What the system actually controls — and what it deliberately refuses to.

The founding split

A trader may make discretionary decisions; capital deployment may not be discretionary chaos.

That sentence is the system's constitution in one line. MARS is a discretionary-quant framework: it accepts that trade selection is a human judgment involving pattern, context, and experience that no rule set fully captures — and it refuses to accept the same for position sizing. The two questions get opposite treatments because they have opposite error profiles. A bad trade selection costs you one trade's risk. A bad sizing decision, repeated under emotional pressure, costs you the account. So discretion keeps its territory and gets defended there, while sizing becomes a lookup routed through quantitative governance.

The operator's territory

What stays yours — and it's more than most systems leave.

The system never tells you which setup is valid, which branch a trade belongs to, whether volatility context is acceptable, whether a specific trade is worth taking, or how to manage it once open. That's not an oversight or a gap awaiting a future module — it's the deliberate retention of the part of trading where human judgment genuinely outperforms rules. What MARS demands in exchange is honesty about those judgments: the branch labeled correctly, the adherence recorded truthfully, the outcome logged as it happened. The discretion is protected; the reporting on it is not optional.

The system's territory

What MARS takes — the boundary conditions of every trade.

The governed side is a short list with enormous consequences: what gate the account currently occupies, what tier the evidence authorizes inside that gate, what cycle pool that tier resolves to, how much of it open carryover exposure has already consumed, and whether any override is being invoked and logged. None of these answer 'should I take this trade?' — they answer 'how much may this trade risk, given everything the system knows about the account's state?' The operator's answer to the first question and the system's answer to the second meet at execution, and neither may overrule the other.

The refusals

What MARS will not do, stated plainly.

The manuals are unusually explicit about the negative space, and the honesty is load-bearing. The deployment engine does not prove edge by itself — edge proof comes from trade results, EV scorecards, and structural diagnostics. The simulation layer does not validate a plan, only stress-test the assumptions given to it. The analytics labs diagnose opportunity but cannot grant permission to exceed capital protection. And nothing anywhere in the stack generates entry signals. A system that claimed all of these would be more marketable and less useful; the refusals are what keep each module's output trustworthy within its actual competence.

• No entry signals — MARS governs the trade you chose, it doesn't choose it.
• No edge proof from any single module — proof accumulates across journal, scorecard, and diagnostics.
• No permission from analysis — diagnostic strength never overrides capital-state authority.

The key idea

Rails don't remove movement. They control direction.

The rail metaphor is precise and worth taking literally: a rail system doesn't decide where the train goes or how fast it's driven — it makes derailment structurally difficult. MARS puts discretion on rails rather than replacing it, which is why it fits traders who already have a model and need an operating structure around it, and fits poorly anyone hoping to be told what to trade. The governance isn't the product's constraint. It's the product.