Why the benchmark moves with your tier path.

The static lie

Fixed-risk simulation benchmarks a rule-breaker.

The standard retail Monte Carlo assumes the same risk fraction on every trade of every path, forever — through a −25% drawdown, through a cold streak, through everything. But the live MARS operator is forbidden from behaving that way: at −13% the gate caps deployment, at −19% it compresses further, and the tier allocator is re-pricing every cycle against evidence. Comparing that governed live account to an ungoverned simulation produces systematic nonsense in both directions — the simulation drawdowns run deeper than yours ever could, and its recoveries run faster, because it never de-risks. A benchmark that models behavior you're not allowed to exhibit is not a benchmark. It's a control group from a different experiment.

What 'dynamic' means mechanically

Every simulated path obeys the same governance you do.

Inside the benchmark, each of the 50,000 paths runs the full year — 52 weeks at the live rhythm of 4 concurrent-trade cycles per week, 16 trades weekly — and after every completed cycle, the path's drawdown state is evaluated against the gate ladder. A path that slips past −7% gets capped at T5. Past −13%, the defense brake compresses it to T3/T4 territory. Past −19%, containment; past −27%, suppression; past −34%, T1 only; past −40%, lock. When the path recovers, the caps release in the same order the live system's would. The simulated equity curves therefore carry the exact fingerprint of governed deployment: shallower adverse tails than fixed-risk would show, slower deep-drawdown recoveries — because the paths were de-risked down there, exactly as you would be.

Why this makes comparison legitimate

Like is finally compared with like.

Once every simulated path obeys the ladder, the percentile bands become genuinely yours: the median is the centerline of governed futures, the adverse percentiles describe drawdowns that governed deployment can actually produce, and gate-dwell statistics — how much simulated time is spent in Buffer, Floor, Deep-Floor — have live counterparts worth checking. This is what makes the entire live-comparison workflow legitimate. When your month lands at P40 on the equity table, that placement means something, because the 50,000 accounts you were ranked against were playing your game under your rules. Against a static benchmark the same placement would be a category error dressed as a statistic.

The ruler doctrine

A ruler, not a motivational target.

The Operating Standard's core doctrine states it flatly: the benchmark is a ruler. The median is not a goal to chase; it is the expected centerline. The bands are not grades; they are the normal range of governed outcomes. Live performance is compared to bands and risk-state behavior — never judged emotionally from one week. And the hard boundary holds in both directions: the benchmark informs interpretation but grants no permission. It cannot raise a gate cap, excuse an override, or soften a lock condition. Drawdown state remains the highest capital authority, and the benchmark's own assumption register says so on its first page.

• Fast compounding is not automatically good if it arrives with abnormal drawdown, casual overrides, or tier usage the gate never authorized.
• Benchmark assumptions are refreshed after approved model changes — never from casual live-week emotion.
• If live assumptions have drifted from the register (branch mix, rhythm, cost model), document the drift before comparing at all.

The key idea

The benchmark simulates your constitution, then asks whether you're living inside it.

The move from static to dynamic simulation is the move from 'what would a generic account with my edge do?' to 'what should a governed MARS account with my edge look like?' Only the second question has an actionable answer, because only the second population shares your rules. Every table downstream — equity bands, drawdown bands, gate dwell, tier usage, throttle efficiency — is a facet of that one comparison, and every one of them inherits its legitimacy from the same source: the ladder ran inside the simulation before it ever ran inside your account.