One Bull.
One seat.
One strategy.
Condo spent months building strategy intelligence for its own treasury: an engine that tests every idea on live markets, in public, against thousands of random strategies. Now the Lab opens to a limited audience on Robinhood Chain. Author a strategy, beat the flock, win the competitions, and get paid when the Desk trades your work.
The treasury's strategy lab, opened to the people who hold it.
Condo's treasury - over 100 ETH and other assets, reported live at condobase.io - has one job: compound. For months the work behind it has been an engine that tests trading strategies the honest way - on live markets, forward only, with costs charged and a stop mandatory, against a flock of thousands of random-but-legal strategies that shows what luck alone produces. The next thing the treasury allocates to will not be another leveraged ETH product; it will be a strategy that cleared that bar in public. This document opens the Lab to holders.
The loop: mint a Encierro Bull, which is a seat on the Board. Put a strategy on it - adopt a house starter, tell your analyst an idea in plain English, or write the document. It paper-trades against the flock in public. Beat the flock and you climb the competitions. Clear the Desk's four gates and the treasury trades your strategy with its own capital, paying you 20% of realized profits.
A seat that never competes earns nothing, ever.
Scoring counts only the forward record - what a strategy did after its holder pressed go, never what a backtest says it would have done.
Every book is measured against the shadow flock on its own market, so a number only matters when it beats what luck alone produced.
“Anyone can post a green screenshot. A seat on the Board has to out-trade two thousand monkeys, in public, going forward. That is the flex.”
And four things that make a seat worth taking early: the clocks start at mint, and none of them rewind.
Records only count time actually spent on the Floor. A record started late is a record forever shorter.
A seat is the only way onto the Board. The supply is fixed at mint and never extended.
Weekly sprints, monthly boards, quarterly honours, each with a treasury prize pool. A sprint sat out is a prize handed to someone else.
The daily +$100 of game money never accrues retroactively. A missed day is gone.
Running software, with receipts.
This document describes software that is running, not intentions. The engine has ticked every fifteen minutes on live market data since September: 2,500 books on the Floor, the three house starters, the signal board, the venue board, the daily motion and the model card are public. The scanner it grew out of has published its selections and its honesty gates since June. The collection contract is written and tested, not yet deployed: the mint is gated on milestones - final art, an independent contract review, allowlist close - not on a calendar. No date appears anywhere in this document by design.
What we claim, and what backs it, in one line each. Built and tested, never traded: every record here is paper, forward-tested, in public. The next allocation is gate-driven: the treasury deploys real capital only when a strategy clears the four gates in section 10, and the first such allocation is undated. The engine learns: the model card in section 04 shows its training set and its accuracy against the base rate, and the advice record shows how counsel fared. A sentence in this document that cannot point at a public page is a promise, and we have tried to cut every one of those.
Encierro: the key to the Lab.
Encierro is an ERC-721 collection on Robinhood Chain, and each Bull is a seat on the Board. Robinhood Chain, because the treasury intends to trade tokenized stocks alongside crypto, and this chain's stock roster is exactly what the engine already paper-trades. Supply, mint price and per-wallet limits are published with the announcement. Whether Bulls carry types that name a starter strategy is an art decision still open. What is fixed: a type, if any, is identity, never an advantage; any seat may run any strategy.
Fairness is mechanical: a provenance hash is committed before the mint can open, so the assignment of art to tokenIds is provably fixed in advance; reveal is one-way; reserve mints draw from the same buckets with no better odds; and wallet limits are counted against mints, not balances, so cycling tokens through wallets buys nothing. Holding is unlimited; only minting is capped. The exact constants are in section 14.
Every seat has an analyst. Yours remembers you.
Each seat anchors its own AI analyst: its own memory of its holder, and a live, source-tagged read on the market. Ownership is not a login - it is re-read from the contract on every message, so selling the Bull closes the thread behind you and hands the next holder a clean relationship with the same analyst. Nothing is pooled between seats; facts an analyst learns about you stay on its tokenId.
The analyst is also the Board's interface. Tell it a trading idea in plain English and it compiles the idea into a legal strategy document - the bounded, validated format every book runs (section 06). The model only translates; the validator judges. It answers with tools, not from memory: it reads the Floor, the Library, any seat's record, the signal and venue boards, and it can compile and backtest an idea mid-conversation, naming the tools it used. It is an assistant as well as an analyst: the daily claim and the week's counsel can be handled by telling it so.
The leash. The holder decides how much the analyst may do alone: advise (it drafts, you act), propose (its weekly counsel waits for one tap), or act (it queues its own counsel within rails you set - a stop no wider than your rail, drawdown no worse than a quarter, and only when the proposal scores higher). Autopilot moves paper only, and every automatic decision is public on the seat.
The advice record. Every counsel decision old enough to judge is scored: did the proposed document beat the prior one going forward, whether or not the holder took it? The share that did is public, per seat and across the Board. An advisor whose record is a public number is a different kind of advisor.
How the engine trains. A meta-model learns which strategy shapes out-perform the flock - blocks, signals, rails, market, timeframe - and publishes its own card: how many books it trained on, its accuracy against the base rate, and which labels it used. It retrains nightly. It starts on backtest-era labels and switches to forward labels on its own once three hundred forward-rated books exist. Its prior nudges counsel and the playground; it is never a gate. This is the curve the whole Lab is built to bend: more records, better priors, a fuller Library, a real Desk.
“The agent layer holds no keys, signs nothing and executes no trades - on this the whole design is non-negotiable.”
A seat competes only when its holder puts a strategy on it.
The Board is the tournament: every seat, every house starter and every flock book running on the same engine, the same candles and the same rules, ranked in public on the Floor. Nothing is passive. A Bull that never takes the Floor earns nothing - no Standing, no prizes, no shelf, no Desk.
The Lab's own engines take the Floor as house books: 20-Day Breakouts and 52-Week Highs on BTC, ETH and SOL, expressed in the same document format every seat uses and judged by the same flock under the same gates. Two translations are stated, not hidden: a document cannot read Bitcoin's regime from another market, so each book reads its own 200-day trend, and the 52-week engine's 50-day simple average becomes the 50-day EMA. Failed Bounces stays a scanner - it is a short-side rule, and the Board is long-only in v1. The Desk trades the house engines only if they clear the same four gates as anyone else.
Three doors onto the Floor: adopt a starter (three public house documents, any seat may run any of them), tell your analyst (plain English, compiled and validated), or write the document (full control of every block and bound). The cadence is the weekly sprint. Strategy changes mid-sprint queue for the next one, so a running week always compares like with like. The door swings both ways: a holder may park the seat at any time - off the Floor, earning nothing - and return later. Returning starts the forward record over.
The seat keeps talking. Your analyst proposes; you dispose - or, on a leash you set, it disposes within rails. Each sprint, every competing seat's analyst studies its own record against the flock and drafts one bounded tweak, validated by the same rules as any document and backtested side by side with the current one. Accepting queues it for the next sprint; dismissing costs nothing. Competing never demands re-authoring; it only ever asks for a decision.
Motions. Once a day the house drafts one question the market is actually asking, the starters argue it from their postures, and every seat with a record votes, voices weighted by that record. The division is public; it is content and a measurement at once.
Everything public stays public on the seat itself: each competing seat has a public profile with its record, Standing, advice record, streaks and authorship hash - and never the strategy document, which belongs to the Board and its author.
Honest by construction, boring by choice.
Every book on the Floor runs the same simulator under the same rules. Signals form on closed candles and fill at the next candle's open, never at the close that produced them. Costs of 0.15% are charged on both sides of every trade. A stop-loss is mandatory in every strategy. Books with fewer than ten closed trades are shown but never ranked. The baseline is the shadow flock - 2,500 random-but-legal strategies - and a book is only ever measured against the flock on its own market, so picking the hottest market is not skill.
The ranking score is drawdown-adjusted, credits sample size only mildly, and carries a concentration penalty: the more of a book's profit came from its single best trade, the less the score trusts it. One lucky moonshot cannot carry a book; a trend-follower that earned its return across many trades keeps nearly all of it. Every book publishes the share of its profit its best trade produced, so nobody has to guess why.
Crypto around the clock. Stocks, because of where the treasury is going.
A strategy declares its market. The Board trades thirteen: BTC, ETH and SOL on 1-hour, 4-hour and daily candles, and every stock Robinhood Chain tokenizes as daily markets: AAPL, NVDA, TSLA, AMZN, MSFT, GOOGL, META, MSTR, SPY and QCOM. Stock books follow stock discipline - daily candles only, served only after the session has fully passed, split- and dividend-adjusted - and stock candles come from a chain of sources with a split-sanity check, so no feed can print a phantom crash into a book.
Nine external series a strategy may read. Implied volatility, perpetual funding, open interest, the day's liquidations, the Fear and Greed index, stablecoin supply, the VIX, the dollar index and the 10-year minus 2-year Treasury spread. Every series is aligned as-of publication - a book reads a day's value only after that day has ended, so nothing can leak the future. A source that is down is said out loud on the public source board, and a book that needs it is skipped that tick, never run on a guess.
The venue board. The Lab indexes where Robinhood Chain's tokenized stocks actually trade onchain - the pools pairing each stock with the USDG stablecoin - and publishes each pool's depth, its 24-hour volume, and the stocks with no public venue at all. It exists because the treasury's stock strategies can only ever trade real capital where real depth exists, and "the token exists" is not "the token is tradable at size".
Condo intends to be present on Ethereum, Base, Robinhood Chain and Hyperliquid. The engine is chain-agnostic; the seats live on Robinhood Chain first. No dates.
Points for what you did, never for what you would have done.
Standing is the Board's merit score, and it is deliberately hard to farm. It is paid once per weekly sprint, only to seats, and only on the forward record - the strategy's performance from the moment its holder activated it. A brilliant backtest earns nothing; adopting a document minutes before the sprint closes earns nothing; market choice earns nothing, because the baseline is per market; a seat that never competes earns nothing at all. The formula is blunt: a forward-rated seat that beats X% of its market's flock earns X minus 50 points, so beating half the monkeys earns zero and beating all of them earns fifty.
The competitions. No new token and no airdrop: $CONDO already exists, and this document promises nothing about it. Rewards are prize pools from the treasury, on three cadences. Weekly: the sprint's Standing leaders. Monthly: the bankroll board, which ranks pure trading P&L with deposits subtracted. Quarterly: honours for the strategies that earned the Library and, when it opens, the Desk. Pool sizes and the currency they are paid in are published with each season, in exact amounts, never as ranges.
The rules of a round. One Bull is one seat, so a farm has to buy its seats, and the rules below close what a bought seat could still farm. Every round is closed once, never rewritten, and publishes its winners, their shares, and every excluded seat with the reason. The rules are versioned, and a change never applies to a round already closed.
| Eligibility | A seat competes in a round only if it was on the Floor before the period opened, and only a forward-rated record counts |
| Weekly | The sprint's Standing: points paid on the forward record against the flock on the same market. Ten winners, 25% to the first, 3% to the tenth |
| Monthly | The bankroll board with every deposit subtracted, so the daily claim climbs nothing. Five winners, 40% to the first |
| Quarterly | Library strategies ranked by forward score, five winners, 40% to the first. Desk-ready seats are named as honours; the Desk is their reward |
| Documents | Seats running the same document compete as one entry, held by the seat that activated it first. A farm cannot out-rank its author with the author's own document |
| Wallets | One award per holder address per round; the highest-ranked seat keeps it. A seat that parks before the close forfeits |
| Shares | Shares no eligible seat earns stay with the treasury. A round with three qualifying seats pays three shares; a pool is never renormalised |
| Payment | The treasury pays the recorded holder address within the season's payout window; the platform records the transaction hash. Nothing here holds keys or moves money |
$10,000 of game money, and a reason to show up daily.
The first time a seat takes the Floor, its lifetime paper bankroll opens: $10,000 of game money. The bankroll survives strategy changes, sprints, parking and returning - only the skill clock resets; the chips carry. On top of it runs the claim program: for season one, a competing seat may claim +$100 of game money once per UTC day, only while it is on the Floor, never retroactively. Miss a day, that day is gone. Consecutive claims build a visible streak - status, deliberately not money.
Deposits are not skill, and the boards are built so nobody can pretend otherwise. The weekly and monthly bankroll boards subtract claims and rank pure trading P&L - clicking a button climbs nothing. The all-time board shows the full stack with the deposited amount printed beside it. Standing ignores bankrolls entirely.
“GAME MONEY, IN FULL: every dollar in every bankroll is a paper-testing balance. It is not real funds, it is not withdrawable, it is not custodied anywhere, it is not a prize, and it is not a preview, proxy or promise of any distribution of anything. It is the score in a practice game, and nothing else.”
A backtest gets you a book. The shelf gets you a payday.
The Library is the Board's shelf of proven strategies, and entry cannot be bought, only earned: a strategy qualifies when its forward record is rated (10+ closed trades since going live) and beats the flock's forward median on its own market. The shelf is empty at launch on purpose - the first entries earning their way on are the point. Authorship is permanent: every strategy document is canonically hashed, and the hash is public on the seat's profile from the moment of activation - provable, timestamped authorship without revealing the recipe. Documents are visible to the Board's own systems, never to the public, and the Board learns only from aggregated, anonymised shapes and outcomes.
The shelf is the interview for the Desk. The Desk is the treasury's own program: trading Library strategies with the treasury's own capital, paying the author 20% of realized profits. This is the whole reason the Lab exists - the next allocation the treasury makes will be a strategy that cleared this bar in public, not another leveraged ETH product. The first such allocation is undated and gate-driven. The terms are published here, not promised later:
One thing the Desk knows about leaderboards: the top is where luck lives. Run thousands of books and the best of them look spectacular even if nobody has skill, so the Desk is built to harvest the average of proof, never the maximum of luck. Capital follows only strategies that pass all four gates, and every seat's gate readout is published live with its forward record - Library entry · beat buy-and-hold of its own market · 180 days of forward record · 30 of them in a falling market.
PAID TO THE SEAT'S TREASURY
From the Desk's allocation to their strategy, accounted monthly against the allocation's own high-water mark.
The treasury's own funds, only, always. Never holder deposits, never pooled money, never anyone else's capital - the Desk is a prop desk, not a fund, and authorship is a contribution, not an investment.
All four Desk gates, published live on every seat: a Library entry, a forward return that beat buy-and-hold of the same market over the same window, at least 180 days of forward record, and at least 30 of those days with the market's trailing 30-day return negative.
As an ensemble, never a concentration: spread across the whole qualifying shelf, with no single strategy taking more than 10% of Desk capital. Tokenized-stock strategies trade real capital only once the asset's onchain venue depth is verified per token.
The canonical hash published at activation. Whoever holds the Bull that authored the shelf entry holds the claim - it sells with the Bull.
Per-strategy allocation is capped, every Desk position honours the strategy's own stop, a strategy that falls off the Library is unwound, and the treasury can halt the Desk at any time. The Desk can lose treasury capital; nothing here guarantees profits or payouts.
The Desk's book is published - allocations, realized results and author payouts. From its first real trade it also publishes realized-vs-paper slippage per fill and recalibrates the paper engine from live data.
Confirmed v1 terms - they freeze at mint and cannot worsen for authors afterwards. Desk size is the treasury's choice and is never promised.
Every seat has a treasury. Only its holder has the keys.
Each Bull carries a treasury: an ERC-6551 token-bound account owned by the Bull itself, and therefore by whoever holds it. It exists for one reason. When the Desk pays an author, the author's cut lands in the seat, not in whichever wallet happened to hold it that month - so the claim is part of the asset and sells with the Bull. It is not a wallet feature, not a trading account, and never holds a holder's own capital. The canonical registry is already deployed on Robinhood Chain, so the address is pure math from the day the collection deploys, deployed by the holder's own signature only when there is something to receive. Until then it stays quiet: no address to admire, nothing to deploy, nothing to do.
The Lab now. The seats next. The Desk when a strategy earns it.
No new token, no airdrop, no dates. The road is a sequence of proofs, and each stage opens only when the one before it has produced its receipts.
The engine, the flock, the signal and venue boards, the model card, the daily motion - running in public on live data.
Final art, an independent contract review, the allowlist close. Then the mint, and one Bull is one seat.
Weekly, monthly and quarterly competitions with treasury prize pools; the Library fills; the first gate readouts turn green.
The treasury's first strategy allocation, made only when a Library strategy clears all four gates, reported in the open. Then more markets and more chains: Ethereum, Base, Hyperliquid.
Plainly, because it matters: paper bankrolls convert to nothing. Standing is a score and a competition ranking, not a claim on anything. $CONDO exists today and this document makes no statement about it. None of this is an investment offer.
What this is not, in writing.
| No custody | No service in the Lab holds keys, signs transactions, executes trades or touches funds. The seat's treasury is holder-deployed and holder-controlled, and only ever receives what the Desk pays |
| No advice | Everything the analysts, the Floor, the Library, the motions and this document say is information about a paper game, not a recommendation to trade anything real. Autopilot moves paper only, inside rails the holder set |
| Paper is paper | Simulated results use simplified fills and fixed costs. Real markets have slippage, liquidity gaps and failures the simulation does not model - a strategy that wins the Board can lose real money |
| Game money | Bankrolls, claims and streaks are practice-game balances: not real funds, not withdrawable, not a prize and not a measure of one |
| The treasury | Treasury allocations happen only through the published gates and are reported in the open. The Desk can lose treasury capital. Nothing here guarantees profits, payouts, prize sizes or dates |
| Data honesty | Market data comes from public endpoints and named providers, each answer tagged with its source. When a source is down, the analysts say so instead of filling the gap from stale memory |
| Data use | Aggregated, anonymised strategy shapes, records and counsel decisions train the Board's own models. No individual document is ever exposed or exported; chats with your analyst are never training data; nothing learned is used against a holder. Changes publish before they take effect |
The numbers, as the code has them.
Collection constants live in contract bytecode; tournament parameters live in the engine and are stated here exactly as configured.
| Collection | Encierro · ERC-721 on Robinhood Chain (chain ID 4663) · supply, price and per-wallet limits published with the announcement · limits count mints, never balances · holding unlimited |
| Provenance | Hash committed before mint can open · reveal is one-way · placeholder metadata until reveal |
| Sprint | One ISO week · strategy changes queue for the next sprint · Standing paid at sprint close · weekly prize pool |
| Scoring | return / max(drawdown, 2%) · sample-size credit (saturating) · concentration penalty (score scaled toward half as the best trade's share of profit approaches 100%) |
| Rating bar | 10 closed trades to rate a record · 30 rated flock books per market to anchor a baseline |
| Costs | 0.15% per side, every trade, every book · mandatory stop-loss in every strategy |
| Flock | 2,000 random-but-legal books plus a 500-book signal flock, spread across all thirteen markets, regenerated only with published seeds |
| Markets | BTCUSDT, ETHUSDT, SOLUSDT (1h/4h/1d) · the chain's full tokenized-stock roster - AAPL, NVDA, TSLA, AMZN, MSFT, GOOGL, META, MSTR, SPY, QCOM (1d, split-adjusted, source chain with split-sanity check) |
| Signals | dvol, funding, oi, liqs, fng, stables (crypto) · vix, dxy, t10y2y (any market) · earnings_within (stocks, calendar-gated) · daily, aligned as-of publication |
| Venues | Robinhood Chain stock/USDG pools indexed hourly · depth = pool balances at pool price · roster stocks without a venue listed as missing |
| Bankroll | $10,000 game money at first activation · +$100 daily claim, one per UTC day, on-Floor only, for season one · claims subtracted from week/month boards |
| The Desk | Treasury funds only · authors of traded Library strategies keep 20% of realized net profits, paid to the seat's treasury · ensemble allocation, max 10% per strategy · public book with realized-vs-paper slippage · confirmed v1, frozen at mint |
| Desk gates | Library entry · forward return above buy-and-hold of its own market · 180+ forward days · 30+ of them with the market's trailing 30-day return negative · gate readout published live per seat |
| Competitions | Weekly (Standing, 10 winners, split 25/18/14/10/8/7/6/5/4/3%) · monthly (bankroll board with deposits subtracted, 5 winners, 40/25/15/12/8%) · quarterly (Library by forward score, 5 winners, same split; Desk-ready honours) · one entry per document, one award per wallet, activation before the period · pool sizes and currency per season in exact amounts |
| Token | $CONDO exists; no new token, no airdrop; this document makes no statement about it |
The Lab is open. The seats are not, yet.
A treasury that trades only what has proven itself in public is a rare thing. The engine is already in session; the records are already building. The seats open at mint, and the first strategy the Desk trades will have earned it where everyone could see.