# T012 — Conway Semiannual LPR Report Version Divergence Two productions contain documents with the same July 1, 2025 title, author, recipient, and opening deployment narrative but different statistical paragraphs. Internal consistency strongly suggests that the later-produced copy carries the intended H1 2025 figures and that the earlier copy retains an H1 2024 statistical block, but no produced version history, correction notice, or author statement designates an authoritative final version. ## Statement A The `PD-2026-354` copy states that, from January 1 through June 30, **2024**, six stationary LPRs read 6,316,810 plates, generated 2,007 hits, and logged three successful outcomes ([[LPR Report First Half 2025]], `LPR Report First half 2025.docx`, statistical paragraph). ## Statement B The `PD-2026-477` copy states that, from January 1 through June 30, **2025**, fourteen stationary LPRs read 7,458,633 plates, generated 3,327 hits, and logged four successful outcomes ([[LPR Report First Half 2025 (PD-2026-477 Copy)]], `Memo first half of 2025.docx`, statistical paragraph). ## Why it matters The figures feed program-volume and outcome-rate analysis. Treating both as the same period would be false; declaring one the final agency-approved version without provenance would also overreach. The safest use is to attribute each statistical block to its exact produced file and period while preserving the version conflict. ## Resolution status **Open.** The 2025 block is internally consistent with the shared opening paragraph's fourteen-camera deployment description; the 2024 block is internally consistent with the earlier six-camera system. That supports a working reconciliation, not an authoritative version designation. Closure requires a version history, correction/transmittal record, or confirmation from the report author/custodian. ## Discovery Surfaced during the 2026-05-22 supplemental ingest and formalized as a tension during the 2026-07-20 contamination/dependency lint. ## Notes - Quote and analyze the two files separately. - Do not combine the figures into a single success/error rate.