You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
SCE-002 — Scenario Composition, Parameterization, and Randomization
Outcome
Extend the pre-run experiment-authoring surface selected by #652 so a study declares how scenario-family variation points are fixed, enumerated, balanced, sampled, or assigned to controlled factors.
Scope
Binding experiment factors and conditions to named scenario variation points.
Selection policies selected by the design, including fixed, exhaustive, Cartesian, balanced/stratified, weighted, bounded random, without-replacement, and t-way coverage where supported.
Explicit distinction among controlled factors, nuisance variables, ordinary configuration, and runtime stochasticity.
Bounded distribution parameters and validation.
Root stochastic-control and random-stream suite references.
Cross-artifact validation against the referenced expanded scenario family.
Authoring examples, schema/contract generation, fixtures, and MCP validation/scaffolding.
Acceptance criteria
The SDL declares what may vary; the experiment declares how and why it varies.
Every policy resolves to a declared compatible variation point and bounded domain.
Allocation counts, factor assignments, and sampling policies are mutually consistent.
Requirements
Outcome
Extend the pre-run experiment-authoring surface selected by #652 so a study declares how scenario-family variation points are fixed, enumerated, balanced, sampled, or assigned to controlled factors.
Scope
Acceptance criteria
Non-goals