The loophole and the lesson

Around 2010, the betting exchange was the future of betting. Betfair let people bet against each other instead of against a bookmaker, and lay a horse to lose as easily as back it to win. When it floated in London in October 2010 it was valued at £1.4bn. Its advertising promised customers they could cut out the middle man, until the advertising watchdog banned the line in 2011 because Betfair takes a commission on winnings.¹

Sixteen years on, I'm hearing the same things about prediction markets. Kalshi and Polymarket let people trade on whether something will happen, sport above all. Kalshi's trading in October was more than 20 times what it was a year earlier,² and at the end of the month it was valued at $40bn.³

I think the exchange story tells us how this one ends. The exchanges changed betting for good, yet they never won over the mass market. Prediction markets will change betting too, and my bet is that they end up as bookmakers themselves. But they have also shown the industry what an easy bet looks like to a fan, and that lesson will outlast the legal fight.

The last revolution

Betfair was founded in 2000 on a clever idea. Customers set their own prices and matched each other's bets. It took a commission on winnings, and the prices were better: Betfair claims they average about 20% better than a bookmaker's.¹

Yet plenty of customers kept going to bookmakers. In February 2013 Betfair launched a fixed-odds betting site of its own to compete with them, and its chief executive explained why. The exchange, he told investors, "does not fully address the mass market, due to perceived and real complexity". The new site gave newcomers something simpler and more familiar. It covered markets where the exchange was thin, especially in-play, and won back bets its own customers had been placing with rivals. It helped Betfair sign up 65% more new UK customers that year.⁴ In 2016 Betfair merged with Paddy Power, a bookmaker, in the deal that created what is now Flutter.¹

To be fair, Betfair Exchange is a success. It is licensed, and it handles about 90% of the money traded on betting exchanges.⁵ But by the end of 2024, the average amount traded in its win markets on British racing had fallen every year since 2016.⁶ The exchange found its customers: professionals, traders and serious punters. The mass market chose the easy bet.

Who was on the other side

On average, the exchange gave good value. A study published this year by Karl Whelan of University College Dublin looked at more than 200,000 football matches on Betfair. Before kick-off, people taking the prices on offer lost 2.5% on average, and the prices were accurate forecasts of results.⁵

The average hides who was winning. The people posting prices did better than the people taking them. Late in matches, takers lost about 70% of what they staked on the least likely outcomes. Much of the trading is done by bots through Betfair's software interface.⁵ Some firms had agents courtside sending scores back before other punters could see them. One such agent was arrested at the 2014 Australian Open, though the charges were later dropped.¹

Betfair saw the imbalance itself. In 2008 it began charging its most successful customers extra, fewer than 0.5% of them.¹ In 2025 that became an expert fee of up to 40% of a year's profits.⁶

That is what I mean when I say casual customers were cannon fodder for the whales. Nobody needed to cheat them. They were playing a traders' game against traders, and the traders were faster and better informed. Betfair also works to keep its markets clean, sharing information with about 30 sports bodies.¹ But an exchange is still built for people who trade, and most people who like a bet on the match don't want to.

What the exchange left behind

The exchanges left a lasting mark on the bookmakers they were meant to replace. Punters could see what a sharp price looked like, and bookmakers had to live with the comparison. Betfair counts in-play betting and cash out among the firsts it brought to market.⁷

Cash out is the clearest example. On an exchange you could always lay your bet back to lock in a profit or cut a loss. Cash out turned that trader's move into a single button. Betfair described it as a way of making a complex idea simple, and 38% of its football customers used it in the year to April 2013.⁴ By 2015 Sky Bet, William Hill, bet365 and BetVictor all offered it.⁸ The exchanges improved the whole industry, and the improvements that spread were the simple ones.

The new revolution

A prediction market sells contracts that pay $1 if something happens. The price runs from 1 to 99 cents, so it reads like a chance out of 100. Kalshi has been registered with the Commodity Futures Trading Commission (CFTC), America's derivatives regulator, since 2020, and Polymarket's American arm since July 2025.⁹

They list contracts on everything from politics to economic data, but sport and parlays, which British readers know as accumulators, make up the vast majority of trading.¹⁰

The leagues have come on board. The NHL was first, in October 2025, with deals for both Kalshi and Polymarket.¹¹ In March, Major League Baseball made Polymarket its official prediction market exchange and signed an integrity agreement with the CFTC, the first between the regulator and a major league.¹² On 26 October the NBA named Kalshi and Polymarket its preferred exchanges. It limits which markets they can offer, bans league and team staff from trading and has signed an agreement of its own with the CFTC.¹³ The NFL is the big holdout. It has urged the Supreme Court to let states regulate these markets.^{13, 14}

The money has followed. At the end of October Kalshi raised $1bn at a valuation of $40bn, up from $22bn in May.³ A month earlier, Flutter and DraftKings were worth less than $23bn between them on the stock market.¹⁵

What they got right

Anomalies are worth studying because they point at a gap. This one points at a big one. Each market asks one plain question, such as whether a team will win. The price reads like a chance out of 100, with no odds formats to learn, which solves the algebra in French problem in one line. And Kalshi and Polymarket don't push a casino at you the moment you sign up. Fans noticed: during the World Cup, Kalshi had more daily American app users than DraftKings or FanDuel.¹⁶

Not all of it is healthy. Prediction markets take customers from 18, an age licensed bookmakers can't serve in most states.¹⁰ And a price that reads like a chance can flatter a long shot. Economists at University College Dublin who studied more than 300,000 Kalshi contracts, across many types of market, found that cheap contracts won less often than their prices suggested. People who took the prices on offer lost almost 32% on average, against about 10% for those who posted them, and the heaviest losses were on long shots.⁹ Simple needs protection built in, or it just makes losing easier.

Same game, two rulebooks

Part of their advantage is the product. The rest is the rulebook. Kalshi lists a contract by certifying to the CFTC that it complies with federal law, and it needs no prior approval.¹⁰ It argues that this puts its sports contracts under federal law alone, beyond the reach of state gambling law. The CFTC agrees. Its chairman, Michael Selig, says Congress gave the agency sole responsibility for these markets. The CFTC has sued 10 states that tried to rein them in, most recently Utah.¹⁷ In October it took bets placed with licensed casinos and bookmakers out of the federal definition of a swap, and proposed writing event contracts, sport included, into it.^{18, 19}

The states say a bet on a game is a bet on a game, and the courts are split. The Third Circuit appeals court sided with Kalshi in April. The Ninth Circuit in August and the Sixth Circuit in September went the other way.²⁰ New Jersey has asked the Supreme Court to settle it. Its backers include 39 states and Washington DC, the American Gaming Association²¹ and the NFL,¹³ along with former senator Chris Dodd and former CFTC chairman Gary Gensler, who both helped shape the 2010 law at the heart of the case.¹⁴ Kalshi's response is due on 7 December, and the Court has not yet said whether it will hear the case.²¹

Two other moves show where the pressure is coming from. Utah has brought criminal charges, which Kalshi calls a "crazy overreach".²² And Senator Ted Cruz, who chairs the Senate Commerce Committee, has said he plans a bill after the midterms that would need a state or tribe to approve sports contracts before they can be offered.²³

Side by side, the two rulebooks look like this.¹⁰

Kalshi's answer is that it is a federally regulated exchange that "answers to federal law, and no single state should have a veto over it".¹⁷ That is a serious legal argument, and the courts will settle it. My view is about the business: whatever the courts decide, the federal route works like a loophole. It lets one group of companies offer sports contracts without the state licence, the state tax or the state rules that licensed bookmakers accept. Those bookmakers paid dearly to get in: New York charged each of its online betting platforms a $25m licence fee.²⁴ I find it inconceivable that one side keeps a free pass while the other pays. Congress's own researchers put it more carefully, saying sports event contracts "arguably create a form of regulatory arbitrage".¹⁰

Apples and oranges

In July, Bloomberg reported an estimate from H2 Gambling Capital that prediction markets took about 27% of legal American sports betting during the World Cup, up from about 9% in January. As the reports said at the time, the comparison doesn't use identical measures.¹⁶ The figure has been widely quoted since, including by Congress's research service.¹⁰ I think it compares apples with oranges, for two reasons.

The first is how the volume is counted. Kalshi counts every contract at its $1 face value, whatever the customer paid, so a 10-cent contract adds $1 to volume. In the 30 days to 30 October, Kalshi traded $93.0bn of contracts at face value, and its customers paid $21.0bn for them.² Even the lower figure counts a position again each time it is resold, while a bookmaker's handle counts the stake once.^{2, 16}

The second is where the money comes from. Kalshi doesn't publish state figures, so the best we have is an estimate. In April, Eilers & Krejcik Gaming estimated that 43% of sports contract volume came from California and Texas, home to about one American in five, and 69% from the 19 states without legal online sports betting. In mature betting states, it put Kalshi's share of handle at about 2%.^{25, 26} Those figures are modelled estimates, and they predate the World Cup and the NFL season. But they fit what Flutter told investors last November: prediction markets were having a "negligible impact" in states where FanDuel already took bets.²⁷

If the estimates are right, most of this volume comes from places with no licensed bookmaker to take the bet. California and Texas collect no betting tax on it.²⁶ New York, with about 20 million people, collected $1.32bn in sports betting tax in 2025.²⁵

Outside America

Elsewhere, regulators have mostly decided what prediction markets are. Britain's Gambling Commission says their core is "akin to" a betting exchange, so serving British customers would need the betting intermediary licence that covers exchanges, which have existed in Britain since 2000. Operating without the right licence is a criminal offence.²⁸ Most of the action abroad has been against Polymarket's international site.

Polymarket says it will challenge the French decision in court.³⁰ Still, the pattern is clear. Where a country already licenses betting, it treats a prediction market as betting.

My bet

Exchanges changed the industry and then became part of it. Betfair launched a fixed-odds betting site 12 years after it opened and merged with a bookmaker three years later. I think prediction markets will go the same way, only faster.

You can see it happening already. On Kalshi, combos, which are accumulators by another name, took a third of the money customers paid in the 30 days to 30 October.² The leagues sign prediction markets up the way they sign bookmakers, with official data, logos and integrity rules.¹³ In August the CFTC told the exchanges that showing "bookmaker-style odds" was likely to mislead people, which tells you how close the two products have become.¹⁰ And the bookmakers have met them halfway, with FanDuel and DraftKings running prediction markets of their own.¹⁰

My bet is that prediction markets end up as licensed bookmakers, state by state. The irony is that they can now afford it. Kalshi is valued at $40bn,³ and PitchBook estimates it will make about $1.2bn of adjusted earnings this year, before share-based pay, which would be more than DraftKings' own guidance.¹⁵ Licence fees that once kept newcomers out are small change to a business like that. Cruz's bill would push them that way faster.

There is another way this could go. The CFTC has proposed defining event contracts, sports included, as swaps in its own rules.¹⁹ If that becomes final and the Supreme Court sides with Kalshi, the federal route could hold for years. I still wouldn't bet on it.

The lesson

The lesson is the one the exchanges taught. Most people don't want to become traders, and they don't want to be treated like gamblers from the first tap. They want an easy, fun bet on the sport they love, in a place they trust.

Betting inside the sport can give fans all of that. The fan picks what they think will happen in the app they already use and sees, in plain words, what a small stake costs and what it could bring back. A licensed bookmaker takes the bet, taxed and regulated like any other. The bet stays a bet on the sport, and nobody needs a loophole.

The courts will decide where prediction markets can operate. Fans will decide where betting belongs.

About the figures: figures are as published by each source on the date shown. Prediction market volumes count contracts at their $1 face value unless stated otherwise. The state shares are modelled estimates. 18+. Free support is available 24 hours a day from the National Gambling Helpline on 0808 8020 133.

Keith O'Loughlin is CEO of Funteron and former CEO of OpenBet. He chairs Christchurch Casino in New Zealand.