You click "Buy YES" at 47 cents on a prediction market. Polymarket shows you an order book, a fee, an order confirmation. Two months later the market resolves and your account credits $1 per share.
What actually happened in the middle? The answer involves smart contracts, oracle disputes, market makers, and the surprisingly tricky question of "what counts as truth."
This is the under-the-hood walkthrough of an event contract — the mechanics nobody explains, and the part that matters when something goes wrong.
The five-second version
An event contract is a tradeable promise. It says: "If event X happens by date Y, you get $1. If it doesn't, you get nothing." The price between zero and $1 is the market's probability estimate.
Buying YES at 47 cents means: you're paying 47 cents to receive $1 if the event happens. Your maximum loss is 47 cents. Your maximum gain is 53 cents. The price moves continuously as new traders place orders.
That's the layer everyone sees. Below it, the mechanics are intricate.
What "buying YES" actually does
When you click Buy on Polymarket, you're not buying from "Polymarket." You're buying from another trader who is willing to sell YES at 47 cents. They might be selling because they think the actual probability is lower than 47%. They might be selling because they're hedging. They might be a market maker providing liquidity.
The matching happens via a central limit order book. Your buy order at 47 cents either matches against an existing sell order at 47 cents, or sits in the book until someone else hits it.
On Polymarket specifically, the order book lives on the Polygon blockchain. Every trade is a smart contract execution. The matched trade results in YES tokens flowing to your wallet and USDC flowing out.
You now own one YES share. If the event resolves YES, the smart contract pays you $1 in USDC. If it resolves NO, your YES share is worthless.
The NO side, simultaneously
Every YES share has a corresponding NO share. The seller of your YES share is, in effect, holding the NO side of the same contract.
If you buy YES at 47 cents, the NO side is by definition trading at 53 cents (47 + 53 = 100). This identity holds across the order book.
The implication: anyone can construct the opposite side of any trade by buying NO. If you think a contract is overpriced, you don't need to short YES — you just buy NO. This is cleaner than equity short-selling and is one reason prediction markets work better than stock markets for binary questions.
The market maker problem
For tight spreads to exist, someone has to be willing to quote both YES and NO continuously. These are market makers — professional liquidity providers.
On Polymarket, market makers are typically running algorithms that maintain a spread (say, buying YES at 46 and selling at 48 simultaneously). They earn the spread (2 cents per share) and take inventory risk on net positions.
The catch: when sudden news arrives, market makers widen spreads aggressively to protect themselves. A market that was 46-48 might suddenly be 30-65 for thirty seconds while market makers reassess. This is why prices can look wild around news events even when the underlying probability hasn't moved much.
If you're trading event contracts, never market-buy during a news spike. The market makers are intentionally giving you a worse price than the equilibrium.
How the market resolves: the oracle question
The contract says "pays $1 if BTC closes above $100K on December 31." But who decides whether BTC actually closed above $100K?
This is the oracle problem and it's where event contracts get philosophically tricky.
Polymarket uses a UMA Optimistic Oracle. Here's how it works: when the resolution date arrives, anyone can propose a resolution ("YES, BTC closed at $103K"). If nobody disputes within a window (usually 2-48 hours), the resolution stands. If someone disputes, the question goes to a vote among UMA token holders, who collectively decide the truth.
Most resolutions are uncontested. The price closed where it closed; everyone agrees. But for markets like "Did candidate X concede the election?" or "Did event Y count as a war?" the resolution can become genuinely subjective. UMA voting has been controversial multiple times, including in 2024 election markets.
Kalshi avoids this by using strict, contract-specified data sources. Every Kalshi market lists exactly which source resolves it. This is cleaner but less flexible — Kalshi can't list markets where the resolution source is ambiguous.
The trade-off: Polymarket lists more markets but has occasional resolution disputes. Kalshi has zero disputes but a narrower catalog.
The fees you don't see
Polymarket charges 0% trading fees on the official platform. But there are three hidden costs.
Spread cost: the difference between buy and sell prices. On liquid markets this is 1-2 cents per share. On illiquid markets it can be 10-20 cents. Spread cost is your actual transaction cost.
Gas fees: each trade is a Polygon blockchain transaction. Currently around $0.01-$0.10 per trade, but can spike during network congestion.
Bridging cost: getting USDC onto Polygon costs roughly 1-2% via centralized exchanges, sometimes higher via DEXs. If you're funding a new account, this is a real cost.
Total round-trip cost on Polymarket: typically 2-4% on a liquid market, 10-25% on an illiquid one. Compare to a stock trade at near-zero round-trip cost. Event contracts are more expensive than they look.
Kalshi charges 1-2% explicit fees but no gas or bridging cost. Net round-trip cost is similar to Polymarket on liquid markets, much better on illiquid ones.
What happens when the event hasn't resolved yet
Most event contracts let you trade out before resolution. You bought YES at 47 cents. A month later, news breaks and YES is now trading at 80 cents. You can sell your YES share at 80 cents and lock in your gain immediately.
You don't have to wait for resolution. The market is liquid through the life of the contract (assuming there's volume).
This is why prediction markets are useful even if you don't care about being right at resolution. They're real-time information aggregators. The price reflects current consensus, and you can take a view, trade it, and exit before the event happens.
For Bitcoin price contracts, this means you can effectively get directional exposure to BTC via a YES/NO trade rather than spot. For elections, it means you can express a view on the political shift without holding through the election itself.
The risk nobody talks about: counterparty failure
What happens if Polymarket fails? The contracts are smart contracts on Polygon, so technically your YES share survives even if Polymarket's interface goes dark. You can claim resolution through the underlying smart contract.
In practice, this requires technical skill most users don't have. If Polymarket's interface goes down, most retail users couldn't withdraw their positions even though the on-chain assets still exist.
Kalshi is centralized but regulated, with SIPC-equivalent protections for user funds. Manifold uses play money, so counterparty risk is moot.
For meaningful position sizes, this is the under-discussed risk. Smart-contract-resident assets are safer than custodial accounts in theory but harder to access in practice.
The three trade types you should know
Beyond the basic Buy YES / Buy NO, sophisticated traders use three more patterns.
Mint and split: deposit $1 USDC, mint one YES and one NO simultaneously, then sell one side immediately. This is how arbitrageurs lock in spread profits when the YES + NO total is not exactly $1.
Liquidity provision: deposit USDC into a market's liquidity pool. You earn the trading fees but also bear inventory risk. On Polymarket this is automated; on Kalshi it's done by institutional market makers only.
Spread trades: simultaneously buy YES on one market and NO on a correlated market. Example: buy YES on "Fed cuts in September" and NO on "S&P closes 2026 below 6,000." If you're directionally right on Fed policy, you win twice.
What this means for new users
The mechanics matter less than the discipline. Three rules that protect new users:
Never market-buy when news has just broken. Wait at least 5 minutes for the market makers to settle the spread.
Never put more in than you can afford to lose. Even well-resolved markets can take weeks to pay out, and your capital is locked the whole time.
Always read the resolution criteria before buying. The single most common loss for new traders is misreading what the market actually requires for resolution.
What Juno makes simpler
Juno is being built to abstract away the mechanics that make Polymarket intimidating to new users. Native fiat onramps. Clear resolution sources. Single-currency settlement. No bridges, no gas optimization, no UMA voting.
The mechanics are still there underneath — it's still a real-money event contract — but the user experience is closer to a stock trading app than a DeFi protocol.
The market itself is the same. The friction to participate is what changes.
The 47-cent lesson
Behind every 47-cent contract is a smart contract, an order book, a market maker, an oracle, a resolution dispute mechanism, and a settlement layer. Knowing how it all fits together turns prediction markets from a black box into a tool you can actually use.
The price you see is the price the market makes. The mechanics behind it determine whether the market is right to trust.
You don't need to be a developer to trade event contracts. You do need to know the order book is real, the oracle is human, and the price is yours to take or leave.