How Prediction Market Prices Work

A prediction market contract pays $1 or $0. That is why the price can be read as a probability: traders will not keep paying 70¢ for something they think happens 40% of the time, and they will not keep selling it at 40¢ if they think it happens 70% of the time. The price is where those disagreements clear.

This sits in prediction market basics, next to what prediction markets are and how they resolve. The research steps that turn a price into a trade decision are in How to Research a Prediction Market Before You Trade.

Implied probability

If a "Yes" contract pays $1 when the rules are met and $0 otherwise, then:

Implied probability ≈ price in dollars.

"Rough odds" here are the break-even odds of the contract before fees, not a sportsbook line. A 25¢ contract returns $1 if it wins, so you need it to hit at least one time in four to break even, ignoring costs.

The same logic runs in reverse. If "Yes" is 62¢, "No" is the other side of a $1 payout, so it should trade near 38¢. Paying 62¢ for Yes risks 62¢ to make 38¢. Paying 38¢ for No risks 38¢ to make 62¢.

A worked example

Suppose a market asks whether a bill will be signed by December 31, and "Yes" is offered at 40¢.

• You buy 100 Yes contracts.
• You pay 100 × $0.40 = $40.
• If the rules are met, the contracts settle at $1 and you receive $100. Profit before fees: $60.
• If they are not met, the contracts settle at $0 and you lose the $40.

The market is saying this outcome is about 40% likely. Your break-even probability, before fees, is 40%. If you think the true probability is 55%, the contract looks cheap. If you think it is 30%, it looks expensive.

Now put a fee on it. Kalshi's general event-contract trading fee, from the fee schedule effective July 7, 2026, is:

fee = round up(0.07 × contracts × price × (1 − price))

for an order that matches immediately (a taker), on the standard schedule. At 40¢ for 100 contracts:

0.07 × 100 × 0.40 × 0.60 = $1.68

You pay about $41.68 to receive $100 if you win. Your break-even probability is now about 41.7%, not 40%. The edge you thought you had has to clear that gap, plus the spread if you want to exit early.

Polymarket US uses a similar shape. Its fee schedule (checked October 28, 2026, reflecting the October 19 volume-tiered rebate schedule) is fee = 0.0695 × contracts × price × (1 − price) for a standard taker. At 50¢, 100 contracts cost $1.74 in taker fees. Novig's live and futures taker fees use the same price × (1 − price) shape with different coefficients, and pre-game straight trades on the app are currently fee-free. Details live on each platform's fee page, linked below, and on Kalshi vs Polymarket.

Bid, ask, and the midpoint

You rarely trade at a single printed price. The bid is the highest price someone will pay. The ask is the lowest price someone will sell. If the best bid is 60¢ and the best ask is 64¢:

• A market buy lifts the ask and pays 64¢.
• A market sell hits the bid and receives 60¢.
• The midpoint, 62¢, is the cleaner read of the crowd's probability.

The last trade might have been 60¢ an hour ago. If nothing has traded since, that print is not the market. This is why a research note should record the bid, the ask, and the time, not a single number copied from a headline.

What moves the price

Prices move when someone is willing to trade at a new level. The usual reasons:

• New information. A data release, a vote, an injury, a court docket. Liquid markets reprice in minutes. A two-day-old story is usually already in the price.
• A change in who is trading. A large order walks through the book and moves the price even if no fact changed. In a thin market, that move can be the whole story.
• Time. As a deadline approaches, uncertainty often collapses toward 0 or 1. A 50¢ contract the night before a known announcement is a different object from a 50¢ contract three months out.
• The rules getting clarified. If an exchange posts a clarification about what counts, the probability of the contract can change even though the real-world event did not.

None of these tells you the price is "right." They tell you why it changed. Your job is to decide whether the new price matches your estimate. Write the estimate as a range before you anchor on the screen. If the midpoint sits inside your range after fees, there is no trade. That comparison is step 6 of the research guide.

When the price is a bad probability

Treat the number with suspicion when:

• The spread is wide. A 40¢ bid and a 70¢ ask is not "55%." It is a market where you cannot transact near the midpoint in size.
• Volume is tiny. The last trade can be one person. It is not a crowd.
• You cannot exit. If you need the price to be a probability you can act on, you also need someone on the other side later.
• The rules are ambiguous. Traders may be pricing different interpretations of the same title. The price is then a mix of the event and a dispute about the contract.

A clean probability is a tight spread, recent trades, a named resolution source, and a fee you have already subtracted.

Prices are not advice

A 70¢ contract is not a recommendation to buy, and it is not a forecast with a confidence interval attached. It is the level at which the last traders agreed to exchange risk. Sometimes that level is a sharp summary of public information. Sometimes it is stale, thin, or wrong about the rules. Reading it well means separating those cases, which is what resolution rules and the research checklist are for.

Sources

Checked October 28, 2026.

• Kalshi fee formula, schedule effective July 7, 2026: Fee schedule PDF and Fees help center
• Polymarket US taker coefficient 0.0695 and October 19, 2026 rebate tiers: Fee schedule
• Polymarket international category fees: Fees
• Novig formula and the pre-game / live / futures split: Fees on Novig