Ship the tiny version

Set a hard cap at the start and hold it. A small finished game beats a big plan.

Solhex grew every round and never shrank. A small drone-base game became a plan for 25 chapters and more than 20 hours (about 50 items, 110 researches, a fleet, enemy bases) in 9 days and 171 builds. Then Kraig said "it's just getting to be too much" and lost interest.

The first half was already good. It closed as Part One at Chapter 10, about four hours, with an ending. That's the game you can play today.

What changed after

The next games started with a cap written down before any code:

• Sol Factory: 6 levels, 13 items, 4 buildings, one map.
• Swarmlight: playable the same evening. The first build was one loop that tests the idea: fly, mine, one tower, one swarm that copies itself. Plain shapes, no art.

Swarmlight's first build was playable within the hour, and Kraig's play changed the design every round after that. A thin build he could play beat a full spec he couldn't.

Rules worth keeping

1. Write the cap down at the start: length, item count, chapters, a date.
2. Every playtest round cuts or simplifies something, not only adds.
3. Finish a small game (a Part One) before planning a sequel.
4. Ship the first build tonight, ugly, if the idea needs testing.