MYCELIUM

A Game of Living Worlds

A cooperative board game where you play a Spirit of Nature — each one a real, scientifically studied phenomenon — growing a biodiverse, interconnected world across a century of climate, and defending it against everything that century brings.

Cooperative · 2–4 players 12 asymmetric Spirits 117 cards 6 Ages · ~24 rounds Every card grounded in real science
The idea

The science is the mechanic

Not a theme painted over an engine. Biodiversity literally is resilience; connection literally is strength; a keystone species literally holds up its web. Where the popular version of a story is overstated, the card says so — and those honesty flags are a feature, not an apology.

Keywords are guilds

Abilities reference functions — Pollinator, Keystone, Nitrogen-Fixer — not card names. Ecologists group species by what they do, and substitutable species make a deck that scales.

Connectors are food webs

Every card declares what feeds it and what it feeds. Actions resolve bottom-up the chain, so the resolution order teaches the structure.

Bonds are mutualisms

A scarce set of obligate pairs — otter and kelp, acacia and ant. While both live, neither can be broken. Split them and both are exposed.

The arc

Doom, then bloom

A slow, near-hopeless opening. A middle where collapse feels certain. Then a late game where matured ecosystems and interlocking Spirit powers detonate together and clear the goal by a hair — a win that should feel planned, not lucky.

Where things stand

Built so far

117 cards

Full data model — biome, climate, points, cost, trophic role, keywords, bonds. Every one renders from its record.

A working scoring engine

Biodiversity, Connection and hazard mitigation compute live, and show their arithmetic.

Playable rules

Economy, loss condition, terraforming, hazards and scoring are all defined. Never yet played — see Balance for what is resolved and what is still open.

The rules

Player's Manual

Everything needed to actually play, from setup to the final scoring. Numbers here are the current design values — they are meant to be playtested and moved.

Systems, worked

Mechanics

The two engines that carry the game, both computed live rather than asserted: the terraform cascade that turns a desert into forest, and the ecology scoring that makes a keystone species worth far more than its own points.

Engine 1 · The signature move

The terraform cascade

Biome is never chosen — it is read from Water × Warmth. Raise a hex to Elevation 4 and it becomes a Mountain: the windward side gains water, the leeward side loses it, and every changed hex is re-read on the table. Click a hex to raise or lower it.

Engine 2 · Why a keystone matters

Ecology scoring, and what withers when one card goes

Six cards forming a complete web. Toggle the Sea Otter out and watch what a 3-point card is really worth.

The kelp cannot hold without a keystone predator — the same rule real urchin barrens follow.

Read this as an illustration, not as the scoring rules. The region model below sums a whole region each time, which is how victory scoring used to work — and it was replaced precisely because it cannot be done by hand. Victory is now two incremental tracks (Biodiversity ≥ 24, Connection ≥ 16) that only tick on discrete events. What this demo still shows truthfully is how much of a web hangs off one keystone card.
Nothing dies here. Remove the Sea Otter and the kelp withers — it drops to zero Vigour with its ability switched off, but it stays on the board and recovers the moment a keystone returns. The art shows an urchin barren, which is exactly what a kelp forest degrades into, and exactly what it grows back from.

A card only ever leaves the board through direct destruction — a Pulse hazard, or another player. Never through a dependency you failed to draw. That rule exists because a card that can be dead on arrival is a card nobody picks up.

Biodiversity

Rewards distinct function, not distinct species.

Connection

What each card contributes, plus bonds that are actually intact.

Vision boards

Reference

The source images this design is being built toward. Click any board to open it full size. Read together they define the look — and one of them describes a game meaningfully different from the rules as currently written, which is called out below.

Board 03 · The physical target

This is the game on a table

Low-poly stacked hexes with snap-on sculpts, elevation reading as literal layers of soil beneath each tile. This validates the protected element directly: elevation is not an abstraction here, it is the toy. Volcanoes, peaks, lakes, oases, settlements — all snap-on, all readable across a table.

Low-poly stacked hex board with sculpts Open full size ↗
ConfirmsElevation stacking, low-poly direction, snap-on features, and the visual grammar of the terraform cascade.
CostsEvery hex here is a moulded tile plus risers. This is the 80-riser manufacturing question in physical form.
Board 01 · The presentation sheet

The living world, Spirits, and building pieces

The modular hex world, ten Spirit cards with finished art, the full set of building archetypes, and a worked player-area example.

Game presentation sheet Open full size ↗
ConfirmsBiome tiles with built-in features, natural feature tokens, condition tokens, and the human building set already authored as cards.
DivergesShows ten Spirits; the roster is now twelve. Card anatomy here is Cost / Prerequisites / Effects / Scoring / Flavour — close to the Field Guide face but not identical.
Board 02 · The systems sheet

A complete alternate ruleset

This one is not just art direction. It specifies core systems, actions, turn structure, player components, victory conditions and card anatomy — and several of them contradict the rules as currently written.

Game elements summary and components Open full size ↗

⚠ This board and the current rules are two different games

Not a small drift — a structural one. Both are coherent; they are not compatible. This needs a decision before more rules get written on top of the wrong foundation.

QuestionBoard 02 saysCurrent rules say
How do you pay for things? "No Traditional Resource Costs." You meet prerequisites — terrain, network, unlocks — and play the card. A Nutrient economy. Income from Presence on living hexes; cards cost 1–5.
Where do cards come from? Hand or a shared Market row, plus unlockable specialist decks (Travel, Water, Industry, Ecology, Civic, Energy). Hand only, with Age-gated succession packs.
What do you own? A personal tableau: Charter, 3 Institutions, 2 Policies, Developments, and an Impact/Reputation track. No tableau. Everything lives on the shared board.
How do you win? Points from synergy — adjacency, networks, conditions — plus personal objectives and five shared tracks. Two tracks only: Biodiversity ≥ 24, Connection ≥ 16.
Is it co-op? "Cooperative or Competitive Modes." Fully cooperative; everyone wins or loses together.

My read. Board 02 is a Euro engine-builder — closer to Ark Nova or Terraforming Mars. The current rules are a cooperative puzzle — closer to Spirit Island, which is what the original vision named. The prerequisite system in particular is elegant and worth stealing regardless: "you may build a Watermill where there is flowing water and elevation" is more thematic than "pay 3 Nutrients," and it removes a whole layer of bookkeeping.

What I would do: keep the cooperative structure and the terraform cascade, and replace the Nutrient economy with Board 02's prerequisites. That keeps your tile-laying core, drops the resource accounting the physical-first brief dislikes, and makes terraforming even more central — because terraforming is how you create the prerequisites. It needs your call, though; it is your vision, not mine.

Decisions needed

Open Questions

Seven forks that are genuinely undecided. Each is written so you can form a view without reading the specs, and each carries a recommendation with its reasoning shown — a recommendation you cannot argue with is not much use.

Most of these come from a real divergence: the reference boards describe a Euro engine-builder, while the written rules describe a cooperative puzzle. Both are coherent. They are not compatible.

Blocking other work sits downstream of it Open real fork, not yet urgent Decided direction agreed, recorded here
The deck

All 117 cards

Every card below is rendered from its data record in the Field Guide face — no card is laid out by hand. Art is procedural: 60 composable scenes, so the deck can grow without the illustration becoming a bottleneck.

The question

Does the game actually work?

Short answer: on paper, yes — but nobody has played it.

Three weeks ago this page said the game could not be played end to end, because it had no economy, no loss condition and no workable scoring. All three are now defined, along with a terraform system, biome fitness, two speeds of hazard, and a defence for each hazard scope. A full round can be walked through from Foresight to Renew without hitting an undefined rule.

What remains is a different class of problem. The rules are complete enough to play but entirely unvalidated — every number is a design value, not a measured one. And two structural questions are still open, because your own reference boards describe a different game from the written rules.

Below: what got resolved and how, what is still open, and what has held up throughout. The single highest-value next step is a paper prototype, not more design.

The arc, modelled

Pressure against power

Computed from the balance model, not drawn by hand. Three play qualities, to test whether the intended arc survives contact with how well people actually play.

Findings

What stands between here and a playable game