How to Bet on a Prediction Market: Prices, Fees and Access

A contract priced at 40 cents pays 1 dollar if it wins: decimal odds of 2.50, before the fee.

A referee seen from far behind on the centre circle of a hushed floodlit stadium, the ball on the spot and the players frozen far away

A prediction market sells contracts on outcomes, and each price reads as a probability.

A sports contract is a Yes or No share that pays 1 dollar if the result goes its way and 0 if it does not, so a price of 0.40 reads as a 40 percent chance. You pay a fee set by a formula, plus the gap between the best price to buy and the best price to sell. Whether you may trade is decided twice: by the rules of the place and by the law of your country. A price is an estimate of a probability and promises nothing about the result. Adults only, 18+.

How does a price become odds?

A prediction market looks like sports betting from the outside. Instead of a bookmaker quoting odds, people trade contracts that pay 1 dollar if the outcome named on them happens and nothing if it does not, and the price is what the crowd pays for that chance. If order books and back and lay are new to you, how a betting exchange works covers them. The example here is one contract at 40 cents (Polymarket calls a contract a share).

  1. Read the price as a probability

    A contract at 0.40 costs 40 cents, and the market puts the chance of the outcome at 40 percent.

  2. Turn it into fair odds

    Decimal odds are 1 divided by the price: 1 / 0.40 = 2.50, before any fee.

  3. Know what the contract pays

    Polymarket's documentation says winning shares redeem for 1 dollar each and losing shares become worthless.

  4. Add the taker fee

    Polymarket publishes the formula, shares × rate × price × (1 - price), with a sports rate of 0.05. At 0.40 the fee per contract is 0.05 × 0.40 × 0.60 = 0.012, or 1.2 cents, and Polymarket's own table gives 1.20 dollars for 100 shares at that price. A taker (an order that matches one already waiting in the book) pays it; according to Polymarket, makers never do.

  5. Read the effective odds

    With the fee added to what you pay to enter, the cost is 0.40 + 0.012 = 0.412. A win pays 1 dollar, a net gain of 1 - 0.412 = 0.588; a loss costs 0.412. The effective odds are 1 / 0.412 = 2.43, and you break even only if the outcome happens more than 41.2 percent of the time. Polymarket does not say whether the fee comes out of your shares or your balance, so 2.43 is a calculation under a stated assumption.

Kalshi publishes its own fee schedule, which could not be read at the source.

What do fees do to the odds?

In dollars the fee peaks at 0.50 and is symmetric around it, per Polymarket's documentation. As a share of the price it does the opposite: calculated for a taker from the published formula and the 0.05 sports rate, it is 4.5 percent at 0.10, 2.5 percent at 0.50 and 0.5 percent at 0.90. A longshot at 0.10 with fair odds of 10.00 pays about 9.57 if the fee is added to the entry cost.

Polymarket taker fee as a share of the price, sports rate 0.05

Published formula, calculated
View the calculation table and its assumptions

Assumptions.

  • One taker order, the sports rate of 0.05 and the step 4 formula at every price; the pages do not separate leagues.
  • The fee is paid once at matching and added to the cost of entry, whatever the result. Polymarket says only that fees are calculated in USDC and applied at match time, so this is an assumption.
  • The contract pays 1 dollar or 0, with no 50-50 result and no postponement. Left out: the spread, liquidity, network fees and a taker sale before the match ends.
  • Fees are rounded to 5 decimal places (the smallest is 0.00001 USDC), which changes nothing at two decimals; 1.106 at 0.90 keeps it apart from the fair 1.11.
  • Sensitivity at 0.40: 2.425 if the fee came out of the shares you receive, 2.470 if it came out of the winnings of a winning contract only. Polymarket's pages confirm none of the three.
Polymarket taker fee and effective odds at seven prices, sports rate 0.05
PriceImplied probabilityFair oddsFee per contractFee as share of priceNet gain if it winsLoss if it losesBreak-even probabilityEffective odds
0.1010%10.000.0045004.50%0.8955000.10450010.45%9.57
0.2525%4.000.0093753.75%0.7406250.25937525.94%3.86
0.4040%2.500.0120003.00%0.5880000.41200041.20%2.43
0.5050%2.000.0125002.50%0.4875000.51250051.25%1.95
0.6262%1.610.0117801.90%0.3682200.63178063.18%1.58
0.7575%1.330.0093751.25%0.2406250.75937575.94%1.32
0.9090%1.110.0045000.50%0.0955000.90450090.45%1.106
Calculated from Polymarket's published fee formula and sports rate, under stated assumptions; the bars are not market data. They leave out the spread and liquidity, and they do not depend on how the fee is collected.

Set it against a bookmaker by the probability you need to break even. A bet at -110 on both sides of a match is decimal odds of 1.909 each way and needs a win 52.38 percent of the time; the two sides together imply 104.76 percent, a margin of 4.76 percent. A Polymarket taker at 0.50 needs 51.25 percent on the fee formula alone, the price plus a fee of 0.0125.

The gap of about 1.1 points does not show which route costs less, and the formula cannot settle it. Three things are missing: the spread and the liquidity; the zero rate for makers, which applies only to an order waiting in the book that may never be filled; and the true probability, which nobody knows. The -110 above is only an example, since real prices vary by bookmaker and by day.

Kalshi's fee schedule could not be read at the source, so no Kalshi figure appears here.

How is a prediction market different from a sportsbook and from an exchange?

A sportsbook quotes prices with its margin built in, whereas each of the two places says your counterparty is another user, so what you pay is a fee formula plus the gap between the best buying price and the best selling price. Kalshi's API documentation says a bid to buy Yes at 0.40 equals an offer to sell No at 0.60, so the two sides are one market.

Against a classic exchange the mechanics are close; the product and its settlement rules differ. The guide to the best betting exchange sites compares classic exchanges, and whether you can use Betfair from your country covers that side.

Kalshi and Polymarket side by side

The cards run in alphabetical order, and no ranking is implied.

  • Kalshi

    Event contracts

    Kalshi describes itself in its Member Agreement as a CFTC designated contract market in the United States, and states that your counterparty is always another member.

    Fees. Kalshi publishes a fee schedule; details not confirmed here. Its help centre says the fee is charged on the expected earnings of a contract and that some markets have different fees.

    Access. The Member Agreement prohibits trading event contracts for anyone domiciled in, organised in or located in a list of named jurisdictions, currently 55, or in a place under US sanctions; the list includes the United Kingdom, Canada, Australia, France, Ireland and India. It says these restrictions apply to trading and do not in themselves prohibit membership or non-trading access, where applicable law and Kalshi's policy allow it; Kalshi also keeps sole discretion to grant, deny, condition, suspend or revoke access.

    What it documents

    • A rulebook and contract terms, with a settlement value of 1 dollar for its sports game contracts
    • A named list of restricted jurisdictions

    Open points

    • Fee schedule not confirmed
    • A position accountability level of 25,000 dollars per strike for each member in its American football, basketball and soccer game contracts: above it Kalshi may ask for information and require a reduction
  • Polymarket

    Site and US app

    Polymarket says the counterparty to your trade is another user.

    Fees. Its documentation and help centre publish the taker formula and a sports rate of 0.05. Only takers pay, and there is no Polymarket fee to deposit or withdraw USDC.

    Access. Polymarket lists restricted locations in two places, its help centre and its documentation. The two name largely the same countries but do not always describe them the same way: the United Kingdom is fully restricted on one page and close-only on the other, and close-only means you can close existing positions but not open new ones. The help page states a count of fully restricted countries that does not match the countries it lists.

    What it documents

    • A fee formula and a sports rate, so a taker fee can be calculated before an order
    • Resolution rules written market by market

    Open points

    • The fee pages do not say whether the fee comes out of your shares or your balance, or which product they cover
    • Polymarket US, its separate app for US residents, publishes a different fee schedule

Smarkets, an exchange licensed in the UK and Malta among other places, brands its site "Smarkets Predictions" and says on its investor page that its US application to run prediction markets is in review.

Each place's own page covers identity checks and funding: how to bet on Kalshi and whether you can use Polymarket.

Regulated in the United States is not the same as allowed for you

Being regulated in the United States does not make a platform lawful for you, whatever the platform's own list says: its rules and the law where you live are separate questions, and both need checking.

  • Kalshi: it bars trading event contracts from named jurisdictions, and by opening an account a member states that they are not in a place where doing so is prohibited by law or by Kalshi's policy.
  • Polymarket: its pages sort places into fully restricted and close-only ones and do not always agree on a country. It says restrictions follow where you are physically, and prohibits using a VPN or similar tool to bypass them.
  • Kalshi says it must verify who you are. For Polymarket, no readable page says what applies, and silence is no proof that no check exists.
  • The place decides, at the moment you act, and a broker lifts no restriction that applies to you.

Legality varies by country, so read the rules that apply to you. You must be 18 or older.

What can go wrong?

You can lose the whole amount you pay for a contract: one bought at 0.40 that ends at 0 has cost 0.40, or 0.412 with the fee, under the assumption above.

On liquidity, Polymarket says its order book has no trading size limit by design and that it cannot promise a desired amount will trade without moving the price, or at all. The spread (the gap between the best price to buy and the best price to sell) sits outside the fee formula, and the documentation of neither place gives a real sports spread.

The price a screen shows may be the last trade or a midpoint, and the best price you can get right now may differ from either. When a team sheet is published, a contract can move, and what you noted at breakfast may describe a match that has since changed.

Settlement rules differ. In its American football, basketball and soccer game contracts, Kalshi says a postponed game played within 48 hours of its original start settles on its result, while a game cancelled or not played within that window settles at the last fair price set by Kalshi at its sole discretion.

Polymarket writes its rules market by market. In some sports and esports markets, a cancelled match, a tie or a match not played by a set deadline resolves 50-50, which pays 0.50 dollars per share. A share bought at 0.62 that settles 50-50 returns 0.50, a loss of 0.12; bought at 0.25, the same result is a gain of 0.25. Each rule belongs to the contract or market it comes from: the same sport can be settled differently on different platforms, so read the rules of the one you trade.

Polymarket's documentation says markets are resolved through the UMA Optimistic Oracle, in about two hours when nobody disputes the result and four to six days when someone does. Kalshi says markets typically settle shortly after the event ends, though the data source and manual review can change the timing.

Which brokers list Kalshi or Polymarket?

A broker listing is a line in the broker's own list: no place is shown here confirming it. Whoever opens a broker account becomes that broker's customer, with its identity checks, excluded countries, fees and withdrawal rules on top; the guide to betting brokers covers that layer.

The access lists below show BetInAsia's BLACK platform for both places, and nothing more can be said about it here; AsianConnect is in neither list. MadMarket's page says its Edge platform lists Kalshi and Polymarket among its exchanges, without saying how the two places' own rules apply to its customers. Sportmarket's page says it lists Polymarket among the prediction markets on its platform, that settlement follows each platform's own rules and that a cancelled match may not return your full stake; Kalshi is not on that list.

The broker layer adds risks of its own. Public reviews of BetInAsia describe accounts left in review for weeks and withdrawals blocked, and a published reading of its terms describes withdrawal caps left to its discretion. MadMarket's own pages state its withdrawal fee in three ways that cannot all hold, and no independent test of it is on record. Sportmarket's own pages make a withdrawal free of fees only after you have wagered five times each deposit, ask for an identity check before any withdrawal and exclude several countries outright. Complaints are louder than praise online, and many published tests come from sites with a commercial interest in the operator.

Should you trade at all?

  1. Does the place let you in? Read its current rules for your country on its own site, then the law where you live.
  2. Can you accept the counterparty and the loss? A contract can go to zero, and settlement can follow rules you did not expect.
  3. Does your sport exist, and how deep is the market? Read the best prices and the sizes waiting there before you stake anything.

The edge cases catch readers out: a place open to your country while your law says otherwise, a match postponed after you paid, a close-only status that lets you exit but not add, or a contract your regulator treats differently from a bet.

A neighbouring product has its own guide: fantasy and pick'em betting.