Each tool enforces the session gate separately, and nothing checks that they agree.

Why the window is everywhere

Session is a precondition for each tool's claim, not a filter on its output.

The window is not a convenience for avoiding quiet hours. Each tool depends on it for a different reason. The regime switch needs participation behind a volatility reading before that reading can mean expansion rather than thin drift. The continuation stack's momentum and participation components are measuring order flow that barely exists outside the main sessions. The convergence read is sharpest where structure actually resolves, which is around the opens. Three separate arguments arriving at the same window — which is precisely why it ended up implemented three times.

The mechanical fact

The tools cannot read each other's settings, and there is no way to make them.

Chart scripts run in isolation. One indicator has no access to another's inputs, which means the session window cannot be defined once and referenced everywhere — it has to be entered separately in each tool and kept identical by the operator. The manuals say this directly, and the recommended practice is to match the values once and save a preset. That is a sensible workaround and it is still a manual process with no verification behind it. The system's only guarantee that the three windows match is that someone typed the same thing three times and did not change one of them later.

How the drift happens

Nobody edits three tools at once. That is exactly the problem.

The realistic failure is not a wholesale misconfiguration. It is a single adjustment made for a good local reason — extending the window on one tool to examine an early move, widening another during a week of unusual hours — and then not reverted, because reverting requires remembering that the setting was shared. Each tool continues to behave correctly according to its own configuration. Nothing malfunctions. The tools have simply stopped answering questions about the same market, and the divergence has no observable signature on the chart.

What lands in the record

Fields written under mismatched gates are recorded as though they matched.

This is where a settings problem becomes an evidence problem. The journal receives a regime label from one tool and setup context from another, with no indication of the window each was computed under. Grouping those trades by session afterwards produces a slice whose members were not, in fact, all measured the same way. Nothing in the row reveals it. The slice will still compute, still produce a number, and still be wrong in a manner that is invisible to every downstream check — because every downstream check reads the fields, and the fields are internally consistent.

• The mismatch is silent at the chart, silent at capture, and silent in the analysis.
• A single unreverted adjustment is enough to contaminate a review period.
• Comparability across trades is what is actually at stake, not any individual read.

The defence

One written record of the window, checked on a schedule rather than trusted.

Since the tools cannot enforce agreement, the enforcement has to sit outside them: the window recorded once in the plan, the tools brought back to it deliberately rather than assumed to be at it, and a check folded into the weekly routine — where the review is already looking at settings and the market is closed. The point is not that drift is likely in any given week. It is that drift is undetectable after the fact, so the only moment it can be caught is before the fields it would corrupt have been written.

The key idea

A constant duplicated across instruments that cannot see each other is a maintenance obligation.

Every shared value held in more than one place, with no mechanism to reconcile the copies, is a promise that someone will keep them aligned by hand. The chart layer contains exactly this, unavoidably, because the platform provides no way to link the settings. Naming it as an obligation rather than treating it as a solved detail is the whole of the remedy — the tools will never announce that they have drifted, so the announcement has to be a habit.