Monday, May 7, 2012

My Approach To Game Development

Creation Painting, Michael Parks
SandWyrm here. CaulynDarr, who I also know off-net, wrote something in the comments to my last Turn Structure post that I think needs addressing. He thinks that my development strategy is too complex. Fair enough. Let's discuss it.
CaulynDarr wrote:

I think you're taking an over complicated approach to design here. Basically you are assuming that you need to do X things, and if you approach things that way you'll end up justifying the need for X things in some way. That sidesteps the question of if you really need to do X things and not Y things instead.

You're throwing everything including the kitchen sink into the scope of your design. This is a very waterfall approach to a development process. You run the risk of creating a bloated tightly coupled behemoth that's is more complex and more expensive than you need.

You should approach the problem in a more agile iterative manner. Don't try to define everything up front. Instead, work it out step-by-step.

Basically start with the core element of your game. It could be shooting, assaulting, or moving; whatever you want to generate 75% of the tactical depth to the game. Get that working, and then add in the next part. Every phase of the development should produce a fun game by building off what you already have. If any new addition makes it too complex or less fun, you back off and redesign that addition.

You might discover things like, "Man smoke adds complexity to the game, but not that much tactical depth or fun, let's ditch it" This is opposed to, "My grand design says we must have a smoke phase, lets go redesign our perfectly good shooting system to make it work"

I get that you see problems with 40K or FoW and are trying to fix those while incorporating the good elements right off the bat. Your going to get into a fighting the last war situation. You may fix past mistakes, but your not going to create the next thing in miniature rules design that way. You won't create a successful rules system by patching everyone else's holes, you'll do it by creating something that stands well on it's own.
If M42 was a pioneering project for a gameplay experience that nobody had ever seen before, I would agree with you. But we're not. We're derivative of wargames in general and of 40K/Flames in particular. That means our structure, the one everyone has experience with and likes the feel of, is fairly fixed. So our differences to the competition matter much more than our similarities.


If you're making a buddy-cop movie, certain things are a given. The audience expects them. One cop will be more wild and the other will be more serious. There will be a grumpy police chief that yells at them, and a cool villian they'll have to beat. At some point the 2 cops will get stressed, yell at each other, and make up before going off to kick the villian's ass.
Beverly Hills Cop
Alien Nation
Beethoven
Se7en
Miami Vice

They all follow the same basic structure. Writing and filming the individual stories is still a huge amount of work, but you don't have to sweat the structure. You can fiddle with it here and there, do a few things different, play with expectations, etc. But the structure is a given. It supports you and lets you concentrate on what's important. Which is how you're DIFFERENT from what came before.


We're creating a series of systems here, but this is not a software project. If you're designing software to book airline flights, success is easy to gauge. The flight is either booked or it isn't. The employees at the gate can either figure out how to use it or they can't. You can look at the result of the system and say 'yes this works' or 'no it doesn't'. You can file bug reports and approach it in a very disciplined way. You can even write contracts that specify the exact outcomes expected.

But in a game, or an entertainment product of any kind, you can do all of that and still have a product that doesn't sell. That nobody wants to watch or play. Because only one factor is ever really important in this business, and it's called APPEAL.


When I worked on the movie Antz, the management of the production was a thing of beauty. We produced a quality film that came in under budget and ahead of schedule. Everyone was happy at how smooth everything went. We then released the movie and... it barely made back the money it took to make it. Because a certain other bug movie made by Pixar had more APPEAL.

Better story. Better characters. No creepy child molester in the cast. Nobody knew the development hell Pixar went through though. The wasted time, the changes, the overly-complex full-3D backgrounds that took weeks to render. I heard about it from friends that worked there. But did the audience care? Nope.


Then Shrek got dropped on us by Dreamworks. It was the very definition of the project from hell. A film that had already cost them $40M (the entire budget of Antz) with not a single thing to show for it. Katzenberg would fly up from LA and spend hours pouring over the shape of this stupid ogre's neck or ears and completely ignore the story problems we all wanted him to make decisions on. Who could even take that stupid character seriously with Myers voicing him just like Fat Bastard? What? They just scripted a mud-fight when we don't have the tech to render that?!!!

I left 6 months into production and expected that movie to never be released. Go look on Wikipedia at all the re-writes, redone animation, and everything else that happened.


But... Shrek has earned $484M to date. Why? Because Katzenberg and Myers made every crazy-sounding decision based on one all-important thing: APPEAL

Is this <whatever> appealing? Is it funny? Is it what we want? No? REDO IT. Repeat until we get it right. It took me over 10 years to finally understand that their success wasn't accidental. Though I couldn't see it at the time, it was the result of a deliberate creative process.

Turns out, they're not alone:

James Cameron does it that way.
Apple Computer does it that way.
id Software does it that way.
Valve does it that way.
Blizzard does it that way.

Every single one of them consistently puts out products that are huge hits. Because they always put APPEAL first.

I interviewed once at Blizzard. I know they spent 3+ years making Diablo II. I know the last year of that was 12+ hour days, 7 days a week. I know they threw away 75% of their finished work. Because they would try something, develop it fully, and then test it. If it didn't have appeal, they trashed it.

Is that an efficient way to work? Yes, it is. Don't think so?

Antz cost around $40M, and made $92M. Which is break-even after marketing and distribution.

Shrek cost around $100M in total, and made $484M.

One is a cultural icon that earned back nearly 5 times it's total production cost. The other is selling for $2 a copy in the bargain-bin of your local supermarket and, from a studio standpoint, might as well have never been made.

So What's The Take-Away Here?


Appeal comes first. Appeal stays priority number one. Everything else is secondary to appeal.

A huge part of what I've been doing lately is judging the appeal of our ideas. Is there something folks want to see in the game? Is there something that has always seemed odd to them? If so, we need no other reason for putting those ideas into M42. Because at the end of the day, THAT is what will make them choose us instead of the competition. Because we have something players want that they can't get elsewhere. Because we make things work like everyone always thought they should work.

We can put loads of energy into making the perfect document format, or coming up with perfectly structured development practices, or whatever. But if we put too much of our energy there, then we won't make a perfect GAME. Because you can't optimize for everything at once. You have to choose what's most important. And I know, from hard experience, what that is.

So we'll iterate, iterate, and then iterate some more. The rough-draft game we'll have in 3-6 months won't be the same game we finally release. Because we'll examine every piece individually to see if they work and are appealing. Then we'll step back and see if the whole game works and is appealing.

If we discover that our Fluff isn't appealing, we'll re-write it. If nobody likes realistic psychic powers, we'll redesign that part of the system. If everyone hates the 2 die roll comparative system I've been championing for 9 months now, I'll be the first to throw it away and start over.

Nothing is sacred. Nothing is set in stone. Everything must be appealing to our audience or we fail.

No comments:

Post a Comment

Popular Posts