Game Development Reference
In-Depth Information
I go through several of these cycles, changing the level and testing it
again and again. Since I can iterate so quickly, I don't bother to spend time
analyzing. I just drop things in the level and try them within minutes.
Since everything is still rendered in simple gray blocks, the work stays
fluid. This is good, since the changes I'm making are large conceptual
jumps like changing a tower to a bridge or replacing the main threat with
another kind of enemy. I'm not worrying about details yet.
Within a few hours, I've gone through several iteration loops, and I've
probably changed the overall concept several times. Perhaps I started with
enemies in a tower (actually just a tall block with snipers on top), but found
that didn't work. I might try a bridge (a long, wide block over a long hole
in the floor). I might have tried minefields, snipers, trenches, artillery, and
any other broad strokes I could think of. Extremely rough versions of any
of these can be executed in minutes.
After between three and eight loops of trying different concepts, I
land on one that works. And here the process begins to change. The loop
lengthens, and the changes I'm making start to shrink. Instead of test-
ing every hour, I test every two or three. Instead of ripping and replacing
entire buildings, I'm adjusting positioning on walls and pillars. As always,
changes aren't made in response to imagined problems, but real ones that
I observed in testing. Every test shows me new, obvious changes that need
to be made.
This is where I get a level artist involved. He's probably not working on
the space directly—it's still too early for that—but he is consulting on its
artistic feasibility. If my overall concept is artistically nonsensical, I might
restart from the beginning. More likely we discuss ways to adjust the space
to make it art-friendly. For example, the level remains in gray, but a tower
or a bridge might be given a specific shape that suggests a style, theme,
world story, and mood. He might mock something up or make an art test
level to explore artistic ideas for the space.
The iteration loops continue. The fight becomes more refined and
balanced. Sometimes the space changes in response to artistic or narrative
concerns, but most changes are still driven by balance, pacing, clarity, and
depth concerns as I observed them in my self-tests.
Eventually, though, I hit a wall. There's a point where testing your
own work no longer teaches you anything new. By this time I've made a
fight that works well when I play it—but the game isn't being made for me.
It has to work for all its players. And the only way to understand how well
it is working with real players is to watch them play.
