Operator brief · 143

Four research tools, four production rules — the boundary drawn precisely.

The key idea

The frame

Research and validation layers — not authority layers.

The manual's classification is unambiguous: the EV Lab, EV sensitivity tables, scenario profiles, and Monte Carlo tools are research and validation layers. They help test whether a rule change, profile change, branch weighting, or management tweak is statistically reasonable. They should not directly rewrite the live production system on impulse. Note what is being restricted — not the tools' quality, and not their conclusions, but their route into production. A sensitivity table can be entirely correct and still have no standing to change a live rule this afternoon.

The four rules

Each component's boundary is stated in its own terms.

The specificity matters because each tool tempts a different violation. EV sensitivity tables map where probability combinations turn green, yellow, or red — and their production rule is to serve as a guardrail map, not a live trade-entry trigger. The EV Scorecard Lab tests scenario profiles, payoff assumptions, fat-tail distributions and Monte Carlo — and its rule is to test changes safely before migrating them to CP3 or the Scorecard. Scenario profiles compare alternate branch-weight profiles — reviewed weekly or monthly, and explicitly not switched because of one emotional week. Monte Carlo benchmarks produce distributions of equity paths and drawdowns — compared against actual performance during review, not during a single trade.

FigureThe R&D components and their production rules

Component	What it tests	Production rule
EV Sensitivity Tables	Where probability combos turn green/yellow/red	Guardrail map — not a live entry trigger
EV Scorecard Lab	Profiles, payoffs, fat tails, Monte Carlo	Test before migrating to CP3/Scorecard
Scenario Profiles	Alternate branch weights, blended EV behavior	Weekly/monthly review — never one emotional week
Monte Carlo Benchmarks	Distributions of equity paths and drawdowns	Compare during review — not during a trade

Straight from the architecture manual's §11 table. Each rule forbids a different specific misuse — which is why the boundary is stated four times rather than once.

The common thread

Every rule is about timescale as much as about authority.

Read the four rules together and a pattern surfaces: three of them forbid a research output from being consulted at trade time. Not a live entry trigger. Not during a single trade. Not because of one emotional week. The boundary is not merely that research cannot rewrite production — it is that research is a review-cadence instrument, and consulting it inside the trading day converts it into exactly the impulse-driven override the directive exists to prevent. A Monte Carlo benchmark consulted mid-drawdown is not being used as a benchmark; it is being searched for permission.

Where R&D sits in a dispute

Last, when modules disagree — and that placement is deliberate.

The QA section's resolution order runs hard plan rules, gate, throttle, cycle exposure, volatility, checklist, scorecard, structural diagnostics, then R&D. R&D is at the bottom of a nine-item list, below even the structural diagnostics that themselves cannot grant live permission. This is the correct placement for a layer whose entire competence is hypothetical: the EV Lab's output is a claim about what would happen under assumptions, and a claim about a hypothetical cannot outrank a measurement of the actual. The tools test the future, in the manual's phrasing. The future does not get a vote on today's size.

The honest reading

The sandbox protects the research as much as the production.

The directive is usually framed as protecting live trading from untested ideas, which is true and incomplete. It also protects the research from contamination. A scenario profile that has been quietly half-adopted — consulted for some decisions, ignored for others, never formally promoted — is no longer testable, because the live results it would be validated against are now partly produced by it. The sandbox keeps the comparison clean in both directions: production stays a controlled environment, and research keeps something honest to compare itself to.

The migration path

“Test before migrating” names a destination, and the destination matters.

The EV Scorecard Lab's production rule is not merely that changes be tested — it is that they be tested before migrating to CP3 or the Scorecard. That phrasing identifies where a promoted change actually lands: in the workbooks that carry live authority, not in the lab where it was validated. A change that has been proven in the sandbox and never migrated is not in production, however convinced the operator is by it. This distinction sounds pedantic until an operator is trading against a profile that exists only in a lab tab and in their intention, at which point the live Scorecard is computing blended EV under weights nobody is actually using.

Connected inside MARS

The R&D Sandbox →

Prime Directive →

Advanced EV Analytics Lab →

Monte Carlo Lab →

Scenario Manager →

Every brief documents the same shipped system.

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

Get MARS Pro — $497 Explore the Platform