So, yesterday, I posted this,

which is the first screenshot of the game that I could find. There are probably earlier versions, which didn’t have the shadows rendered. But it’s close to what our starting point was. If you’ve played (or even just seen) the final version of the game, you’ll notice there are a LOT of things missing. No fares, no timer, no money, no jumpers, ice, etc.

I come from a background of making console games, where there’s a huge push to make a “vertical slice” of a game - a small portion of the game that looks and plays like the final game. The rationale is that since you’re only making a little bit of it, it shouldn’t be too much work, and you’ll learn a lot in making those “final” art assets that will help when you go to make the rest of the game.

A “vertical slice” is helpful in a traditional console game, where you have investors to appease, who aren’t going to be able to sit down, play the thing, and imagine what the finished project will look like. If you show them something like the screenshot above, they’d freak out. After all, they’re paying for something that looks state-of-the-art, right?

And a lot of the reasons to do a vertical slice do actually make sense.  But building vertical slices of games have a lot of problems as well. Take a standard third-person action game, for instance - if you want to show off a vertical slice of the game, the first thing you have to make is the main character, with a huge amount of polish, animation, etc. Getting that right is one of the hardest things you’ll have to do in a game - it requires the combined effort of texture artists, modelers, animators, riggers, engineers and designers. But because it’s the first thing that the players see, if you’re trying to make a vertical slice, you have to make that look really good.

But what happens during the course of development, of course, is that things change. Lots of things change. For instance, if your animators assumed a particular run speed, and animated to that, then you realize you need to run faster, all the animations need to get rebuilt. If you realize your character needs to be taller, shorter, fatter, whatever - all the time you invested getting your character ready for the vertical slice all goes out the window!

…or does it? What seems to happen as often as not is that what exists has a certain inertia. If you know that the character runs too slowly, but it’ll take three weeks to change the run speed, you think, “Well, maybe it’s good enough as it is…” And maybe it is. But those things all start to cascade. Little decisions that take longer than they should to iterate on all build up, until what you’re left with is a game that just doesn’t work quite right.

In other fields of design, you often hear about the “Rule of 10.” Basically, the idea is this - the earlier you fix a mistake, the cheaper it is. If I have an idea, think about it for a while, and realize it’s garbage, it’s free for me to say, never write that idea down. If I we decide that something’s wrong with a game when we’re writing up the design documentation, it’s quick to change. If we make a quick, 1-day prototype of a system and realize it sucks, we’ve lost the day it took to test it. If we took that 1-day prototype and had an artist work on it, we lose the day prototype, and all the work the artist put into it. And on and on and on. It sounds simple, but it’s incredible how easy it is to have lousy things propagate on and on until you’re a month from launch, and you’re *still* making changes. At every stage, a mistake gets 10x more expensive to fix. That’s the Rule of 10, and it’s why Vertical Slice development stinks - you invest a lot before you have a chance to fix the mistakes.

So, our first screenshot is hideous, because it was really cheap to build. We rolled the ball around, made sure it felt right, and made sure that there was something fun in rolling a ball around, bouncing off walls, and avoiding obstacles like water. It was a way to take the central mechanic - the thing that players would be doing all the time - and make sure that worked really well.

And the funniest thing of all? Even though we did all that up front, halfway through development, we showed it to people and they hated it.

What happened?

How’d it get fixed?

Why’s it so awesome now?

We’ll talk about that in a later post.