Skip to content
Montex Alpha Rail System
$497
← Back to Forex Strategy Evaluation

Operator brief · 432

Keep, adjust, retire — and why adjust is the dangerous one.

The key idea

The three exits

Every evaluation terminates in one of three places.

The pipeline is built to produce a verdict, and the verdict set is deliberately small. Keep means the edge claim survived and deployment continues at its current authority. Adjust means something specific and identified is wrong, and a defined change is made to address it. Retire means the strategy no longer earns capital and stops receiving it. The narrowness is the point: a verdict set that includes 'watch it for a while longer' is a verdict set with an infinite loop in it. Watching, in this architecture, is not a verdict — it is the default state a strategy occupies while evidence accumulates, and calling it a decision is the most common way an evaluation is quietly abandoned rather than completed.

FigureThe verdict, and what each one obliges
evidence in, obligation out ↓Evaluation completesBranch EV, stability, robustness, benchmark placementKEEP — authority unchangedDuty: keep grading weekly; nothing is settledADJUST — one named changeDuty: the evidence clock restarts from zeroRETIRE — capital withdrawnDuty: record why, so the reasoning outlives it

Every exit carries a duty. The one with the heaviest duty is the middle one.

The trap

Adjustment restarts the clock, and almost nobody prices that.

When a strategy is modified, the evidence gathered under the previous configuration no longer describes the thing now being traded. Strictly, the sample resets. A trader who adjusts every time results disappoint therefore never accumulates enough evidence about any single configuration to reach a verdict on it — while experiencing every adjustment as decisive action. This is the mechanism by which a strategy can be traded for two years and evaluated for none of them. The activity feels rigorous. The evidence base stays permanently newborn.

What earns an adjustment

A named cause, not a disappointing outcome.

The discipline that makes adjust usable is a requirement that it be attached to an identified structural cause, not to a result. A red grade is a symptom, and the Scorecard is structurally incapable of naming what produced it. If the diagnostic work has located something specific — capture efficiency collapsing on one branch, a volatility regime the envelope did not anticipate, friction that grew — then adjustment is targeted repair and the sample reset is a price knowingly paid. If the only input is that the last stretch was poor, the change is not repair. It is churn with a rationale attached. The practical filter is to require the cause to have been named before the outcome was known, or at minimum to be identifiable in the diagnostic record rather than reconstructed from the disappointing period. A cause discovered by looking backwards at a bad month is usually a description of the month.

Why retire is easier than it feels

The verdict is about the capital, not about the trader.

Retirement is resisted because it is read as a judgement on the person who built the strategy. Mechanically it is nothing of the kind: it is a statement that this configuration, on this evidence, no longer justifies capital. The strategy is not destroyed. Its record is kept, the reasoning is written down, and it remains available for re-examination if conditions or the evidence change. What ends is its claim on deployment. Framed that way, retire stops being an admission and becomes ordinary portfolio hygiene.

The threshold problem

Publish the retirement condition before the drawdown, not during it.

A retirement threshold chosen while underwater is not a threshold; it is a negotiation with a losing position, and it will move. The condition has to be defined in advance, in the same act that justified deployment — the level of sustained expectancy decay, or benchmark placement, or structural finding at which capital is withdrawn. Written early, it costs nothing and is easy to state honestly. Written late, it is written by the part of the operator most invested in the answer being no.

Keep is not passive

The verdict that changes nothing still obliges something.

Keep is often treated as the null outcome — the evaluation happened, nothing moved, back to trading. It carries the same obligation as the others: the weekly grading continues, and the verdict stays open to revision on new evidence. A strategy that was kept in March has not been cleared for the year. It has been cleared until the next reading disagrees, which is the only kind of clearance an evidence-based system is able to issue. Treating keep as a settled matter is how a strategy ends up carrying capital for a year on the strength of one favourable review conducted at the start of it.

Connected inside MARS

Forex Strategy Evaluation →
Weekly Trading Scorecard →
Structural Diagnostic Engine →
EV & Edge Validation →

Every brief documents the same shipped system.

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

Get MARS Pro — $497Explore the Platform