Historical Stock Data Quality Report: Comparison with NYSE Daily TAQ

Published 2026-10-22. Version 1.0; analysis prepared 22 October 2026. Read the HTML or PDF and claims.json together.
results.json reports exact field concordance, not full-archive accuracy.

Reproduce using Python 3.10+ (standard library):
python reproduce.py --reference-trades EQY_US_ALL_TRADE_20260701.gz --reference-master EQY_US_ALL_REF_MASTER_20260701.gz --cboe-zip cboe-equity-sample.zip --hdn-root /path/to/licensed/files --hdn-symbols /path/to/symbols.csv --work-dir /path/to/scratch --output reproduced-results.json

The HDN root contains day_by_symbol and minute_by_symbol folders. Every file in published-file-identities.json must match the recorded uncompressed bytes, SHA-256 and rows; a .csv.gz is accepted if it decompresses to the exact CSV. Obtain external inputs separately with appropriate rights. No raw third-party trades or licensed price rows are included.

Verification is source identity plus implementation checks, not independent certification. The temporary source folder on OVH is removed only after complete local evidence preservation.
