Updated 2026-08-25

What I Got Wrong Building an Arcade Game

I would not build an arcade game the way I built the first version of this one. Almost everything I added early, I later removed.

Arcade games look like the easy category. There is no story to write, no progression system to balance, no economy to model. You fly, you shoot, you die, a number goes up.

That is exactly what makes them hard. There is nothing to hide behind. If the core loop is not enjoyable in the first ten seconds, nothing else in the build will save it.

The features that made it worse

The early version had more in it. Upgrades between runs, multiple weapon types, a currency. All of it reasonable on paper, and all of it removed.

Upgrades were the clearest mistake. The moment a run’s outcome depends on what you unlocked earlier, the score stops being a measure of how well you just played. A player who does badly cannot tell whether they misplayed or were simply under-equipped, and that ambiguity quietly kills the “one more go” instinct the whole genre runs on.

Cutting them made the game smaller and considerably better. Every run starts identical. The only variable is you.

What actually needed the work

Readability, almost entirely.

An asteroid field on a dark background is a legibility problem long before it is a difficulty problem. Players were not dying because the game was hard — they were dying because they could not tell which rock was on a collision course. Adding a faint spatial grid did more for perceived fairness than any amount of tuning to spawn rates.

The second thing was feedback. A hit needs to be unmistakable within a frame or two: debris, a score number that appears where the thing died rather than at the top of the screen, and enough contrast that it registers in peripheral vision while you are looking somewhere else.

On building for the Mac

Making an arcade game for a desktop rather than a phone changes the assumptions. Sessions are longer, the screen is bigger, and the input is more precise — but the player is also one keystroke from something else entirely.

That argues for very short runs and an immediate restart. Any friction between dying and playing again is where a session ends.

The honest state of it

Astro Belt does one thing. It is a score-attack game about not hitting rocks, and it does not pretend to be more than that. If you want progression and unlocks, this is the wrong game — deliberately.