The one idea you need
A market asks a yes/no question — "Will X happen by date Y?" — and issues two kinds of share, YES and NO. They always cost $1 together. If YES trades at 63¢, the market is saying there's roughly a 63% chance. Buy YES at 63¢ and you make 37¢ per share if it happens, lose 63¢ if it doesn't.
That's the whole mechanism. Everything else — order books, liquidity, resolution — is plumbing around it.
A worked example
You put $100 into a market where YES is 40¢. You get 250 shares. If the event resolves YES, those shares pay $250 — a $150 profit. If it resolves NO, you lose the $100. You don't have to wait for resolution either: if news pushes the price to 55¢, you can sell your 250 shares for $137.50 and book the gain immediately.
How prices are set
- Polymarket runs an order book. Your order sits at your price until someone takes it, or you cross the spread and take an existing order.
- Placing a resting limit order makes you a maker; you get a better price but no guarantee of a fill.
- In thin markets, the gap between best bid and best ask can be several cents — that gap is a real cost, and often larger than any fee.
Deposits, fees and settlement
- You fund with USDC, a dollar-pegged stablecoin, held in a wallet you control.
- Trading has historically been fee-free on most markets, with the cost showing up as spread rather than commission.
- Markets resolve against a specified source. Once resolved, winning shares redeem for $1 each and you can withdraw.
Where the edge actually comes from
1. Reading the resolution text, not the headline
Market titles are marketing; the resolution criteria are the contract. Half of all mispricings exist because casual traders priced the headline and ignored a clause about the data source, the cutoff time, or what happens if the event is postponed.
2. Time decay on "will X happen by" markets
Markets that require something to happen by a deadline decay toward NO as the window closes with nothing happening. Traders systematically overpay for YES on long-shot deadline markets — the same behavioural bias that makes lottery tickets sell.
3. Cross-market inconsistency
Related markets often don't add up. If "X wins" trades at 55¢ and "X wins by more than n" trades at 58¢, one of them is wrong by construction. Scanning families of related markets for arithmetic contradictions is more reliable than forecasting the event itself.
4. Sizing
Edge means nothing if you blow up. A fractional Kelly rule — stake a fraction of the edge-implied optimum, typically a quarter — keeps you solvent through the inevitable run of correct-but-losing calls.
Common beginner mistakes
- Buying at market in a thin book and paying 4¢ of spread on a 3¢ edge.
- Confusing your confidence with a probability. If you can't say 'I think this is 70%, the market says 58%', you don't have a trade.
- Holding to resolution when you could have taken the same profit weeks earlier and recycled the capital.
- Ignoring the opportunity cost of money locked in a market that resolves in eight months.
Keep reading
Compare venues in Kalshi vs Polymarket, or see how the same mechanics apply to Bitcoin price markets.