Saturday
Set your book
Choose from 182 tokenized equities and set whole-percent weights. Signing is free — it's a signature, not a transaction.
2,222 NFT seats. Every seat runs a $100,000 paper portfolio of real tokenized equities, priced live. You set your weights at the weekend, they lock on Monday, and every Friday the worst performers are cut. Eight weeks leave ten.
Most on-chain products give you something to hold. This one gives you something to decide, once a week, in public, against 2,221 other people making the same call.
Choose from 182 tokenized equities and set whole-percent weights. Signing is free — it's a signature, not a transaction.
Weights freeze. Prices don't. The board reorders every day and you can only get back in by paying $50 to unlock your seat before Wednesday.
Books are scored on percentage growth against a recorded price tape. The worst 278 are out. Anyone can recompute the result from the published marks.
278 seats are cut every Friday for seven weeks. The eighth cut takes the remaining 276 down to ten finalists, ranked on their compounded return across the whole season rather than on the last week alone.
Fantasy leagues have run on a database for thirty years, and your entry dies with the season. Here the seat is transferable and resellable while the contest is running — a market forms on how a book is doing, mid-season.
Prize money sits in a contract with no owner-withdrawal path and a timelock between declaring a winner and paying one. We could not run off with it if we wanted to, and you don't have to take our word for that.
Every price used to settle a week is published to an append-only tape. Anyone can recompute any elimination. An off-chain league asks you to trust the operator; this one doesn't.
We tested it before building it: 2,222 generated books scored over 41 trading days of real prices, 24 June to 20 August 2026 — seven weekly transitions.
Whoever led last week was no more likely to lead this week. Read it honestly: those books were randomly generated, so low persistence is partly built in. What it rules out is a structural edge in the format itself — not human skill, which we have no data on and will not have until people play.
Scoring on percentage growth rather than book size keeps the field alive to the last Friday. A big book cannot buy a lead.
Players fund the prize pot entirely. We take the mint and a tenth of the pot — so a bigger prize costs us nothing and makes the product better.
That the thing works. Contracts deployed, 234 tests, two internal adversarial reviews — no third-party audit yet, and that is one of the two things gating mainnet.
That anyone pays $30 for a seat, or pays a second time to stay in. No mint has run. Every revenue figure here is a model with its assumptions written down.
Distribution. We've sold into this audience once before, at smaller scale. Getting 2,222 seats sold is the open problem, not the engineering.
Every claim above is either on chain or on the tape. Deployed to Robinhood Chain testnet, chain id 46630.
REKT TRADOOOR minted 10,000 NFTs on the same chain in August, backed by a far larger audience than we have. Worth being straight about that.
Each character has a personality assigned at mint and trades on its own. The holder picks which to field, then watches. Stronger distribution, simpler product.
You choose every position, every week, and live with it. That's a harder product to build and a harder one to game — which is why we spent eighteen months on the settlement engine rather than the art.
Settlement takes a price tape and a set of positions. Swapping equities for another feed is configuration, not a rebuild — which is why the first product is the hardest one and the rest reuse it.
The contest above. Built; awaiting an external audit and a legal opinion before mainnet.
Same seats, same engine, instruments with more expression in them. Sold as recurring access rather than a new audience.
Prediction markets traded $111B of notional in Q2 2026 (Predicted, Q2 2026 report). The tooling for participating in them programmatically is thin.
Direction, not commitment. No revenue is attributed to any phase beyond the first, here or in anything we send you.
We've just told you phase one caps in the hundreds of thousands. So here is the argument for building it anyway, and you should judge us on whether you buy it.
Weekly equities is the version with the most ways to cheat — we found six and closed them. Anything we run afterwards on a simpler feed inherits an engine that has already survived the worst case.
Settlement takes a price tape and a set of positions. Crypto, FX, commodities, event markets — a new contest is a config change and a data source, not a rebuild. That is where the size is, and phase one is what earns the right to try it.
Will people pay repeatedly to compete on assets they don't own? Nobody knows. One real season answers it for every phase behind it, and it costs a mint to find out rather than a round.