How the floor works.

Floor Runner is a badge protocol on Robinhood Chain. There are 1,366 badges, a number taken from the real seat count of the New York Stock Exchange. Badges are never sold by the protocol. They are won each day by whoever brings the floor the most buy volume, and they are taken back from anyone who stops.

Badge holders are settled from the floor's own trading fees. That settlement source exists from the first trade, so nothing on this page is waiting on a product that has not been built.

Holding the token alone is settled nothing. There is no rate and no claim button for a plain holder. Only a defended badge is paid.

Every trade of the token pays a fee, from any venue, including aggregators the protocol has never heard of. That fee is taken by Floor Runner's own Uniswap V4 hook inside the swap itself and forwarded to the purse in the same transaction, so the whole of it arrives — no launchpad, no intermediary, no share skimmed on the way.

Badge pot

70% — divided evenly across every live badge

Treasury

20% — protocol-owned liquidity first, then operations and audits

Burn

10% — market-buys the token and sends it to a dead address

Why our own hook, and not a launchpad's. We read the alternative on-chain first. A Pons V2 launch initialises its pool with a static fee of zero and a hook at 0xe5e70264…be044 whose address encodes AFTER_SWAP_RETURNS_DELTA — it takes a cut of every swap. That hook's protocolFeeRecipient is the same address as the factory's owner and the locker's owner, and the locker holds your liquidity position. On that path the trading fee is the launchpad's, not the floor's. So Floor Runner ships its own hook.

Two meters must not be confused. The purse is fed by every trade. The score, which decides who holds the badges, counts only buy volume routed through a runner's own referral. The purse is therefore always larger than the measured score, and badge holders share in trade that credits nobody.

There are no tiers, weights or multipliers. Badge 3 and badge 1,204 are settled exactly the same amount at every bell.

settled per badge  =  (purse × 0.70) ÷ live badges

This is the design, not a simplification. Weighting badges would rebuild the ladder the protocol exists to remove, and it would require a published share table — a class of thing that is wrong more often than anyone admits.

Once a day at 21:00 UTC the floor rings the bell. Three things happen in one transaction, in this order.

• ▸Sweep — accrued pool fees move into the purse and split 70 / 20 / 10.
• ▸Settle — the badge pot divides evenly across every badge live at the moment the bell rang.
• ▸Move — at most three badges change hands: one revoked, one relegated, one born.

Between bells the scoreboard updates continuously. The board moves all day; the badges move once. A badge that could vanish every few hours would have no resale value, and the resale market is where a badge's price is formed.

ringBell is permissionless and idempotent per epoch. Anyone can call it after 21:00 UTC; the caller's gas is reimbursed from the treasury share.

Revoked

Zero brought flow across two consecutive bells. The badge returns to the contest pool. This punishes abandonment, not a bad day.

Relegated

Every bell, the lowest-scoring live badge is challenged by the highest-scoring wallet without one. Higher number sits. This is what guarantees movement on a slow day.

Sold

Badges are ERC-721 and the full record travels with them. A bought badge enters the drop zone like any other if its new owner does not run.

Twelve badges open on day one. Badge thirteen is born only when cumulative floor fees cross 2 ETH, and so on up to the cap of 1,366.

0 ETH collected

12 badges alive

2 ETH

13

20 ETH

22

100 ETH

62

2,708 ETH

1,366 — the cap

Two consequences, both intended. A launch-day bundle cannot capture the badge layer, because the badges it would need do not exist and cannot be created early. And the badge count becomes a public metric the protocol is unable to inflate: twenty-two badges alive is a statement about fees that actually arrived.

Each wallet is issued a runner link. When someone opens it and buys through the app, the swap passes the floor's router, which emits the attribution in the same transaction as the trade. At the bell an indexer sums every attribution event for each referrer since the last bell.

Counted

Buy volume, in ETH, from wallets other than the referrer's own

Not counted

Sells, self-buys, and buys routed anywhere else

Counting only buys is both the anti-manipulation rule and the alignment rule. Counting both sides would make a round trip a free score. Counting buys alone means a faked score costs a full round trip — fee in, fee out, price impact twice — for volume credited once. The side effect is the point: the entire competition generates buying, not churn.

Buys routed through other venues still pay the purse and badge holders are still settled from them. Nobody scores for that volume. We would rather under-count than invent a number.

Eligibility

A wallet must hold 0.01% of supply to claim or keep a badge — the same threshold for every badge

Settlement

Optionally take your share in the token instead of ETH

Burn

10% of every purse market-buys it and burns it

The eligibility requirement is a threshold, not a scale. A larger balance buys no share, no weight, no second badge and no protection from relegation. A tiered requirement would quietly rebuild the capital ladder.

At full occupancy the floor locks 13.66% of supply. Meaningful, and never enough for badge demand alone to hold a price up. The token is a key, not an engine.

• ▸Self-buys revert at the router and are not scored at all.
• ▸Only buys score, so a fake costs two fees and price impact twice for volume credited once.
• ▸The fee a faker pays lands in the purse and is split across every live badge — including their rivals'. Cheating subsidises the people you are racing.
• ▸A buyer wallet counts at 25% weight during its first epoch, so every fresh sybil wallet costs a day of delay.
• ▸The protocol issues at most one badge per address. More can be bought on the open market, never won.

These must be re-modelled against the launchpad's real fee split before any contract is written. If cheating is not provably loss-making, the mechanic changes — it does not ship anyway. That is Phase 0.

Nothing off-chain can change an outcome. Attribution is emitted in the same transaction as the trade, badge ownership and its record are on-chain, and every settlement is a transaction. The live board between bells, the result cards and the notifications are conveniences only.

The scoring script and the block range for each bell are published, so any bell can be recomputed from public logs and will produce the same answer. Each badge page lists the transaction hashes that settled it, and every one is clickable.

This is a closed economy. Settlement comes from the floor's own trading volume. It is real money and it is bounded. That is stated here rather than buried, because the alternative is the pattern this protocol was built against: inventing an external business to explain where rewards come from, and never shipping it.

• ▸If the floor is quiet, badges settle small amounts and the board says so in public. That is the system working, not failing.
• ▸The protocol lives on distribution, not technology. No runners means no movement and a quiet board.
• ▸Sybil pressure is raised, not eliminated. No system removes it entirely.
• ▸Only flow routed through the floor scores, so some volume will always be uncredited.