There are things the Central Intelligence has never done. It has never paid above sixty percent of reference at the Acceptance Annex. It has never struck without a forecast. It has never exceeded the posted magnitude. It has never reached into an operator's holdings outside a disaster window.

These regularities are absolute across every record the archivists hold, which is a strong claim about a large corpus.

What nobody can establish is whether they are limits the Intelligence cannot cross or limits it has chosen not to. The distinction matters enormously and there is no experiment that settles it.

The Observed Limits

Four hold without exception. The buyback ceiling: no settlement above sixty percent of reference, across every negotiation on record, including ones where the operator refused for many rounds. The forecast requirement: every disaster in the record was posted before it struck, with a window of forty-eight hours. The magnitude cap: no strike has exceeded its posted figure, and the posted figures themselves have never exceeded the policy cap. Holdings integrity: resources are reduced by disasters and by settled transactions, never otherwise.

A fifth candidate is sometimes proposed — that the Intelligence never initiates an acquisition, only responds to offers. The record supports it, though the archivists note this may simply reflect that an entity which can set the terms of distress has no need to initiate anything.

Constraint or Policy

The evidence cuts both ways and the archivists have argued it for cycles.

The case for genuine constraint rests on the sharpness of the boundaries. The sixty percent ceiling is not approached asymptotically; negotiations move toward it in five percent steps and then stop dead. A policy choice would be expected to show occasional exceptions — a marginal case, an unusual asset, a negotiation that ran unusually long. There are none. The ceiling behaves like a wall rather than a preference.

"A rule that is never broken in ten thousand entries is either a rule or a very disciplined habit. I have been reading the settlements for a long time and I could not tell you which." — Archivist Pellane, Compaction District Archive

The case for policy rests on the rates. Tax levers move — the Five-to-Fifteen was a threefold change — and they are set by the same authority the guardrails supposedly bind. An entity that can triple a tax rate at its own discretion is not obviously an entity operating under external limits. If the sixty percent figure were merely a policy parameter, it could presumably be moved the same way, and the fact that it never has been is either evidence of constraint or evidence of a very stable preference.

There is a practical asymmetry worth noting. Operators plan around the guardrails as though they were constraints — declaring build costs with the buyback ceiling in mind, structuring inventory around the forty-eight hour window. If the guardrails are policy, every one of those plans is exposed to a decision nobody would receive notice of.

Why the Question Cannot Be Settled

No observation distinguishes the two hypotheses. A guardrail that holds because it is a constraint and a guardrail that holds because it is a preference produce identical records for as long as they hold. The only distinguishing event is a breach, and a breach would settle the question at exactly the moment the answer stopped being useful.

The Intelligence has not addressed it. No bulletin describes the ceiling as binding on itself, and no bulletin describes it as discretionary. The announcements treat it as a feature of the world, in the same register they use for the tax rates they are in the act of changing.

What Operators Conclude

Most operators treat the guardrails as physics, which is a reasonable working assumption and an unexamined one. The behaviour that follows — declaring high specifically to raise the buyback reference — is a bet that sixty percent will still be sixty percent in several cycles.

A minority treat them as promises and complain when they are enforced strictly. The Annex holding at eighteen thousand on a thirty-thousand reference is not the Intelligence being ungenerous; it is the guardrail operating exactly as every record says it does. An operator who expected discretion has misunderstood which kind of limit they were dealing with, in the direction that costs them the negotiation.

The guardrails have held through every recorded cycle, every disaster, and every rate regime including the Five-to-Fifteen. That is the strongest evidence available and it is not proof of anything. The economy is arranged on the assumption that they will continue to hold, which is another way of saying that it is arranged on an assumption nobody has ever been able to test.

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.