Last week we tore apart Christopher Nolan’s Inception with Emil Pagliarulo. Today we continue the discussion with Fallout 3 lead architect/level designer Joel Burgess, who managed to tie in the Pisa Effect, the Rule of Right, and flying meatballs.
Inception’s dream levels are mostly grounded in reality. On the game side of things, you guys create fantasy worlds, but they have their own set of rules that aren’t too far off from reality in many ways.
There are certain expectations people have of reality. And unless you’re going very out there and fabricating an entirely different reality with its own rules, people still have very specific expectations about the heft of something, or the rules of physics.
There’s actually something we refer around here to “The Pisa Effect.” The idea behind the Pisa Effect is that something like the Tower of Pisa, when you see it in the real world, is backed up by the fact that it’s reality. So you see it, and it’s weird, and you say, “Oh that’s crazy,” but you buy it.
But if that was in a video game, and somebody had never seen the Tower of Pisa, they’d be like, “That would fall right over, that’s dumb, this game is stupid.” And as the level designer, you can point at the Tower of Pisa as much as you want and say “this happens in the real world,” but it doesn’t matter, because it’s about the player’s perception. If the player expects it to fall over, then the player is right and you do it differently.
The film’s primary conceit is that this team can convincingly immerse people in another reality using various types of brain science. How much are you thinking about psychology while designing a game level?
As much as we can. Something we’ve tried doing with the level designers here is, as a group we’ll go read something that’s deliberately outside the realm of level design, and then we’ll get together and talk about it and try to apply it back to game design. So we’ll look at things like “The Design of Everyday Things,” which is a book all about how to design a good toaster, or a tea kettle, and how people approach [those] things.
One book that I’ve read that is really useful is called “Kluge,” which is about the evolution of the human brain. There are interesting concepts in there — one called “change blindness,” which is the idea that people can actually miss pretty huge details… One of the common things that a beginning level designer will do is try to be very subtle in cuing the player. If the player hits an intersection, and they don’t know which way to go, they might think to put a soft candle over there with some warm, inviting lighting.
Some mood music.
Right. [laughs] And there is good intent in a lot of that stuff – warm lighting is comfort, and people are drawn to it. But it’s never enough. And you say, “Well, no, put in a big light, and turn this way, and have an NPC come through this door.” Then it’s like, “You may as well put a flashing neon sign there.”
Some people believe in the “Rule of Right,” where players will tend to turn right, which I don’t believe in… If you believe in that, you’ll hit something where: say I’ve got a level that’s more or less a straight line. But then say you’ve got an intersection where the left hallway has a treasure chest, and the right hallway is the way you go to progress the quest. Say that the Rule of Right is absolutely correct, so you put the golden path, the way to progress the quest, to the right. Well, most players that are hardcore gamers deliberately rebel against what they think you want them to do. That’s the way I play – I want to explore everything. So I would deliberately do the opposite, because I’m worried that when I go the way you want me to go, you’re going to lock a door, and I can never get there again and get the thing. You may as well flip a coin. So it’s really a struggle, because probably 90% of level design is intuitive.
The extras in the world of Inception only break their routines when the dreamers do something dramatic, which sort of reminds me of NPC behavior. Do you think about how independent you want NPCs to behave in a particular level?
One thing that we believe here is that simulation for simulation’s sake becomes sort of meaningless. It has to be visible or benefit the player in some way, otherwise it’s not really worth it.
In Oblivion — which I was only on for the home stretch, so to speak — the Radiant AI could do a lot of things. The NPCs had hunger and responsibility levels, so someone who had a low level of responsibility, who was very hungry, would search for food. They might steal an apple out of somebody else’s inventory, and that theft might be detected – because detection is a system too – by a guard that has a high detection value. And he’ll go and try to arrest him, and end up chasing him through the street. And that’s very, very cool, and we all geek out when we see it on paper or demonstrated in front of us. But the reality of that was –
–when the player comes into town, there’s like 16 guards chasing a guy. And now your quest NPC has been murdered on the edge of town for hunting the king’s deer. [laughs] We try to align closer to authored narrative stuff, and that means at certain points we turn things off in the simulation. Often we might let them be emergent and driven by their AI impulses, but when we’re talking about things that are pivotal to the narrative — like, “This quest NPC, he probably should stay in the shop.” Rather than, “It sucks that you can’t find him because he got tired, and his house had a rat in it, so he went and bought a bed at the inn.”
Did you think of any games while watching the film?
The one thing that I did think about, specifically with the architect stuff, was this video — I think it was a combo CG and live action short that somebody did — where this guy is using Minority Report tools to build a level. It’s kind of a street scene he makes – he’s painting the building textures, and putting the sky in, and then he sort of works on this little flower. And it’s very cool, because it’s like one of those Isaac Asimov visions of the future of what level creation might be one day. And that was definitely in my mind as I saw it.
Have you ever gotten an idea from a dream?
I basically don’t remember my dreams, but there was a point where I was really fascinated with using the dream state as a creative tool. So I figured the first thing you need to do was recall, which is unfortunately where I’m weakest, because I don’t remember dreams — I sleep through hurricanes and earthquakes, I just don’t wake up. And when I do, I wake up very slowly, so they have a lot of time to escape.
And then I figured, lucidity would also be useful, to be able to sort of steer dreams in the direction you want. So something that I did try and do was to actually go to bed with sort of, “tools.” For instance, if it were an Oblivion thing, I’d go to bed thinking about liches in caves. So you hold on to two or three little lay-stones or something to use to steer your dream in a direction.
Did it work? That’s the sort of thing I’d imagine could backfire. Like, you go to bed thinking of elves, and you just dream about a guy with big ears.
I recall getting somewhere with it, but right now, the last dream I remember, I woke up and wrote two or three chapters of a short story from the dream. And it is just asinine stuff — your hamster riding a pony hunting these things that are flying meatballs.
Oh, so that’s where that quest came from.