Scenario Sensitivity Sweep¶

Robustness assessment — instantiates Incompatible Requirement Set Resolution

Varies the uncertain inputs across plausible scenarios to learn whether the incompatibility is robust or an artifact of one assumption — and which assumptions, if they moved, would flip the verdict.

A "no joint solution" verdict is computed under assumptions — a demand forecast, a latency budget, a cost figure — and those assumptions may be soft. Scenario Sensitivity Sweep stress-tests the verdict against its own inputs. It re-runs the feasibility question across a range of plausible values and named scenarios, and watches whether the conflict holds everywhere or dissolves in some corner of the input space. Its defining move is to make the conflict's dependence on assumptions the object of study: it does not choose a resolution, it tells you how much the resolution can be trusted — flagging the assumptions the verdict rides on, the thresholds at which it would flip, and the uncertainty that no choice here can remove.

Example¶

A bank's treasury faces three commitments that its planning model says cannot all hold next year: a regulatory liquidity-coverage ratio floor, a return-on-equity target the board has promised, and a lending-growth commitment to key clients. Before anyone starts sacrificing one, the treasury runs a sweep. It varies the inputs the verdict depends on — deposit run-off rate, the yield curve, and the credit-loss rate — across a base case and several named macro scenarios (a mild recession, a rate spike, a funding squeeze).

The pattern is telling. Under most of the range the three commitments are jointly infeasible, so the conflict is real and worth resolving. But the sweep also finds that the whole verdict pivots on one number: the assumed deposit run-off rate. Hold run-off below ≈6% and a feasible plan exists; above it, no plan does. That single finding reshapes the work — the binding uncertainty is deposit stability, so the resolution should be built to survive high run-off, and a recheck trigger is set: if observed run-off crosses 5%, re-open the whole appraisal. The sweep resolved nothing on its own; it told the team which assumption their resolution must not depend on, and what to watch.

How it works¶

The method is a structured re-run of the feasibility question over its uncertain inputs:

List the load-bearing assumptions. Enumerate the uncertain inputs the incompatibility verdict depends on, with a plausible range or set of scenarios for each.
Sweep and re-evaluate. Vary them — one at a time to attribute effect, and in joint scenarios to catch interactions — re-running the feasibility check on each and recording whether the conflict persists.
Find the flip points. Locate the threshold values at which the verdict changes from infeasible to feasible (or vice versa); the assumptions with nearby flip points are the ones the resolution is hostage to.
Emit triggers and residue. Translate the most dangerous flip points into recheck triggers to monitor, and report the uncertainty that remains after the sweep — what the analysis still cannot pin down.

Tuning parameters¶

Range width — how far each input is varied. Narrow ranges risk declaring a fragile conflict "robust"; implausibly wide ones make everything look uncertain. The range should match genuine forecast error.
One-at-a-time vs. joint scenarios — single-factor sweeps give clean attribution (which assumption matters); joint scenarios catch the compounding cases a one-at-a-time sweep misses.
Scenario selection — whether scenarios are evenly gridded, drawn from history, or chosen as adversarial worst cases; adversarial framing surfaces fragility that averaging hides.
Flip-point resolution — how finely the threshold is located near a verdict change; finer bracketing sets a sharper recheck trigger at the cost of more runs.
Trigger tightness — how much margin to leave between the observed value and the flip point before the trigger fires; tight triggers catch drift early but fire often.

When it helps, and when it misleads¶

Its strength is that it separates a robust incompatibility — one worth the pain of resolving — from a fragile one that a single soft assumption is manufacturing, and it names exactly which assumptions the resolution must be hardened against.^[1] The recheck triggers it emits also give the eventual decision a shelf life instead of a false permanence: the resolution is re-opened when, and only when, a load-bearing assumption moves.

Its limits are the usual ones for a what-if study. A sweep only explores the inputs you thought to vary, so an unlisted assumption sits silently at its base value and can hide the real fragility; and one-at-a-time sweeps understate risk when several assumptions move together. The tidy "flip point" invites false precision over inputs that are themselves guesses. And it is easily run to reassure — sweeping only over ranges narrow enough to keep the conflict looking robust (or looking benign), to justify a resolution already chosen. The discipline is to register the assumptions before running, include the uncomfortable ranges, and carry the residual uncertainty forward into the decision rather than letting the base-case verdict masquerade as certain.

How it implements the components¶

Scenario Sensitivity Sweep realizes the robustness-and-assumptions machinery of the archetype — the parts that qualify how far the verdict can be trusted:

assumption_register — it enumerates and stress-tests the uncertain inputs the incompatibility verdict rests on, surfacing which are load-bearing.
recheck_trigger — it converts the dangerous flip points into monitored conditions that re-open the resolution when an assumption drifts.
uncertainty_residue — it reports the uncertainty that remains after the sweep: the part of the answer no choice at this stage can remove.

It probes robustness to input values, not to the scope and meaning of the commitments — that is Scope-Boundary Stress Test; nor does it build the feasibility model it re-runs, which is Constraint-Satisfaction Solver Pass.

Instantiates: Incompatible Requirement Set Resolution — it qualifies the incompatibility verdict with its dependence on assumptions and the triggers to revisit it.
Consumes: Constraint-Satisfaction Solver Pass — each scenario re-runs the feasibility check on re-valued inputs.
Sibling mechanisms: Scope-Boundary Stress Test · Constraint-Satisfaction Solver Pass · Compatibility Matrix · Requirements Traceability Matrix · Proof Checking · Constraint Relaxation Experiment · Pareto Frontier Analysis · Stakeholder Frontier Review · Impossibility-Theorem Instantiation Review · Minimal Unsatisfiable Core Extraction · SAT/SMT Satisfiability Check · Weighted MaxSAT or Soft-Constraint Optimization · Decision Record with Residue

References¶

[1] Sensitivity analysis — systematically varying a model's inputs to see how much each moves the output. A one-at-a-time (OAT) sweep attributes effect to single inputs (often visualized as a tornado diagram ranking them by influence); joint or scenario sweeps capture the interactions OAT misses. Here the "output" is the feasibility verdict itself. ↩