Structural Health · Verdict Logic
Hierarchy, not averaging.
Diagnostic categories
6
Readout fields
5
Blends disagreement
NEVER
Read this first
What the verdict says
01Five fields, and the last one is the only instruction.
The readout names the overall system state, whether the entity is scale-eligible or maintenance-only or throttle-biased or audit-biased, which metric is driving the state, what the primary weakness is, and finally a plain-English posture: press the edge, maintain, scale selectively, throttle, audit, or protect capital. Everything before the last field is evidence. The last field is the only part that is meant to change behaviour, and it is still subordinate to the gate.
Reading order
02Classes first, gaps second, scores last.
The interpretation flow is fixed and it deliberately defers the composite. Read the direct class outputs first: efficiency, profit quality, risk context, expectancy, hit rate, matrix state, functional role. Then read the directional gaps to see whether live deployment is helping or diluting the benchmark reading. Only then read the normalised and composite scores, and only once their helper scores are stable and validated. The narrative metrics come next, and the final readout comes last, as a summary rather than as an override.
- A composite read before its components is a number without an explanation attached.
- Composite scores are only trustworthy once the helper scores feeding them have been validated.
- The readout summarises the condition; it never overrides a governance rule.
Disagreement handling
03Split threads produce a flagged verdict, not a mushy middle.
When threads disagree, expectancy stable while efficiency decays, the engine issues the more cautious verdict and names the dissenting thread rather than averaging the two. The reasoning is that a moderate average of one healthy subsystem and one failing subsystem describes a machine that does not exist, and the operator would then manage the imaginary average instead of the real failure. Naming the dissenter is what makes the verdict actionable.
Falsifiability
04A verdict that cannot be wrong is not a verdict.
Every verdict is written into the monthly review with the reading that produced it, which creates a health history the operator can actually see trending. That record is what makes the engine accountable to itself: a verdict of holding that is followed by three months of deterioration is a fact about the engine, not only about the account, and it is visible only because the verdict was written down at the time rather than reconstructed afterwards.
How MARS uses this
Each closed trade logs its worst adverse excursion and best favorable excursion. The lab plots them into quadrants and pairs the picture with capture efficiency, giveback, and fee drag - then the EV Driver Diagnostics rank which repair (entry location, coefficient, trail logic, cost avoidance) would move net expectancy most.
How it benefits you
Review produces one concrete fix instead of a vague resolution to 'trade better'. You see whether stops are too tight, entries are chased, exits leak the runner, or fees quietly eat the edge - and whether a profitable week was actually clean or just lucky.
Every trade plotted by pain taken (MAE) versus opportunity generated (MFE). The four quadrants diagnose entry precision and exit management at a glance.
Reference
The six wiring categories behind the verdict
| Category | Metric type | How it is wired |
|---|---|---|
| 1 | Direct spectrum and lookup reads | A live value resolves to a class through the spectrum, matrix, or lookup tables |
| 2 | Directional comparison | Two values compared, returning positive, neutral, or negative meaning |
| 3 | Normalised band scores | Helper scores on a zero to one scale, mapped onto a meaning ladder |
| 4 | Semantic modifiers | Class and context labels pulling narrative meaning from the lookup layer |
| 5 | Rule composition | Several diagnostics combined into a stance or narrative sentence |
| 6 | Composite rollups | Weighted composites of normalised sub-scores, on a granular ladder |
Edge cases & failure modes
Where it can mislead
- !Erratic strength scores below steady adequacy, because stability is weighted rather than assumed.
- !A strong average carrying one critical thread is a warning with good camouflage, which is why survival reads can cap the verdict outright.
Further illustration
Six structural inputs are scored monthly into a composite, and the composite maps onto the regime ladder. MARS reads results in context: the same numbers mean different things earned in Expansion tailwind versus survived through Instability, and posture is adjusted to the regime rather than to the last trade.
Twelve months classified across the regime ladder - the system's own weather record, from Expansion down toward Drawdown Control.
Connected inside MARS
This module doesn't work alone.
Go deeper
Operator briefs on this territory.
Deep dive — 01
60/30/10: the weights are the doctrine, written as arithmetic.
Compliance 60, compression 30, gate improvement 10. Read the weights as a ranking and the doctrine is right there.
Read the full brief →
Deep dive — 02
How three threads become one answer — and what the answer throws away.
One number from three threads. Compression is why it is readable — and why the same score can mean four things.
Read the full brief →
Deep dive — 03
Thresholds published in advance, or the verdict is just a mood.
Fix the bands before the month, or the month will choose the bands. The order is the whole safeguard.
Read the full brief →
Every module ships in the complete MARS package.
One price. Eleven workbooks, three TradingView indicators, and the full manual library — $497.