Skip to content
Montex Alpha Rail System
$497

Live vs Benchmark · The Comparison Stack

Five metrics, plus behaviour.

Metric rows

5

Behaviour checks

4

Read as

ONE SET

How MARS uses this

The five-metric comparison stack ends with behavior, and tier usage is where behavior hides. Live EV and drawdown can both pass while tier dwell quietly drifts aggressive — which is why the stack includes this audit: the model's expected tier profile against what the operator actually deployed.

How it benefits you

Aggression creep gets caught while it is still a pattern on a chart rather than an oversized loss. Equally, unnecessary timidity shows up as a measurable gap, so you deploy the edge you have actually earned instead of leaving modeled expectancy on the table.

0%10%20%30%T1T2T3T4T5T6T7BENCHMARKLIVE

The behavioral row of the comparison stack: modeled tier dwell versus live deployment, T1 through T7.

Why five and not one

Any single row can flatter.

A hot expectancy row with degrading tier discipline is a worse report than a flat expectancy row with clean behaviour, and only the full stack read in one sitting makes that ranking visible. Reading rows individually, in whatever order they happen to be encountered, reliably produces the most comfortable available conclusion. The stack is therefore read as a set, at a scheduled slot, with no cherry-picking permitted before the set is complete.

The behavioural half

Behaviour is where drift hides while the metrics still look fine.

Gate dwell, tier usage, throttle behaviour, and open exposure are the four checks that catch a system whose numbers have not moved but whose habits have. Elevated tier usage during a defensive gate history, frequent full-tier overrides, or open risk that failed to compress fresh deployment are all findings even in a profitable month, because each one describes authority being exceeded rather than performance being poor.

  • Overrides are permitted but must be deliberately logged, with a written reason.
  • Capacity compression is normal behaviour and is not a missed opportunity.
  • Under-usage of authorised tiers is also a finding: unmonetized modelled expectancy is still a leak.

Where each row is sourced

The stack is fed from live workbooks, not from memory.

Compliance Panel 3 supplies equity, return, drawdown, gate state, and capital metrics. The Cycle Command Console supplies cycle-level closed R, open risk, floating R, and smart-capacity inputs. The Throttle Control Panel supplies tier usage, authorised pool, and exposure behaviour. The Weekly Scorecard supplies expectancy context. Each row of the stack therefore has a named source, which is what allows a disagreement between rows to be investigated rather than argued about.

The authority boundary

04

The comparison informs interpretation. It grants nothing.

A favourable stack does not permit exceeding the current gate cap, ignoring a lock condition, or using a full-tier override casually. This is stated as a hard boundary rather than a preference, because a benchmark that can authorise deployment has quietly become a second risk authority competing with the first, and the two will eventually disagree during the exact week when clarity matters most.

Reference

The stack: each row, its evidence, and what it proves

RowQuestionEvidence
ExpectancyIs live EV equal to or better than modelled EV?Return distribution by reporting period
DrawdownIs live drawdown equal to or better than modelled?Adverse bands and gate threshold breach probabilities
Profit qualityIs risk-adjusted profit factor confirming the result?The RAPF pipeline and its spectrum classes
Risk conversionIs risk converting as efficiently as assumed?The RAER pipeline, cumulative PnL over cumulative risk
AccelerationIs growth velocity tracking the modelled path?The acceleration pipeline and recovery velocity
Gate dwellIs time in each gate normal?Expected dwell and worst gate reached
Tier usageIs deployment inside the modelled profile?Tier usage by period and tier-by-gate conditionals
Throttle behaviourIs deployed risk converting into return?R per deployed pool, exposure efficiency, throttle value-add
Open exposureIs carryover compressing fresh deployment as designed?Carryover, active risk burden, remaining pool capacity

Doctrine

Band13 wk26 wk52 wkOperator read
P90+64.2R+131.8R+268.4Rhot path — do not extrapolate
P75+47.5R+96.3R+197.0Rstrong but ordinary
P50+31.1R+63.9R+130.6Rthe median story
P25+16.8R+35.4R+74.2Rslow — still inside the model
P10+4.3R+11.7R+27.9Rthe path sizing must survive

Cumulative R by horizon · 50,000 resampled paths

Further illustration

MARS reduces the fan to its actionable rows. Each percentile band carries a defined operator read - P50 is the planning story, P90 is explicitly quarantined from expectations, and P10 is the row the tier structure must survive. Reviews quote these rows directly instead of eyeballing a chart.

The percentile bands rendered as an operator table: cumulative R by horizon, and what each band is allowed to mean.

Connected inside MARS

This module doesn't work alone.

Connected — 01

Live vs Benchmark (full page)

Open system page →

Connected — 02

Classifying Deviation

Open system page →

Go deeper

Operator briefs on this territory.

Deep dive — 01

Table K: live results against the simulated envelope, by the book.

Three live numbers, seven bands, one fixed workflow — the monthly comparison cockpit by the book.

Read the full brief →

Deep dive — 02

Every row in the stack can flatter alone. None can flatter together.

A hot expectancy row with degrading tier discipline is a worse report than a flat one with clean behaviour.

Read the full brief →

Deep dive — 03

Profit quality and risk conversion ask different things of the same month.

One asks what the profit is made of. The other asks what it cost in deployed risk.

Read the full brief →

Every module ships in the complete MARS package.

One price. Eleven workbooks, three TradingView indicators, and the full manual library — $497.

Get MARS Pro — $497Explore the Platform