Dwell, velocity, squeeze: how long, how often, how hard.

Three axes

Duration, frequency, and severity are genuinely independent.

Dwell time records how long the account has remained in its current gate, and its diagnostic weight is directional: long dwell in a stress gate suggests persistent damage rather than a one-day event. Transition velocity counts movement between gate states across the period, describing churn rather than direction. Squeeze depth describes how severely the capital envelope was compressed while the system sat wherever it sat. A month can be high on any one of these and low on the others, and each combination is a different condition requiring a different response — which is precisely why the pressure score keeps them as separate inputs rather than collapsing them earlier.

The pairings that matter

Long dwell with low velocity is not the same as short dwell with high velocity.

The informative combinations sit in the interaction. Long dwell with low velocity describes a system parked — settled in one state, not moving, which is reassuring in Growth and alarming in Floor, since the same reading means stability at the top of the ladder and entrenchment at the bottom. Short dwell with high velocity describes thrashing: a system oscillating between states, never settling long enough for any capital posture to be meaningfully applied. Long dwell with deep squeeze is the compound stress case — parked in a constrained state under real pressure. And high velocity with shallow squeeze is the mildest of the adverse patterns, a system moving a lot without any single move costing much, which usually points at threshold proximity rather than deterioration.

Why velocity is not direction

A transition count says movement happened, never which way.

This is the most common misreading of the three. Transition velocity is a magnitude and carries no sign — a month with six transitions could describe a system deteriorating steadily down the ladder, one recovering steadily up it, or one that moved down and back three times and finished where it started. All three produce the same velocity reading, and the operator's instinct is to interpret high velocity as bad because deterioration is the salient case. Direction lives elsewhere: in gate improvement, which compares start and end state, and in dwell, which says whether the system settled anywhere. Velocity's genuine contribution is stability — it says whether the capital state was a stable posture or a moving target, and a moving target undermines every downstream decision that assumes a settled envelope.

• Six transitions ending higher and six ending lower are the same velocity and opposite months.
• High velocity is a signal about the reliability of the gate state, not about the direction of travel.
• Read velocity with gate improvement, or the reading acquires a sign it does not have.

Why they are not averaged into one

The pressure score combines them and the components stay visible.

The monthly pressure score does weight these into a composite, and it deliberately does not replace them. The reason is the same one that governs every composite in the system: a mid-range pressure score can be produced by long dwell alone, by heavy churn alone, or by moderate readings on all three, and those are different months. Keeping dwell, velocity, and squeeze individually available means the composite can serve as a trigger for attention while the components serve as the diagnosis. An operator who reads only the pressure score has a number that tells them the month was heavy without telling them in which of three quite different ways.

The key idea

A month has a duration, a frequency, and a severity, and no one of them is the month.

Compressing gate behaviour to a single figure is necessary for comparison across months and destructive if it is where the reading stops. The three axes exist because the same aggregate pressure can describe a system that sat still under strain and one that never sat still at all — and the correct response to those two months has almost nothing in common.