Slay the Spire
What it is. Climb a branching map, fighting turn-based card battles. After each fight, draft one card of three (or skip). Your deck is your build; the run is one long refinement of it against enemies whose damage you can read in advance.
Player fantasy / why it’s fun. I am sculpting a machine and watching it come alive. The pull is the deck as an evolving strategy: a pile of weak cards slowly becomes an engine, and the moment two cards click into a combo is the whole genre’s dopamine. Every choice — which card, which path, which relic — compounds.
Design DNA
Make the growing toolkit the game. The player doesn’t level a character; they curate a deck, and every reward is a fork: take this card (and dilute your draw) or skip it (and stay lean). Wrap that in a route of pre-visible risk — a branching map where you choose your fights, elites, shops, and rests before you take them — and escalating enemies whose intents are shown, so losing is a planning failure, not a coin flip.
The magic is subtractive as well as additive: a good deck is often a small deck. Thinning is a strategy, not just adding. This is the counter-intuitive core — most progression systems only grow, so “the reward you decline” and “the card you delete” are dead concepts. Slay the Spire makes both live decisions: the skip has teeth because every card you add dilutes the odds of drawing the ones you already want, and card-removal at shops is often the strongest purchase. Power is a function of consistency, and consistency is a function of restraint.
Every layer — the draft, the map, the relics, the energy budget — is a decision made against incomplete future information but perfect present information. You never know the next reward, but you always know exactly what this fight will do.
Load-bearing structures
| Structure | Why it works |
|---|---|
| Draft-of-3-with-skip | Every reward is a real decision — including the decision to not dilute. The skip is what makes drafting have teeth. FUN.md truth. system-reward-schedules. |
| Deck as build; thinning matters | Power comes from composition, not stats. Removing weak cards is as strong as adding good ones. system-build-diversity. |
| Pre-visible branching map | You choose the route (elite for a relic? shop? rest?) before committing — pattern-risk-reward made spatial. system-procgen-design. |
| Telegraphed enemy intents | Each enemy shows next-turn damage/effect; combat is a solvable puzzle, not a gamble. Intent honesty is CI (FUN.md). system-telegraphs. |
| Relics = passive multipliers | Persistent modifiers that warp the whole run’s math and reward committing to an archetype. pattern-emergence. |
| Energy as the real constraint | Cards are cheap; the per-turn energy budget forces every hand into a small optimisation. system-economy. |
| Win-rate window as the balance instrument | Tune to a band (e.g. 11–19/20), not a point — both “too hard” and “too easy” break CI. |
What to steal
- Draft-of-N with a skip. The skip converts “free reward” into “meaningful decision.” Steal this into any progression, not just cards.
- Make composition the power, and let subtraction be a strategy. A leaner deck/loadout should sometimes beat a bigger one.
- A pre-visible route of risk. Show the map’s fights/rewards ahead; make the player commit to a path. This is portable to any run structure.
- Telegraphed intents so every fight is solvable. Assert intent honesty as a one-line audit (FUN.md).
- Balance to a win-rate window (e.g. 11–19/20), not a point — both edges break CI, since “too easy” is as broken as “too hard.” Assert the draft delta: same pilot drafting-off must lose far more (FUN.md’s 17→9).
- A per-turn budget (energy) as the real constraint. Cards being cheap to
acquire but limited to play per turn means every hand is a small
optimisation puzzle — the budget, not the card pool, is where tactical depth
lives. See
system-economy.
What’s just theme (drop it)
- The spire / the Ironclad-Silent-Defect fiction. Cosmetic; the loop is theme-agnostic.
- Playing cards specifically. The structure is “draft-refine a personal
toolkit of single-use actions.” It could be spells, dishes, dance moves —
see
system-build-diversity. - The relic art and names. Passive run-warping modifiers is structural; the trinkets are flavour.
- Fantasy monsters. Any enemy that can show its next move satisfies the
telegraph truth —
system-telegraphs.
Composes into
genre-deckbuilder— its canonical anchor; the draft-refine loop lives there.system-build-diversity— deck-as-build with meaningful thinning.system-procgen-design— the branching, pre-visible map.system-reward-schedules— draft-of-3-with-skip cadence.system-telegraphs— intent-driven, solvable combat.
Twist seams
- Slay the Spire but the deck is shared and you draft against an
opponent (perspective) — every skip becomes a card denied to them; the
route is a contested board. Pushes toward
system-coop-and-competition. - Slay the Spire but the map is the deck (mechanic-swap) — you draft
rooms onto a path and then walk them; placement is the build. This is the
anchor-loop-herobend. - Slay the Spire but you can only ever hold 10 cards — every add forces a cut (constraint) — makes thinning mandatory, not optional; the whole run is a rolling replace-decision. Sharpens the subtractive pillar into the core verb.
- Slay the Spire but tonal — the deck is a diplomatic hand and fights are
negotiations (tonal) — recolor combat as conversation; “damage” is
persuasion, “block” is patience, intents are the other party’s next
argument. Keeps the draft-refine spine, swaps violence for stewardship.
Pairs with
genre-narrative-decisions.
See also
genre-deckbuilder·anchor-balatro·anchor-loop-hero·system-telegraphsdesign/FUN.md#11-·-roguelike-deckbuilder— win-rate window + draft-delta + intent audit.examples/lumen/— generate-and-prove content: every board solver-verified before it’s served.sandboxes/procgen-lab/—Rng/pickEntryfor weighted card & map generation.