Skip to content
← Back to MON — Monte Carlo

Operator brief · 392

Expectations are manufactured before exposure, never after.

The key idea

The ordering claim

Before a single live trade is sized, the adverse bands are already drawn.

The sequence is the doctrine. Evidence is resampled, the distribution is produced, the adverse bands are read, and the tier is set to survive them — and only then is capital committed. Nothing in that chain is unusual as a list of activities. What is unusual is the insistence on the order, because the same activities performed in a different sequence produce a fundamentally weaker instrument.

The contamination

A benchmark built after the result is a rationalisation with a chart.

Construct the distribution after a drawdown, and the parameters get chosen — unconsciously and in good faith — by someone who already knows what happened and wants it to be normal. The window gets extended slightly, the outlier gets reconsidered, the branch mix gets adjusted, and the resulting fan comfortably contains the recent experience. Every individual choice is defensible. The output is worthless as a check, because it was fitted to the thing it was supposed to evaluate. The contamination is also difficult to detect after the fact, because the resulting fan is internally consistent and the individual parameter choices remain defensible when examined one at a time.

  • Built before: the bands are a prediction the live result can fail.
  • Built after: the bands are a description the live result cannot fail.
  • Only the first can tell you anything you did not already believe.

What it buys during the drawdown

The value arrives at the moment the operator wants to change something.

The practical return on pre-commitment is entirely concentrated in bad stretches. A trader eleven percent down consults a benchmark drawn before the drawdown began and reads that the P10 path descends to nineteen percent — so this is inside the envelope, unpleasant, and expected. That reading is only load-bearing because it predates the pain. The identical number, computed that morning, carries no authority whatsoever, and the operator knows it, which is why it fails to settle anything.

FigureThe order that keeps the benchmark honest
Recordreal trades, real frictionResampleinto thousands of orderingsDraw the bandsadverse path read firstSet the tiersized to survive P10Deployinside the stated envelopeComparelive against the drawn bandsBEFORE CAPITAL

The bands are drawn before exposure. Regeneration happens on schedule, never in response to a result.

Regeneration discipline

The benchmark is rebuilt on a schedule, never in response to a result.

It follows that when the ruler is remade matters as much as how. Regeneration is scheduled and triggered by evidence accumulation or structural drift — never by a disappointing month. A benchmark rebuilt because live results were uncomfortable is the contamination problem arriving through a side door, and it is more insidious because it feels responsible: the trader believes they are updating with new information when they are moving the goalposts. The schedule exists to make that impossible to do accidentally.

The wider principle

Anything that judges a result must predate it.

The rule generalises across the whole system and appears in several places wearing different clothes. The gate ladder is written while calm and applied under pressure. The override reasoning is logged before the trade, not after. The weekly questions are fixed in advance so they cannot be replaced with kinder ones. All three are the same discipline: an evaluative standard authored after the event it evaluates is not a standard, and the temporal ordering is what makes the difference. Stated generally: the authority of any standard comes from when it was written, not from how strict it is, and a severe rule authored after the fact carries less weight than a lenient one authored before.

Why it earned the first syllable

Everything downstream is a comparison, and comparisons need the reference first.

Expectancy grades live performance against something; alpha is excess over something; the rails keep behaviour inside something. All three presuppose a reference that already exists. Putting MON first in the name records that dependency — the benchmark is not one component among four, it is the one the other three are measured against, and a system that skipped it would still have expectancy and rails while having no way to say whether either was producing anything.

Connected inside MARS

Every brief documents the same shipped system.

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