So this is really a lame conversation, not a lame discussion. It’s a follow up to last week’s discussion (Part 1, Part 2) that begins with me doubting the completeness of the Hawkins Memory-Prediction Framework but then develops into something more interesting.
Jay: OK, so I was arguing that there’s more to immersion than the predictability thing, because it doesn’t completely take things like dialog and art into account.
Stefan: I think immersion is all still the predictability thing. But that graphic style is a result of the same basic mechanics of memory-prediction. Or rather, the impact that graphic style has is a result of those. Immersion is achieved through all sensory input, and at all levels from simple perception of shapes to complex understanding of plot and character.
Jay: So some games I will predict well written dialog and others I won’t?
Stefan: It’s not always even that you will predict the dialog, but that’s part of it. Another part is that when dialog that you didn’t predict occurs, it has to be able to be fit into your conceptions of the game world. That’s the difference between a good plot twist and a cheap gimmick or random reversal.
Jay: I don’t think I agree with you. Can you give me any tangible examples of how a more generic art style in Shadow of the Colossus would diminish immersion because of how it affects predictions?
Stefan: With more generic art, my brain would notice and categorize fewer features of the graphic style.
Jay: But then what of an even more stylized art that looks terrible? You’d categorize even more nuances perhaps, but if it looks bad you probably will be less immersed.
Stefan: Part of the reason terrible art looks terrible is because it jars you out of categorization. But I’m oversimplifying. Because good art isn’t immediately 100% categorizable, instead it requires re-categorization. The same way a good plot isn’t 100% predictable, but has twists. The key bit is that after each twist, it all still makes sense.
Jay: You’re proposing bad art is by definition something that gives you less to predict? I’m not sure you’re wrong but it seems like you’re working up a tautology. I could be wrong, though.
Stefan: No, mediocre art gives you less to predict (but doesn’t violate its own internal rules). Bad art violates the rules it sets up, jarring you. And good art gives you rules to predict, and ways to reshape those rules as the style surprises you with new things.
Jay: But good art can be minimalist.
Jay: Then isn’t there is bad art that will give more to predict than some good art?
Stefan: More details in the art doesn’t actually equate to more predictions in your mind (or doesn’t have to). We may be interpreting prediction differently. And I agree that bad art can offer a lot to predict, but it also often jars you by violating those predictions.
Jay: OK, how is better dialog going to lead to a more cohesive universe? How do a slightly better and slightly worse sentence that say the same things change anything?
Stefan: One makes you sit back and say, “That’s a crap sentence.” And the other makes you feel whatever the writer wanted you to feel.
Jay: Then good dialog can be jarring and less immersive.
Stefan: How so?
Jay: If I read a “standard” sentence and don’t think twice and then read a sentence and stop and say “damn that was well written.”
Stefan: Yes, but that isn’t jarring in the sense that I’ve been using it. Although I think it could break immersion.
Jay: When a level in Psychonauts is cool enough I will pause it and call Lou (my girlfriend) in to see it.
Jay: We have slightly shifted direction and I’m now arguing that quality and immersion aren’t always hand in hand, which is sort of scary. Perhaps Shenmue was more immersive than better looking and “cooler” games because so much of it was sheer detail, which supports the predictions thing. That quality can break immersion actually supports your position, only better things don’t lead to more affirmable predictions necessarily. Something too good or too bad can violate your expectations, even if its in a pleasing way
Stefan: And I think quality breaking immersion only happens when that quality is not consistent. If the whole game looked as good as the best level in Psychonauts, you wouldn’t pause at a given point and call anyone in. But speaking of which, that black velvet level was amazing.
Jay: There are some games that are consistently amazing, the quality doesn’t drop and rise much in Shadow of the Colossus but each beast deserves to be seen by all your friends and is distracting in how cool it is. Things can be great but different, though. Psychonauts can be cut down to its three best levels and each one may still make you go “holy crap.”
Stefan: True. Each one is still something new. Which then gets to another argument I wasn’t expecting to make. It may be worth sacrificing some of the immersion to achieve that “holy crap” through quality of one thing or another. Sometimes you may want to stun people into non suspending disbelief.
Jay: Yeah, strange. Immersion is almost at odds with boredom. Like you don’t want to violate expectations too much but then how do you keep the player from falling asleep? A walking around town simulator may be very immersive but is something actually immersive if you get bored in 5 minutes then stop playing?
Stefan: Yeah. The same way a mediocre, generic graphics style might still be immersive But lacking flare, it’s boring. And the same with predictable plot, music, etc. We need shifts in our expectations, without being jarred out of them. Changing goals, changing perceptions of who is with us/against us, how well we think we’re doing…A lot of the joy I get from games is pulling out of situations where I expected to die/fail.
Jay: The idea that very exciting games are not usually very immersive games is new to me. Designers tend to talk about how important immersion is without really thinking much beyond its surface.
Stefan: I think the two could be merged. And you could have an exciting, immersive game. You just need to set the ground rules for the world in such a way as to keep it from being boring when compared to the player’s life.
Jay: But excitement would have to be controlled well. Even in a game, say Panzer Dragoon 5, that you establish that people have dragons and that they’re awesome, the first time the player gets a dragon he will pause the game and crap his pants if the up close graphics of his particular dragon are amazing. Controlling quality and excitement to such an extent to make your game never jarring, for good or bad, sounds nigh impossible.
Stefan: True. But that might be what the character would do too.
Jay: The character craps his pants and says, “god damn!”
Stefan: In which case, the player is still immersed, so long as they’re thinking, “Holy crap…a dragon!” and not, “Holy crap, look at those polygons!”
Jay: That’s actually a decent way to lessen the blow.