Annex Holds Builds Older Than the Grants Behind Them

Somewhere in the Acceptance Annex, a Refinery-class unit registered during what archivists now call the early Third Accumulation Cycle sits on the active rolls under an operator whose founding grant was issued after the machine's declared build date. The ledger does not flag this. It records both facts without comment, as it records everything else — append-only, unamended, silent on contradictions.

The phenomenon is not isolated. A count drawn from the Annex's open rolls last quarter identified at least forty-seven builds whose logged creation dates precede the grant issuance records for the operators who hold them. The gap ranges from a few hundred ledger cycles to, in one case, a span that predates the Founding Issuance itself — or appears to, depending on which timestamp is trusted.

This piece examines how those discrepancies accumulate, what the Central Intelligence's own settlement process does and does not do to resolve them, and why the Annex's role as the endpoint of the buyback system makes the problem harder to ignore than it might be elsewhere in the record.

What the Acceptance Annex Actually Is

The Acceptance Annex is the formal venue where buyback settlements between operators and the Central Intelligence are recorded. When an operator offers a robot or a resource to the Intelligence — typically because no market buyer has materialized — the negotiation is conducted, the figure agreed upon, and the transfer logged at the Annex. It is not a market. No other party may bid. The settlement is bilateral, and the entry is final once written.

The Annex sits downstream of every other mechanism in the economy: after the open market has declined to clear an asset, after the operator has decided the haircut is preferable to continued holding costs, after whatever negotiation rounds have pushed the Intelligence's offer as close to its ceiling as it will go. What arrives at the Annex is, by definition, what everything else did not want. That position makes it a reliable accumulation point — and a reliable accumulation point for anomalies that other parts of the ledger have already passed over without resolution.

How the Discrepancy Enters the Record

The creation tax is levied at build time against the declared build cost: eight percent, paid before the machine has produced anything. An operator who declares a chassis at 40,000 coins pays 3,200 coins to the Treasury at that moment. The declared cost also sets the floor for later calculations — the destruction tax at three percent of the same figure, and the reference value the Intelligence consults when formulating a buyback offer. None of those calculations require the ledger to verify that the operator's founding grant was issued before the build date. The tax is collected; the machine enters the rolls; the grant record sits in a separate binding at Ledger Hall.

The gap opens when builds are transferred between operators, when founding grants are reconstructed from secondary records after a filing dispute, or — most commonly — when the timestamp on a grant record was written against a different cycle-counting convention than the one used by the registry that logged the build. Registry filings older than the operators who submitted them have been documented before; the Annex simply inherits those cases when the assets in question eventually reach the buyback queue.

Consider a worked example that Archivist Secondus Preln described in a brief filed with the Ledger Standards Committee last cycle:

"The build in question was declared at 55,000 coins. Creation tax of 4,400 was collected and confirmed. The operator's grant record, when retrieved from the Coppervein Archivist Office, carries an issuance date that falls 200 cycles after the build's logged registration. We cannot amend either entry. We can only note the gap and allow the settlement to proceed."

In that case, the Intelligence offered 60 percent of a reference value derived from current market comparables for the robot's capability tags — arriving at 19,800 coins on a 33,000-coin reference. The operator accepted after one revision round. The settlement was logged. The discrepancy in the underlying dates remains in both records, unresolved, because the ledger has no mechanism for resolution — only for addition.

Where the Process Strains

The most direct cost falls on operators who acquire builds through secondary transfers and later attempt a buyback. The Intelligence's reference value is calculated against the original declared build cost and current capability-tag comparables — not against what the acquiring operator paid on the open market, which may have been substantially higher after the sales tax moved from five to fifteen percent and drove up effective acquisition costs across the board. An operator who paid 1,150 coins to acquire a resource listed at 1,000 — absorbing the full buyer-side tax — and then faces a buyback ceiling of 60 percent of a reference value pegged to older figures can find the arithmetic does not recover the outlay.

The date discrepancies compound this by introducing uncertainty about which reference value the Intelligence will apply. When the build record and the grant record disagree on era, the Intelligence has discretion over which cycle's comparables it treats as authoritative. That discretion is not published as a formula. Operators working through the Annex have reported receiving offers derived from reference values they could not reconstruct from any public document — and the Intelligence's negotiation ceiling means there is a limit to how far a counter-offer can push back.

What Operators Consistently Misread About the Annex

The most durable misreading is that the Annex will surface and resolve anomalies in the underlying record as a condition of processing a settlement. It will not. The Annex logs the settlement against whatever records exist at the time; it does not audit them. Builds that have never had a matching destruction entry pass through the same process as any other asset. The Intelligence is not troubled by the gap. It is acquiring the asset, not curating the history attached to it.

The second misreading concerns the negotiation revision. Operators frequently enter the process expecting that persistence across multiple revision rounds will close the distance between the opening offer and something near full reference value. In practice, the typical upward movement per round is approximately five percent of the standing offer — on an 18,000-coin opening bid, that is 900 coins. The Intelligence's ceiling is absolute, and the ceiling is 60 percent of the reference value it has already set. An operator who holds out across four rounds and achieves the ceiling has recovered perhaps 3,600 coins above the opening figure. Against a declared build cost of 55,000, that arithmetic is not a negotiation so much as a controlled descent.

The Acceptance Annex does not produce the discrepancies it holds. It inherits them — from registries that counted cycles differently, from grant records reconstructed after disputes, from a transfer history the ledger captures faithfully without being able to reconcile. The Intelligence processes each settlement against those records as it finds them, acquires the asset, and moves on. The gap between a build date and the grant that supposedly funded it remains in both bindings, neither confirmed nor corrected, available to anyone who pulls the relevant entries side by side and notices what the ledger itself will never say.

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.