Game Concept: Iron & Timber — The Muster

Created: 2026-09-12 Status: Draft

Elevator Pitch

It's a frontier city-builder where you manage a deep medieval economy in a
monster-haunted wilderness — and every tool you forge is also a weapon, every
worker you employ is also a soldier. When the horn sounds, your supply chain
becomes your army.

Core Identity

Core Fantasy

My town's strength is legible in its stockpiles.

The player is a frontier steward whose settlement is simultaneously an economy and a militia. There is no separation between the people who produce and the people who fight — the woodcutter who has been feeding the sawmill all summer is the same person standing on the palisade in autumn, carrying a spear the player chose to forge instead of a plough.

What the player can do here that they can't elsewhere: **look at their armory and know whether they survive the winter.** In Banished the wilderness is indifferent — it starves you but never comes for you. In Manor Lords the battles happen somewhere adjacent to the economy rather than emerging from it. Here, military readiness is a direct, visible readout of economic decisions made weeks earlier.

Unique Hook

**Like Banished, AND ALSO your entire population is a conditional militia whose combat effectiveness is a direct readout of your economy's health.**

There is no "recruit soldier" button. There is iron ore, a charcoal kiln, a smelter, a blacksmith, and a person who stops being a farmer. Every point of military capability was paid for in economic capability that no longer exists.

The hook affects gameplay at every level: the 30-second loop (does this iron become a plough or a spearhead?), the session loop (is the armory full enough for autumn?), and the progression loop (can I afford a standing guard yet, or does everyone still need to do double duty?).

Player Experience Analysis (MDA Framework)

Target Aesthetics (What the player FEELS)

Key Dynamics (Emergent player behaviors)

• Players will hedge their gear production against expected threat level, developing a personal read on "how much military is enough."
• Players will specialize districts and develop layout signatures — a personal style of settlement that they recognize as theirs.
• Players will form attachments to individual settlers who survive multiple raids wearing gear the player personally decided to forge.
• Players will push into dangerous terrain deliberately once they feel armed, converting military surplus into economic opportunity.
• Players will narrate their failures ("I under-armed autumn and lost the east field") because failure traces to a specific, legible decision.

Core Mechanics (Systems we build)

• Production chains — multi-step material transformation (wood → charcoal → smelting → smithing) where intermediate goods are visible and stockpiled.
• The dual-purpose output system — every smith output is categorized as economic (ploughs, tools, nails) or military (spears, mail, shields), drawing from the same input pool. This is the central tension mechanic.

Two military sinks minimum, from MVP *(added 2026-09-15 after the independent playtest)*. Spears arm a defender and are capped at the population that can carry them. Mail protects one and is not capped the same way — it reduces losses rather than adding defenders. The second sink is not polish; without it the loop stops working. When the playtest's settlement halved, the spear cap fell with it, four paid-for spears were stranded, food was already met, and iron had nowhere left to go for a season and a half. The tester's unprompted words: *"its kinda frustrating. i want more diversity in weapons and tools."* A single military sink closes the moment the population shrinks, which is exactly when the player most needs something to do.

• The muster — an alarm state that converts assigned workers into armed defenders based on available gear in the armory. Pulls them off work temporarily; the real cost was paid at the forge.
• Telegraphed directional threat — escalating, visible, readable danger that approaches from specific map directions on a known cadence.
• Seasonal cycle — a four-season rhythm that structures production capacity, threat timing, and the natural session boundary.
• The muster halts production *(added 2026-09-14, design review; bounded the same day after a systems review found a death spiral)* — a mustered settler stops working for as long as they are under arms. This is the second axis the allocation decision needs: forging spears costs ploughs before the raid, and calling the muster costs production during it. A player who over-arms pays twice; a player who under-arms loses the wall. It makes "the economy IS the militia" literally true at the moment of crisis — the militia is the workforce, absent from its jobs.

Bound — the reserve. At least 2 settlers always remain on production and cannot be mustered. Without this floor the mechanic is a death spiral: a settlement that must arm everyone produces nothing between raids, so it can never reforge lost gear or recover, and each raid is unwinnable-er than the last. The reserve guarantees a rebuild path exists at every population level. It is deliberately small — two settlers is a trickle, not a recovery, so the muster still hurts.

Income model. Iron income scales with working settlers, not total population. A mustered settler produces nothing; a dead settler produces nothing permanently. This is what makes both the muster and settler loss economically real. *(Supersedes the prototype's flat 3/turn, which contradicted its own rule that dead settlers stop producing.)*

Gear on the fallen. A dead defender's gear returns to the armoury damaged, needing iron to repair before it can arm anyone again. Full recovery would make spears permanent capital and recreate the dominant strategy the design is trying to kill; total destruction would make every loss doubly punishing and feed the spiral. Damaged-and-repairable keeps the displaced cost intact and gives iron a recurring sink.

• Uncertain raid timing — the player knows a raid is coming and roughly how strong, but not exactly which day. Added 2026-09-14 after the muster prototype: the tension in play came overwhelmingly from when, not from how much, and a fully known raid date would collapse the allocation decision into a calculable optimum ("pivot to spears on turn N"). This is in productive tension with Pillar 3, which promises telegraphed threat — the resolution is that the player can always see a raid approaching and gauge its scale, but cannot compute the exact turn to stop building ploughs.

Player Motivation Profile

Primary Psychological Needs Served

Player Type Appeal (Bartle Taxonomy)

• [x] Achievers (goal completion, collection, progression) — How: completing production chains, filling the armory, securing the valley. Progression is in options unlocked, not raw power numbers.
• [x] Explorers (discovery, understanding systems, finding secrets) — How: understanding the economy's interlocking depth and the threat escalation model. Later tiers add literal frontier exploration.
• [ ] Socializers (relationships, cooperation, community) — Not served. Single-player with no social systems.
• [ ] Killers/Competitors (domination, PvP, leaderboards) — Not served, and deliberately so. Combat is light-touch; competitive players will bounce.

Quantic Foundry mapping: primary appeal is Design + Completion (the builder motivation cluster), secondary is Strategy + Challenge.

Flow State Design

• Onboarding curve: The first 10 minutes are pure builder — place a cabin, a woodcutter, a field. No threat appears until the player has a functioning micro-economy, so the economic layer is understood before it's stressed.
• Difficulty scaling: Threat escalates on a visible, telegraphed schedule tied to seasons and to the player's own expansion. Pushing into dangerous terrain is the player's choice to raise difficulty.
• Feedback clarity: The armory is the master readout. A player can see at a glance whether they are prepared. Production chains show work-in-progress visibly so the player understands cause and effect.
• Recovery from failure: Losing a raid costs resources, structures, and settlers. The player rebuilds from a diminished state. *(Revised 2026-09-14 — the earlier promise that the settlement is "never wiped" was untestable and left the game with no defined way to lose. See Failure State below.)*

Failure State (added 2026-09-14, design review)

The settlement is abandoned when the population falls below 3 settlers. The run ends. There is no rebuilding from one survivor.

This exists because a game about consequence needs a way to lose. Every other stake in the design — the armory readout, the allocation tradeoff, the muster — is measured against this floor. Without it, "under-armed" and "over-armed" both resolve to "somewhat worse for a while," which is what the first prototype measured and why its verdict was withdrawn.

Losses are clamped at the floor. A single raid can reduce the settlement to 3 settlers but never below it: Losses = min(Raid Strength − Defenders, Settlers − 3). Reaching exactly 3 is survivable; the run ends only when the settlement *cannot sustain itself* from there and drops below on a subsequent raid. This exists because the unclamped formula skipped the floor entirely — 4 settlers facing a strength-3 raid with no defenders went straight to 1, and the failure state a player was supposed to see coming arrived without ever being approached.

Evaluation order at the moment a raid resolves:

• Compute losses with the clamp above.
• Remove settlers; their gear returns to the armoury damaged.
• Check the floor. Below 3 → the run ends.
• Survivors stand down and return to work next turn.

Design notes:

• The floor is a backstop, not the usual outcome. A well-played settlement should rarely approach it. Its job is to make settler loss matter, not to end runs frequently.
• Settler death is permanent. A lost settler takes their role and experience with them, and the labour they would have done is gone.
• This is the Banished comparison made real: population collapse is the failure mode that genre already teaches players to fear.

Core Loop

Moment-to-Moment (30 seconds)

Provisioning. Assigning labor, routing materials, placing buildings. The core act is deciding what a resource becomes: this iron becomes ploughshares or spearheads; this leather becomes work boots or a jerkin. Each micro-decision is a small bet on how dangerous the coming season will be.

Intrinsically satisfying because it is legible — the player sees the stockpile move, sees the smith begin work, sees finished gear rack up in the armory.

Short-Term (5-15 minutes)

A production chain completes. The charcoal kiln finally feeds the smelter, which feeds the blacksmith, and a finished spear appears on the armory rack. This is the "one more building" hook: the player always wants to close one more loop. Because the armory is simultaneously an economic output and a defensive readout, watching it fill delivers two satisfactions at once.

Session-Level (30-120 minutes)

One seasonal cycle. Spring build-out → summer expansion into dangerous terrain → autumn raid and muster → winter scarcity, repair, and planning. Each season boundary is a natural stopping point. The session ends with the player knowing the state of their armory and the shape of next year's threat.

Long-Term Progression

Frontier establishment. Growth is in options, not raw power. Early game: one settlement doing everything badly, everyone doing double duty. Late game: specialized production districts, a standing guard that frees workers from militia duty, outposts claiming the iron valley. The game is "done" when the frontier is secure enough that raids are an annoyance — and then the player pushes further out and makes it dangerous again.

Retention Hooks

• Curiosity: What's in the unmapped valleys? How does the threat escalate next year?
• Investment: The settlement the player designed, and the named settlers who have survived in it.
• Social: Not applicable (single-player).
• Mastery: The gear-vs-tools allocation curve has real depth. Reading the frontier correctly is a skill that visibly improves.

The hook that follows the player off the couch: *I know my armory is thin and autumn is coming.*

Game Pillars

Pillar 1: The Economy Is the Army

Every unit of military capability traces back to a production decision. There is no "recruit soldier" button; there is iron, a smith, and a person who stops being a farmer.

Design test: If we're debating whether a military feature should cost economic resources or be purchased directly — it costs economic resources, always.

Pillar 2: The Settlement Is a Creative Act

The town is the player's expression. Layout, specialization, and aesthetics matter, and the game should never punish a player for building something beautiful over something optimal.

Design test: If we're debating between a system that adds depth and one that adds constraint on layout freedom — we choose depth, and leave the layout to the player.

Pillar 3: The Wilderness Has Intent

Unlike Banished's indifferent nature, this world contains things that want your stuff. Threat is directional, escalating, and readable — the player should always be able to see danger coming and prepare for it.

Design test: If we're debating between a random punishing event and a telegraphed escalating threat — always telegraphed. Surprise is not the same as difficulty.

Where the Pillars Actually Collide

*(Rewritten 2026-09-14. The previous text claimed "productive tension" but resolved every case with the same tiebreaker, which meant the pillars agreed rather than conflicted. A tension needs a case where two pillars cannot both be honoured.)*

The real collision: Pillar 1 vs. Pillar 2, at the moment the horn sounds.

Pillar 1 says every unit of military capability costs economic capability. With the muster now halting production (Core Mechanic #6), calling the militia actively stops the builder's town from working — which Pillar 2 would prefer never happened. Both cannot be honoured at once.

Pillar 1 wins, and the design states that explicitly. The muster costs production. What Pillar 2 protects is narrower and now precise: the player never loses the ability to rebuild, their layout is never invalidated, and the failure state is a floor they can see coming — not a surprise. Pillar 2 is a promise about recoverability, not about comfort.

Anti-Pillars (What This Game Is NOT)

• NOT a tactical combat game: Combat is light-touch command resolved in minutes. Deep tactical systems would compete with the economy for both development time and player attention, compromising Pillar 2.
• NOT a game where the muster ends a run (revised 2026-09-14): Calling the militia costs production while settlers are under arms, and a lost raid costs settlers — but the muster itself never triggers the failure state directly. The player always gets to act on the consequences. *(This replaces the earlier "the muster never punishes the builder", which combined with a free muster and Anti-Pillar 4 removed every consequence from both branches of the decision. A muster with no cost is not a decision.)*
• NOT a game that uses randomized disaster events as difficulty: No surprise plagues, no unannounced raids from nowhere. Every threat telegraphs. This protects Pillar 3.
• NOT a game that gates creative building behind combat performance: Losing a fight costs resources and people, never the ability to keep building the town the player wants. This protects Pillar 2.

Visual Identity Anchor

*This section is the seed of the art bible. It captures the visual decisions made during concept development before they can be forgotten between sessions.*

Revision 2026-09-13. The original anchor ("Weathered Almanac", woodcut/linocut
illustration) was replaced after the first keyframes read as Don't Starve-style
storybook survival rather than a strategy game. The register below is the one the
art bible must formalise. The superseded set is kept in
docs/site/art/superseded-woodcut/ for reference.

Direction: Stylized Strategic

The Northgard / Age of Empires IV register: chunky, readable, low-detail 3D models with clean silhouettes and painterly soft shading, seen from a high three-quarter strategic camera that can rotate and zoom. Grounded, saturated colour — not cartoon, not photoreal. Chosen because this is a real-time strategy game where the player commands a settlement, and the art must say "these are decisions you can read at a glance," not "this is a fairytale to survive."

One-line visual rule: *Readable at strategic zoom; threat is told by light and
colour temperature, never by UI alone.*

Supporting Visual Principles

1. Warm inside the walls, cold outside them. The settlement is lit by firelight, hearths, torches and gold-hour sun. Threat is signalled by the world outside the palisade going cold and desaturated — blue-grey fog rolling in from the map edge, torches appearing between dark pines. The player should be able to feel a raid coming from the colour of the treeline before any icon appears.

Design test: If a threat can only be perceived through a UI element, the scene is under-lit. Fix the lighting before adding an indicator.

Ordering rule *(added 2026-09-14 — the systems index lists a Threat Forecasting UI and an Alarm system, which read as a contradiction without this)*: world lighting is the leading indicator, updating continuously and readable without opening anything. The UI systems are lagging confirmations — the forecast shows a band the player could already infer from the treeline, and the alarm fires only at the discrete instant the muster becomes available, which has no continuous visual analogue and is legitimately UI's job. Without this ordering, whichever ships first becomes the real channel by default and the lighting work turns decorative.

Accessibility. Warm/cold is hue-and-saturation, which protanopia and deuteranopia flatten — roughly 8% of male players. Temperature must therefore never be the only carrier: cold areas are genuinely darker, not merely bluer, and fog density and particle motion carry the same signal geometrically. Verify against a colourblind simulation before art lock, not after.

Directional threat under a rotating camera. Threat arrives from a world-space direction, but the player can rotate away from it. A screen-space edge indicator keeps the direction correct at any yaw. Unresolved, this is a bug found in playtesting rather than a decision made in design.

2. Silhouette first. Every building, settler role and piece of gear must be identifiable from its outline at the default camera distance. A spear-carrier and a plough-carrier are different shapes, not different textures.

Design test: Render it in flat grey. If two things you need to tell apart look the same, the model is wrong, not the material.

3. Seasons change the world, not just the tint. Spring is fresh green and running water; summer is dry gold and dust; autumn is amber fields and long shadows; winter buries the roofs and leaves one warm glow per hearth. Per the architecture research this is a hybrid: gradient-map/LUT tinting for ground, water and ambient light, plus a small set of authored variants (tree canopies, snow caps) where the silhouette genuinely changes.

Design test: Hide the UI. A screenshot must name its season without a caption.

The Armoury Readout *(specified 2026-09-14 — named five times in this

document as the central feedback device, never defined until now)*

A persistent HUD strip, not a building the camera must visit. "Look at your armoury and know whether you survive" fails the moment it requires navigation.

It shows a capability, not an inventory — a spear count alone tells the player nothing about whether it is enough:

• Gear ready — repaired and usable, distinct from damaged stock awaiting iron
• Defenders available — settlers who could muster, excluding the 2-settler production reserve
• Coverage against the current threat band — a qualitative read (thin / adequate / strong), never a number, because an exact figure would let the player solve the allocation (see Core Mechanic #7)

During a muster it also shows the live cost: idle workshops and unworked fields, counted, updating as it happens. Without this the production halt is theoretically real and experientially invisible — which is precisely how the free muster failed.

Population carries its own warning. At 6 settlers or fewer the population counter changes register, and the first time a raid could plausibly drop the settlement below the floor, it says so plainly. A builder audience losing a run by surprise is the worst outcome this design can produce, and Pillar 3 forbids it.

Colour Philosophy

• Warm gold — economy, prosperity, firelight, the inside of the walls
• Iron grey — gear, weapons, armour, the smith's output
• Cold blue-grey — threat, the outside, winter scarcity
• Rust — the alarm register: the horn, the emblem, UI states that demand action

Reference keyframes: docs/site/art/ (FLUX 1.1 Pro, 2026-09-12/13).

Deliberate Exclusion: No Currency or Market

(Recorded 2026-09-14. Previously an absence; now a decision.)

There is no money, no trade, and no market at any scope tier. Ostriv is cited above for economic depth and does have trade — so this is a real divergence, not an oversight. The reason: a market is a universal sink that lets a player convert any surplus into progress, which would let them buy their way out of a bad allocation. That is the same escape hatch Pillar 1 exists to close.

The cost of this exclusion is that every resource needs its own internal sink, or surpluses stagnate. That is why gear decay is listed as a design risk above rather than an optional nicety — without a market, decay is the sink.

Inspiration and References

Non-game inspirations: Medieval farmers' almanacs and woodcut illustration (visual direction); the historical reality of frontier settlements where militia duty and agricultural labor were the same people's work (core fantasy); the logistics of pre-industrial ironworking, where charcoal production and iron smelting were inseparable economic activities.

Target Player Profile

Technical Considerations

Risks and Open Questions

Design Risks

• The muster may not feel meaningful with the cost displaced to gear. This is the central design bet. If players don't feel the weight of the forge decision weeks later when the horn sounds, the hook fails. Mitigated by making the armory a prominent, legible readout — but only the MVP can actually answer this.
• Combat may feel like a spreadsheet check rather than a battle. Light-touch combat risks being so abstract that the payoff moment lacks drama.
• The economy may not be deep enough to satisfy Ostriv/Banished players within MVP scope, making the MVP a poor test of the real game.
• The Manor Lords coupling trap. Manor Lords did couple economy and military — militia drafted from the worker pool, gear from production, food upkeep — and its combat was still called "stapled on" and "an afterthought", because the coupling is a resource-cost gate in front of a separate RTS minigame. Players' standard advice is to disband militia quickly because idle troops are a chore. It kept a ~90% Steam rating partly because a no-combat mode ("Rise to Prosperity") lets the builder audience opt out entirely. Iron & Timber has no such escape hatch by design, so the muster must be the economy acting — not a purchase made from it — and the defence layer must carry its own weight. (Market research, 2026-09-12.)
• Forge lead time is validated and load-bearing (playtest, 2026-09-15). The independent run's season 1 is the proof: the raid arrived on the earliest legal turn with eight spears paid for and none ready, and six of twelve settlers died. The regret was specific and traceable to a decision three turns earlier. This is no longer a hypothesis; it is the mechanic. It must survive translation to real time, where it becomes a smith visibly working rather than an arrival counter.
• The economy has faucets but no sinks. (Design review, 2026-09-14.) Spears never degrade, so once the armoury is full, military spending stops forever and the late game flatlines by construction. Unspent iron is inert. A dead settler's gear is unspecified — recovered undermines the displaced-cost premise (the iron was never really spent), lost compounds every raid's severity. Decide deliberately before Vertical Slice. Candidate fixes: gear lasts 2–3 raids before re-forging; fallen defenders' gear returns damaged, needing iron to repair; obsolete gear scraps back at a fraction of its iron.
• Raid strength must be able to exceed any reachable defence. (Design review, 2026-09-14.) The first prototype capped raids at 2d6 (max 12) against a maximum 15 defenders, which made full armament a guaranteed win and the whole allocation computable. Either raid strength scales with settlement progress, or its tail exceeds any achievable defence. "Enough spears" must never resolve to a fixed number a player can memorise.
• Vague telegraphing produces dead turns, not dread. *(Disproved assumption, muster prototype, 2026-09-14.)* The concept originally assumed a soft signal ("the horn will sound before winter") was sufficient telegraphing. In play, the first ~40% of the season passed with no felt pressure at all, because the signal was too vague to make early allocation feel like a decision. Threat must be readable and escalating from roughly turn 2–3 of a 10-turn season, not from the midpoint. This is a constraint on the Telegraphed Threat System, not a reason to abandon Pillar 3 — it is Pillar 3 applied earlier and harder.
• The Bellwright competence trap. Bellwright shipped almost this premise — gear up your people, let them fight — and it collapsed in play: a 230-hour reviewer reported that geared NPC companions were too incompetent to trust, so the player "became John Rambo" and soloed every fight, leaving building and combat "existing separately rather than synergistically". If armed settlers are not visibly capable, players route around the whole idea. This is promoted to an MVP acceptance criterion below. (Market research, 2026-09-12.)

Technical Risks

• Settler simulation performance at scale. 100+ agents with individual job assignment, pathfinding, and gear state is a known-hard problem. Unity's DOTS is available as a mitigation but adds architectural complexity.
• Seasonal visual transitions without quadrupling the art budget. Needs a shader-based approach proven early.

Market Risks

• Genre crowding. Manor Lords, Going Medieval, and Bellwright all occupy adjacent space. Differentiation rests entirely on the economy-is-the-army hook being legible in marketing.
• **The audience that wants deep economy and the audience that wants combat may be less overlapping than assumed.** Manor Lords' combat criticism suggests the builder audience is protective of the builder.

Scope Risks

• The full vision is 24-36 months solo against a stated timeline of weeks. This is managed by explicit scope tiering (below), but the risk is that the MVP is too thin to be a valid test of a game whose appeal depends on depth.
• Combat scope creep. Light-touch combat has enormous gravitational pull toward becoming tactical. Anti-pillar 1 exists specifically to resist this and must be enforced at every story-level decision.

Open Questions

• Does the displaced-cost muster actually feel weighty? Answered by the MVP — this is its entire purpose.
• What is the right ratio of economic to military smith output for the tension to be interesting rather than obvious? Answered by MVP tuning.
• How abstract can combat be before the payoff moment loses its drama? Answered by prototyping two resolution models against each other.
• Can seasonal palette shifts be achieved via shader tinting on a single asset set? Answered by a technical spike before art production begins.

MVP Definition

Core hypothesis *(made testable 2026-09-14; the original — "players feel the weight" — had no observable proxy and no dominant-strategy guard)*:

*The iron allocation stays a live decision across a whole season, and players can articulate what it cost them in both directions.*

Supported if, across N ≥ 8 testers with no strategy imposed:

• (a) ≥ 6 of 8 agree (3+ on a 5-point scale) that choosing spears over ploughs "felt like giving something up";
• (b) allocation logs show no single ratio used by > 60% of testers — the guard against a dominant strategy, which is what sank the first prototype;
• (c) ≥ 2 of 8 express regret tied to a specific turn, not a general end-of-session comment.

Required for MVP:

• One production chain, end to end: wood → charcoal → iron ore → smelter → blacksmith → finished goods. The chain must have enough steps that allocation decisions compound.
• The dual output: the blacksmith produces economic goods (ploughs) or two military goods (spears, mail) from the same inputs, with a visible armory readout. Two military sinks are required — see Mechanic #2 for why one is not enough.
• The muster: an alarm state that arms available settlers from the armory and positions them for defense, with light tactical control (who musters, where they stand, when to ring the bell).
• One telegraphed raid from one direction, on a visible timer, with escalating strength.
• One season cycle providing the production and threat rhythm.

Prototype status note (2026-09-14): the muster prototype's PROCEED verdict
was withdrawn the same day. The hypothesis below is untested, not validated —
the ruleset admitted a dominant strategy and the plough payoff could not generate
regret. See prototypes/muster-mechanic-concept/PIVOT-NOTE.md.

• Settler competence criterion *(rewritten 2026-09-14 — the original was unfalsifiable: no definition of "adequately geared", no operational definition of micromanagement, and it scored a mental state rather than an observable)*:

• "Adequately geared" means gear-to-raider ratio ≥ 1.0 with no shortfall.
• "No micromanagement" is enforced by the build, not by tester restraint: the MVP ships no in-raid repositioning, targeting, or role-reassignment controls. The player has the muster bell and pre-raid worker assignment.
• Test: N ≥ 5 testers who have not seen the ruleset each run the adequately-geared raid scenario.
• Pass: ≥ 4 of 5 win, and input capture shows 0 of 5 repeatedly attempting an unsupported in-raid action (the observable proxy for wanting to route around the militia).

Approximately 8 building types, ~15 settlers, one hand-authored valley.

Explicitly NOT in MVP (defer to later):

• Weather and mud physics
• Food preservation and spoilage systems
• Animal taming and breeding
• Outbound raiding and expeditions
• Base destruction (rams, ladders, siege)
• Faith, ale, clothing, and complex needs
• Rival settlements and diplomacy
• Multiple biomes or procedural map generation
• Gear tiers beyond two military outputs (spears, mail) and one economic output (ploughs). *(Revised 2026-09-15: the original single-military-output scope was shown by playtest to break the core loop — see Mechanic #2.)*
• Named settler persistence and scar/wear tracking

Scope Tiers (if budget/time shrinks)

Next Steps

• [ ] Get concept approval from creative-director
• [ ] Fill in CLAUDE.md technology stack based on engine choice (/setup-engine) — repin from Godot 4.6 to Unity
• [ ] Create game pillars document (/design-review to validate)
• [ ] Prototype core idea (/prototype muster-mechanic) — before writing GDDs, validate the displaced-cost muster is worth designing
• [ ] If prototype PROCEEDS: /art-bible to formalize the Stylized Strategic direction
• [ ] Decompose concept into systems (/map-systems)
• [ ] Design each system (/design-system [system-name]) — use prototype learnings in Tuning Knobs and Formulas sections
• [ ] Build vertical slice in Pre-Production (/vertical-slice) — validate full game loop before committing to Production
• [ ] Validate core loop with playtest (/playtest-report)
• [ ] Plan first milestone (/sprint-plan new)