I run two D&D campaigns. They’re set in two different homebrew worlds: Nuvem and Irafeen. I’m also writing a fantasy novel. The novel is set in Nuvem.
Irafeen is a great campaign setting. I’d never write a book in it.
The difference between those two statements is almost the entire interesting question of worldbuilding for me. So: two worlds, two purposes - and what I’ve learned trying to figure out which is which.
Nuvem: a world that has to hold a story
Nuvem is the world I built when I started writing fiction seriously. The campaign came later. The fact that the campaign happens in the same place as the novel is deliberate - I wanted the table to feed the page. When something interesting happens in a session, I have a place to put it. When I work out a piece of cosmology for a chapter, I get to see how it lands when players poke at it.
The constraint on Nuvem is that it has to hold a long-form story. That changes everything about how I build it:
- Coherent geography that you can walk across. No “and here’s a continent we don’t talk about.” Every region has to have a reason it exists in the same world as the others.
- History that compounds. The events of a hundred years ago should still be exerting pressure on the present. If they’re not, they’re decoration.
- Characters who survive their introduction. A novel needs characters who can carry weight across hundreds of pages. So even table NPCs in Nuvem get tested for that - if they can’t, they don’t make it into the book.
The discipline of “this has to be publishable” is what makes Nuvem feel like a real place. And paradoxically, that’s also what makes it a good campaign world: my players are wandering through a setting that has to make sense at every scale.
Irafeen: a world that has to break
Irafeen is the opposite project. It’s a city built on the corpse of a dead god, where corporate factions harvest divine energy as a resource, where a vast cosmic intelligence is gestating beneath the streets, and where a psionic alien species the players don’t fully understand is being secretly farmed for its accumulated knowledge.
If I tried to write a novel in Irafeen, it would collapse. Too much escalation, not enough room to breathe. Every faction has the volume turned to eleven. The cosmic horror is too literal. The named threats are too on-the-nose.
That’s also exactly what makes it work at the table.
A campaign world doesn’t need internal consistency at the level a novel does. It needs dramatic surface area. Every faction in Irafeen exists to give the players a different kind of conflict - corporate intrigue with Vitalis, occult investigation with the Cognoscenti, body-horror dread with the Crimson Shadow, cosmic dread underneath all of it. Players can pick which thread they’re pulling on, and pulling on any of them produces interesting results.
The constraint on Irafeen is that it has to support an improvised story. That changes everything in the opposite direction:
- High-density factions, low-density geography. Players spend most of their time in one city, so I built the city deep and the rest of the world shallow.
- Hooks faster than coherence. I introduce things before I know what they mean, and discover the meaning by watching what players do with them.
- Failure states baked in. Some factions are supposed to win some of the time. The cosmic intelligence is supposed to gestate. The pressure of “it’s getting worse” is the engine.
A novel can’t run on those rails. A campaign can’t run on anything else.
The actual lesson
For a long time I treated worldbuilding as a single skill. It isn’t. A world for the page and a world for the table are doing different jobs, and the things you’d do to improve one will often make the other worse.
The single biggest mistake I see in worldbuilding advice (and the one I’ve made a hundred times myself) is treating “more detail” as universally good. A novel needs detail that compounds. A campaign needs detail that opens. Detail that does the wrong job in either context is dead weight.
What I do now: when I’m thinking about a piece of worldbuilding, I ask which world it’s for first. If it’s for Nuvem, I ask whether it can hold up across hundreds of pages. If it’s for Irafeen, I ask whether it gives players something to pull on right now.
Two worlds. Two purposes. The discipline of keeping them separate is what makes both of them better.