The rail keeps the trader from destroying the edge while extracting it.

The tension

Harvesting and damaging are the same activity performed differently.

An unharvested edge produces nothing, so extraction is not optional — and extraction requires taking risk repeatedly, which is precisely the activity through which edges get destroyed. This is not a paradox to be resolved but a tension to be managed, and it explains why a governance layer sits between the trader and the market rather than beside them. The rail is not protecting the account from the market. It is protecting the edge from the process of using it. It also reframes what a good week means: a week in which the edge was harvested without any of the four damage routes activating is a better outcome than a more profitable week in which one of them did.

Four destruction mechanisms

How extraction damages the thing being extracted.

The damage arrives through four routes, each of which is a normal response to a real situation:

• Size inflation — a favourable run reads as a larger edge, so risk rises, so a normal losing run lands at the wrong size.
• Rule churn — a losing run reads as decay, so the method is altered, and the altered method has no evidence behind it.
• Concentration — the branch that recently paid attracts deployment, and the diversification that smoothed the curve quietly disappears.
• Friction creep — more trades, more markets, more activity, and the per-ticket costs consume a rising share of a fixed edge.

Why none of them feel like errors

Each is a reasonable response to a real observation.

The four routes share a property that makes them durable: at the moment of commission, every one is defensible. The trader who sizes up did observe a strong run; the trader who alters the method did observe losses; the trader who concentrates did observe one branch outperforming. None of them is being reckless. They are drawing conclusions from small samples — the single most natural inferential move available — and the rail's job is to be unimpressed by small samples on their behalf.

The shape of the damage

Ungoverned extraction outperforms first, and that is the trap.

The chart's most important feature is the left-hand section, where the ungoverned line sits above the governed one. That ordering is not incidental — the behaviours that eventually destroy the edge produce better results while conditions are favourable, which is exactly what makes them so hard to argue against in month six. Any honest account of governance has to concede this: the discipline underperforms visibly during the period before it is needed, and its case cannot be made from recent results.

What the rail does not defend

It cannot protect an edge from genuine market change.

A boundary worth being precise about. If a method stops working because the market it exploited changed, no guardrail recovers it — the rail defends against self-inflicted damage, and market-inflicted decay is a different category requiring validation work and possibly retirement. Conflating the two leads to a specific error: a trader whose edge genuinely died concludes their discipline failed and tightens the rails, which improves nothing because the rails were never the problem. Distinguishing the two is exactly what the structural diagnostics exist for, which is why a governance layer without a diagnostic layer tends to produce traders who blame their discipline for a problem their discipline never caused.

The measurable version

Extraction quality is a number, not a feeling.

The whole argument reduces to a quantity the system reports: how much of the available edge actually reached the account. Prescribed size against deployed size, gross expectancy against net, offered excursion against captured. Each gap is a place where extraction consumed some of what was being extracted, and the sum of them is the cost of the process. That figure is what the rail is trying to minimise, and stating it as a measurement rather than a virtue is what keeps the discussion honest.