How to Read Prediction Market Odds: Price, Spread and Fees

A prediction market price is an implied probability. A YES contract trading at 62¢ pays $1 if the event happens and nothing if it does not, so the market is pricing that event at about a 62% chance. To read the number correctly you also need three adjustments: which side of the spread you would trade at, what the fee adds, and how well prices on that kind of market have matched outcomes in the past.

The fee examples below use Kalshi’s fee schedule effective July 7, 2026 and Polymarket’s fee documentation as of November 10, 2026.

Why does the price equal a probability?

Every contract settles at $1 or $0. If you think an event has a 70% chance and the contract costs 62¢, you expect it to be worth 70¢ on average, so buying looks cheap to you. Traders who think the chance is lower sell. The price settles where buyers and sellers balance. Polymarket’s documentation states it directly: a share at $0.25 implies a 25% chance, $0.50 a 50% chance, $0.75 a 75% chance.

YES and NO are two sides of one contract and sum to $1. Kalshi’s API docs spell out the mirror: a YES bid at 60¢ is the same as a NO ask at 40¢, and a NO bid at 30¢ is the same as a YES ask at 70¢.

Which number should you read: bid, ask or the headline price?

The headline price is a summary. On Polymarket, the displayed price is the midpoint of the best bid and best ask, and if the spread is wider than 10¢ the site shows the last traded price instead. Polymarket’s own example: best bid 34¢, best ask 40¢, displayed price 37¢. A buyer pays 40¢ and a seller receives 34¢. Nobody trades at 37¢ in that book.

Read it this way:

• The midpoint is the best single estimate of what the market thinks.
• The ask is your entry price if you buy now.
• The bid is your exit price if you sell now.
• The spread (ask minus bid) is a cost you pay on a round trip. Polymarket’s docs note that tighter spreads mean more liquid markets.

How do liquidity and order book depth change your price?

The best ask only covers the contracts offered at that price. A larger order walks up the book. Kalshi’s documentation includes a sample order book that shows this. Its best YES bid is 42¢ for 13 contracts. Its NO bids, converted to YES asks, are 44¢ for 17 contracts, then 55¢ for 20, then 56¢ for 29.

Worked example: buying 30 YES contracts in Kalshi’s sample book

1. The first 17 contracts fill at the best ask, 44¢: 17 x $0.44 = $7.48.
2. The next 13 fill at the next ask, 55¢: 13 x $0.55 = $7.15.
3. Total cost before fees: $14.63, or an average of 48.8¢ per contract.

The headline spread was 2¢ (42¢ bid, 44¢ ask), but a 30-contract order paid an average of 4.8¢ more than the best ask. This is why volume and depth matter: a price set by a few small trades says less than one backed by many contracts resting near it. Polymarket’s docs say its order book has no trading size limits but that large orders “may move the price significantly,” and they advise checking depth before trading in size.

How do fees shift the break-even probability?

Both exchanges charge takers (orders that fill immediately against resting orders) a fee shaped like P x (1 - P), which peaks at 50¢ and shrinks toward 1¢ and 99¢.

• Kalshi, general markets: fee = round up(M x 0.07 x C x P x (1 - P)), where C is contracts, P is price in dollars, and M is a multiplier that defaults to 1. The fee is rounded up to the nearest centicent ($0.0001). Maker fees use 0.0175 with a default multiplier of 0, so resting orders in most markets pay nothing. There is no settlement fee. (Kalshi fee schedule, effective July 7, 2026.)
• Polymarket: fee = C x feeRate x p x (1 - p), charged to takers only. The taker rate is 0.05 for sports, economics, culture and weather; 0.04 for politics, finance, tech and mentions; 0.07 for crypto; and 0 for geopolitics. (Polymarket fees.) These docs describe Polymarket’s international exchange. CryptoSlate reports the US app’s taker fee as 0.0695 x C x P x (1 - P).

Break-even probability = (price + fee per contract) / $1.

Worked example 1: 100 YES contracts at 62¢ on Kalshi, as a taker

1. Fee: 0.07 x 100 x 0.62 x 0.38 = $1.6492.
2. Total cost: $62.00 + $1.6492 = $63.6492.
3. Break-even: the contract must win 63.65% of the time. The fee moved the bar up 1.65 points.
4. If YES wins, you receive $100.00, a profit of $36.35 (57.1% on $63.65).
5. If NO wins, you lose $63.65.

Had you posted a resting order at 62¢ in a market with no maker fee and been filled, the cost would be $62.00 and the break-even 62%.

Worked example 2: the same trade on a Polymarket sports market

1. Fee: 100 x 0.05 x 0.62 x 0.38 = $1.178. Polymarket’s own table for 100 sports shares lists $1.20 at 60¢ and $1.14 at 65¢, which brackets this figure.
2. Total cost: $63.178.
3. Break-even: 63.18%.
4. On a geopolitics market the rate is 0, so the break-even stays at 62%.

Why fees matter most on cheap contracts. The dollar fee is largest at 50¢, but as a share of what you pay it is largest at the extremes. Using Kalshi’s taker formula on 100 contracts:

Kalshi’s fee schedule page lists the same $0.07 to $1.75 range for 100 contracts in most markets.

Why is a 62% price not a forecast that something will happen?

A 62% price says the event fails about 38 times in 100 comparable cases. When a 62¢ contract loses, that alone does not show the market was wrong. You can only judge prices across many contracts, by checking whether events priced near 62% happened about 62% of the time. That property is called calibration.

How well calibrated are prediction market prices?

Two studies give a sense of the size of the errors.

• Kalshi, 300,000+ prices. Bürgi, Deng and Whelan (University College Dublin, January 2026) studied Kalshi trades from 2021 through April 2025. Prices were informative and grew more accurate as markets neared close, but showed a favorite-longshot bias: contracts costing under 10¢ lost over 60% of the money put into them, while contracts above 50¢ earned small positive returns. Their illustration: a 5¢ contract that wins 3% of the time returns minus 40% before fees, while a 95¢ contract that wins 98% of the time returns 3.1%. Makers averaged minus 9.64% and takers minus 31.46%.
• Kalshi and Polymarket, 292 million trades. An arXiv paper by Nam Anh Le (February 2026) analyzed 327,000 binary contracts and found calibration errors differ by topic and time horizon. The most persistent was underconfidence in political markets, where prices are pulled toward 50%, on both exchanges.

Our read: treat a price as a good starting estimate, and apply the most caution to cheap longshots.

How to read a prediction market price in five steps

1. Note the midpoint as the market’s estimate, and check the venue and time, since prices move.
2. Find the ask (if buying) or the bid (if selling). That is your real price.
3. Check depth at that price against the size you have in mind.
4. Add the fee for that venue and category to get your break-even probability.
5. Compare your break-even to your own estimate, keeping in mind that longshots have historically won less often than their prices imply.

This guide explains how prices work. It is not a recommendation to trade.