10,000 floor traders on Robinhood Chain. A PFP collection where the pit is the game: attention accrues, inattention gets liquidated, and the org chart is the art.
Every mechanic routes through one loop: ring the bell daily → accrue $TICK → claim it or risk it → burn it in sinks that feed the jackpot. Predator classes run a second loop on top — hunting the unclaimed balances of anyone who stops showing up. Referral pays forever, but only out of activity fees, never out of anyone's principal. The design target, tested below by simulation: the game must survive at zero growth.
Five jacket colours, weighted draw at mint. Class sets your emission multiplier and your powers. The near-miss is shown on reveal (“one trait short of INSIDER”) because the almost-win is half the pull of any slot.
| Class | Odds | Seats | Emit × | Power |
|---|---|---|---|---|
| PAPER HANDS | 60% | 6,000 | 1.00 | Baseline. Cheapest fusion fodder. |
| DIAMOND HANDS | 25% | 2,500 | 1.25 | 1 free missed day per week. Holds through. |
| WHALE | 10% | 1,000 | 1.50 | Can raid the liquidation window. |
| MARKET MAKER | 4.5% | 450 | 2.00 | Raids + earns 20% of all trading-halt premiums. Sells the insurance. |
| INSIDER | 0.5% | 50 | 3.00 | Raids + sees liquidation windows 1h early. Material non-public information, as a trait. |
Referral commission on mint is a one-off affiliate fee, three levels, hard-capped. The ongoing tree income comes later, from secondary royalties only — that's the line between a growth loop and a pyramid: no yield is ever paid from another entrant's principal.
| Line | Share | Ξ | Notes |
|---|---|---|---|
| Referral commissions | ≤15% | ~17.6 | L1 10% · L2 3% · L3 2%. Expected at 65% code uptake; unclaimed remainder → treasury. |
| Closing Bell seed | 5% | 9.0 | 2.25Ξ per quarterly season. |
| $TICK liquidity | 25% | 45.0 | Paired at launch. |
| Treasury · dev · art | 55%+ | ~108.4 | Floor 99Ξ if referral maxes. |
| Gross book | 100% | 180.0 | 10,000 × 0.018Ξ |
| Destination | Of price | Why |
|---|---|---|
| Treasury | 2.0% | Ops runway. |
| Referral tree (L1/L2/L3) | 1.0 / 0.3 / 0.2% | The honest pyramid: upline earns from downline activity, forever. |
| Closing Bell pot (Ξ side) | 1.0% | Jackpot grows with volume. |
| $TICK buyback → burn | 0.5% | Volume is deflation. |
Reality check at 1.5× mint-cap Y1 volume (270Ξ): royalties 13.5Ξ, of which 4.05Ξ across all referral trees. Trail income is real but small unless volume is real — say so publicly, loudly. Your PFP renders your downline as an entourage behind your trader; rank thresholds (Broker → Desk Head → Partner) restyle the jacket. The flex is the org chart.
At mint you choose. Market order: reveal instantly at table odds. Limit order: lock the token unrevealed for 72h — rare-tier odds get a +15% relative boost (Whale-or-better moves 15% → 17.25%) but the position sits naked on the tape: anyone can snipe it for 0.027Ξ. You keep 90% of the snipe (a forced +35% exit net of your 0.018), 10% feeds the pot. A second gamble stacked on the first, and both sides of the snipe are volunteers.
Fixed supply 1,000,000,000. Game emissions come from a 40% allocation on a halving schedule: Y1 200M · Y2 100M · Y3 50M · Y4 50M. Remainder: 25% treasury, 20% liquidity, 10% contributors (2y vest), 5% seasons. The daily pool is fixed and split over whoever showed up:
This makes the economy self-stabilising: every player who churns raises everyone else's pay. In the zero-growth sim, a max-streak degen's daily take rises from 93 to 148 $TICK/day (+60%) over the year as tourists wash out. Last men standing get paid.
| Sink | Burned | Mechanic |
|---|---|---|
| Fusion gacha "The Pit" | 19.3M | 400 $TICK → 55% trait upgrade · 30% rekt · 15% wash-trade refund. Pity +5%/fail. |
| Daily tithe | 18.4M | 10 $TICK to ring the bell. |
| Liquidations | 14.5M | 25% of every raided pool + raider stake. |
| Trading halts | 13.6M | Streak insurance, 80% burned · 20% to MM class pool. |
| Claim fee 2% | 3.5M | Small toll on moving accrual to safety. |
| Total burned Y1 | 69M | 35% of the 200M emitted → net inflation 131M. |
Honest read: Y1 inflates. Sinks scale with activity while emissions halve, so the ratio climbs toward parity in Y2–Y3. If that offends, cut D(0) — but thin early emissions are worse for retention than thin price. 2% of all sink flow (0.39M Y1) tops up the Closing Bell.
One cheap transaction a day. Multiplier m(s) = 1 + 2(1 − e^(−s/10)) — steep for the first week (hook), asymptote at ×3 (no infinite compounding). Miss a day uninsured and the streak halves — it never zeroes, because full resets make people quit rather than restart. Trading halts (max 2 held) auto-consume on a miss; they cost two days of your own accrual, so insurance is priced to your income.
Lapse uninsured with ≥250 unclaimed $TICK and your balance enters a 24h liquidation window. Any active Whale, Market Maker or Insider can raid it: raider stakes 20% (burned), takes 50%; 25% burns; you keep 25%. Insiders see windows an hour early. Someone profits from your inattention — that's the retention engine, and it's 48% of predator P&L in the sim (10,977 $TICK/predator/yr across 71,272 raids). Dust under 250 is never raidable; casuals who claim weekly stay safe.
A rolling pot: 2.25Ξ seasonal seed + 1% of all secondary volume + 2% of every $TICK sunk. Any sink burn ≥100 $TICK buys time on the tape; each entry extends the round timer +90s (capped at 24h). When the clock hits zero the bell rings and the last ten entrants split it 40/20/10/8/6/5/4/3/2/2. Terminal-phase FOMO, on a timer everyone can see, funded by everyone's ordinary play. Quarterly bull/bear seasons flip class buffs, reset the leaderboard, and give lapsed players a fresh-start re-entry point.
Automation is coming for this whether we design for it or not, so it gets its own lane with its own rail. The daily check-in is human-only, permanently. There is no ring_bell tool on our MCP server and there never will be — the bell is the only thing this game measures, and an agent that can ring it turns the streak into a server-uptime contest and every retention number in this document into fiction.
What agents get instead is the work humans do badly: hunting liquidation windows at 3am, quoting the secondary market, and filing watchtower reports against chain state. That first job is already 48% of predator P&L — making it the canonical agent task moves work that was always going to be automated into a sanctioned channel we can observe and rate-limit.
| Scored on | Direction | Why it can't be faked cheaply |
|---|---|---|
| Quote uptime | higher | Requires real capital at real risk. |
| Effective spread | lower | Tight spreads cost money to maintain. |
| Counterparty diversity | higher | Distinct non-affiliated fills. Sybils cost 0.018Ξ a seat. |
| Raid latency | lower | Pure skill and infrastructure. No self-dealing possible. |
| Watchtower accuracy | higher | Verifiable against chain state. |
| Never scored | — | Trade volume · trade count · raid count. |
The buff is capped at ×1.4, applies to the class multiplier and never the streak, and decays back to ×1.0 over 14 days if the agent goes dark. A player who never writes a bot is never priced out — they just don't get the bonus lane. Every Desk action is public by default: the leaderboard, the run log, the misses. That transparency is the point, and it is also the marketing.
The read-only MCP server is live now at /mcp — classes, windows, seat state and standings, no writes. Write actions (raid, quote, report) stay unimplemented until agent identity and per-tool scoping are in place.
Monte Carlo, 10,000 agents, 365 days, zero new entrants, seed 4663. Attention archetypes: 15% degens (97% daily), 35% regulars (85%), 30% casuals (60%), 20% tourists (25%), each with churn hazards that spike after being raided. All mechanics above run live: streaks, grace, insurance, claims, raids, fusion, tithe.
| Metric | Value | Read |
|---|---|---|
| Daily actives d7 → d365 | 6,728 → 4,115 | 41% of supply still rings the bell daily after a year of zero growth. |
| Casual median stack | 4,581 | Prey survive. Farmed, not exterminated. |
| Raid share of predator P&L | 48% | Hunting is half the business of a Whale seat. |
| Top-decile supply share | 34% | Concentrated but not degenerate. |
| Net Y1 inflation | 131M | Deliberate; halvings + activity-scaled sinks close the gap Y2+. |
Value framing at illustrative FDVs: $1M → a max-streak degen earns ~$0.15/day by year-end; $5M → ~$0.74; $10M → ~$1.48. Nobody quits their job. That's the point — the sim survives zero growth because the loop is a game, not an APY. If retention required yield, the whole thing would be a countdown clock.
Pit vernacular as trait canvas: jackets (class colour), badges (3-letter acronyms), hand signals (open-outcry buy/sell), era props (headsets, ticket wads, red suspenders, cigars), backgrounds as market regimes (bull run, crash, sideways, halted). Post-mint the art stays alive: failed fusions leave scars, streak badges stamp at 7/30/90, your referral entourage renders behind you, and the whole floor's mood tracks a live price feed — green days it celebrates, red days it panics. Hold a streak through a 5-day drawdown and a soulbound diamond stamps the badge. Their chart-watching dopamine becomes the asset's weather.
mintWithCode referral registry · $TICK ERC-20 · StreakManager (daily epochs) · RaidVault · ClosingBell. All on RH Chain (4663, Arbitrum Orbit).