Data reference · NFL / NBA / WNBA

Use and cite draft ranking data

Choose a fixed capture, understand its coverage, and check the file before using it in a notebook, chart or article. This guide describes the ranking exports actually published by Ultimate Big Board.

Choose a saved capture Download the Python example Read the 14-field dictionary

1. Choose live data or a fixed capture

Use the CSV download on a board's methodology page when you want its current ranking. That download follows the live board: a later request may return different rows, a new snapshot, or a new capture time.

For a reproducible analysis or a precise citation, choose a page in saved ranking captures and download its fixed rankings.csv. Retain the page URL, CSV, snapshot identifier, calculation version and SHA-256 checksum with your analysis. The collection preserves the listed captures; it does not promise every intermediate update.

The capture catalog records the fixed CSV URL, checksum, row count, capture timestamp and source count for each published capture. Existing captures stay fixed when new entries are added. The catalog is a useful integrity reference from this site, not independent certification of the source opinions.

The first saved capture set

Captured October 2, 2026 at 18:27:11 UTC. These historical counts describe these files, not today's live boards.

Draft board	Exported rows	Eligible mocks	Snapshot
2027 NFL	102	24	`69`
2028 NFL	32	2	`55`
2027 NBA	85	8	`c6b07b467c0b3c8a`
2028 NBA	49	2	`09c970f42d2418ce`
2027 WNBA	60	7	`41b70009de53907d`
2028 WNBA	60	2	`b355fc3864978a67`

2. Read the CSV field dictionary

Each fixed CSV uses the same 14-column header below. Empty position or school values are unknown, rather than zero. Source labels beginning with spreadsheet formula characters may carry a protective apostrophe; preserve the original CSV bytes when checking its checksum.

Field	Format	Meaning and limits
`Rank`	Positive integer	Published ordering in this board. It is neither an actual draft pick nor an exported average-pick value.
`Player`	Text	Published prospect label. Use the profile URL as a within-board identity when comparing files; names alone can collide.
`Position`	Text, possibly empty	Position label recorded in the export. An empty value remains unknown; do not fill it by inference.
`School`	Text, possibly empty	School label recorded in the export. It is not a permanent history of a player's enrollment.
`DraftYear`	Integer	Draft cycle: 2027 or 2028. Do not mix years in a ranking comparison.
`League`	Text	NFL, NBA or WNBA. Each league/year board has its own eligible sources and calculation.
`Snapshot`	Text identifier	Identifier of the underlying ranking snapshot. Preserve it as text; do not interpret it as a date or compare identifiers across leagues.
`SnapshotGeneratedAt`	Timestamp with timezone	When the underlying snapshot was generated. This differs from when the CSV was captured.
`CapturedAt`	Timestamp with timezone	When the CSV bytes were captured. Record this with the fixed URL and checksum.
`BoardURL`	HTTPS URL	Live big-board destination for the same league/year. Its current rankings may differ from this capture.
`MethodologyURL`	HTTPS URL	Explanation of the ranking method for that league/year. The linked page is live, rather than a frozen copy of the method.
`ProfileURL`	HTTPS URL	Live profile destination within this league/year. Profile content can change after capture.
`CalculationVersion`	Text	Version of the ranking calculation. Preserve it when documenting comparisons; a changed method needs separate interpretation.
`EligibleMocks`	Positive integer	Total eligible mocks supporting the snapshot. This is not each player's appearance count or the number of independent publications.

This export does not include individual mock selections, per-player appearance counts, confidence components, average-pick values or actual draft results. Follow the matching board, profile and original-source pages when your question needs those details; do not reconstruct them from the rank column.

3. Verify a capture with Python

The downloadable example uses Python 3.10 or later and its standard library. Save validate_capture.py, then run this command from the folder containing it. No package installation, API key or account is needed.

python validate_capture.py --capture-url "https://ultimatebigboard.com/captures/nfl/2028/20261002T182711Z-9680aa40d8a1/"

The example fetches only this site's public catalog and the selected fixed CSV. It refuses redirects and oversized files, checks the SHA-256 and row count, and verifies the exact header, league, year, snapshot, timestamps, calculation version, source count and within-board profile identity. It rejects duplicate or out-of-order ranks. A failed check stops the analysis instead of silently substituting the live board.

After verification, it prints the capture's identity and the first five published ranks. Its verify_capture(catalog, capture_url, csv_bytes) function returns the checked metadata and rows for further analysis. Keep that metadata beside any derived chart. Verification establishes file consistency with the catalog; it does not validate every upstream projection.

4. Compare carefully

Keep league and draft year separate. NFL 2028 and NBA 2028 are different datasets, even when their rank columns contain similar numbers.
Check both timestamps. Capture time tells you when this CSV was saved. Snapshot generation time tells you when its underlying ranking was calculated. Neither is the publication date of every source.
Preserve missingness. A prospect missing from one file may reflect source coverage or list depth. Do not assign a bottom rank or an undrafted outcome without additional evidence.
Inspect support before claiming consensus. The first 2028 captures each have only two eligible mocks. That is a preliminary sample. Total eligible mocks are not independent-publication counts or the support count for each prospect.
Check the calculation version. Rank changes can reflect a different method, changed source eligibility or source coverage as well as changed projections. This export alone cannot explain the cause.
Use identity links within the same board. Compare ProfileURL rather than a name-only join, and preserve unmatched rows. A live profile may later change; the URL is not a cross-year identity guarantee.

5. Cite the capture and credit original sources

A useful citation gives the league/year, snapshot, calculation, capture time and stable capture URL. Include the checksum in your reproducibility notes. For example:

Ultimate Big Board. 2028 NFL consensus rankings. Snapshot 55; 2 eligible mocks; calculation live-2.3-source-independence. Captured 2026-10-02T18:27:11+00:00. https://ultimatebigboard.com/captures/nfl/2028/20261002T182711Z-9680aa40d8a1/ CSV SHA-256: 9680aa40d8a192074bb6540f9a451789be2047daf305425df447f82bd36dd395.

Use the linked league/year methodology to interpret that board. When discussing a named analyst's projection, link to the original mock through the source library, rather than crediting an aggregate as the original author.

These files preserve aggregate rankings, not the original articles. Free access does not establish an open reuse license for upstream writing, images or other material. No license is asserted by this guide; retain attribution and check the applicable rights for your intended use.