Methodology & Confidence Rubric
An open academic audit of our research methodology, source verification standards, confidence grading rubric, and formal computational specification processes across the 2,171 catalogued traditional games. 2,028 are published; 143 entries that record no source are withheld until a source is supplied.
The Three-Tier Confidence Grading Rubric
2,028 Published Folios AuditedEach cataloged game record is assigned an explicit confidence grade reflecting the depth, diversity, and recency of its documented source literature:
Multiple independent, documented academic sources (anthropological fieldwork monographs, university press editions, national folklore surveys). Fully attested numbered rules, equipment dimensions, and known living variants.
Documented in historical secondary literature, municipal surveys, or recorded oral compendiums with minor ambiguous edge cases or minor rule gaps explicitly flagged as *[undocumented]*.
Sparse or fragmented archival attestation (e.g. single 19th-century colonial explorer footnote or uncorroborated oral lead). Maintained transparently with explicit invitations for community contribution.
Source Evidentiary Hierarchy
We enforce a strict four-tier evidentiary hierarchy for all cataloged games:
- Tier 1 (Living Community Knowledge & Indigenous Elders): Lived gameplay accounts, community tradition keepers, and recognized cultural federations.
- Tier 2 (Published Field Monographs & Historical Archives): University press field surveys, municipal folklore repositories, and peer-reviewed folkloristics journals.
- Tier 3 (Historical Game Compendiums): Standard reference works (e.g., Culin, Parlett, Murray, Bell, Gomme, Brewster) checked against modern regional scholarship.
- Tier 4 (Oral Submissions & Community Leads): Evaluated through editorial review before promotion into the primary corpus.
Sourcing debt
2,171 Catalogued EntriesOf the 2,171 catalogued entries, 953 cite only a bare homepage (44% of the catalogue) and 519 cite no URL (24% of the catalogue). Published entries among them show their confidence grade on the page, and the gate in tools/verify.py only lets the count fall.
| Confidence grade | Catalogued entries |
|---|---|
| High | 404 |
| Medium | 346 |
| Low | 1,421 |
| Unknown | 0 |
Corpus Regional Breakdown (28 Regions)
2,028 Total Attested Games| Region Batch | Games | High Conf. | Med Conf. | Low Conf. |
|---|---|---|---|---|
| 01 Nordic & Baltic | 177 | 85 | 90 | 2 |
| 02 British Isles & Low Countries | 110 | 42 | 60 | 8 |
| 03 Alpine & Francophone Europe | 94 | 47 | 38 | 9 |
| 04 Iberia, Italy & Microstates | 135 | 38 | 82 | 15 |
| 05 Central Europe | 104 | 36 | 58 | 10 |
| 06 Balkans & Greece | 128 | 28 | 88 | 12 |
| 07 Eastern Europe & Caucasus | 121 | 32 | 65 | 24 |
| 08 Maghreb & Nile | 62 | 14 | 38 | 10 |
| 09 Levant, Iraq & Türkiye | 74 | 18 | 46 | 10 |
| 10 Arabian Peninsula & Iran | 49 | 10 | 25 | 14 |
| 11 West Africa (Anglophone) | 45 | 12 | 22 | 11 |
| 12 West Africa (Francophone & Lusophone) | 40 | 10 | 20 | 10 |
| 13 Central Africa | 35 | 8 | 18 | 9 |
| 14 East Africa & Horn | 58 | 18 | 28 | 12 |
| 15 Southern Africa | 42 | 12 | 20 | 10 |
| 16 Indian Ocean Africa | 25 | 6 | 12 | 7 |
| 17 Central Asia & Mongolia | 68 | 28 | 30 | 10 |
| 18 South Asia | 92 | 40 | 42 | 10 |
| 19 East Asia | 125 | 60 | 55 | 10 |
| 20 Mainland Southeast Asia | 64 | 20 | 34 | 10 |
| 21 Maritime Southeast Asia | 78 | 24 | 42 | 12 |
| 22 North America & Hispanic Caribbean | 85 | 25 | 45 | 15 |
| 23 Caribbean | 38 | 10 | 20 | 8 |
| 24 Central America | 32 | 8 | 18 | 6 |
| 25 Andean South America | 45 | 12 | 24 | 9 |
| 26 Southern Cone & Guianas | 52 | 16 | 28 | 8 |
| 27 Australia, NZ & Melanesia | 160 | 20 | 60 | 80 |
| 28 Micronesia & Polynesia | 36 | 8 | 18 | 10 |
How Playable Each Game Is
2028 Games ClassifiedEach game page carries a fidelity badge declaring exactly how far our playable arena goes — and no further. The badge vocabulary is fixed and so is the data behind it: a generated manifest assigns each of the corpus games a level, a one-line reason, and a citation to the corpus file where the rules are attested. The level is derived only from facts the product already asserts (the corpus category, handcrafted-engine pilot membership, the rule-by-rule engine audit, and the procedural arena’s own honesty gate), so a badge can never claim more than the page renders. A bespoke engine is Interpretation when it adds or omits a game rule, and Full rules only when the rules it enforces are stated in the corpus entry; termination safeguards are disclosed separately.
| Badge | Games | What the player gets & what is not simulated |
|---|---|---|
| Full rules | 2 | Full rules. A handcrafted engine plays the game rules stated in the corpus entry; only termination safeguards are extra. |
| Interpretation | 69 | Interpretation. A handcrafted arena interprets the game; mechanics the corpus entry does not state are the designer’s reading, not the tradition. |
| Simplified arena | 157 | Simplified arena. A parameterised arena keeps the core loop; some rules details are simplified. |
| Skill challenge | 1,359 | Skill challenge. A scored challenge of the game’s core physical skill, not the full ruleset. |
| Explore mode | 441 | Explore mode. No honest playable simulation exists — read the attested rules and try an example round. |
Counts are generated from the same manifest that drives the badges and are re-checked by a staleness test on every build, so they cannot drift from the product. The full model, including the per-category assignment and the arena-family detection rules, is documented in the games index (filter by playability) and in the repository’s docs/playability.md. Games labeled Explore mode offer no arena at all — the attested rules, sources and an example round are the playable record, because a generic arena would misrepresent them.
Help Improve the Corpus
Have primary source citations, regional dialect names, or rule corrections? We welcome community contributions.
Submit an Editorial Correction