Point-in-time universe lab

Survivorship-Bias Guard

Choose a scenario, step through bitemporal membership evidence, and observe why a later roster can diagnose omissions but cannot change the historical result.

All identities, events, rosters, and returns are synthetic
Stage 1 of 9Stage 1 of 9

Historical membership stage

Complete preview: A, B, C eligible; D ineligible.

Definition
Synthetic historical and later rostersThe historical membership resolver is on the left. A later, non-causal roster comparison is on the right. Causal historical membershipeffectiveAt: 2023-12-31 | knownAt: 2024-01-02 later Hindsightcausal = false coverage: --not bias magnitude
eligibleineligible / incompleteunknown / ambiguousnon-causal diagnostic

Diagnostics

pendingcausal result
Eligiblepreview: A, B, C
Ineligiblepreview: D
Unknownnone
Ambiguousnone
HINDSIGHT DIAGNOSTIC

Later roster A, C, D would omit historical B. This preview cannot change the causal set.

What to notice now

The initial state is informative: it previews the complete canonical outcome while stage one starts with the universe definition.

supported baseType = named_research_universe

Audit trail

StageCheckEvidenceResultCausal?

Keyboard: Right Arrow = Step, Left Arrow = Back, Space = Play/Pause. Reset is deterministic. The synthetic return illustration is intentionally not computed in this membership lab because it cannot reconstruct missing terminal outcomes.