Stress-Test Margin Check¶

Assessment (audit) — instantiates Safety Margin Design

Applies simulated and historical adverse scenarios to an already-designed margin to check whether it actually survives the cases it is meant to cover.

Stress-Test Margin Check takes a margin someone has already designed and subjects it to explicit adverse scenarios to see whether it genuinely holds. Its defining trait, distinct from every mechanism that sizes a margin, is that it verifies rather than creates: it produces a pass, a fail, or a shortfall against a set of named cases — especially cases where several bad things happen at once — exposing whether the sizing assumptions survive the correlated tail rather than only the average. A margin that looks ample on typical conditions can be dangerously thin when the stresses coincide; this check is what finds out before reality does.

Example¶

A grid operator carries a planning reserve margin — generation capacity above forecast peak demand. On average figures it looks comfortable. The stress-test margin check runs it against constructed adverse scenarios rather than averages: a 1-in-10 summer heat wave that drives demand to an extreme and the simultaneous forced outage of several large units and low wind output — the coincident tail, not each factor alone. It replays historical near-misses as reference cases and builds correlated combinations, then computes whether unserved-load risk stays within the standard (for instance, a loss-of-load expectation no worse than the "1-in-10" resource-adequacy benchmark). If the margin fails the combined scenario, the check reports a shortfall: the reserve that looked fine on averages does not survive the correlated tail, and the operator learns this in a planning study rather than during a blackout.

How it works¶

The distinguishing moves of a check, not a design:

Assemble a scenario set — historical worst cases plus constructed combinations, deliberately including coincident and correlated failures.
Characterize the tail the margin is supposed to cover, so the scenarios probe the right variation.
Anchor on reference cases — real past events and near-misses that keep the scenarios grounded.
Run the existing margin against each and report survival or shortfall — the output is a verdict on a margin, not a new margin.

Tuning parameters¶

Scenario severity and count — how extreme and how many; too mild passes anything, too extreme fails everything.
Correlation assumptions — whether adverse events are allowed to coincide; independence is the flattering, usually wrong, assumption.
Historical vs. synthetic — replayed real events vs. constructed combinations; the strongest sets use both.
Pass criterion — the standard the margin must meet to be judged adequate.

When it helps, and when it misleads¶

It earns its place by catching margins that are adequate on average but fail on the correlated tail — the failures that averages and single-variable checks miss entirely. Its own failure mode is a scenario set too narrow or fighting the last war: a test that only replays known events blesses a margin that the next, unseen combination will break. The classic misuse is reverse-engineering a benign scenario set so the existing margin passes — a rubber stamp dressed as diligence. The disciplines are to include coincident and correlated failures, keep the scenario set independent of the answer you want, and refresh it as new events reveal new tails.^[1]

How it implements the components¶

Stress-Test Margin Check fills the verification-and-scenario side of the archetype — it interrogates a margin but never makes one:

stress_scenario_set — the constructed and historical adverse cases the margin is run against.
variation_profile — the tail and correlated-failure behavior the scenarios are built to probe.
reference_case_baseline — the real past events and near-misses anchoring the scenarios.

It does not create or size the margin — that is Structural Safety Factor or Conservative Estimate — nor assign an owner and trigger to act on a failed check (Minimum Reserve Requirement); the scenarios it builds are consumed by a sizing rule such as Risk Capital Buffer.

Instantiates: Safety Margin Design — Stress-Test Margin Check is the verification that a designed margin survives adverse cases.
Sibling mechanisms: Risk Capital Buffer · Conservative Estimate · Structural Safety Factor · Capacity Headroom · Budget Contingency · Reserve Inventory · Schedule Float · Setback Requirement · Safe Operating Limit Chart · Minimum Reserve Requirement · Premortem Margin Review

Notes¶

A stress test verifies but does not fix. A failed check is only useful if it connects to action — restoring the margin, adding capacity, or escalating — which lives in Minimum Reserve Requirement and the archetype's restoration trigger. Left unlinked, a shortfall finding just files itself away and the thin margin persists until the scenario it failed arrives for real.

References¶

[1] Loss-of-load expectation and the "1-in-10" reserve-margin standard are real resource-adequacy benchmarks used to judge whether a generation margin survives extreme-demand and outage scenarios. ↩