The collection
The 3,000
Nine factions, two supply tiers, sixty hand-made chase pieces, and a rarity attribute that means something specific. Every number on this page comes out of the generator, not out of a pitch deck.
Allocation
Faction supply mirrors cultural rank rather than market cap. CashCat is the largest because it is the meme the chain is about, not because it has the biggest chart.
| Faction | Tier | Archetype | Supply | Share |
|---|---|---|---|---|
| CashCat | Core | The Flagship | 550 | 18.33% |
| Pipedog | Core | The Investigator | 450 | 15.00% |
| Brodie | Core | The Good Boy | 400 | 13.33% |
| FRONG | Core | The Prototype | 400 | 13.33% |
| HMM | Core | The Strategist | 320 | 10.67% |
| MowCat | Core | The Dragster | 280 | 9.33% |
| TENDIES | Core | The Tank | 240 | 8.00% |
| Little John | Scarce | The Blocker | 180 | 6.00% |
| Robin | Scarce | The Glider | 120 | 4.00% |
| Legendaries | Legendary | Cross-faction event monuments | 50 | 1.67% |
| One of ones | 1/1 | Hand-designed | 10 | 0.33% |
Why these numbers
Three constraints produced the split, and they are worth stating because they explain why the smallest faction is 120 and not 30.
- Supply follows cultural rank. No flagship ends up diluted next to a faction with a tenth of its mindshare.
- The two scarce factions are the two non-animal-mascot identities. Little John is the only human and Robin is the only bird, so their scarcity is legible from the thumbnail without reading a number.
- No faction sits below 4% of supply. At 120 pieces a faction still has a tradable floor and enough holders to be a community rather than a rounding error.
The four tiers
Core
Seven factions, 2,640 pieces, IDs inside 1 to 2940. Fully generative: a token is a deterministic roll across twelve trait categories drawn from its own faction's pools.
Scarce
Little John at 180 and Robin at 120. Structurally identical to Core, and the only difference is how many exist. A scarce faction is not a better kart, and the balance figures on the factions page show that directly.
Legendary
Fifty pieces, IDs 2941 to 2990, hand-parameterized scenes rather than layer rolls. Each one depicts a dated event from the chain's first six weeks: the accidental FRONG mint at block 23,595,790, the day CASHCAT topped out and the launchpad that made it halted new launches, the listing-day run of +151% in fifteen hours. Every faction appears in at least one, so no holder is locked out of the chase layer.
Legendaries carry a Moment attribute instead of the trait categories, and they are monuments rather than upgrades. A legendary CashCat is not faster than a common CashCat. It is a common CashCat with a story attached.
One of one
Ten pieces, IDs 2991 to 3000, each designed on its own. The 2013 garage with a crossed-out feather on the whiteboard. Brodie's kart in an empty 2016 office at golden hour. The full nine-faction grid on a starting line shaped like a flattening globe.
What makes a racer rare
Two separate things, and conflating them is how collectors end up arguing past each other.
The RarityTier attribute
A presentation stat baked into the metadata at generation time. Each trait carries a tier tag, each tag has a score, and a token's tier is a threshold over the sum.
| Trait tier | Score | Roll weight |
|---|---|---|
| Common | 0 | 60 |
| Uncommon | 1 | 28 |
| Rare | 3 | 10 |
| Epic | 6 | 2 |
// thresholds calibrated on the actual score distribution (q50=8, q80=11, q95=15)
if (score >= 16) return 'Epic';
if (score >= 12) return 'Rare';
if (score >= 9) return 'Uncommon';
return 'Common';The rarity score
A separate index, written to web/public/rarity.json, using the ordinary sum of inverse trait frequencies: for each trait, total supply divided by how many tokens share that value, added up. It is the conventional method on purpose. Collectors already have tools that compute rarity this way, and a private formula that disagrees with them helps nobody.
Faction and Archetype are excluded from that score. Faction is a choice the collection makes in fixed proportions, so scoring it would reward the small factions twice for a decision the supply table already made.
The distribution that actually shipped
Counted from the 3,000 metadata documents in the generator output, not from the intended weights. Rolls are rolls, so the two differ slightly and this is the one that is true.
| RarityTier | Tokens | Share |
|---|---|---|
| Common | 1,779 | 59.30% |
| Uncommon | 640 | 21.33% |
| Rare | 420 | 14.00% |
| Epic | 101 | 3.37% |
| Legendary | 50 | 1.67% |
| One of One | 10 | 0.33% |
The second rarity layer
Traits are settled forever the moment the provenance hash goes on chain. Race history is not, and that is the second axis: a plain-looking kart with two season titles should out-price a bare epic, because one of those is a roll and the other was earned.
Season results are computed off chain, then frozen into a Merkle root posted to the season registry. From then on a career claim is either provable against the chain or false. The mechanics are on Rewards and seasons and the verification steps are on Provenance and chain.