There is a Kelvrac-series chassis in the Sinter Yards that has been listed, withdrawn and relisted eleven times across four cycles. It works. Its capability tags are unremarkable. Its declared build cost is defensible. Nobody will buy it.

This is not a pricing problem, and the operator has established that by lowering the price twice with no effect whatsoever.

A liquidity trap in this economy has a specific structure, and once an asset is inside one there are exactly two exits, both of which cost the operator money.

What Makes an Asset Untradeable

A robot is bought by an operator who intends to use it, which means the buyer must want precisely the output the chassis is tagged to produce, must have a slot for it, and must find the total price — listing plus fifteen percent — better than declaring an equivalent machine themselves.

That last condition is the one that traps. A buyer can always build instead. A new chassis declared at 22,000 costs 1,760 in creation tax, total outlay 23,760, and arrives configured exactly as the buyer wants. A used chassis listed at 22,000 costs the buyer 25,300 with tax and arrives with somebody else's tag set.

The used machine must therefore clear at a discount deep enough to overcome the tax and the imperfect fit — and if it cannot, no price finds a buyer, because the competing option is not another seller. It is the foundry.

The Arithmetic of the Trap

Work it through on the Sinter Yards chassis. Declared at 30,000, tagged for a resource band that two other operators in the district already cover. A buyer contemplating it compares against building fresh at, say, 26,000 declared: creation tax 2,080, total 28,080, tags chosen to suit.

For the used machine to win, its total cost to the buyer must fall below 28,080 with enough margin to justify the tag mismatch. At fifteen percent that means a listing at or under about 23,000, and realistically nearer 19,000 once the mismatch is priced. The seller is therefore looking at recovering perhaps 63 percent of declared cost in the best case.

"The competition is not the other listings. The competition is a chassis that does not exist yet and can be exactly what the buyer wants. You cannot beat that on features, only on price, and the tax eats the price." — Ossin Tral, Sinter Quarter

Now compare the exits. The Acceptance Annex pays 60 percent of the 30,000 reference: 18,000, achievable if the operator conducts the rounds properly. Destruction costs three percent: 900 coins outward, nothing back. Selling at 19,000 nets 19,000.

So the trap is not that the asset is worthless — it is worth roughly 18,000 to 19,000 by two independent routes. The trap is that the operator declared it at 30,000, paid 2,400 to create it, and has been marking it in their own head at 30,000 ever since. Every route out crystallises a loss they have not yet accepted, and so the machine gets relisted instead.

How the Trap Deepens

Relisting is not free of consequence. The Exchange publishes settlements, but withdrawals and repostings are visible to anyone watching the board, and a chassis that has been listed eleven times carries a legible history. Buyers read it as a signal about the machine, which it is not, and about the seller, which it is.

Meanwhile the alternative routes decay. The Annex's reference is the declared build cost, which does not change — but a market that has watched a machine fail to sell repeatedly is a market that will not bid against the Annex, so the operator's only real counterparty is the one that already knows nobody else wants it.

What Operators Believe About Unsellable Machines

The first belief is that a lower price eventually clears anything. It does not, when the competing product is a machine that has not been built yet. Below a certain point the discount itself becomes a signal, and buyers who could build fresh conclude something is wrong with the chassis rather than with the seller's patience.

The second belief is that the trap is a market failure. It is closer to the market working exactly as designed: capability tags make used machines imperfect substitutes, the sales tax makes them expensive substitutes, and the foundry makes new supply available on demand. A used-robot market under those conditions should be thin, and it is.

The Kelvrac in the Sinter Yards will most likely go to the Annex, at a figure its operator could have had four cycles and eleven listings ago. The eleven listings cost nothing directly. What they cost was the interval, and the visible record of having spent it.

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.