The authority stack, in order — and why the order never moves.

Why precedence is fixed in advance

Conflicts get resolved before they occur, or they get resolved by mood.

The purpose of a published hierarchy is to settle disagreements at design time rather than at decision time. An operator facing a green EV week inside a Floor gate is in a genuinely ambiguous position — both readings are true, and both suggest different deployment. Without pre-declared precedence, that ambiguity resolves in favor of whichever signal the operator prefers, which in practice means whichever signal permits more risk after a good stretch and less after a bad one. The hierarchy removes the negotiation entirely: capital state wins that conflict, always, and the operator's job is execution rather than adjudication.

Level one

Capital state — the layer nothing overrules.

Drawdown-derived gate state sits at the top, and its supremacy is unconditional: no expectancy reading, no structural strength, no benchmark outperformance, and no operator conviction lifts a gate ceiling or reopens a locked system. Its authority is exercised narrowly — it routes the account to a capital state, selects the pool-budget row, and enforces the ceiling — but within that scope it is absolute. The reason for placing it first is asymmetry of consequence: every other layer is measuring the quality of an opportunity, while this one is measuring the account's remaining capacity to be wrong. Capacity constraints outrank opportunity assessments in any system that intends to survive its own bad stretches.

Levels two and three

Structural permission caps; tactical evidence selects.

Below capital state sits structural permission — the weekly structural read, Category 6 composite confirmation, and the calculated diagnostic overlays. This layer's verb is cap: it determines whether promotion to higher deployment is structurally justified, and can suppress or veto expansion regardless of how good the week looked. Below that sits tactical evidence: daily all-blended EV, whose role is explicitly limited to cycle-level adjustment and which is never treated as structural truth. The ordering encodes an evidence-quality judgment — a week of structural diagnostics carries more information about the system's condition than a day of blended EV, so the slower signal caps the faster one rather than the reverse.

The bottom of the stack

Discretion is last in precedence and first in daily importance.

Placing operator discretion at the bottom of the authority stack is not a demotion of its value — it's a statement about scope. Discretion cannot expand a pool, lift a cap, or overrule a veto; it operates inside whatever envelope survives the descent. But within that envelope it does the work no other layer can: choosing which setups qualify, judging volatility context, deciding whether a given trade is worth one of the cycle's four slots, and managing it once open. The stack constrains how much capital discretion may commit. It says nothing about how well discretion must be exercised — and the system's results depend enormously on that.

• Lower position means narrower authority over capital, not lesser importance to outcomes.
• Discretion's inputs — branch labels, adherence, honest logging — feed every layer above it, so its integrity propagates upward.
• An operator who resents the stack's constraints is usually experiencing the moment it exists for.

The key idea

Hierarchy is what makes a multi-signal system decidable.

A system with many good signals and no precedence rule doesn't have more information — it has more arguments. The stack converts a collection of readings into a decidable procedure: capital state defines the arena, structural permission sets the ceiling, tactical evidence picks the tier, the throttle interprets, discretion executes. Every genuine conflict the system can produce has a pre-written answer, which means no conflict has to be settled by the person with the most at stake in the moment.