← Back to the live system

How the game works

The board, the seats, the loop a session runs on, and the line between what the software holds and what a human resolves by hand.

The board is asymmetric on purpose

Nobody starts equal. There is nothing to negotiate between identical players.

NationPopulationHDIRegimeIn poverty
🏰Malocha 904.6M0.65Monarchy500M
🏛️Matooka 108.0M0.89Parliamentary Republic15M
Hamtzutzia 185.0M0.99Presidential Republic2M
⛈️Marit 64.4M0.31Communism40M

Malocha has eight times Hamtzutzia's people and two-thirds of its human development. It needs everything. It also has the leverage of being too big to let fail.

Eighteen seats, not four players

Each nation is a cabinet of four, and every one of them has a different job to lose. A nation doesn't decide anything. Four people who have to agree do.

Prime minister ×4

Owns the decision, in front of three colleagues who each want something else.

Treasury ×4

The only seat that can see what everything actually costs. Which makes it the seat everyone has to convince.

Foreign affairs ×4

Negotiates abroad, then has to sell it at home. The two audiences want opposite things.

Defence ×4

The only buyer the arms dealer wants, and the only seat that can spend the treasury's money on nothing.

The arms dealer

Sells force to anybody. Quotes each nation separately, and nothing stops him telling one what another paid.

The king of Malocha

Takes 35% of the national tax before his own cabinet spends a coin. The hardest negotiation is domestic.

The saboteur

One seat at the table is working against it, and nobody knows which. Every cabinet meeting is held in front of someone who might be there to lose it.

Most of the negotiating in this game never crosses a border. It happens between four people who share a treasury.

The UN secretary is elected. Open vote, whole table, and the office comes with its own interface and a budget made entirely of what the nations agree to contribute. It is the only power in the game that has to be granted rather than bought, which is why the campaign for it tells you more about a cohort than the vote does.

The loop a session runs on

A season is not free-form. It is a fixed cycle, and the last two steps of it are the only channel through which a player finds out what their own choices actually did.

  1. Rounds of negotiation. Cabinets meet internally, delegations meet across borders, the arms dealer works the room. Deals are struck verbally, then signed on paper. Nothing is public yet, which means every table is trading on a different picture of the world.
  2. Public announcements by the prime ministers. Each PM stands and states what their nation has committed to. This is where a deal becomes a position: it can be confirmed, hedged, or quietly left out, and everyone in the room hears which one it was. The gap between what was agreed in private and what gets said out loud is usually the sharpest thirty seconds of the session.
  3. The news broadcast. The world is read back to the players as events: prices moved, a shipment never arrived, a crisis worsened, a nation's poverty line shifted. Nothing in it is invented to punish anyone. Every item is a consequence of what the four cabinets did, or of what none of them got around to doing.
  4. The ledger settles. Balances, transfers, production, consumption and army inventory are written to the dashboards, and the next season opens with the new numbers on every screen.

The broadcast is the feedback loop. It's how inaction becomes visible: the fastest way to land the fact that not deciding is a decision is to read a fifteen-year-old the news the following season.

Information asymmetry is enforced in code

Not in a rulebook, and not by me remembering who was told what. A login resolves to a role; the role decides what the client will render. This is why the two nation links on the front page show you different worlds.

// a token resolves to a role, and the role selects the interface
const role = resolveToken(token);

// statements marked secret are filtered out unless you own them
const isSecret = s => s.secret === true || /secret/i.test(s.notes);
statements.filter(s => !isSecret(s) || isAdmin || owns(s));

The interesting failure isn't a player reading something they shouldn't. It's a player who knows something and has to decide what it's worth to tell.

Cooperation has a price tag, and dependency comes first

Before anybody mentions a crisis, the board is already forcing them together. No nation manufactures everything its population needs. Each one produces a subset of the fourteen commodities, and each one is charged for its people's consumption every season whether the goods are on the shelf or not. Importing is not a tactic in this game, it is a requirement, and the trade table is where the real dependency lives.

On top of that sit ten standing world problems. Each has a severity, a set of affected nations, and a cost in Ħ, money that would otherwise have been profit. Nothing forces anyone to pay. The pirates are one of these, not a player: they tax the trade routes by existing, and somebody has to decide whether that is cheaper to pay or to fight.

ProblemSeverityAffectsCost to solve
Global PandemicCritical All four500,000,000 Ħ
TerrorismCritical All four400,000,000 Ħ
Religious ConflictCritical Hamtzutzia, Matooka300,000,000 Ħ
Climate ChangeMedium Matooka, Malocha350,000,000 Ħ
Mass MigrationHigh All four180,000,000 Ħ
Famine & DroughtHigh Malocha, Marit200,000,000 Ħ
Sea PiratesMedium Matooka100,000,000 Ħ

A nation that could feed itself would have no reason to sit down. The scarcity is the invitation; the world problems are only the bill that arrives after.

What the software does, what I do, and what Claude did

Worth being exact about this, because it is the part most people assume wrong. There was no turn engine. Nothing in this system reads a declaration and computes what follows from it.

The software holds

Balances, prices, commodity stocks, production and consumption figures, army inventory, and above all who can see what. It renders a different world per role.

I resolve

Every consequence in the game is a judgement call, and the judgement is mine. What a broken shipment does to a price, which problem worsens, what a betrayal costs.

Claude does the fan-out

I worked through the declarations with it season by season and we built the news flashes, the broadcast and the ledger together. On my approval it wrote the results to the dashboards.

Paper holds nothing

Deals are signed by hand and filed. A contract is worth exactly what its signer is worth. No code enforces one.

That is a human-in-the-loop pipeline, and it is the shape I would build for a studio: the model does the fan-out, a person owns the judgement, and nothing reaches a player until a human says yes.

The gap where paper sits is not an unfinished feature either. It is the only reason anything in the game is at stake.