Operator brief · 167

Before the engine runs, the simulation must be declared.

The key idea

The shape of the commitment

Nine settings define every run the Lab will ever produce.

The Parameter Console is the canonical control layer, and the settings it holds are the entire identity of a simulation: account preset, target mode and multiple, path count, risk mode, trades per week, horizon weeks, branch profile, cost model, and random seed. Every other tab in the workbook resolves inputs that feed back into these. The design point is that the console wins any disagreement — if the Quick Start panel and the console appear to differ, the console is the truth and the panel is a mirror of a subset. One surface holds the commitment, which is what makes a run citable later. A configuration scattered across a dozen tabs with no precedence rule is not a plan; it is a mood that happens to produce numbers.

The pipeline

Configure, validate, recalculate, run, recalculate, read.

The operating sequence is fixed and its most-skipped step is the dull one. The engine reads cached formula values from the workbook, so any save that drops those caches leaves the engine simulating stale configuration — the manual names this as the single most common cause of blank or wrong results. Hence the discipline: recalculate the input workbook before the run so the configuration formulas are evaluated, then recalculate the output workbook after it so the dashboards pick up what the engine wrote. Two recalculations bracket every honest run. Neither is optional, and neither announces itself when skipped, which is precisely why the sequence is doctrine rather than advice.

FigureThe golden operating sequence — six steps, two of them recalculations

The bracketing recalculations are the steps most often skipped and the ones whose absence is hardest to detect: a stale-cache run fails silently, producing plausible numbers from a configuration that is no longer on screen.

The first gate

Probability integrity: the ladder must be well-formed before it can be sampled.

Zone 3 carries its own gate, and it checks structure rather than optimism: monotonicity across the hit-rate ladder, values inside legal ranges, blank-safety where an operator entered nothing, and branch weight sums landing at exactly 1.00. These are not stylistic preferences. A non-monotone ladder — a higher R-multiple recorded as more likely than a lower one — describes a market that does not exist, and an engine sampling from it produces confident nonsense. The gate exists because the failure is invisible downstream: nothing about the resulting equity fan looks wrong. The corruption is upstream, in a ladder that was never coherent.

The second gate

Expected value: a go or no-go verdict before any path is dealt.

The pre-simulation expectancy check computes per-branch and blended expectancy along with the payoff ratio, and returns a verdict. Its function is to catch the case where the configuration describes a losing system before fifty thousand paths are spent demonstrating it in detail. This is worth stating plainly because it inverts the usual assumption about what simulation is for: the Lab is not the instrument that tells you whether your edge is positive. Arithmetic tells you that, in advance, for free. The Lab tells you what happens to a positive edge when order, friction, and governance get involved — which is a different and much harder question, and one that only becomes worth asking after the cheap check passes.

The master verdict

One aggregate reading, and a flagged row naming the zone at fault.

Zone 12 aggregates every zone's own validation into a single readout: cleared to simulate, or blocked. When it blocks, it does not merely refuse — it names the zone responsible, so the operator is debugging a location rather than a mystery. The architectural choice here is worth noticing. The workbook could have let a run proceed with warnings, which is what most analytical tools do, and the results would look identical to a clean run. Instead the gate is binary and it sits between the operator and the button. A simulation that cannot state its own readiness is not permitted to produce a number that will later be quoted in a review.

The key idea

The configuration is the experiment; the paths are only its consequence.

Everything the page above describes — modeled branch probabilities, payoff structure, tier allocation, throughput — is a set of declarations made before the engine starts, and the fifty thousand paths are nothing more than those declarations played out to exhaustion. This is why two simulations that disagree almost never disagree about mathematics. They disagree about what was declared, and the declaration is recoverable only if it was pinned to one surface, gated by integrity checks, and cleared before the run. Treat the console as the experiment and the fan as the result, and every simulation the Lab produces stays auditable a year later. Treat the fan as the experiment, and none of them do.

Connected inside MARS

Monte Carlo Lab →

R&D Boundary →

Quality Assurance →

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