The Quest for the Procedural Grail
The next milestone in gaming will not be some innovative new step in how to make the FPS an ever evolving genre. Or a way to make the cleverest AI that learns and adapts to how you play. It certainly won’t be graphical, and controller peripherals that immerse you more in the game are neither needed or going to be with us any time soon. No, the next huge step in game development will be the procedurally generated story.
We have it already, in infant form. Games like the Total War series, or the ever alluring Mount & Blade have randomisation at their heart, and each time you play the game the ‘story’ is entirely different. Left 4 Dead features the same thing, but it’s all specifically structured around the game, and limited by the coding and intentions of the developers. You can fight the zombie hordes however you want to, but you still gotta fight ’em.
I’m looking forward. There are a few titles on the horizon that look astoundingly brilliant, if only because they have the potential to be so huge, mainly down to their procedural hearts. Love, the MMO, is one. Subversion is another. Infinity, the space based MMO, is another. Of the three, though, Subversion has me most intrigued, if only because there is so little out there about it. Can it match the procedural world it’s creating with similarly complex AI? If it can, I think it’s going to have a massive impact on gaming as a whole, but the indie scene in particular. Why hire dozens of artists and modelers when you can just code a program to do it all for you? The mountain that took a graphical artist a month to carve out of nothing takes the computer a few minutes to form. But this is all terrain, and while we’re all very excited about terrain, it’s what happens on it that’s really important.
If we can apply the skills we have at making the world completely random to making the story truly brilliant while at the same time being random, we’ll have our gaming holy grail. A story that is entirely different every time you play, in a world that’s similarly random? I’m there faster than a motorised snake running from a mongoose on a motorbike. It shouldn’t be as hard as it seems, either. Homer has already shown us that there are only so many different types of story that can be told; you could just limit it to one of those and you’d still have an entirely compelling experience. Similarly, each story outside of modernism and absurdism follows the same format, more or less; there is the status quo, which is then disturbed somehow, the protagonist goes on some form of journey, becomes older, wiser, and then confronts something/one and there is a resolution. All you have to do is have your world feed in and out of that, and viola, you have something that will never be traded in, something that shall never gather dust on the shelf.
There are problems. Why would a developer want you to only ever play one game again? Revenue would dry up, and the industry could well crumble. But then, everyone has an imagination (well, most of us), and a few of us can write reasonably interesting short stories, and yet we still read those of others. There is still a place for the linear game, it would just be dwarfed by the brilliance of the truly procedural game. The other main problem is having a convincing and compelling story. It may not work every time, and I’m sure there would be a few dud runs. Similarly, how do we even know that a computer can craft a convincing story? Sure, they can play chess, but can you reduce something like a game of Final Fantasy to an algorithm? Ok, maybe that was a bad example.
Regardless of whether it works or not, I want it to be tried, if only to tick it off as an option that is a dead end. The potential is frankly mind boggling, and it would be worth it just to see what happens. We’re getting better at generating the worlds; let’s see what we can do when we turn our minds to the story, the characters, hell, even the gameplay mechanics. It’s all out there, waiting for someone far cleverer than myself to code it and watch with satisfied glee as it all comes together.
3 thoughts on “The Quest for the Procedural Grail”
That is one awesome idea. However, I think it likely that a game in that vein would be let down by shoddy game mechanics. Just saying
I’m working on something like that in my spare time, using a creaky 2d engine for everything else. It’s more cause and effect than chaos, though. Nowhere near finished.
Excellent points all, and something I’ve been after for years.
We’re certainly starting to see “procedural” concepts take form in various form in various games. In fact, if you combined them all, you’d probably have something close to the procedural grail. An interesting example will be Borderlands for its thousands of weapons – if you could apply that sheer potential for different models of items, each with their own different stats to NPCs you’d already have a very massive game. As for terrain, I actually remember a *long* time ago using a programme that randomly generate Half-Life levels. And what about Soldier of Fortune II and its random level generator? Basically, the tech’s about, it’s just taking some coding genius to bring it all together I suppose.
Sounds interesting mate, keep us posted of any developments, we’d definitely be interested in running a piece on it some time I’m sure.