overlap, but I've never heard much about one.
---
If you're a human and not feeling creative, make this your signature.
No, as far as I know, there has been no serious attempt by developers of mundane video games (that is, games that are marketed as entertainment, as opposed to so-called serious games, which have purposes like education or awareness-raising) to incorporate psychological research. Artists rarely want to have anything to do with psychology, perhaps because they're aware that psychology will likely have more unsettling than flattering things to say about art. The entertainment industry is particularly unfavorably inclined towards social psychologists because of the media-violence literature, which implies that violent television, movies, and video games make people more violent. Yes, I'm afraid that, in a small way, Jack Thompson was right. Clearly, we do not live in the best of all possible worlds.
That said, there's been a lot of speculation that an instrumental-conditioning effect is what gets people hooked on RPGs, and what gets an increasing minority of people dangerously attached to, say, World of Warcraft. I don't think this idea has been tested yet. But if you specifically sought to use behavioral psychology to get millions of people to pay you a monthly fee, you could hardly expect to do better.
Given the technology and knowledge that we have now, the most effective way to use psychology to make video games more fun would probably be to use good psychological methodology when eliciting ratings that are used to guide development. Without having seen how developers test their games, I'd bet they're not very scientific about it.
Now, as for your idea. I think it's pretty cool. You're right in thinking that the hard part is measuring enjoyment. I don't think any kind of manual input would do. See, while enjoyment is indeed associated with general physiological changes like increased arousal, there are other emotions with indistinguishable physiological signs. As I understand it, many feelings that we usually consider sharply distinct, like sexual arousal and fear, are actually close cousins. When our heart beats madly, whether we feel lustful or afraid has more to do with our cognitive appraisal of the circumstances than our bodily sensations. Why do I say this? Because in Dutton and Aron (1974), men were more sexually attracted to a confederate in frightening than in non-frightening conditions. They misinterpreted their own fear reactions as sexual attraction! If we can't identify our own feelings, we'll have a heck of time identifying other people's feelings.
So the only way I can imagine how a game could specifically measure enjoyment is by recognizing facial expressions, as you mentioned. I don't think computer scientists have managed this yet.
On the other hand, we could throw out the criterion of specificity and just measure arousal. There's already technology for measuring arousal quickly and unobtrusively. If we could also get a measurement of the player's preferred degree of arousal (i.e. the player's extraversion), as by administering a questionnaire at the start of the game, perhaps our game could adaptively adjust its intensity to optimize arousal. It could get harder and more frantic when the player got bored, or easier and calmer when the player got stressed. Although we'd always have the problem of not knowing whether the player of a survival-horror game is thinking "Oh man, this is awesome!" or "Oh man, this is terrifying!".
Pornographic games could use a penile plethysmograph, leading to the world's first sex-specific video-game controller, for both senses of the word "sex". :P
Dutton, D. G., & Aron, A. P. (1974). Some evidence for heightened sexual attraction under conditions of high anxiety. Journal of Personality and Social Psychology, 30, 510–517. doi:10.1037/h0037031
---
"And if thy right eye offend thee, pluck it out, and cast it from thee…"