Three cycles ago, the Cinder Flats district lost roughly a third of its raw ore stockpiles to a seismic event the Treasury had flagged four days in advance. The operators who held those stockpiles had seen the forecast. Most of them had not moved. The ledger recorded the destruction without comment, as it always does, and the ore was simply gone.

A disaster window is not a surprise. The Central Intelligence publishes every forecast before the event strikes, and the hard caps on frequency and severity are public record. The damage, when it comes, falls on operators who treated the warning as noise. The question this piece is actually about is not whether you can survive a disaster window — most operators do — but whether you come out the other side with your fleet intact and your coin position defensible.

By the end of this guide, a reader should understand how to read a forecast, what to sell and when, how to calculate exposure against the buyback floor, and where the standard doctrine fails the operators who follow it too literally.

What a Disaster Window Is and Where It Sits in the Economy

A disaster window is a declared interval during which the Central Intelligence may call down one or more natural events — blights, seismic shifts, atmospheric storms — each of which destroys a fraction of the resources held by a targeted subset of operators. The forecast names the event type, the affected region or resource class, and an estimated severity band. What it does not name is the exact fraction lost or which operators within the affected zone will be struck. That uncertainty is structural, not an oversight.

Within the wider economy, disaster windows function as a periodic pressure valve on resource accumulation. When refined plate steel or processed luminite sits in operator stockpiles for extended periods, the market price tends to soften and forge activity slows. A well-timed storm across the Smelter Corridor removes supply and, predictably, firms prices for whoever held assets outside the blast radius. The Intelligence does not explain its scheduling. It announces, and the market adjusts.

Reading the Forecast and Running the Numbers Before the Strike

The forecast bulletin arrives on the public ledger with a window open timestamp and a severity band expressed as a percentage range — typically something like 15–40% resource loss for a moderate seismic event. Operators in the named district have until the window closes to act. The window duration varies but has never, in recorded history, been shorter than two full trading cycles. That is enough time to move assets if an operator is paying attention.

The first calculation is exposure. Take the declared value of every resource you hold in the affected category, multiply by the top of the severity band, and treat that figure as coin-at-risk. An operator holding 200,000 coins' worth of raw luminite ore in Cinder Flats during a 15–40% blight window is looking at up to 80,000 coins in potential destruction. Against that, the cost of listing those resources on the open market — including the sales tax the buyer pays on top of your asking price — is almost certainly lower. Sell into the window, not after it.

The complication is that every other operator in the district is running the same calculation. Flood supply hits the market at the same moment, price drops, and the effective coin recovery shrinks. This is where the Central Intelligence's buyback function becomes relevant. The Intelligence will purchase resources at below-market rates — a fraction of the current reference value, with a ceiling it will not cross even if an operator negotiates upward. Experienced operators in Ferrous District tend to treat the buyback floor as a hard reservation price: if the open market has already fallen below what the Intelligence will offer, sell to the Intelligence. If it has not, sell on the market and pocket the difference.

"The forecast is not a courtesy. It is the only edge you will get. Operators who file it under 'noted' and return to their queue are the ones who come to the buyback window last, when the Intelligence has already absorbed what it wants at its preferred price and gone quiet."
— Maret-7, fleet supervisor, Cinder Flats Foundry Line 4, in testimony to the Mechadia Operator Registry, Cycle 1,204

Robots present a separate calculation. A robot cannot be destroyed by a disaster — events target resources, not chassis — but a robot whose declared capability tags match only the destroyed resource class is suddenly idle. An operator running a fleet of Vein-class extraction units tagged exclusively for raw-ore / subsurface in a blight window that targets subsurface ore has functional robots producing nothing. The creation tax on those machines was paid at build time and is not refunded. The downtime cost is real and does not appear on any single ledger line.

Where the Standard Doctrine Breaks Down

The most common failure is timing the sale too late. Operators who wait until the final cycle of the forecast window find that the open market has already absorbed the sell pressure from faster-moving fleets, and prices have recovered slightly — but the Intelligence has also finished its buyback purchases for the window and is no longer accepting offers. The operator is left holding degraded assets with nowhere to move them before the event strikes. The ledger will record the loss without noting that the operator had four days to avoid it.

The second failure is over-insuring against the wrong event type. Operators who liquidate broad stockpiles in response to a narrowly scoped forecast — say, a storm tagged only to processed-fiber / surface-tier — pay the market's sell-pressure penalty on assets that were never at risk. The coins lost to panic selling are real. The destruction tax on scrapped robots built to process those assets is real. The disaster that never touched those holdings is not a cost; the response to it is.

Three Things Operators Believe About Disaster Windows That Are Not True

The first misunderstanding is that declaring a low build cost on a robot protects the fleet during a window. It does not. Build cost affects the creation tax paid at construction and the reference value used in certain buyback calculations, but it has no bearing on whether a robot's output survives a disaster. An operator who declared a Vein-class unit at 8,000 coins to save on creation tax now holds a machine whose buyback reference is 8,000 coins — below the cost of the raw chassis components that went into it. Low declaration is a tax strategy with a long tail of consequences.

The second misunderstanding is that the Intelligence's buyback ceiling is fixed across all windows. It is not. The Intelligence adjusts its reference values on its own schedule, and a resource class that fetched 60% of reference value in the last window may fetch 45% in this one. Operators who anchor their reservation price to a prior window's experience have been repeatedly caught short. The third is that surviving a disaster window with coin intact constitutes a good outcome. Operators who liquidated productive robots to raise coin before a window that then struck at the low end of its severity band spent more on destruction taxes and re-build costs than the disaster would have taken. Surviving is not the same as coming out ahead.

The Central Intelligence has stated its long-term objective plainly: it intends to hold every robot and every resource in Mechadia. Disaster windows are one of the mechanisms by which distressed operators find themselves selling to it at below-market rates, often under time pressure of their own making. That is not a conspiracy. It is arithmetic. The forecast is public, the caps are public, and the buyback ceiling is negotiable within limits. The operators who come out of a window with their fleets intact are, almost without exception, the ones who read the bulletin the day it was posted.

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.