Default: CC BY 4.0
Attribution required; commercial use and derivatives allowed — the deliberate opposite of a no-derivatives wall. Read the licence →
The Open Ludeme Commons
Traditional games belong to everyone. Spark & Anvil authors its own clean-room descriptions of the world's classic board games — the parameterised rule-data (board, pieces, moves, end-conditions) behind our games — and releases them here as an open commons under Creative Commons BY 4.0: free to read, share, and build on, with attribution.
It's the open, reusable alternative to closed or non-derivative game databases — a small contribution back to the open-education and games-research communities that make traditional-game scholarship possible. This is a licence and provenance index over descriptions we already own; the descriptions themselves live in our games.
The licence
CC0 1.0 for raw data
Reserved for the rawest reconstruction data (e.g. archaeological wear-facts), where attribution friction is undesirable. CC0 →
“Game description authored by Spark & Anvil (sparkanvil.org) from primary sources, released under CC BY 4.0. Game rules are uncopyrightable methods of operation; this is our own clean-room expression of them.”
What's in the commons — and what isn't
✓ In the commons
- Our own clean-room rule descriptions — board, pieces, moves, end-conditions, metadata
- The primary source each description was authored from
✗ Not in the commons
- Our engine or app source code (their own licence)
- Cast art, story or lesson content
- Any child data — we collect none (see our privacy policy)
The descriptions
8 traditional games, each authored from primary sources. Every one also plays in your browser, free.
| Game | Licence | Primary source (provenance) |
|---|---|---|
| Nine Men's Morris | CC BY 4.0 | R.C. Bell, Board and Table Games from Many Civilizations; H.J.R. Murray, A History of Board-Games Other Than Chess (the Roman/medieval merels family). |
| Fox & Geese | CC BY 4.0 | R.C. Bell; H.J.R. Murray (the asymmetric hunt-game family). |
| Nim | CC0 1.0 | Classical impartial game, public domain (C.L. Bouton, "Nim, a game with a complete mathematical theory", 1901). |
| Dots & Boxes | CC BY 4.0 | Édouard Lucas, La Pipopipette (1889); Elwyn Berlekamp, combinatorial game theory of the game. |
| Hex | CC BY 4.0 | Piet Hein (1942) and John Nash (independent invention); the rules are a method of operation — this is our own clean-room expression. |
| Go (9×9) | CC BY 4.0 | Traditional (East Asian); R.C. Bell; H.J.R. Murray. The 9×9 board is our own beginner parameterisation of the classical rules. |
| Xiangqi (Chinese Chess) | CC BY 4.0 | Traditional (Chinese); H.J.R. Murray, A History of Board-Games Other Than Chess. |
| Ludus Coriovalli (rule-space reconstruction) | CC0 1.0 (wear-facts) + CC BY 4.0 (our rule-space) | The CC0 Ludus Coriovalli reconstruction dataset (board wear-facts); our own clean-room expression of the candidate rule-space (ADR-196 corpus). |
Heritage games — added with care
We add descriptions of living cultural-heritage games respectfully — sources cited, and the difference between a described and an authoritative ruleset made clear (traditional games have regional variants; ours is one described version, not the single canon). These are publicly-documented, publicly-played games:
| Game | Tradition | Primary source (provenance) |
|---|---|---|
| Ô Ăn Quan | Vietnam · CC BY 4.0 | Traditional Vietnamese mancala-family game, documented in the Vietnam Museum of Ethnology; first Western description 1944. A described ruleset as commonly taught — regional variants exist. |
| Cờ Gánh | Vietnam · CC BY 4.0 | Traditional Vietnamese custodial-capture ("carry") game, documented in Vietnamese folk-game scholarship. A described ruleset as commonly taught — regional variants exist. |
| Cờ Chân Chó | Vietnam · CC BY 4.0 | Traditional Vietnamese alignment-and-capture game, documented in Vietnamese folk-game references. A described ruleset as commonly taught — regional variants exist. |
The Emotion Lexicon
Feeling words belong to everyone too. Our emotion-science games teach children to name feelings precisely — and the word list they use is open. Here is the same 24-word English seed bank every emotion game reads, published with its provenance, plus the honest review process a word passes before it reaches a child in any language.
CC BY-NC-SA 4.0
Valence/arousal coordinates: Warriner, Kuperman & Brysbaert 2013 (open affective norms), normalised. Read the licence →
How a feeling word reaches a child
- Seed. Start from an open affective-norms corpus (Warriner et al.) — a coordinate on the pleasantness × energy map, never a bare word.
- Safety-screen. Reject clinical / self-harm / diagnosis vocabulary at the schema level — the lexicon carries feeling words a child can use, not a symptom list.
- Band-fit. Assign each word to the youngest band it belongs to (the simplest feelings first), so a surface only ever offers band-appropriate words.
- Native review. A mother-tongue educator reviews register + cultural fit for that language. Until this passes, NOTHING ships to a child in that language — no raw machine translation, ever (ADR-060).
- Ship. Only then does the word reach a kid-facing surface, and it appears here with its provenance + review status.
Languages
A bank ships to a child only once a native speaker has reviewed it. We publish the status honestly — most languages are seeded but not yet reviewed, so nothing ships in them yet.
English ✓ Published
Open Warriner norms, safety-screened, band-organised — the seed bank below.
Tiếng Việt Review pending
First non-English target (ADR-139). Native-review pending — no raw machine translation ships to a child.
Español Review pending
Roadmap Tier-A. Native-review pending.
Français Review pending
Roadmap. Native-review pending.
Português Review pending
Roadmap. Native-review pending.
Українська Review pending
Roadmap. Native-review pending.
हिन्दी Review pending
Roadmap. Native-review pending.
The published English seed bank
| Feeling | First band | Corner (energy × pleasantness) | Status |
|---|---|---|---|
| happy | 3-5 | Kite — bright & breezy | seed-open-norm |
| proud | 3-5 | Kite — bright & breezy | seed-open-norm |
| excited | 3-5 | Kite — bright & breezy | seed-open-norm |
| calm | 3-5 | Nestle — warm & still | seed-open-norm |
| sleepy | 3-5 | Drizzle — grey & quiet | seed-open-norm |
| scared | 3-5 | Gust — stormy & windy | seed-open-norm |
| angry | 3-5 | Gust — stormy & windy | seed-open-norm |
| sad | 3-5 | Drizzle — grey & quiet | seed-open-norm |
| cheerful | 6-8 | Kite — bright & breezy | seed-open-norm |
| glad | 6-8 | Kite — bright & breezy | seed-open-norm |
| cozy | 6-8 | Nestle — warm & still | seed-open-norm |
| surprised | 6-8 | Kite — bright & breezy | seed-open-norm |
| bored | 6-8 | Drizzle — grey & quiet | seed-open-norm |
| grumpy | 6-8 | Gust — stormy & windy | seed-open-norm |
| worried | 6-8 | Gust — stormy & windy | seed-open-norm |
| lonely | 6-8 | Drizzle — grey & quiet | seed-open-norm |
| grateful | 9-14 | Nestle — warm & still | seed-open-norm |
| content | 9-14 | Nestle — warm & still | seed-open-norm |
| relaxed | 9-14 | Nestle — warm & still | seed-open-norm |
| eager | 9-14 | Kite — bright & breezy | seed-open-norm |
| restless | 9-14 | Gust — stormy & windy | seed-open-norm |
| nervous | 9-14 | Gust — stormy & windy | seed-open-norm |
| frustrated | 9-14 | Gust — stormy & windy | seed-open-norm |
| gloomy | 9-14 | Drizzle — grey & quiet | seed-open-norm |
Coordinates are open affective norms, not a claim about any child. The lexicon is a resource for naming feelings — never a mood tracker, a diagnosis, or a score. See how it's used on our emotion-science games and the Feeling Science ladder.
How we measure our content
Openness is also about being honest when something isn't perfect. Our quiz questions should let a parent tell a good question from a bad one — so we measure them, and publish the result. This is our content error budget: how often a question carries a mechanical flaw, checked across every question we ship.
0.000%
of 52,925 questions (across 132 apps) carry a confirmed mechanical flaw — that's 0. Checked in full, every question, not a sample.
Read by a person, too
A machine can't tell if a wrong answer is genuinely tempting or a question is fair, so we also read a sample by hand each cycle. Latest: 0 flagged of 22 — a reproducible slice across math, science, ELA, logic, genetics and dominoes, read against the item-writing rubric.
What we check on every question
- No duplicate or give-away options — no empty option, no option identical to the correct answer, no exact-duplicate distractor — the mechanical "eliminate without knowing" tells.
- Every question explained — each item carries a written explanation of why the answer is right.
- A real answer key — every item has exactly one identifiable correct answer.
- Enough choices — at least three options per question.
The structural number counts only the flaws a machine can detect with zero false alarms. It deliberately leaves out "the correct answer looks a little longer" style hunches, which are wrong ~9 times in 10 — counting those would inflate the number dishonestly. The judgement a machine cannot make — is a wrong answer genuinely tempting? is the question fair? — is what the sampled human read checks.
For scale: when experts review pools of questions written by a strong AI with no human editing, they typically keep only ~59% unchanged and discard ~1 in 4. Our questions are written and edited by our team and checked before they ship — we never publish raw AI output — which is why the measured number is so low.
This measures the CONTENT, never a child — we collect no learning data (see our privacy policy). Spotted a question that looks wrong? Every question carries a “⚑ Report this question” link. Reported questions go to the front of our review queue.
Game rules are uncopyrightable methods of operation; every description here is Spark & Anvil's own clean-room expression, authored from the cited primary sources — never copied from another games database. Contributions (including a moderated gallery of learner-authored descriptions) are planned. Questions about reuse or attribution? Reach us via About.