First things first: my English is not too good, so i hope this will be clear enough.

I didn’t want to wrote a whole post-mortem. Since, while thinking about it and try to find stuff to put in it, my mind stop at one thing that, I think, was my principal “revelation”. Astroz turns out to be a kind of action-puzzle game. I define Action Puzzle to be a puzzle (game with a dominant strategy) with an execution difficulty. Portal is probably the best known (and made) kind of action-puzzle game.

I’m not a level designer. So when I first realize that my game turn out to be a puzzle game, my thought was : “Ouch, I need to make at least one level per mechanic in my game, plus some other here and there to let the player practice the mechanic. And, one or two final levels, the icing on the cake.

I made 15 levels. I think some people made it through the twelfth. In the last hour of the LD, I just create a last level, put everything in it, try it, and thought” mey… that should be feasible”…

After all this, my approach of dividing each mechanic in a separate level, I think, was good. But since it’s an action-puzzle, I think it’s also important to create level just for understanding one puzzle mechaninc, but then, there must be another level just to practice execution.

So, in a way, if your solution is hard to find, it should not be difficult to execute. And, if it’s hard to execute, it should not be hard to find. Except maybe the last level (or last chapter, or last level in a chapter… well you see the logic here). What’s important is to isolate the resolution component from the execution component most of the time. And, combining both should be only to test the player in the end of a series, like an exam or midterm exam, to make sure that your player is competent enough to the next chapter, series or what ever you want to call it

In the end, it seems obvious. But I think it’s still a good rule of thumb when you’re in the final rush and you have the wonderful idea to make a action-puzzle game.