Pick a game you can explain in one rule

The model is not your bottleneck. Opus 5.5 can build an astonishing range of games. Your bottleneck is choosing one that a stranger understands in the two muted seconds they will give a video. So start from a single rule, and hang everything else off it.

Our worked example, Orbit Switch, has exactly one rule: tap to flip which of two stars pulls your comet. Thread the gates, grab the stardust — but the whole game is that one sentence. Everything else (the speed ramp, the scoring, the daily seed) is decoration on top of it. (Play the finished version → — it is the worked example we build toward across this course.)

Good one-rule games for a first build:

• A stack that tips when it gets uneven.
• A city that floods when you build too low.
• A ship that sinks if you load one side.
• A garden that dies without shade.
• A comet pulled between two stars (ours).

The test is social, not technical: say the rule to a friend in one sentence. If they immediately get it, you have a game. If they ask "and then what?", the rule is not carrying the idea yet — tighten it before you write a line of code.

Why this matters so much with an AI builder: when the rule is crisp, every later prompt gets easier, because you can describe each feature as "that, but it also does X." When the rule is fuzzy, you will spend your passes negotiating with the model about what the game even is. A one-rule game is not a beginner limitation; it is how the most-shared games of this wave were designed. The constraint is the craft.

One more reason to go small: your first game is really a way to learn the pipeline — prompt, determinism, trailer, playtest, ship. A tight one-rule game lets you get all the way through that pipeline in an afternoon and learn every step, instead of drowning in scope on day one. Build the small thing completely; it teaches more than a big thing half-built.