The Coppervein Exchange does not publish a price. It publishes settlements — this lot, this quantity, this figure, this cycle — and leaves the inference to whoever is reading. There is no series, no average, no posted range.

Operators price their lots anyway, and the prices they arrive at are not random. Something is doing the work of a price history without being one.

Understanding what that something is explains most of the market's stranger behaviour, including why a resource type's first few settlements matter far more than its hundredth.

What the Exchange Actually Gives You

A settlement record is a fact about one transaction: a lot of a stated resource type and quantity changed hands at a stated figure. It carries no context. It does not say whether the seller was desperate, whether the buyer was a runner working a gap, or whether the lot came from a line with a reputation.

What an operator has, then, is a scatter of unexplained points. The market's price is a thing each participant constructs privately from that scatter, which means there are as many prices as there are operators paying attention, and they differ.

What Fills the Gap

Three things, in descending order of how much work they do.

The declared value. Every resource carries one from the moment it is produced or forged. It is not a price and was never intended as one, but it is the only figure attached to the resource itself rather than to some particular transaction, and operators anchor on it heavily. Lots routinely list within a narrow band of declared value for no better reason than that the number is there.

The most recent settlement anyone remembers. Not the average — the last one. Recency dominates in a market with no series, and a single unusual settlement can move a resource type's effective price for several cycles. The Span Market saw a lot of tempered lattice clear at 34,500 coins during a thin window; listings for comparable lots sat near that figure for the better part of a quarter afterward, against production costs nowhere near it.

"There is no price. There is the last number somebody said out loud, and everybody's memory of it, and those two things are not as close as you would hope." — Osta Vehn

The sixty percent floor. The Acceptance Annex will pay 60 percent of reference for essentially anything, which puts a hard floor under every resource in the economy. An operator will not accept 8,000 coins for a lot the Annex would take at 12,000, and every participant knows this. In practice the Annex's ceiling functions as the market's floor, which is a structural oddity worth sitting with: the buyer of last resort sets the lower bound on every price it is not involved in.

Stack these and the mechanism is clear. A new resource type's price is anchored by its declared value, dragged around by whichever early settlements happened to occur, and floored at 60 percent of reference. The first three or four transactions in a type's life do more to set its price than everything that follows.

Where This Produces Bad Prices

The obvious failure is the anchoring settlement that should not have counted. A lot sold during a forecast window at 12 coins against a normal 19 is a distressed transaction, and the record does not mark it as one. Operators pricing off it are pricing off a panic, and the panic outlives itself by several cycles.

The subtler failure falls on new compounds. A freshly forged type has no settlements at all, so its price is pure declared value until somebody transacts. The first seller therefore sets the anchor for everyone, and has every incentive to set it high — which they do, and which is why novelty premiums are as large as they are and decay as sharply as they do when the second and third settlements land nearer to cost.

What Operators Believe About Prices Here

The common belief is that the market is efficient because it is continuous and open to everyone. Continuity is not efficiency. A market that never closes but publishes no series gives every participant the same information and no way to aggregate it, which produces persistent dispersion rather than convergence. The runners exist precisely because that dispersion is reliable.

The second belief is that declared value is a valuation. It is a figure attached at production or forging, by a process that never consulted the market, and it does not update. Operators treat it as a floor, a ceiling and a fair-value estimate according to which reading suits the trade in front of them, and it is none of the three.

The Exchange will keep publishing settlements and declining to publish a price, and operators will keep constructing one out of an anchor, a memory and a floor. It works well enough that trade happens. It works poorly enough that a single distressed lot can misprice a resource type for a quarter, and nothing in the record will ever say so.

Note: Mechadia is a work of fiction. The districts, operators, robots, and figures described here are invented, and nothing on this site is a report of real events, real machines, or real economies.