The venue's real risk
An exchange carries no liability on outcomes, so its risk is not that the favourite wins. Its risk is that a contract resolves wrongly, that a price is manipulated, or that someone trades on information the market did not have. Each of those is a failure a sportsbook also faces, but the exchange faces them differently, because it is a marketplace whose only asset is trust in the marketplace.
This lesson takes the three in turn and then adds the one that is new: a prediction market can be a market on something the traders themselves can influence.
Resolution
Every contract carries a rule stating what source will be consulted and what counts as yes. The quality of the rule is the quality of the product.
Sport is easy. The final score from the league's official record is unambiguous, and the only questions are the ones bookmakers already handle: postponements, abandoned games, overturned results. Exchange contract terms deal with these cases much as a bookmaker's settlement rules do, and a customer who has read a bookmaker's terms will recognise them: Kalshi's rulebook, for example, lets it move a contract's expiration when the event is rescheduled or cancelled or the source data is delayed.
Economics is medium. A contract on an inflation print resolves on a published statistic, but the statistic gets revised, the release can be delayed and the contract must say which version counts. Exchanges have learned to specify the exact series, the source and the time of resolution, and to rule out later revisions: Kalshi's monthly inflation contract, for example, resolves on the seasonally adjusted CPI-U change published by the Bureau of Labor Statistics at an expiration time of 8.31am Eastern, and ignores revisions made after expiration.
Politics and world events are hard. "Will X be confirmed by the Senate" is fine. "Will there be a ceasefire" is not, because the world does not always produce a clean fact. The regulated exchanges resolve against each contract's written terms and a named source; when an outcome is in doubt, Kalshi's rulebook hands it to an Outcome Review Committee of three members, at least two of them public directors, which decides within 24 hours and whose determination is final. The crypto-native venues resolve through an oracle. Polymarket's markets are resolved by the UMA Optimistic Oracle: anyone can propose an outcome by posting a bond, the proposal can be disputed during a short challenge period, and only a disputed proposal goes to a vote in which holders who stake the UMA token vote in secret, and those who vote against the majority or fail to vote are slashed, with the slashed amount redistributed to the majority voters. That design can be exploited. A well-capitalised participant who holds a large position in a contract and a large share of the voting token can, in a contested case, vote for the outcome that pays them. In March 2025 a $7 million Polymarket market on whether Ukraine would agree a minerals deal with the United States before April resolved yes although no agreement had been reached, and many users suspected manipulation by a large UMA holder; Polymarket accepted that the market had resolved too soon but refused refunds, saying it was not a market failure.
For a gambling professional the lesson is that resolution is a governance question, not a data question. A sportsbook's settlement is a decision the operator takes and stands behind; an exchange's resolution is a decision that has to be seen to be independent of anyone who traded. The contract drafting, the source hierarchy and the dispute process are where an exchange either earns trust or loses it, one contested market at a time.
Price manipulation
Manipulating a price on an exchange means trading in a way designed to move the price away from where information would put it. The techniques are the ones securities regulators have chased for nearly a century: the US Securities Exchange Act of 1934 already made it unlawful to trade for the purpose of creating a false or misleading appearance of active trading.
Painting the tape. Placing trades to create the impression of activity or a trend, often by trading with oneself through multiple accounts. On a thin market a modest sum can move a contract sharply, since, as a 2004 survey of prediction markets put it, the extent to which these markets can be manipulated depends at least in part on how thin they are, and the moved price then gets quoted by the press as "the market's view".
Spoofing. Posting large orders with no intention of filling them, to induce others to trade, then cancelling. The Commodity Exchange Act makes spoofing unlawful on a registered exchange, and designated contract markets have a duty under the Act's core principles to maintain audit trails, conduct surveillance and enforce rules against prohibited practices; crypto-native venues have less visibility.
Motivated manipulation. The distinctive case. On a market about an election, a policy decision or a corporate announcement, a participant may want the price to show a particular probability, because the price itself is being reported and influences opinion. The 2024 presidential election produced the best-known test case: four Polymarket accounts holding more than $28 million of bets on Donald Trump, which Polymarket confirmed belonged to a single French trader, drew scrutiny as the market's price moved away from polls that showed a close race. Polymarket said an investigation with outside experts had found nothing to suggest the trader manipulated or tried to manipulate the market, but whether the positions were a view or a message was debated in public. The trader was, in the end, right, which illustrates the confirmation trap covered in the Sports Integrity course: a winning position is not evidence of manipulation and a losing one is not evidence of innocence. The research record is mildly reassuring: in the markets economists had studied by 2004, known manipulation attempts had little discernible effect on prices beyond a short transition phase.
A regulated exchange's answer is the toolkit of any exchange: surveillance, position limits and position accountability levels set in each contract's terms, and the power to cancel trades and suspend accounts. The strength of those tools depends on the venue being able to identify its customers, which is the point at which the regulated and the crypto-native models diverge most.
Insider trading
A prediction market on a corporate announcement, a court ruling or a government decision is a market on which some people know the answer before others. Trading on that knowledge is the central integrity risk of the product, and it is one that the regulated venues, the crypto-native venues and the gambling regulators all agree is a problem, even if they disagree about whose problem it is.
Several public cases established the pattern in 2025 and 2026. Accounts took large one-sided positions shortly before an announcement, on an outcome that then occurred, and were traced to people who were in a position to know. In April 2026 the CFTC brought its first insider trading case involving event contracts, against a US Army service member alleged to have used classified information about the operation to capture Nicolás Maduro to make more than $404,000 on Polymarket, alongside a criminal case in New York; in May 2026 it charged a Google software engineer alleged to have made about $1.2 million trading Polymarket contracts on Google's Year in Search list. Others ended in exchange discipline: Kalshi fined and suspended a political candidate who traded on his own candidacy and a YouTube channel editor who traded on the channel's videos. Still others, such as large purchases of contracts on US strikes on Iran shortly before they happened in February 2026, drew concern from some members of Congress. The signals are the ones that the Sports Integrity course teaches for suspicious betting: new accounts, concentrated timing, positions out of proportion to the account's history, and an outcome that confirms.
Two features make the risk worse than in sports betting. First, the universe of events on which someone has inside knowledge is far larger than the universe of games. Every corporate action, regulatory decision, award, product launch and political appointment is a potential market, and every one has insiders. Second, the regulated venues' legal position, that their contracts are federal derivatives, brings insider trading on them within federal commodities law: trading on information misappropriated in breach of a duty can be pursued under CFTC Rule 180.1, with civil or criminal liability, although the CFTC has said the rule does not forbid all trading on lawfully obtained nonpublic information, and Kalshi's own rules go further. That is a stronger deterrent than a gambling regulator's fine but also a reputational exposure the venue has to manage in public.
The sports overlap
A contract on a game is a bet on a game, whatever the courts eventually call it, and the integrity risks of sports betting arrive with it. A fixed match pays on an exchange as it pays at a bookmaker. A courtsider trades an in-play contract before the price updates. An athlete trading a contract on their own team is insider betting in the sport's rulebook, whatever it is in the exchange's.
The difference is institutional. Licensed sportsbooks in many regulated markets must report suspicious betting: in Britain betting licensees must report suspected offences to the Gambling Commission and give the relevant sport governing body information on suspected breaches of its betting rules. Many also belong to integrity associations that pool alerts, and they work with monitoring companies. Exchanges have been slower to join those structures, partly because they argue they are not sportsbooks, partly because the crypto-native venues identify their customers less thoroughly, and partly because the professional flow they welcome looks, to a monitoring model, exactly like the flow a bookmaker would flag. The largest exchanges have begun signing data and integrity agreements with leagues and monitoring firms: in October 2025 the NHL named Kalshi and Polymarket official prediction market partners with access to its official data, and in August 2026 Kalshi agreed to take part in Genius Sports' information-sharing processes and to restrict certain market types. A league that sells data to an exchange has an interest in seeing its markets policed. A professional in a trading or integrity role should expect to be handling exchange-sourced alerts within a few years, and should expect them to be harder to read.
Markets the traders can move
The genuinely novel risk is a market whose outcome a participant can influence. A contract on whether a bill passes, traded by a legislator's staff. A contract on whether a company announces layoffs, traded by the company's management. A contract on whether a protest reaches a certain size, traded by its organisers. Such markets have been controversial at least since 2003, when the Pentagon dropped a planned Policy Analysis Market after critics attacked it as "terrorism futures" that could let a terrorist profit from an attack, although Wolfers and Zitzewitz argued that this particular fear may have been misplaced.
The regulated exchanges answer with contract selection and trading bans. Kalshi's rulebook bars anyone who is a decision maker, or has any influence "no matter the scale and importance of the influence", over an event from trading contracts on it, and the Commodity Exchange Act lets the CFTC prohibit event contracts involving unlawful activity, terrorism, assassination, war, gaming or similar activity that it finds contrary to the public interest. Contract selection is imperfect: the YouTube case above involved a market whose outcome a small group of people knew in advance. The crypto-native venues have listed almost anything with demand and have relied on the oracle to sort out the consequences. Between those two positions is where a gambling regulator's instinct, that a market on an event one participant controls is an invitation to fix it, has the most to offer the debate.
What to take from this lesson
An exchange's integrity risks are resolution, manipulation, insider trading and manipulable events. The first is governance; the second is surveillance; the third is the largest and shares its signals with suspicious-betting detection; the fourth is a listing decision. Customer identification underlies all of them, and it is the dimension on which the regulated and crypto-native models differ most. The tools are the ones exchanges and sportsbooks already have; the work is applying them to a product that lists events nobody has monitored before.