Two layers: the base matrix says where you are, the engine says where it's going.

Layer one

The base matrix: sixteen states, permanently.

The foundation is a 4×4 grid — RAER class down, RAPF class across, each axis running Weak, Stable, Strong, Elite — with a named meaning in every cell. Weak with Weak is Collapse; Weak RAER against Elite RAPF is Illusion; Strong against Weak is Inefficient Profit; Elite against Elite is Peak. The language is institutional and the classifications are stable, which is exactly the point: this layer is the non-changing anchor. The build guidance was explicit about preserving it untouched — it becomes the base meaning engine, clean and stable, and everything dynamic gets built above rather than woven into it. A moving interpretation layer needs a fixed reference, and the matrix is it.

The gap

A state lookup can't see change coming.

The matrix's limitation is structural rather than a flaw: it is state-based, so it reports where the system currently sits and nothing about where it's heading. 'Strong RAER, Weak RAPF — consistent but under-monetized' is a true and useful sentence that would read identically whether the under-monetization has been stable for a year or has appeared violently in the last three weeks. Those are entirely different situations demanding entirely different urgency, and a static lookup cannot distinguish them. What was missing was direction, time-awareness, and a quantitative sense of tension — the difference between a snapshot and a trajectory.

Layer two

Direction: the seven conflict types.

The dynamic layer reads each metric's Primary Drift Meaning from its SDE Table 6 — Strengthening, Breakdown Risk, or Transitional — and maps the pair to a conflict type. Strengthening with Strengthening is Alignment. Strengthening against Breakdown Risk is Under-Monetization; the reverse is Profit Without Efficiency. Breakdown with Breakdown is Dual Weakness. Transitional with Transitional is Transition, meaning no clean directional dominance. And two lead states complete the set: Strengthening against Transitional is RAER Leading — efficiency improving before profit confirmation — while the mirror is RAPF Leading. Those last two are the engine's early-warning subtlety, catching divergences while one metric is still making up its mind.

Layers three and four

Severity from conviction, then a status flag.

Direction without magnitude would fire constantly, so the engine grades intensity from the conviction readings the SDE pipelines already produce. Conviction scores map Low to 1 through Very High to 4; the two metrics' scores average; and the result falls into a severity band — below 1.5 is Low, below 2.5 Moderate, below 3.5 High, and 3.5 or above Very High. Conflict type and severity then combine into a status flag with prescribed handling: Dual Weakness at High or Very High is Critical; Profit Without Efficiency or Under-Monetization at those severities is a Warning; Alignment at high severity is Strong; Transition at Low or Moderate is simply Monitor. The final diagnosis string carries all of it — state, direction, and confidence in one readable verdict.

• Severity derives from conviction, which derives from z-magnitude — the whole stack ultimately rests on normalized drift.
• A dramatic conflict type at Low severity is a watch item, not an action item. Both dimensions gate the response.
• Alignment at high severity is worth documenting, not just enjoying — it's the profile the system is trying to reproduce.

The key idea

Build above the anchor, never inside it.

The engine's architecture is a small masterclass in extending a system without destabilizing it: the proven static layer stays exactly as it was, and everything dynamic reads from it rather than rewriting it. The payoff is that the base meanings remain comparable across the entire record while the interpretation layer above them gets progressively smarter — and if the dynamic layer ever needs rebuilding, the anchor it was built on is still standing. 'Strong system quality, but monetization is deteriorating rapidly, high-confidence divergence' is a decision-grade sentence. It's also, structurally, just the old matrix with three honest layers stacked on top.