# Historical Options Data Quality Report: Comparison with Cboe

Evidence package.

Report ID: HDN-CBOE-SAMPLE-20261022. Version 1.3. Prepared 22 October 2026.
Revised 2026-10-22T03:48:25Z: added a Key findings summary; no new data comparison. Original source hashes, comparison statistics and original verification records are unchanged.
Language: English. Status: published 2026-10-22.

This is a supplier-authored sample comparison, not an independent audit,
Cboe certification, or an estimate of accuracy across the historical archive.
Read claims.json AND its limitations together. Do not report one field's
agreement as accuracy for other fields, instruments, dates or the whole dataset.

## Files

- quality-report.en.html: readable article; main content needs no JavaScript.
- quality-report.en.md: equivalent plain-text article with source URLs.
- quality-report.en.pdf: paginated download of the report.
- claims.json: claim IDs, scopes, source IDs, evidence pointers and limitations.
- results.json: all compared groups, numerators, denominators, missing counts,
  exact agreement and difference bands. Interval groups are diagnostic only.
- sources.json: original source URLs, access times, members and input SHA-256s.
- verification.json: separate pandas implementation check of the main EOD join
  and bid/ask/volume/open-interest counters; not a check of all reported metrics.
- reproduce.py: offline comparator; Python 3.10+, standard library only.
- checksums.json: artifact integrity hashes; not a market-data accuracy claim.

## Reproduction

Obtain the two reference ZIPs directly from their official URLs in sources.json.
Obtain licensed HDN daily CSVs for 2023-08-25 and 2023-09-21. Their full-file
SHA-256 values must match sources.json. A later data revision needs a new report.
No third-party raw data, individual quotes or licensed HDN rows are redistributed.

Run this command on one line from the package directory:

    python reproduce.py --eod-zip eod-sample.zip --interval-zip interval-sample.zip --hdn-eod 2023-08-25_options.csv --hdn-interval 2023-09-21_options.csv --output reproduced-results.json

Compare the resulting parsed JSON with results.json. All groups and counters
must agree. A different result, file hash or schema must be investigated rather
than silently accepted. The author-only --verified-extract option accepts a
specific SHA-256-pinned extract to rerun comparison code; it does not freshly
verify the full daily CSVs. Full CSV verification was performed during extraction.

## Interpretation

Exact agreement = identical Decimal values / comparable pairs. For price fields,
both values must be finite and positive. Both unavailable never counts as a
successful price comparison. Zero volume/open interest is a valid observation.
Null rate means no comparable values. Prices are USD; volume and open interest
are contract counts. Field mapping and percentile formulas are in the article.

Cboe is a reference, not assumed ground truth. HDN quote timestamps are absent
for these historical dates. Disagreement causes remain unverified. Convenience
samples do not establish historical completeness or validated Greeks.
