Skip to content
← Back to The Five States

Operator brief · 349

Fourteen functions deep: how two numbers become one diagnosis line.

The key idea

The chain

Every weekly block runs the same fourteen functions.

The engine's calculation grid runs weeks vertically and composite branch blocks horizontally, and each block computes an identical sequence. RAER Cum and RAPF Cum are referenced from Table 1 of their respective SDE pipelines — referenced, not recalculated, which is what keeps the engine from being able to disagree with the pipelines it reads. Each is classified against its own published spectrum thresholds. The two class labels concatenate with an underscore into a matrix key. That key resolves, through the efficiency matrix lookup, into a base state and a functional role. Primary drift for both metrics is pulled from Table 6 of each pipeline, and the pair of drifts determines the conflict type. Conviction from both sides — again from Table 6 — resolves into severity, typically taking the stronger of the two. The final function combines base state, conflict type, and severity into a single diagnosis.

FigureThe fourteen functions and where each value comes from
stagefunctionsourcekind
1 — levelRAER Cum · RAPF CumSDE Table 1referenced
2 — classRAER Class · RAPF Classspectrum thresholdscalculated
3 — keyMatrix Keyclass + "_" + classcalculated
4 — stateBase State · Functional Roleefficiency matrix lookupreferenced
5 — directionPrimary Drift ×2 → Conflict TypeSDE Table 6both
6 — weightConviction ×2 → SeveritySDE Table 6both
7 — outputFinal Diagnosisstate + conflict + severitycalculated

Referenced values are read from the SDE pipelines unchanged. Calculated values are derived inside the engine. Nothing in the chain is entered by hand, which is what makes a state reproducible.

Level, then direction, then weight

The chain answers three questions in a fixed order.

The fourteen functions group into three questions, and the order they are asked in is the substance of the design. Where are we — the two classes and the base state, which describe the current standing relationship between efficiency and profit quality. Which way is it moving — the conflict type, derived from comparing the two drifts rather than from either level. How much should we believe it — severity, derived from conviction. A base state without direction is a photograph of a moving system; direction without conviction is a claim with no strength attached. The final diagnosis line is only meaningful because all three have been established separately and then combined, and an operator reading only the state label has taken the first answer and discarded the two that qualify it.

FigureThe three questions, in order, every week
Level — where are wetwo classes, one base stateDirection — which waytwo drifts, one conflict typeWeight — how firmlyconviction resolves to severityDiagnosis linestate + conflict + severityEACH WEEK, EACH COMPOSITE

Each composite block repeats the same loop each week. Level establishes the standing relationship, direction qualifies it, weight says how firmly it is being asserted — and only then does a diagnosis line exist.

Reference, don't recompute

The engine reads the pipelines; it never re-derives them.

Every value in the chain that could have been calculated independently is instead referenced from the SDE pipelines — the cumulative levels from Table 1, the drift and conviction from Table 6. This is a deliberate constraint with a clear rationale, and it is the same one that keeps the housed engines inside CP3 rather than in separate workbooks: an engine that reads the journal must not be able to disagree with the journal about what the journal says. If the Contradiction Engine computed its own RAER, there would be two RAER values in the system, and any divergence between them would be indistinguishable from a genuine finding. Referencing eliminates that class of failure entirely at the cost of making the engine dependent on its upstream pipelines being correct — a dependency the QA layer exists to police.

  • A referenced value cannot drift from its source, because there is only one value.
  • The cost is a hard dependency: a broken SDE table produces a confidently wrong contradiction state.
  • This is why the workbook's control panel carries an engine status field, and why it was wired last.

Why the control panel came last

A summary layer can only be built after the thing it summarises.

The workbook's build order is itself informative. Section 1, the control panel — selected week, latest available week, selected composite, mode, data source, engine status — was wired last, because it summarises a finished engine. Building the readout first is a recognisable failure pattern in analytical workbooks: the dashboard gets designed around what would look good, and the calculation layer is then shaped to produce it. Building it last means the control panel reports what the engine actually computes, including the fields that are inconvenient to display, such as an engine status that says the data source is stale. The ordering is a small discipline and it determines whether the summary layer serves the calculation or the reverse.

The key idea

A reproducible state is worth more than an insightful one.

Nothing in the fourteen functions is clever. Every step is a threshold, a concatenation, a lookup, or a maximum — which means two operators running the same weeks get the same diagnosis, and the same operator running the same weeks six months apart gets it again. That reproducibility is the property that lets a contradiction state be counted, tracked for persistence, and argued with. An interpretation cannot be any of those things.

Connected inside MARS

Every brief documents the same shipped system.

The complete MARS package — eleven workbooks, three TradingView indicators, the full manual library — $497.