There is no "how far it goes" here. You are buying a claim that will be judged, and at expiry it is worth exactly $1 or $0. The numbers in between — 0.37, 0.62 — are the market's estimate of how likely the claim is to come true. Read this and you should be able to open any board and know what it is saying.
A board asks "will this team win tonight." The screen says 0.62. That is not odds and it is not points. It means: someone will pay 62 cents for the claim that they win.
| You pay | Claim comes true | It doesn't | |
|---|---|---|---|
| Buy YES | $0.62 | get $1 +$0.38 | get $0 −$0.62 |
| Buy NO | $0.38 | get $0 −$0.38 | get $1 +$0.62 |
The two sides add to $1. That is not a coincidence — it is the foundation, and §2 explains it.
One caveat up front"adds to $1" is about settlement value, and that always holds. Quotes on the book need not. There is a spread: the two best bids sum to slightly under $1, the two best asks to slightly over. That gap is one of your costs.
The difference from a stock is the deadline. A stock has no expiry and can print any number. Here, at expiry the price is $1 or $0, necessarily. Everything in between is a temporary state.
What you hold is an outcome token. All mutually exclusive outcomes of one event form a complementary set. One rule: at settlement the winning one is worth $1 and every other is worth $0. So whatever happens, a full set is worth exactly $1.
Both directions are available at any time, without waiting for settlement. Which gives a useful corollary: if you can assemble a full set for less than $1, the difference is free money.
Plenty of people have had that thought, so the lane is picked clean — I censused it and the live opportunity inventory was zero. What matters to you is not the trade; it is that "a set = $1" is the anchor under every price on the venue.
Since the winner pays $1 and the loser pays $0, the fair price of a share is its probability.
Bet repeatedly at that price and you make money above a 62% long-run hit rate and lose below it. Which means "I like this one" is not a statement you can act on. The only actionable question is: is my number higher or lower than 62?
Taking liquidity costs a fee, so your real bar sits a little above the sticker. Counter-intuitively the fee is heaviest around 50¢ and cheaper toward both tails. The formula and a table are in the taker fee.
Posting liquidity pays no fee at all — so a maker's bar is the sticker price. The cost of that is you may never get filled.
In a casino your counterparty is the house, the house sets the odds, and the house's expected value is positive by construction.
Here the venue runs a central limit order book: everyone's bids and offers match against each other, so your counterparty is another trader who disagrees with you. The platform does not take the other side. It charges for the match. Nobody sets the price; it is shouted.
| Sportsbook | Prediction market | |
|---|---|---|
| Counterparty | the house | another trader |
| Price set by | the house | the book |
| Venue earns | the vig | a taker fee (makers pay nothing) |
| Win too much | limited, then banned | nobody has a reason to stop you |
| Exit early | usually no | yes — if someone will take it |
Three costs replace it: the fee (computable), the spread (computable), and the other side knowing more than you (not computable). The first two usually run one to two percent. The third has no ceiling.
A sportsbook at least prints its edge in the odds. Here the edge is inside your counterparty — invisible, but present.
| Step | Watch for | |
|---|---|---|
| 1 | Fund and approve | the contract needs an allowance before orders go out |
| 2 | Post or take | posting is free but may never fill; taking is instant and costs |
| 3 | Fill | minimum 5 shares; partial fills are normal |
| 4 | Hold | you can sell any time — if there is a bid |
| 5 | The event happens | — |
| 6 | Resolution | who decides, and on what, differs per category |
| 7 | Redeem | ⚠ the money does not come back on its own |
After settlement your winning ticket is worth $1 — but it is still a ticket, not a balance. Redeeming is an action you take.
The mirror image is worse: the losing ticket goes to zero and then quietly disappears. It does not leave you a line saying you lost. Anyone adding up their activity feed will therefore overstate themselves. I measured it once and was out by 8× — 589 shares evaporated.
| Commonly said | Actually |
|---|---|
| "0.62 is the platform's odds" | it is what another trader will pay; the venue has no view |
| "being right means making money" | depends on the price. Right at 0.95 earns 5%; right at 0.05 pays 20× |
| "you have to wait for the result" | leaving early is normal — thin books permitting |
| "no house, so it's a fair game" | no vig, but fee, spread, and a better-informed counterparty |
The second one deserves a sentence of its own. How often you are right does not matter; the price at which you were right does. A 55% judgement is a good business at 0.40 and a losing one at 0.70. Same judgement, different sign.
You are not buying something that goes up and down. You are buying a claim that will be judged, at a price that is also your break-even bar. So the question worth asking is never "who do I like" — it is "by how much does my number differ from this price, and is that enough to pay the costs."