Compliance Panel 3.
The operating ledger of MARS.

Layer 01 — Evidence

Journal & Trade Evidence

Every trade becomes structured truth: branch, session, R outcome, fees, MAE/MFE, adherence flags, and context. CP3’s journal is the single source of evidence for the entire ecosystem — if it isn’t logged, it doesn’t exist.

• Trade-level capture across branch, pair, session, and setup context
• R-based accounting so results are comparable across account sizes
• Fee and swap capture feeding net-of-friction expectancy
• Adherence flags that make discipline measurable, not anecdotal
• Manual manual-vs-auto discipline: Date, branch, risk, outcome, hit flags, BE status, fees, MAE/MFE, duration, and notes are entered; Risk $, Outcome $, Fees % of 1R, and Fees % of balance auto-calculate and get verified
• The manual is blunt: if the Journal is wrong, every rollup becomes wrong — treat it as a controlled trade ledger, not a casual diary
• Saturday discipline: entries reconciled from paper capture sheets against broker records before weekly review runs

Layer 02 — Aggregation

Weekly Summary Engine

Trade evidence rolls upward into weekly truth: EV by branch, blended expectancy, hit rates, drawdown movement, and gate dwell. Weeks are the heartbeat of MARS — the unit at which posture decisions are made.

• Branch-level weekly EV with blend weighting
• Week tagging — GREEN / YELLOW / RED expectancy states
• Rolling 4W / 6W / 12W windows for trajectory, not snapshots

Layer 03 — Capital Authority

Gate / Brake State

Drawdown from Equity Peak High routes capital into one of seven gates — Growth through System Lock. CP3 tracks the live gate, brake tone, and dwell time. No lower signal can override this layer.

• Deterministic gate routing from EPH drawdown
• Brake tones defining posture from full authority to full stop
• Gate-dwell tracking — how long capital lives in defensive states

Layer 04 — Expectancy

EV Scorecard

The expectancy engine: EV computed at branch, blend, weekly, monthly, rolling, and structural layers. Profit is what happened; the scorecard shows what the machine is worth.

• Normal, Trend Partial, Trend No-Partial, and Overflow EV
• Blended EV under live weight profiles
• Stability and drift reads that separate edge from variance

Layer 05 — Allocation Quality

Branch Risk Engine

Which branch earns the risk budget — and which quietly leaks it? The Branch Risk Engine scores allocation quality so capital flows toward demonstrated edge instead of habit.

• Per-branch risk consumption vs. contribution
• Quality context feeding the Risk Deployment Quality diagnostics
• Attribution that survives blended P&L camouflage

Layer 06 — Compounding

Capital & Compounding Dynamics

Equity path, growth rate, and acceleration — CP3 reads whether compounding is gaining momentum, flattening, or decaying beneath a healthy-looking curve.

• Equity Peak High tracking and recovery mapping
• Growth-rate and ACCEL layers across rolling windows
• Friction accounting — what fees and giveback cost the curve

Engine 07 — Structure

Structural Diagnostic Engine

Six-table pipelines — current value, cumulative truth, rolling condition, stability, drift, z-score — run across EV, DD, RAER, RAPF, and ACCEL. The SDE turns raw metrics into structural meaning and feeds the Interpretation Cockpit’s final readout.

• Maturity-aware interpretation — no verdicts on immature data
• Drift and z-score layers that catch decay before P&L does
• Interpretation Cockpit: classified signals → decision-grade readout
• Diagnoses EV, DD, RAER, RAPF, and ACCEL across six structural tables using rolling windows, stability, drift, and z-scores
• Analyzes seven entities — four primitive branches plus three composites — so weakness is localized before it spreads
• Middle authority by design: it sets review posture, and it can never override Gate/Brake capital protection

Engine 08 — Context

Regime Classification Engine

A monthly diagnostic layer — six gate-behavior metrics scored −2 to +2, composited from −12 to +12, and mapped to ten regimes from Expansion to Critical State, each with a driver tag. Context, not control: the regime describes the month; it never sizes a trade.

• Ten regimes with four sub-labels each — forty nameable states
• Gate Pressure, Structural Health, Volatility, Transitions, Persistence, Compression
• Dashboard-ready output: Main Regime — Driver Tag
• Ten regime labels from Expansion to Critical State, each with four sub-labels and a call-to-action
• Six metrics scored −2 to +2 into a −12…+12 composite from monthly gate and brake telemetry
• Classifies the system's own internal condition — explicitly not a market-regime detector

Engine 09 — Truth Testing

Contradiction Engine

RAER and RAPF must agree for confidence to be justified. When efficiency and profit quality diverge — under-monetization, profit without efficiency — the Contradiction Engine names the state and routes the review.

• Fixed 5.5% RAPF calibration rail vs. live risk readings
• Named contradiction states with routing guidance
• The sharpest special case of signal coherence in MARS
• Five states: alignment, under-monetization, profit-without-efficiency, dual weakness, and transition
• Catches the silent divergence where blended EV still looks fine while quality metrics stop confirming each other