# Complete constructed story review

Source: four tasks T1–T4 in case-rows.csv, minutes per teaching round. Complete pairs T1–T3: before 12, 10, 14; after 8, 9, 11. Medians 12 and 9; difference -3 minutes, relative25 percent reduction from 12. T4 only has after 10. All four after values would have median 9.5; a different selection. No causal effect, customer performance or real person was measured.

S1: private use only; hold public story.
S2: prior V1 file allowed, current V2 changes the number and bytes; new review needed.
S3: current English file allowed, German outside scope; hold German draft.
S4: both current files match the fictional public agreement; ready for a deliberate publication step, not published.

Complete current drafts are story-current-en.md and story-current-de.md. Prior drafts are explicitly superseded counterexamples. review-register.json identifies bytes and grants no real permission.