# Master of the Court — Game Design Document **One line:** You are the power behind the throne of a vain king. Keep his trust above zero for a whole reign by trading favours between four powers — knowing every favour hands someone a foothold in your palace. This doc describes the rules as shipped. The game is config-driven: every number below lives in a data table (start state, deck, sources, crises, tuning constants), so balance changes are data edits, not rule changes. --- ## 1. Win and lose - **Lose:** the **King's Trust** reaches 0 at any moment. Nothing else can kill you. - **Win:** survive **80 beats** of court time (a "beat" ≈ 1.1s of real time; the clock **pauses whenever a decision is open**, so a reign is as long as your deliberation). There is exactly one terminal meter. Everything else exists to push it. ## 2. The meters All meters except the Treasury run **0–100** and clamp at both ends. | Meter | What it is | Why you care | | --- | --- | --- | | **King's Trust** (starts 48) | The lose condition | Drains every decision; below 0 = death | | **Treasury** (starts 120, floor 0, no cap) | Coin | Gates options; emptying it with a spend costs **−6 Legitimacy** (once per emptying) | | **Nobility / Clergy / Merchants / Commoners** (each 50) | The four factions' standing with you | Any faction **≤ 24** presses the king every decision | | **Legitimacy** (50) | The crown's lawful standing | **≤ 24** presses the king | | **Prestige** (50) | The realm's glory — the king's obsession | Main driver of his mood | | **Stability** (50) | Order in the realm | **≤ 20** presses the king | The court layer (section 4) adds three **needs** — Provisions 64, Security 64, Splendour 54 — and a **palace control** % per supplier faction (Nobility 14, Clergy 14, Merchants 10 at start). ## 3. Loop one — Decisions A decision card surfaces from a 45-card deck **every 12 beats** (deck reshuffles when exhausted), plus event cards from the court layer (section 5). Each card offers 2–3 choices; a choice carries some mix of: faction/realm effects (`fx`), a coin delta, a need refill/drain, a control shift, and a follow-up card it can chain into. **Resolution order for a choice:** 1. **Coin** applies (Treasury floors at 0). 2. **Meter effects** apply (clamped 0–100). 3. Any **need/control** side-effects apply. 4. **Bankruptcy check:** if a *spend* left the Treasury at 0 → **Legitimacy −6** (with an outcome line; fires once per emptying). 5. **The king's mood** (below) adjusts his Trust on top of any direct `king` effect. 6. **Ambient erosion:** Trust −2 flat (his favour must be re-won every decision), **plus −3 per meter currently in its danger band** (any faction ≤ 24, Legitimacy ≤ 24, Stability ≤ 20). The outcome line names the offenders. **The king's temperament.** He is vain: money never moves him, glory does. `mood = round(0.6·ΔPrestige + 0.3·ΔLegitimacy + 0.3·ΔStability)`, and **losing prestige stings extra** (a further 0.4·ΔPrestige when negative). So a choice that drops Prestige by 5 costs ~5 trust on mood alone, while gaining 5 earns only +3. Designers: this asymmetry plus the flat −2 is the anti-pandering engine — you cannot coast on his approval; you can only slow the bleed. A practical consequence worth internalising: **the net trust change of a "neutral" decision is −2 or worse.** Trust is a resource you spend by existing; the game is about choosing what to buy with it. ## 4. Loop two — The Court (real time) Beneath the decisions, the household runs on the clock. Each beat: **Needs decay.** Provisions −4, Security −3, Splendour −4 (Splendour decays an extra 0.6 per bare or empty palace room, and gains back +0.25 × the furnishing prestige of each seated room — a furnished palace partly self-maintains its glitter). **Sourcing — the central bargain.** The *only* way to refill a need is to pick a supplier from its row. Every supplier is a trade: a refill, a coin price, faction effects, and **palace control handed to whoever helped**. The full table: | Need | Source | Refill | Coin | Control | Other effects | | --- | --- | --- | --- | --- | --- | | Provisions | Merchant grain | +34 | −18 | Merchants +6 | Merchants +3 | | | Noble estates | +30 | free | Nobility +8 | Nobility +3, Commoners −2 | | | Church tithes | +28 | −6 | Clergy +7 | Clergy +3 | | | Crown granary | +42 | −32 | none | Legitimacy +1 | | Security | Noble guards | +34 | free | Nobility +9 | Nobility +4, Stability +2 | | | Church men | +30 | −8 | Clergy +7 | Clergy +3, Stability +2 | | | Hired swords | +36 | −26 | none | Stability +3 | | Splendour | Merchant luxuries | +34 | −22 | Merchants +6 | Prestige +3, Merchants +2 | | | Noble artisans | +32 | −10 | Nobility +6 | Prestige +3, Nobility +2 | | | Plain & pious | +24 | −4 | Clergy +6 | Clergy +3, Prestige −2 | Design intent: the free options are the trap (they feed Nobility control fastest); the independent options (Crown granary, Hired swords) are control-clean but expensive; "spread your custom" is the core skill. **Seating & furnishing.** The palace has 4 rooms — 2 inner, 2 outer — around the throne. You may seat Nobility, Clergy, or Merchants in a room (Commoners cannot be seated). Each beat, a seated faction gains control: `+0.45 × ring weight × (1 + 0.4 × furnishing level)`, where inner ring weighs 1.6 and outer 1.0. Unseated factions ebb −0.2/beat. Rooms cycle bare → fine (25 coin, +3 Prestige on upgrade) → lavish (55 coin, +6 Prestige) → back to bare (free). Seating is the slow, controllable dial on both control and Splendour; sourcing is the fast, lumpy one. **Tick pressures.** Each beat: Provisions at 0 → Commoners −1 and Stability −1. Security at 0 → Stability −1. Splendour ≤ 12 → Prestige −1 **and King's Trust −1** (the shabby court insults him directly — this is the fastest passive death in the game). Splendour ≥ 72 → every 4th beat Prestige +1 and Trust +1 (the only passive trust income that exists). Merchants' control ≥ 70 → every 3rd beat Treasury −3 (they're skimming). ## 5. Event cards from the court The court loop summons cards on top of the scheduled deck (cards queue if one is already open): - **Crises** — fire once when a need hits 0; re-arm when it recovers to ≥ 30. Each offers a costly fix, a fix that hands a faction +10–12 control, or a "ride it out" that takes meter damage. - **Overreach** — fires when a faction's control reaches **70**; re-arms below 55. Nobility's version can chain into **open rebellion** if ignored (the one card that can swing −22 Nobility / −70 coin); Clergy's strips your Splendour; Merchants' is a shakedown. Control ≥ 70 is also surfaced to the player as a threat line — treat 70 as the game's loudest red line. ## 6. Pacing shape of a run 80 beats ÷ 12 = **6–7 scheduled decisions** per reign, plus crises/overreach you let happen, plus the guided first-run card. With ~−2 trust minimum per decision from 48 starting trust, a clean run spends roughly a third of the king's trust on ambient erosion alone — the rest is yours to win or lose. The intended difficulty: a player who tends Splendour, spreads suppliers, and never lets a meter into its danger band wins comfortably; two simultaneous danger-band meters (−6/decision extra) is usually a death spiral. ## 7. First run A guided tutorial card ("A Petition at Court") opens the first game: a deterministic 2-choice decision (Grant: Nobility +8, Commoners −5, Prestige +3 → mood +2 · Refuse: Nobility −5, Commoners +5, Legitimacy +2 → mood +1) used to teach effect tags and the mood line. The court clock is paused throughout the tour; it's replayable from the header. ## 8. Notes & loose ends for designers - **`armedNobles` is currently cosmetic.** The Noble guards source sets this flag, but nothing reads it — rebellion only triggers via the Nobility-overreach "look away" chain. Either wire it (e.g. rebellion severity, or gate the overreach card) or cut it. - **Commoners control exists in config (starts 6) but is unreachable** — they can't be seated and no source grants it. Vestigial; safe to ignore or repurpose. - **All tuning is data.** Start values, the deck, sources, crisis/overreach cards, furnishing tiers, decay, and the temperament weights are flat config tables. The reskin (`family/`) proves the rules carry to any setting untouched. - **Button convention:** controls display *current state*, not the pending action.