Give Every Cricket Data Table a Verifiable History
Label every cricket table with its source, data coverage, retrieval date, calculation date, method and any exclusions or later corrections.
A published Test cricket table should state where its data came from, which matches it includes and when its figures were last recalculated. Put those details directly below the table. A lone “Updated” label does not establish the source, coverage or calculation date.
Choose how the table was produced and changed; the tool lists the fields its source note needs.
Required Source-Note Fields
For a derived, complete table receiving a routine refresh, publish:
- Source: provider, product or files, and a reproducible link.
- Role: identify the source values used in your calculation.
- Coverage/data through: matches represented by the table.
- Source retrieved: date of the downloaded or queried snapshot.
- Table updated: date the figures were recalculated and checked.
- Method: formula, denominator and material inclusion rules.
- Routine refresh: update the coverage and calculation dates; no correction label is normally needed.
Basis: W3C data-provenance guidance and the publication examples in this article.
A reader who copies, screenshots or cites the table should take its context with it. Do not leave essential provenance only on a general methodology page or in the website footer.
Use a Source Block, Not One Ambiguous Date
For a table that changes as new Tests are played, use this pattern:
Source: Cricsheet men’s Test-match JSON files.
Coverage: Files in the retrieved snapshot for men’s Tests completed from 1 January 2001 through 26 September 2026.
Exclusions: Matches absent from that snapshot. At retrieval, Cricsheet reported 13 Test matches withheld.
Source retrieved: 26 September 2026.
Table updated: 26 September 2026.
Method: Team runs divided by wickets lost. An innings ending with fewer than ten wickets lost contributes its actual wicket count, not ten.
The dates can coincide, but they describe different events:
| Label | What It Means | Why It Matters |
|---|---|---|
| Coverage or data through | Period represented, with exclusions stated separately | Shows whether a completed Test should be present |
| Source retrieved | Date the underlying data was downloaded or queried | Identifies the source snapshot used |
| Table updated | Date the figures were recalculated and checked | Identifies the calculation version |
| Page updated | Date the surrounding article changed | Does not prove that the data changed |
Do not replace “data through” with “updated.” A table recalculated today may still stop at the end of 2025. Conversely, the prose may receive a spelling correction without any change to the table.
The W3C Data on the Web Best Practices recommends publishing metadata and provenance, assigning each dataset a unique version number or date, and providing a version history. Those principles support separate, explicit labels instead of one date expected to carry several meanings.
Name the Files or Product Actually Used
“Source: internet,” “source: database” and “source: scorecards” are too vague. Name the provider and the relevant product, files or table, then link to the closest reproducible source page.
For a table calculated from Cricsheet, link to its Test-match download rather than only its homepage. Cricsheet says its principal data is available as JSON and provides downloads split by match type, including Tests. Its downloads page also identifies withheld matches.
Each JSON file has a meta section containing the format version, file-creation date and revision number, as described in the Cricsheet JSON format. Those fields describe the source file. They do not replace the publication’s retrieval and calculation dates.
If several sources are used, state their separate roles:
Sources: ICC standings for points and deductions; scorecards for match results; calculations by Test Match Stats.
That is clearer than presenting a derived table as though every displayed value came directly from one database. The official World Test Championship standings publish matches played, wins, losses, draws, deductions, points and points percentage, and state the points system. A republished table should say whether those values were copied or independently calculated.
The same distinction can appear in machine-readable metadata. Google’s dataset structured-data guidance recommends sameAs for a simple republication and isBasedOn when a dataset has been significantly changed or assembled from several originals.
Put the Calculation Beside the Provenance
A source identifies the inputs; it does not define the statistic. The note must also state the denominator and any inclusion rules that can change the result.
For common Test cricket measures:
- Batting average: runs divided by dismissals, not innings.
- Bowling average: runs conceded divided by wickets credited to the bowler, not overs.
- Run rate: runs per over, with the treatment of legal balls, extras and differing historical over lengths made explicit.
- Team results: define whether abandoned matches, forfeitures and neutral venues are included.
The source note on a team batting-average table should therefore name both the ball-by-ball files and the runs-per-wicket method. A team run-rate table should state the balls behind its rate. A dynamic World Test Championship table additionally needs a cut-off date because results and point deductions can change during a cycle.
Record Corrections and Revisions Separately
Correct a wrong value promptly, but distinguish a correction from a routine refresh. A short change note is enough:
Correction, 28 August 2026: Pakistan’s 2024 wicket total was understated by two because one innings was omitted. The average and ranking have been recalculated.
Adding newly completed matches is a routine update. Reprocessing old matches after adopting a better method is a revision. Fixing a transcription, input or code mistake is a correction.
The UK Office for National Statistics makes the same distinction: revisions incorporate improved methods or newly available data, while corrections amend errors. Its revisions and corrections policy also calls for stating when and why affected material changed.
Keep the current source block directly below the table and a compact, dated change log below the methodology. For downloadable spreadsheets, repeat the metadata inside the file. Government Analysis Function spreadsheet release guidance says a cover sheet should include the data source, publication date and expected next update. For a multi-sheet file, its table of contents can include the source and the dates of the last and next updates.
The minimum publishable line is:
Source · coverage/data through · source retrieved · table updated · method
Add an exclusions field when the source is incomplete. Add a dated correction or revision note whenever previously published figures change for a reason other than newly completed cricket.