Kicked In The Butt By A Metal Cow
Arcade / Action Games are a pain in the butt to create.

Okay, ANY game is a pain in the butt to create. But some games you can balance pretty well with a mathematical analysis. In an action game, it's more about how loaded (or overloaded) the player is with action-choices in a given length of time, and how much leeway is given for errors in either action or timing of those actions. Given the number of variables present in most games, that's something that can usually only be roughly estimated.

Or designed by "gut feel." And then tested.

And there's my problem. As the designer and developer of my game, I am uniquely UNQUALIFIED to test it for playability. I'm not even sure if I'm good, bad, or "okay" at playing my own game.

When I design a level, I start out with an idea. In many cases, especially if it involves modeling interiors, I sketch it out on paper. Other times I just go with it from my head onto the computer. I play through it in "god mode" - an invincibility mode turned on with a single keypress. I desigh, I redesign, I tweak, I note (and usually fix) bugs, I implement new code as it needs coding, and basically get it to a rough level of playability. As the final step, I turn off god mode and play through the level and try to gauge if it's difficulty is about right.

So I'm working on the new game, "Apocalypse Cow." Well, not all that new to me, really. New as in "Newer than Void War", I guess. I'm working on the "Mecha-Robo-Cow" boss level, which occurs 4/5ths of the way through the game. Said boss is a giant metal cow walking on two legs. Her udder fires anti-aircraft shells that explode, saturating an area with lethal explosions. Her left foreleg is an autocannon. She has a missile launcher on her right shoulder. She's supposed to be hard, as the penultimate boss of the game. And she is totally kicking my butt.

Maybe "kicking" isn't the right term. More like "grinding it into a powder and scattering it into hurricane-class winds."

I've been worried that this gal would be too easy. She can be destroyed in about twelve seconds of sustained fire. The previous boss, from about five levels back, is too easy, and needs a minor overhaul. THAT one has a critical weakness that's embarassingly easy to exploit (hover directly over him, far enough that he's barely off-screen below you, and he can't touch you). A weakness is a good thing (in fact, it's practically REQUIRED in making a good boss battle), but not one that is too easy to exploit.

So I made the Mecha-Robo-Cow a bit tougher. She performs her attacks in deliberate, predictable cycles, with a very deliberate pause between the cycles. She changes her behavior when the player is directly above her, but she's also forced to halt her march towards a mission-failing finish line. Tested from the security of an invulnerable god-mode, the Mecha-Robo-Cow seemed to be perfect. Then I turned off invulnerability, and tested the level out "live."

The first dozen times I played, I couldn't even get close enough to scratch the Mecha-Robo-Cow's paint. In sixty tries, I've managed at best to take her down to about 80% health. A hair over two seconds of the twelve of sustained fire. Clearly, she's too tough, and needs to be made easier.

Sure. But how? And by how much?

Every time I find my helicopter doing catching fire, spinning, and crashing into the ground, it's very clear that I made some kind of error. I zigged when I should have zagged. I made this error in judgement at point A, which compounded to leave me stuck in this situation and point B, which I nevertheless could have escaped if I'd done C. It's kind of a high-speed puzzle. It has the desireable quality of feeling that success is just one more try away. This time, I can do it right, make no mistakes, and win. And so I go on to make another mistake - or the same one - fifteen seconds later. It feels like a fair challenge --- one that I keep losing over and over again.

There are probably a dozen ways or more to fix the problem. From reducing the rate of fire, to introducing a pronounced gap between its attacks, to giving the player access to some kind of "second chance" forcefield, to simply dropping the amount of damage the Mecha-Robo-Boss can take before dropping. Or a combination of several.

But which is best? And how much easier should I make it? I don't know. Again, I'm UNIQUELY UNQUALIFIED to answer just how over-the-top this level is. I'm hoping I can get it somewhere in the ballpark, so when I finally open this game up to external testers they can provide me with good feedback and let me know if I erred too far on one or the other extreme on the difficulty scale. If I get half the players saying "too easy" and the other half saying "too hard," then maybe it'll be just about right.

I don't really have a point to this, other than maybe sharing what's vexing me in this day in the life of a clumsy indie game developer.

Though I admit it is kinda funny having my butt kicked repeatedly by a metal cow.

(Vaguely) related pointless mutterings:
* You Can't Design Fun On Paper
* Ways to Be A Better Game Designer
* Elements That Make More Believable AI
* Apocalypse Cow Status
* Apocalypse Cow Boss

Labels: Apocalypse Cow, Game Design

- posted by The Rampant Coyote @ 6:24 PM (Permalink)