I deleted the whole world and started again
The island had grown for months without a plan. I threw all of it away and rebuilt it one district at a time.
I deleted the entire game world and started over, building it district by district instead of everywhere at once. The old one had no plan behind it, so every new addition made the next one harder. Starting again cost me weeks and was still the cheaper option.
- A world that grew without a plan made every change cost ten times what it should.
- Almost nothing I valued lived in the world file. The systems were untouched.
- Districts get finished before the next one starts. Slower each, much faster overall.
- The week spent deciding to delete was more wasted than the months of work I deleted.
How it got bad
The island grew the way most hobby projects grow. I would sit down, think that a bit of coastline looked empty, and fill it. Then a road, because the thing I put there needed reaching. Then a building at the end of the road, because a road to nowhere looks wrong.
None of it was decided. Each piece was a reaction to the piece before it, and after a few months I had a world where nothing was anywhere for a reason. It looked fine in a screenshot. It was miserable to work on.
The tell was how long changes took. Moving one building meant fixing a road, which meant re-cutting terrain, which broke a spawn, which broke a save. A ten-minute job was reliably a two-day job. That is the sound a project makes when the structure underneath it is wrong.
Deleting it
I sat on the decision for about a week, which in hindsight was a week wasted. Everyone who has done this will tell you the same thing: the moment you start seriously wondering whether to throw it away, you have already found the answer and are just working up to it.
So I deleted the world. Not the systems, not the plant work, not the code. Just the terrain and everything sitting on it. Months of placing things, gone in an afternoon.
I will not pretend that felt good. It felt like admitting I had wasted the time. What made it bearable was noticing that almost nothing I actually valued lived in the world file. The plant pipeline, the lighting model, the save system, none of that was in there. I had been mistaking the visible part of the project for the important part.
Building it back, one district at a time
The rule now is that a district gets finished before the next one starts. Finished meaning terrain, roads, buildings, lighting, and the thing you actually do there. Not blocked out, not roughed in. Done.
This is slower per district and much faster overall, for the obvious reason. When a district is finished it stops changing, and anything I build next can lean on it without being invalidated later. The two-day ten-minute jobs have stopped.
It also means that at any point there is a part of the game that is genuinely playable rather than a whole game that is uniformly half-finished. That turns out to matter enormously for morale, which for a solo project is a real resource and not a soft one.
What I would tell myself
Decide what a place is for before you build it. I know how that sounds. It is the most obvious advice in the world and I ignored it for months, because building terrain is fun and planning is not.
The other one: the sunk cost was not the terrain. It was the week I spent deciding whether to delete the terrain.