Skip to content

Format Forks — What the World Wants to Be

Status: PROPOSED — NOT IMPLEMENTED. Idea parking, 2026-08-30. Owner: delan Scope: conversation net after a from-scratch ask. Story/bible is the only locked input. Genre, card-ness, and the old 5×5 are open. Not a plan. Nothing here is scheduled. Laser tag does not drive these forks; it is a mild tiebreaker only, parked in docs/proposals/laser-tag-crossover.md.

The constraint

The world is good. That is the thing not to throw out.

A first game, made by someone who is not a 3D designer, who likes roguelike card games, and who can ship systems in 2D, is still the sensible factory. Fun is the test. “Never seen in STS or MTG” is not.

Chasing a card verb those games missed is how you get a sixth energy type and a stranger who still says “it’s like Slay the Spire.” The games that felt new in a crowded genre changed the question, not the suit.

  • Balatro — what if the deckbuilder is poker?
  • Inscryption — what if the card game is lying to you about being a card game?
  • Monster Train — what if the board is a three-floor defense?
  • Papers, Please — what if the job is the game?

EchoSpire already has a question the genre does not:

What if a run is both a job and a document, and the next run has to live inside the official version of the last one?

That is the uniqueness hunt. The toy you put in the player’s hands is whichever 2D system makes that question fun.

What I would not do

  • A clean 3D action game, live-service, 4X, or open-world Spire. Wrong factory.
  • Inventing a card resource STS “forgot.” Block, energy, orbs, poison, stances, locations, companions — the nouns are taken. A new noun is not a new game.
  • Shaping the video game so a laser tag venue can exist. The venue might never exist. If two ideas are equal and one would also play on a pack, that may sway a later pick. It does not pick the genre.
  • Throwing out the story to chase a mechanic. The story is the mechanic’s floor.

Four forks

Pick a question. The factory (cards, map, 2D) can serve any of them.

A. The run writes the lie (stay a card roguelike)

You do a House job. Combat can stay cards. At the end of the run you file. The House (or the Censors, or the sim) writes the official account. Your next run’s map, modifiers, and even which enemies exist are seeded from that filing, not from what you actually did.

You are fighting inside a history you helped falsify.

Fun: the distortion is a toy. File clean and the next rift is fair. File a Valerii victory over a mess and the next map has walls that were never there. Salvage the true log from downstairs and you can spend it to correct one node — the humble record as a card you steal from yourself.

Why this is new: most roguelikes forget the last run except for meta unlocks. Here the last run is the content. The bible’s second pillar is the loop.

What you keep: rift map, combat engine, Houses, one craft.
What you throw out: the idea that a seed is innocent. 584 cards as the identity.
Hard part: the filing has to be readable. If the player cannot see what they falsified, it is noise.

B. Company (stay a card roguelike)

Every death already wants to leave an echo. Lean in until that is the game.

A failed run does not restart the world. It leaves a body of possibility in the next one. You can salvage it (Salvari), redact it (Censors), freeze it (Valerii), or fight beside it (the Double craft). The horror of that night was not debris. It was company. The roguelike loop becomes “die and meet yourself.”

Fun: Hades does this as conversation. Almost nobody does it as a unit you can use or that can kill you. Thickness-and-flip (from four-crafts.md) is one toy. It does not require four crafts. It requires one ghost that matters.

Why this is new: the death is not a reset. It is a spawn.
What you keep: runs, cards or not.
What you throw out: “you wake up at the start with nothing but unlocks.”
Hard part: UI. A ghost the player cannot read is a gotcha.

C. Four physics, thirty cards (still cards, not STS)

Stay cards. Stop being a deckbuilder in the STS sense.

No 584-card economy. No shop full of Strikes with hats. Four crafts, ~30 cards that are the physics: Counterweight, Berth, Stand-in, Strain. A run is a short toolkit of overrides, not a constructed deck. The puzzle is the night’s puzzle: locally safe moves that globally tear.

Fun: every card is a rule-break. There is nothing generic to hide behind. Turns are short and loud.
Why this is new: it looks like a card game and plays like a puzzle about a machine that obeys too many people.
What you keep: cards as the input device (easy to code, you like them).
What you throw out: most of cards.json as identity. Relics-as-STS. The 15+5 starter-deck law.
Hard part: enough variety across a campaign without becoming a constructible deck again. House price on a tiny kit has to bite hard or the Houses vanish.

D. Grease (not a deckbuilder)

You are not a hero with a deck. You are downstairs. Jobs arrive as chore lists. The map on the briefing is wrong. Combat, if it exists, is small — a few cards, a simple clash, or even a timing check — because the toy is noticing: wear that contradicts the corridor, a log the Censors did not think to take, an echo nobody else looks up at.

Fun: Papers, Please and Return of the Obra Dinn are fun. They are a different fun. If the itch is “build a deck and slam,” this fork will feel like a betrayal. If the itch is the world, this is the most honest shape.
Why this is new: the genre neighbors are not STS.
What you keep: the bible, 2D, authored jobs, Houses as employers.
What you throw out: the combat engine as the product. Most of the card work.
Hard part: first-game scope in a noticing game is writing and level-craft, which is the skill you said you do not have. AI volume will make this one mushy faster than it will make a card kit mushy.

A fifth, if the factory must stay exactly what it is

Keep the current loop. Put the uniqueness in the opening hour.

The orientation is a complete, fair, beautiful golden-age game — the Spire working. Fun on purpose. Then the real campaign is the same systems with the lights wrong, the map lying, the ghost in the room. Players feel the loss because they played what was lost.

That is not a new genre. It is the pillar as structure. It works on A, B, or C. It does not work on a slideshow that explains the Fall.

Recommendation (not a ruling)

Do not leave 2D systems. Do not leave “a run.” Do not start a 3D project.

Do A or B on top of C’s diet.

  • Cards stay, because they are the factory you can actually run.
  • The card pool gets small and radioactive (four crafts, or even one craft in the demo).
  • The new thing is either the filing (A) or the ghost (B). Both grow out of canon. Both are fun to imagine. Both can ship without a venue.

If those two ever score a dead heat, B is slightly easier to show in a trailer, A is slightly more the pillars. Laser tag would slightly prefer B (a ghost on the floor is already a light). That is not a reason to pick B.

D is the honest world-game and the wrong first game.

Open

  1. Which question is the product: the filing, the ghost, the tiny physics kit, or downstairs noticing?
  2. Does the existing combat engine stay the verb, or become the punctuation?
  3. Demo still “orientation plus inductions,” or demo is one of these questions in twenty-five minutes?

Parking only. Companion to docs/proposals/four-crafts.md (the combat verbs if we stay cards) and docs/proposals/laser-tag-crossover.md (venue ideas that must not drive this file).