01A market opens on every launch
Thousands of tokens are created on the chain every day. On each one, a market opens by itself and asks a question the contract wrote from a template: will it graduate before six o'clock, will its reserve still be above a number tomorrow, will its market cap cross a line.
Nobody writes the question and nobody answers it. At the hour, the contract reads the state of the chain and settles. That reading can be reproduced by anyone, at any time, without trusting us.
The token is never held, never deposited, never asked for permission. The market only reads it.
02Yes and No are two halves of one deposit
Put one unit of collateral in and you get one Yes and one No. Hand back one of each and you get your unit back. After settlement, the winning side is worth one and the losing side is worth zero.
Because a Yes and a No can always be recombined, their prices always add up to one. That is what makes the price readable: a Yes at 0.62 means the market thinks sixty-two out of a hundred. There are no odds to convert.
Selling is handing your side to someone else. The collateral never leaves the market, so the market cannot be emptied.
A market never goes all the way to one side. The more a side is abandoned, the more it pays, so it always finds a taker.
03A trade pays two percent
There is one place where the protocol takes anything: a trade inside a market. Nothing on the way in, nothing on the way out, nothing on winnings.
Two percent is more than a plain swap, and the reason is worth saying out loud rather than hiding. Holding a prediction market open is a losing job before fees, because whoever seeds it ends up holding the side that lost. Without a fee at that level nobody seeds, and with nobody seeding there is no market at all.
04Where those two go
The fee is cut in three, in the same transaction, and none of the three is an address the team controls.
Nothing in that picture reaches the team. A team allocation can exist in the initial supply, vested and announced, but it never draws from the flows.
The seeder is paid for putting collateral at risk and for bringing people to the market. The burn removes tokens for good. The settlement fund pays whoever calls the settlement, and on an auto-indexed market that costs almost nothing, so the surplus is burned too.
05Trading freezes while the chain is measured
A market runs in phases. People trade freely, then trading stops, then the measurement window runs, then anyone can call the settlement and the winning side withdraws.
The freeze is the most important safety line in the whole design. If trading stayed open while the measurement was accumulating, somebody could watch the result take shape and trade knowing it. The freeze removes that asymmetry completely: the second the window opens, nobody can move.
The freeze time is printed when the market is created, so you know to the second when you can no longer leave.
06The number to watch is what is locked
Every open market holds collateral that cannot be sold while it is open. Add all of them together and you get one figure, readable on chain, impossible to dress up: the open interest.
It is not a promise and it is not a projection. It is the amount of token that real activity has taken out of circulation right now, and it moves only because people are using the thing.
07The token is the collateral, and that is the whole role
Every position is opened with it, every seed is made of it, part of every fee is burned out of it. Take the token away and there is no position to open, no market to seed, nothing left standing.
One thing has to be said plainly. Making the token necessary does not create demand for it. It turns demand for the product into demand for the token, and it works just as hard in the other direction when nobody is using the product. Anyone telling you a burn guarantees a price is lying, and the people whose opinion matters will read it that way.
Holding a position also means holding the token while the market runs. Over a few hours that barely matters, which is the real technical reason markets are short.
08What the design cannot do
- 01It cannot make a market insolvent. Collateral only leaves by merging a pair or by a redeem that pays exactly what is there.
- 02It cannot stop you selling. Selling is handing your side to somebody else, and nothing in the market can refuse it.
- 03It cannot be argued with on an auto-indexed market. The answer is a reading of chain state, not an opinion.
- 04It cannot pay the team from the flows. There is no address in the split that anyone controls.
- 05It cannot give you a reliable probability. Prices carry momentum as well as conviction, and this is traded, not measured.
09The flaw
The protocol rides the rate at which tokens are launched on the chain. That rate is very high today and it is cyclical by nature. If it slows, the feed thins out, and nothing in the design prevents that.
Free questions, written by people instead of by a template, are the only cover for it. That is their real strategic purpose, and it is why they exist at all.
Two more things that have to be said rather than buried. There is no technical moat here: a competent team can rebuild this in a few weeks, and only being early, being used, and having the token already distributed defends it. And this is betting. In a lot of places that puts it in the most restricted category there is, so serious geo-blocking applies, no regulated party will ever promote it, and any plan that counts on being adopted by one should be dropped on day one.