I wanted to write this post because I’ve been thinking a lot about what I really enjoy in the tabletop roleplaying hobby. I’m trying to get more clarity on the things I really like so I can have a more satisfying experience while playing. I also find it especially helpful to reflect on the sessions that felt miserable and ask myself why I didn’t enjoy them.

Last week I played Dream Askew. It’s a game where you play as members of a queer enclave in a post-apocalyptic world, but in this case, the apocalypse didn’t hit all at once. It first struck the most vulnerable parts of society, while the wealthy and powerful remained largely intact.

We built the apocalypse collaboratively, “writers’ room” style. The game text gives a list of evocative elements, and we picked the ones we liked, then worked all together, sometimes with difficulty, to make everything fit.

We ended up with an Amazonian jungle where our enclave lives in shacks made of recycled plastic. Electricity is scarce and comes from small turbines powered by nearby streams. The enclave is tangled in damp, exposed wires. The jungle is full of a root that, when brewed into tea, lets people share dreams. This has created a serious issue of lack of “psychic privacy,” because people can force their way into others’ dreams… essentially a form of psychic abuse. So a sanctuary was built in the treetops, above where the spores from the root can’t reach. Sleeping there shields you from dream invasions. The sanctuary is run by a character named Charlie, a kind of spiritual leader who oversees the rituals tied to the drug. He’s convinced the enclave that no “outside world” exists, no “intact society” remains.

The other characters were:

Jackbird, a scavenger who keeps the turbines running and trades with increasingly unstable and violent jungle gangs;

Monsoon, highly sensitive to the dream-root and capable of entering shared dreams even without the tea, who’s started building a cult around their dream-realm: some enclave members prefer to live in Monsoon’s dreams rather than reality;

Rutger, a fugitive from a prison-arcology in the intact society, whose arrival threatens Charlie’s teachings by revealing that an intact society does in fact still exist.

Gameplay alternated between high-level “writers’ room” segments where we decided collaboratively what was happening in the enclave, and zoomed-in scenes where we played out specific moments. Each player had a character, but also took charge of a particular “setting element” (like violent gangs, scarcity, the land itself, the psychic maelstrom, the intact society, etc.). Whoever held an element decided how it behaved, but if your character directly interacted with that element, you had to hand it off to someone else, to avoid playing against yourself.

We also decided to assign NPCs to the various setting elements for simplicity. That way, whoever held an element also played the NPCs tied to it. This isn’t part of the game text, just something we added.

The session was fine. But it kinda clarified for me that this isn’t the kind of play I’m looking for. I’ll try to explain what I mean. Maybe it’ll spark a useful conversation, or at least help me better define what I value in role-playing.

(I might misuse some terminology. Feel free to correct me)

The core issue for me was that I never felt anyone truly had a clear grasp of the situation we were playing in. Because the responsibility for the world is fragmented and constantly shifting hands, no one really knows (or is responsible to know) what is going on in the bigger picture. That made the out-of-scene world feel static until it was “lit up” by a scene, and I constantly felt like I was fumbling around in the dark… a bit like a séance with a ouija board, where you only see the next step after you take it and it kinda always makes some sense.

This also made my input feel less impactful. It almost didn’t matter what I said: the group-ouija setup would generate something that “worked” anyway, but not in a way that really felt organic.

For example, I was in charge of an NPC named Jordan, head of the enclave’s armed security. She asked Rutger to take her to the intact society, fed up with the life in the jungle. Rutger agreed, even though she explained it was a horrible place (an idea invented by Rutger’s player; none of us knew what the intact society actually was). Once they arrived, the player holding the setting element for the intact society described a Matrix-style arcology where people live in tubes inside a digital world run by machines.

But that place only “existed” because we decided to go there. It wasn’t part of the setting before that moment. I think that’s part of why my input felt like it lacked weight: because there was no solid context to “push against.” Worse, it felt like no one had a solid enough grasp on the situation for any input to really matter. Our contributions just bounced back like we were throwing things at a rubber wall that adapted no matter what you hit it with, what I’d call “democratized intuitive continuity.”

It’s the same feeling I get when playing with a generative AI: it’ll respond no matter what you throw at it, and sometimes the result is cool, but there’s no underlying organic structure, no clarity about the broader situation (at least, not with current LLMs).

In that session, I also realized that intuitive continuity isn’t limited to the GM, it can also happen when players improvise their characters entirely as they go, starting with no real concept and hoping to discover it entirely through play. That’s subtler, because it only affects one piece (the character, not the whole situation), but it leads to the same core issue: if no one has clarity about the current situation, it’s hard to give meaningful responses. Whatever the other players give you, you’ll just bounce something back that “works,” but without any internal cohesion… because there is nothing to be coherent about!

Over time, you can sort of build coherence through accumulation of outputs, like a GM who improvises enough that they eventually create fixed points in the situation to build from, but it’s exhausting, and often lacks consistency. And honestly, that kind of cognitive effort just isn’t fun for me.

Yes, you can also have incoherence with prep, but it’s easier to avoid and usually doesn’t touch the entirety of the situation. That’s why I believe prepping enough to have a clear enough situation has value. If both the situation and the characters have some initial prep behind them, we start play with boundaries and constraints that help shape our authority, while honoring what’s already there.

That’s when inputs feel impactful to me. If I throw something at you, and you have to bounce it off your prep and your constraints (rather than just making something up that works), I feel that my input matters. I’m not saying everything has to be prepped 100%. Sometimes you discover new stuff along the way, but when you do, it’s framed by what already exists. You add it to the fiction in a way that keeps it internally consistent.

That “bounce” is what makes play feel true to me. It forces me to rethink my assumptions, rethink where the sessions is going, and I feel like I’m really playing with other people.

If instead, I can just respond however I want, because there’s no solid wall to bounce against, only a soft, adaptable one, then ironically, your input has less impact. Because no matter what you gave me, I would’ve given you the same kind of answer that worked well anyway.

It’s like when your character comes up with some wild plan in the middle of a session, and it works, not because it should, but because the GM just rolls with it to keep things moving, I believe it’s a form of control driven by performance anxiety.

To me, that’s the difference between playing with people and playing with a generative AI. With an AI, anything goes, and it’ll give you a coherent-sounding answer no matter what you say. That can be cool for solo introspection, but it lacks the specific qualities of roleplaying I enjoy: lack of consensus, working with constraints, and the discovery of where we end up when our independent authorities collide without negotiation.