Which gate you are in is one fact. How long you have been there is another.

What the snapshot omits

The same gate is reached by two entirely different histories.

An account that dropped into Buffer on Thursday after a sharp week and an account that has been in Buffer since February hold the identical tier ceiling and the identical operating directive. The gate is correct in both cases — the capital state genuinely is what it is. What differs is everything about what the state means: the first is an episode inside a functioning operation, the second is a condition the operation has failed to leave. Dwell time is the field that separates them, and without it the two are indistinguishable from the gate row alone.

The trap in long dwell

Compression protects capital and slows the recovery it is protecting.

The uncomfortable dynamic in an extended defensive gate is that the mechanism keeping the account alive is also the mechanism keeping it from climbing out. A reduced tier means smaller gains against the same peak, so the drawdown percentage closes more slowly than it opened. That is correct behaviour and it produces a state that can persist far longer than the underlying problem justifies. Dwell is what makes the situation visible as a distinct condition rather than as an unusually long stretch of bad luck, which is how it reads without the duration field. The distinction matters because the two situations call for entirely different responses and the gate, correctly, offers the same one to both.

What long dwell is evidence of

It is a question about the method, and the diagnostics own the answer.

A quarter spent in a defensive gate is not itself a diagnosis, and reading it as one is the mistake in the other direction. What it establishes is that ordinary variance has become an insufficient explanation and the structural layer should be consulted — the six-table pipeline against the account's own history, and the benchmark for whether this dwell is unusual for the branch mix. Dwell escalates the question; it does not answer it. The gate continues to do exactly what it did on day one throughout, which is the correct division of labour.

The other tail

Very short dwell repeated many times is its own finding.

The reading that gets missed entirely is a pattern of brief dwells: an account that touches Recovery, exits within days, and does so repeatedly over a quarter. Each individual episode is unremarkable and none produces a long dwell figure, so nothing in the snapshot or the duration alone flags it. What it describes is an operation running consistently close to a gate boundary — which means deployment is being decided near a discontinuity, and small variations in outcome are producing large variations in authority. That is an instability worth naming even though every individual reading was fine. The usual cause is deployment sized so that a normal losing sequence crosses a boundary, which is a sizing question rather than a gate question.

• Long dwell escalates a question to the structural layer; it is not itself a verdict.
• Compression slows the recovery it protects — a state can outlast its cause.
• Repeated brief dwells describe an account operating on a gate boundary.

Why it lives in this layer

Dwell is computed here because it is a property of the state, not of the results.

It would be possible to derive dwell from equity history elsewhere, and it belongs here because it is a fact about capital authority rather than about performance. CP3 tracks live gate, brake tone and dwell together for that reason — the three describe the account's governing condition as a unit, and the weekly read begins with all three rather than with the gate alone. Dwell also feeds upward: gate dwell is among the telemetry the regime engine consumes monthly, where persistence in defensive states is one of the inputs distinguishing a transitional condition from a degrading one. Reading dwell weekly therefore also means the monthly regime label arrives with context the operator has already seen rather than as a verdict from nowhere.

The key idea

Duration converts a state into a trajectory, and only trajectories can deteriorate.

A gate reading on its own is a photograph and can only ever say where the account stands. Adding the time already spent there turns it into a direction, and direction is what distinguishes an operation working through a drawdown from one that has settled into it. The field costs nothing to maintain — it is a count — and it carries most of the difference between a defensive gate that is functioning as intended and one that has become the account's normal condition.