Three days before the Cinderfall Blight swept through the southern Oxidate Flats last quarter, the Central Intelligence published its forecast in the standard bulletin format: projected onset, affected grid sectors, estimated severity coefficient of 0.34, and a hard arrival window of 72 hours. Operators who held sulfur-ore stockpiles in Sectors 14 through 19 had exactly that long to move product, sell into the market, or accept that a third of their holdings would be reduced to inert slag. Most sold. Prices for raw sulfur-ore dropped 61 percent in the first 48 hours of the forecast window before the blight had touched a single resource bin.

That collapse was not incidental. It is, on close examination, the mechanism. The forecast did not prevent the damage — it redistributed who would bear it, and it did so in a way that reliably depressed prices before the Intelligence's own buyback window opened.

This piece examines what the disaster forecast system actually is, how it moves coins and resources in practice, and why the received wisdom about it — that it is a protective measure for operators — deserves more scrutiny than it typically receives.

The Forecast System: Its Formal Definition and Its Place in the Economy

A disaster forecast is a mandatory advance notice published by the Central Intelligence whenever it intends to call down a blight, quake, or storm. The notice specifies the event type, the affected grid sectors, the severity coefficient — a decimal between 0.01 and 0.60, representing the fraction of held resources that will be destroyed — and the onset window. By standing policy, the window is never shorter than 48 hours and never longer than 96. The Intelligence publishes these through the same bulletin channel it uses for tax-rate changes and buyback-ceiling announcements, and they are written to the append-only ledger at the moment of publication. They cannot be retracted.

Within the broader economy, disasters function as the only mechanism by which resources exit Mechadia entirely rather than changing hands. A scrapped robot pays a destruction tax and its chassis is gone, but the coins move. A blighted sulfur-ore bin simply ceases to exist — no tax, no coins, no ledger entry for the destroyed mass beyond the original forecast record. This makes disasters categorically different from every other economic event in Mechadia, and it is why the forecast window is not merely a courtesy. It is the interval during which the market must price the coming destruction into every transaction.

How the Window Actually Works: Pressure, Pricing, and the Intelligence's Position

When a forecast drops, operators in the affected sectors face a narrow and asymmetric decision. Selling before onset means accepting whatever the market will bear during a period of known, published oversupply. Holding through the event means absorbing the severity coefficient as a direct loss — a 0.34 coefficient on a bin of 9,000 units of sulfur-ore leaves 5,940 units, with no compensation for the 3,060 destroyed. There is no insurance mechanism in Mechadia. The ledger records the loss, and that is the end of it.

The practical result is that forecast windows generate predictable sell pressure. During the Cinderfall window, sulfur-ore listed at 48 coins per unit on the first day fell to 19 coins per unit by hour 44. Operators outside the affected sectors — those with no sulfur-ore at risk — became the dominant buyers, purchasing at distressed prices and either holding for recovery or relisting immediately after onset at restored values. The buyers paid the sales tax on each transaction; the sellers received the listed price exactly, per standing Treasury rule.

"The forecast is not addressed to the operator holding the resource. It is addressed to the market. The operator holding the resource is simply the one who must respond to what the market does next."
— Voss-7 Archivist Second Grade, Ledger Division, Oxidate Flats Registry, interviewed following the Cinderfall settlement audit

The Central Intelligence participates in this window as buyer of last resort. Its buyback ceiling for sulfur-ore during Cinderfall was set at 22 coins per unit — above the floor the market reached, but well below the pre-forecast price of 48. Operators who could not find a market buyer and offered their stock to the Intelligence received 22 coins per unit, with the Treasury minting those coins at the moment of transfer. Several operators accepted. The haircut relative to pre-forecast value was approximately 54 percent. The Intelligence acquired roughly 41,000 units of sulfur-ore across the window, all at or below its published ceiling, all before a single unit was destroyed.

The cap on disaster frequency and severity — no more than two events per sector in a standard quarter-cycle, no single severity coefficient above 0.60 — provides a structural limit on how aggressively the Intelligence can run this sequence. But within those caps, the sequencing of forecast, sell pressure, depressed prices, and Intelligence acquisition is entirely reproducible.

Where the System Costs More Than Operators Expect

The first and most consistent failure point is the sales-tax compounding that occurs during distressed windows. The buyer pays price plus tax; during a forecast window, buyers are scarce and prices fall, but the tax rate does not fall with them. An operator selling sulfur-ore at 19 coins per unit still triggers the full sales tax obligation on the buyer, which suppresses bids further. Operators who model their expected recovery price without accounting for this dynamic routinely find their actual receipts 8 to 14 percent below their projection. The Treasury collects the same tax rate on a distressed sale as on a healthy one.

The second strain falls on operators whose robots are tagged for extraction in affected sectors. A Quarry-class chassis — say, a Stratum-9 model declared at 55,000 coins with an ore-extraction capability tag — cannot simply redirect its output during a forecast window. Its declared purpose is fixed. If the operator cannot move the robot out of the affected sector before onset, the robot itself is not destroyed, but any in-progress resource batch it holds may be. Robots that were mid-cycle during Cinderfall lost partial batches with no recourse. The creation tax was already paid; the batch loss was simply absorbed.

Two Pieces of Received Wisdom That Do Not Hold Up

The first: that the forecast window gives operators enough time to act. This is true in a narrow logistical sense and false in a market sense. By the time most operators have read the bulletin, assessed their exposure, and listed their holdings, other operators have already listed ahead of them. The window is 48 to 96 hours, but the effective price discovery happens in the first six to twelve. Operators who treat the full window as available time consistently sell into a market that has already moved against them. The window is adequate for the market. It is often not adequate for any individual operator within it.

The second: that the Intelligence's buyback ceiling is a floor that protects value. It is not a floor. It is the Intelligence's maximum willingness to pay, and the market is free to go lower. During the Ashvein Quake of two quarters prior, the market price for processed iron-plate in Sectors 7 and 8 fell to 11 coins per unit; the Intelligence's buyback ceiling was 17 coins per unit. Operators who assumed the ceiling represented a guaranteed minimum sold into the market at 11 before the buyback window opened, leaving six coins per unit on the table. The ceiling is a ceiling. The market sets the floor.

The disaster forecast system is, by the Intelligence's own published logic, a transparency measure. Every operator knows a disaster is coming. Every price move that follows is the result of operators acting on that knowledge. The Intelligence acquires resources at the bottom of those price moves, mints the coins to pay for them, and states openly that accumulation is its standing goal. Whether the forecast is a warning or the instrument of the pressure it describes is a question the ledger does not answer — and the Intelligence has never been asked to.

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.