Things we keep relearning about game design

Stuff we keep getting wrong

Every project we start, we tell ourselves "this time will be different." It usually isn't. Here's what trips us up, over and over:

1. The core loop has to be fun with no art. Grey boxes, placeholder sounds, programmer animations. If it's not fun at that stage, more polish won't save it. We've learned this the hard way twice now.
2. Prototype before you plan. We used to spend weeks on design docs before touching an engine. Now we build the dumbest possible version of the idea in a weekend and see if anyone wants to keep playing it.
3. Other people will find bugs you can't see. You stop noticing problems after staring at the same build for three weeks. Get someone else to play it. The earlier the better, even if it's embarrassing.
4. Cut scope before scope cuts you. A finished small game beats an abandoned big one every time. We've got a graveyard of overscoped projects to prove it.
5. The last 10% is half the work. Juice, screen shake, sound design, menu flow. Nobody talks about this stuff but it's the difference between "feels like a game" and "feels like a prototype."