The Media-Betting Complex

Introduction: The Same Day

On November 6, 2025, PENN Entertainment and ESPN announced that they had agreed to end the largest of the broadcaster-sportsbook deals in this chapter. The deal had been signed in August 2023 to run ten years. It lasted two. PENN would pay ESPN $38.1 million “in respect of all remaining fees.” ESPN would keep warrants on 7,957,210 PENN shares, and the unvested balance of its 31.8 million warrant shares was “forfeited and canceled for no consideration” (PENN Entertainment, “Form 8-K”; PENN Entertainment and ESPN, “Mutually Agree”). The warrants it kept carried strike prices of $26.08 and up (PENN Entertainment, “Form 8-K”).

The same day, ESPN announced its next partner. DraftKings would become the network’s exclusive official sportsbook and odds provider and would power a betting tab inside the ESPN app. The ESPN BET name would survive only as a content brand, anchored by a betting show (Manzo). The network that had set out to own a sportsbook ended the day renting its audience to one.

For five years the relationship between American sports media and gambling was told as a merger story. The broadcaster, the story went, had become the bookie: it bought equity in sportsbooks, lent them its name, and took a share of what its viewers lost. The fear was reasonable, and the deals behind it were real. Article 2.3 ended by turning to what happens when the people who control the delay also own the microphone. The industry tried that arrangement from both ends. Broadcasters bought into sportsbooks, and PENN, which ran one, bought microphones of its own in Barstool Sports and theScore. But the story had a second half that the fear did not anticipate. Nearly every one of those ownership deals failed. FOX Bet closed in 2023. NBC’s sportsbook partner was sold for less than NBC’s marketing contract with it had been worth. PENN bought Barstool Sports for about $551 million and sold it back to its founder for $1. ESPN Bet, the largest of them, was taking less than three percent of handle in the states reporting when it was shut down. The two companies that never belonged to a broadcaster, DraftKings and FanDuel, ended up with about 70 percent of it.

This article is about what happened after the bust, because the relationship between the screen and the sportsbook did not end when the equity did. It changed form twice. The broadcaster stopped trying to be the bookie and became the billboard. Then, since mid-2025, a third arrangement has appeared, in which news networks and newspapers display live prediction-market prices beside the news as though the price were part of the reporting. Each change has moved the conflict of interest somewhere harder to see.

The claim this article tests is the one the broadcast invites every viewer to hold: that the information on the screen, the insider’s scoop, the odds in the broadcast, the prediction-market ticker beside the news, is handing the bettor an edge.

The Buy: Five Years of Owning the Audience

The logic of the buy was easy to state, and for a while it sounded like the smartest idea in sports business. A sportsbook entering a newly legal state has to find customers and pay to acquire them. A sports network already has them, by the million, watching the very events the sportsbook prices. Put the two together and the network’s audience becomes the book’s customer list, and the book’s profits flow back to the network.

Fox moved first. In May 2019, a year after the Supreme Court struck down the federal ban, Fox Corporation paid $236 million for 4.99 percent of The Stars Group, with an option on up to half of its American business if licensed, and the two launched FOX Bet (Fletcher). In February 2020 CBS Sports named William Hill its official sportsbook and data provider, without disclosing the financial terms (Fischer, “CBS Sports Names William Hill”). That August NBCUniversal signed PointsBet to a deal that combined $393 million in committed marketing with a 4.9 percent interest in the company and options on more (“A Look Inside”). In November Sinclair Broadcast Group sold the naming rights to its regional sports networks to Bally’s for a reported $85 million over ten years and took warrants for just under 15 percent of Bally’s, with the right to buy about 15 percent more (Carp).

PENN Entertainment ran the same play from the opposite direction, buying media to feed a sportsbook. It paid about $163 million for 36 percent of Barstool Sports in 2020 and about $388 million for the rest in 2023 (PENN National Gaming, “Completes Acquisition of 36% Interest”; PENN Entertainment, “Completes Acquisition of Barstool Sports”). In October 2021 it closed the purchase of the Canadian sports-media company theScore for about $2.0 billion in cash and stock (PENN National Gaming, “Completes Acquisition of Score Media”).

The culmination came in August 2023, when ESPN licensed its name to PENN for ten years. PENN agreed to pay $1.5 billion in cash over the term and granted ESPN warrants for about 31.8 million PENN shares, valued at roughly $500 million, plus bonus warrants on up to 6.4 million more if the new book hit its market-share targets (PENN Entertainment, “Long-Term Exclusive Strategic Alliance”).

The warrants mattered more than the cash. A warrant is worth something only if the share price rises above its strike, and PENN’s share price depended in part on whether its sportsbook made money, which a sportsbook does when its customers lose. ESPN held a financial instrument that paid off when its audience lost bets, as Fox’s and NBCUniversal’s stakes had before it. That was the conflict the critics saw, and they were right to see it.

The Bust: How Every Broadcaster’s Book Lost

The ownership model did not work, and nearly every deal built on it failed.

ESPN Bet launched on November 14, 2023, and PENN reported more than 1 million first-time depositors since the launch (PENN Entertainment, “Fourth Quarter Results,” 2024). It never came near the share its backers had talked about. PENN projected a 20 percent share by 2027 (Glasspiegel); at the 2023 announcement its chief executive said both companies would be excited at around ten percent (Waters). When the partnership was ended, ESPN Bet’s share was 2.8 percent of handle and 2.6 percent of revenue year to date, by state revenue reports (Waters). PENN does not report ESPN Bet’s results on their own. Its Interactive segment, which ran ESPN Bet alongside theScore Bet in Canada, online casino and other digital operations, reported adjusted EBITDA losses of $499.5 million in 2024, the only full calendar year of the partnership, and $267.5 million in 2025 (PENN Entertainment, “Fourth Quarter Results,” 2025; PENN Entertainment, “Fourth Quarter Results,” 2026).

ESPN did not fail to deliver an audience. ESPN’s chairman said the network had driven more than 2.9 million new users to PENN (PENN Entertainment and ESPN, “Mutually Agree”). The audience arrived, and the book still finished with less than three percent of the money wagered in the states reporting. Whatever those users did next, most of their betting did not stay with the network’s book.

The other deals ended the same way, only sooner. Table 2.4.1 sets them side by side.

Table 2.4.1The broadcaster-sportsbook deals of 2019-2023 and how they ended

Deal	What the media company took	How it ended
Fox and The Stars Group (FOX Bet), 2019	4.99 percent stake for $236 million; option on up to half of the U.S. business	FOX Bet closed August 2023; Fox retains an unexercised option on 18.6 percent of FanDuel
CBS Sports and William Hill, 2020	Official sportsbook and data partner; financial terms not disclosed	William Hill bought by Caesars, April 2021
NBCUniversal and PointsBet, 2020	$393 million marketing commitment; 4.9 percent interest plus options	Spending cut, 2023; PointsBet U.S. sold to Fanatics for $225 million, 2024
Sinclair and Bally’s, 2020	Reported $85 million naming rights; warrants on just under 15 percent of Bally’s, with the right to buy about 15 percent more	Networks renamed FanDuel Sports Network, 2024; Operator announced a conditional wind-down, April 2026; operations wound down and streaming shut, September 2026
PENN and Barstool Sports, 2020-2023	About $551 million paid for 100 percent	Sold back to its founder for $1, August 2023
ESPN and PENN (ESPN Bet), 2023	$150 million a year in cash; warrants on about 31.8 million shares	Ended December 1, 2025; $38.1 million payment for remaining fees; unvested warrants forfeited

Source: Fletcher; Flutter Entertainment, “Close of FOX Bet”; Butler; Fischer, “CBS Sports Names William Hill”; Caesars Entertainment; “A Look Inside”; Fischer, “NBCU and PointsBet”; Fanatics Betting and Gaming; Carp; Winslow; Lewis; Katz; PENN National Gaming; PENN Entertainment; Spangler; PENN Entertainment and ESPN, “Mutually Agree.”

FOX Bet, one of the casualties recorded in article 1.3, shut down in stages between July 31 and August 31, 2023 (Flutter Entertainment, “Close of FOX Bet”). PointsBet renegotiated NBC’s deal in January 2023, cutting its annual marketing commitment from about $90 million to about $58 million and shifting some of the obligation to regional NBCU assets (Fischer, “NBCU and PointsBet”), and in April 2024 PointsBet’s American business was sold to Fanatics for $225 million (Fanatics Betting and Gaming). CBS’s partner, William Hill, was bought by Caesars in April 2021 (Caesars Entertainment). Sinclair’s regional networks passed through the bankruptcy of Diamond Sports Group, became FanDuel Sports Network in October 2024 (Winslow), and in April 2026 their operator announced that it would wind them down after the NBA and NHL seasons unless it reached a strategic transaction (Lewis). Main Street then wound down operations, and its streaming service shut down in September 2026 (Katz). The rented name did not outlive the channels it was painted on.

Barstool was the starkest. In August 2023 PENN sold 100 percent of it back to its founder, Dave Portnoy, for “nominal cash consideration,” $1, and told investors it expected a pre-tax, non-cash loss of $800 million to $850 million (Spangler).

Meanwhile the market consolidated around the two companies that had never been owned by a broadcaster. As of August 2026 DraftKings held 36.7 percent of mobile handle outside Nevada and FanDuel 33.0 percent, 69.7 percent between them (Smiley).

The one piece of broadcaster equity from the whole era that is still a live option is Fox’s, and it is not a sportsbook. When The Stars Group was folded into Flutter, Fox kept an option to buy 18.6 percent of FanDuel. An arbitrator ruled in November 2022 that the price would be set from a $20 billion valuation as of December 2020, rising five percent a year, with the option open until December 2030 (Flutter Entertainment, “Arbitrator Finds”). On those terms, 18.6 percent of $20 billion is $3.72 billion, and five years of five percent increases bring it to about $4.75 billion at the end of 2025. Fox still holds the option, unexercised, though Butler reports that exercising it is “unlikely at present” (Butler). The only broadcaster still holding a position in American sports betting holds it as a right to buy into the company that beat its own book, at a price that may no longer be worth paying.

Why the Audience Was Not the Customer

No operator has published the numbers that would settle why the ownership model failed, and this article will not pretend otherwise. But the shape of the failure is consistent with what the earlier articles in this series found about what a betting customer actually buys. Article 2.2 showed that the market-share war is fought with promotional cash, and that the leaders spent more of it in 2025 and 2026, not less. A bettor chooses a book on price, promotions, speed of payment and the app in his hand. None of those is a television brand. A network can put a logo in front of 10 million viewers, and those viewers can download a different app during the commercial break.

That is the lesson the ownership era paid to learn. The audience was never an asset the broadcasters owned. It was one they could only rent.

The Rent: The Billboard Business

Renting is what the networks now do, and the terms of the new deals show how deliberately they changed position. In September 2025 NBCUniversal signed DraftKings as its exclusive sportsbook integration and sponsorship partner across the NFL, including Super Bowl LX, the NBA, the Premier League and Telemundo’s coverage of the 2026 World Cup (NBCUniversal). Sportico reported that the deal was “structured solely as an advertising commitment,” where NBC’s 2020 PointsBet deal had carried ad commitments, equity options and affiliate fees (Novy-Williams). ESPN’s DraftKings deal replaced a $150 million annual cash payment and a block of warrants (PENN Entertainment and ESPN, “Mutually Agree”) with a marketing partnership whose financial terms neither company disclosed (Manzo). Amazon’s Prime Video now lets Thursday Night Football viewers link a FanDuel account and follow their own bets on screen during the game, though not place them there (Mitchell).

The odds graphic itself predates the rental era. In June 2021 NBCSN covered the Ally 400 at Nashville Superspeedway, the network’s first NASCAR broadcast of the year, and threaded betting through both the pre-race show and the race. As the pre-race show opened, odds ran across the ticker at the bottom of the screen. Asked for their picks, the analyst Dale Jarrett offered Kyle Busch as a value play at 10/1, and Kyle Petty, asked for a long shot, took Tyler Reddick at 35/1. On Lap 194 the odds to win appeared on the left of the screen, beneath the leaderboard. All of it was branded by PointsBet, then NBC’s official sports betting partner (DiNitto, “How NBC Sports”). Nothing in the segment was hidden, and none of it was news. The analysts gave picks, the graphic gave prices, and the logo told the viewer whose prices they were. That is the template the newer deals inherited: the odds sit inside the picture, in the place a viewer looks for information about the event, under the name of the company that takes the bet.

For the network, renting is a better business than owning. A sportsbook that buys advertising pays whether its customers win or lose, and the network carries none of the losses of running a book. For the viewer, the change is smaller than it looks. The broadcaster no longer needs him to lose at its own sportsbook. It needs him to bet at someone’s, often enough that the sportsbooks keep buying airtime.

They buy less of it than they did at the peak, though the decline is easy to overstate. The American Gaming Association counted nearly 50 percent fewer sportsbook television ad units in 2025 than in 2021 (American Gaming Association, “2025 Sports Betting Advertising Trends”). Measured in dollars, betting brands spent $477 million on national television in 2024 and $442.5 million in 2025, according to iSpot (Meyers). Then 2026 reversed the trend. Through August 31, betting brands had spent $269.8 million on national television, 40 percent more than a year earlier, and the growth came from a new kind of advertiser: Kalshi spent $41.3 million, and Polymarket, which had spent nothing on national television in 2025, spent $7.4 million (Cassillo).

The clearest sign that the rental market is a real market came from the NFL. The league’s five-year sportsbook sponsorships, worth reportedly just under $1 billion, expired at midnight on March 31, 2026, leaving the NFL with no official sportsbook. The talks had bogged down over a price increase for the official streaming data that Genius Sports sells (Florio, “NFL Currently Has No”; DiNitto, “Why DraftKings”). On August 27, the league announced deals with DraftKings, FanDuel and Fanatics, and none with Kalshi or Polymarket (National Football League; Florio, “NFL Will Have Three”). Nearly five months without a partner ended with the same two companies that dominate the market, plus the one that bought PointsBet.

The Reporter: When the Scoop Is the Product

The rental model has a second tier, and it reaches the viewer in the form of journalism.

On the day of the 2023 NBA Draft, June 22, Shams Charania, then one of the league’s best-known news breakers, with more than 2 million Twitter followers (Taddeo), posted that “Scoot Henderson is gaining serious momentum at No. 2 with the Charlotte Hornets in tonight’s NBA draft.” Charania was at the time a FanDuel partner and said so in his Twitter profile (Taddeo); the same profile, as Legal Sports Report quoted it, also called him a co-host of Run It Back on FanDuel TV (Mazzeo). The betting market moved at once. By Sports Illustrated’s account, Henderson went from plus 400 to minus 700 to be the second pick, and Brandon Miller, the favorite at minus 650, drifted to plus 400; by the time the pick was made the prices had swung back to Miller (Taddeo). Legal Sports Report relayed theScore Bet’s prices for the same tweet, plus 275 to minus 450 (Mazzeo), a reminder that in a market like this there is no single price, only the one on the app in front of you. The Hornets took Miller. This account does not suggest that Charania bet on the draft or that FanDuel saw his report before he posted it.

FanDuel said it was “not privy to any news that Shams breaks on his platforms” (Mazzeo). There is no reason to doubt it, and it misses the point. The value to a sportsbook of a partnership with a news breaker does not depend on the book knowing the news first. It depends on the news moving bettors, and on bettors moving money.

Charania joined ESPN in October 2024 (Rajan), while ESPN Bet was live. ESPN had issued betting guidelines to about 5,000 employees before its book launched, among them the rule that “no story should be reported, delayed, influenced or withheld with the intention of impacting betting lines” (McQuillan, “ESPN Issues Employee Betting Guidelines”). The rule is sound, and it governs intention. The market does not care about intention. A scoop moves a line whether or not its author meant it to, and a network with a sportsbook partner gains from the traffic either way.

At the far end of this tier the scoop no longer needs to be true, or the insider real. In December 2025 Awful Announcing documented accounts on X that presented themselves as sports insiders while displaying prediction-market branding. One, “Scott Hughes,” carried a Kalshi partnership icon until X reclassified the account as parody, though Awful Announcing said his exact relationship with Kalshi was unclear; another, “Emma Vance,” billed as a “lead Polymarket reporter,” posted coaching-move scoops such as Penn State’s supposed talks with Matt Campbell (Keeley).

And at the extreme is the insider who does not report the news but makes it. In October 2025 federal prosecutors in Brooklyn charged the NBA player Terry Rozier in a betting case that alleges he told a friend he would leave a game early and that the tip produced more than $200,000 in wagers on his “under” statistics (United States Attorney’s Office, Eastern District of New York). A superseding indictment in May 2026 added sports bribery charges (Associated Press); Rozier has pleaded not guilty (Senkiw, “Terry Rozier Pleads”). Part 9 of this series returns to those cases. For this article one line from the NFL is enough. When the league asked Kalshi and Polymarket in March 2026 to pull markets it regarded as open to abuse, its list included markets on broadcast commentary, and its executive vice president Jeff Miller explained why: “Some people are going to have that information ... that they can then share” (Purdum).

The Newsroom: The Fifth Mirror

The newest arrangement is the strangest, because it moves the betting price out of the sports broadcast and into the news.

In June 2025 X named Polymarket its official prediction-market partner (Kirkland). In November Google announced it would begin showing Kalshi and Polymarket probabilities in search results and Google Finance (Shubnell). On December 2 CNN made Kalshi its exclusive prediction-market partner, with a real-time Kalshi ticker on its linear channel; per Kalshi’s chief executive, CNN is not paying to license the data (Fischer, “CNN Strikes”). On December 4 CNBC signed a multi-year deal with Kalshi, which Kalshi called exclusive, that puts a ticker on programs including Squawk Box and Fast Money (Kalshi, “CNBC and Kalshi”). In January 2026 Dow Jones made Polymarket the exclusive prediction-market partner of The Wall Street Journal, Barron’s, MarketWatch and Investor’s Business Daily, down to an earnings calendar showing “market-implied expectations” (Polymarket and Dow Jones). In April Fox News Media followed with Kalshi (Senkiw, “Fox to Integrate”).

The leagues have signed too. In October 2025 the NHL signed multiyear partnerships with both Kalshi and Polymarket, with their branding on virtual signage during game broadcasts (NHL Public Relations). Major League Baseball made Polymarket its exclusive prediction-market partner in March 2026, while signing an information-sharing agreement with the Commodity Futures Trading Commission and restricting markets on individual pitches (Major League Baseball). On October 26, 2026, the NBA named both Kalshi and Polymarket its “Preferred Prediction Market Exchange of the NBA,” in a deal that required a memorandum of understanding with the same regulator (Helin). The league keeps hold of the product: the agreements give it control over which trades on NBA games the markets may offer, and the markets get official NBA data from Sportradar (Helin). Of the four major leagues, only the NFL, as noted, has signed none.

Not every partnership has lasted. Yahoo Finance made Polymarket its exclusive provider in November 2025, took its prediction-markets hub down in April 2026 and by September had ended the partnership, without giving a reason (PYMNTS).

This series has so far catalogued three manufactured numbers and one manufactured word. Article 1.3 traced the $150 billion pre-2018 illegal market to a single 1999 guess. Article 1.4 showed prediction-market “volume” reported as if it were sportsbook “handle.” Article 2.1 found a dopamine story borrowed from cocaine research. Article 2.3 showed that the word “live” describes the picture, not the bet. The fifth is the market price reported as news.

When a news channel runs a Kalshi ticker beside a story, it presents a trading price as a fact about the world. Three things about that price are not on the screen. The first is who sets it: whoever trades, and on Kalshi the traders are overwhelmingly sports bettors. The industry analyst Dustin Gouker has put sports at 85 to 90 percent of Kalshi’s volume (Nagy). A thin political or economic market can be moved by a few people with money and a motive, and the move then gets reported as a shift in expectations. The second is why it is there. Per Kalshi’s chief executive, CNN is not paying to license the data. When the network does not pay for the data, the data is the advertisement. The third is who benefits most from the network’s credibility. Klaudia Jaźwińska of the Tow Center argues in the Columbia Journalism Review that the platforms need visible non-sports contracts to support their claim to federal rather than state regulation (Jaźwińska). Newsroom partnerships put those contracts on screens audiences trust.

The people with the strongest reason to trade on a price are the people who already know the answer. In December 2025 Representative Abe Hamadeh called the CNN-Kalshi deal “a built-in structural conflict of interest” in a letter to the acting chairwoman of the CFTC (Hamadeh). In May 2026 the House Oversight Committee opened an investigation into insider trading on prediction markets, citing the April 2026 indictment of an Army master sergeant accused of using classified information about a military operation to make more than $409,000 in profit (House Committee on Oversight); he has pleaded not guilty (“US Soldier Pleads”). This article found no evidence that any news organization partnering with Kalshi or Polymarket has changed its coverage to move a market. The concern is structural. It is not a documented case.

What the Evidence Says About Exposure

The research linking betting marketing to betting behavior is stronger than the research on any other part of this story, and almost all of the research this article found comes from Australia. The most useful study followed 597 regular bettors over three weeks, 6,843 observations in all, and found that exposure to advertising and inducements was “reliably linked to a greater likelihood of betting, higher intended and actual betting expenditure, and spending more than intended.” Among the forms of marketing that mattered most were brands promoted during sporting events (Browne et al.). A companion study found that text messages from operators were associated with a higher likelihood of betting and higher spending within a day (Russell et al.), and a survey of 1,813 sports bettors found that heavier users of inducements were more likely to place impulse bets during play, especially problem gamblers and frequent viewers of sport (Hing et al.).

The American evidence is thinner and less direct. A 2025 study in JAMA Internal Medicine found 23 percent more internet searches for gambling-addiction help nationally after the Supreme Court decision that opened the market, with larger increases where online sportsbooks opened (Yeola et al.). Searches are not diagnoses, and the study does not isolate advertising. No study this article found has measured the effect on American bettors of odds integrated into a broadcast, or of a prediction-market ticker on a news channel.

The rules are thinner still. American betting advertising is governed mainly by the industry’s own code, which in 2023 barred the phrase “risk free,” college partnerships and endorsers under 21 (American Gaming Association, “New Updates”), and by state regulators. Massachusetts bars sportsbook advertising at events where 25 percent or more of the audience is expected to be under 21 and prohibits “saturation” (Massachusetts Gaming Commission). In January 2023 Ohio regulators notified DraftKings that it could face a $350,000 fine over more than 2,500 mailers sent to people under 21 (Ingles); the next month DraftKings agreed to pay $500,000 in fines covering the mailers and a second advertising violation (“DraftKings, Barstool”). The SAFE Bet Act, which would ban betting advertising during broadcast events between 8 a.m. and 10 p.m., was reintroduced in March 2025 and had not been enacted as of October 2026 (Evans, “SAFE Bet Act”). The GAME Act, introduced in May 2026, would bar large platforms from accepting betting or prediction-market ads aimed at minors (O’Connor). Abroad, Italy banned gambling advertising in 2018 (Osborne Clarke), British operators adopted a voluntary whistle-to-whistle ban in 2019 (Gallagher), and Australia announced in May 2026 that, subject to legislation, wagering ads would be barred from free-to-air broadcasts of live sport between 6 a.m. and 8:30 p.m. from January 2027 (Evans, “Crackdown”).

Every one of those regimes regulates the advertisement. As enacted, none regulates the scoop, the odds graphic inside the broadcast, or the ticker beside the news, which are the three forms this article has followed the complex into.

The Mirror Test

Information on the Screen

VerdictIllusion

The Mirror Test is this series’ standing verdict: every strategy, product and claim is judged on whether it creates a real edge or only the appearance of one. The verdicts are EDGE, ILLUSION and CONDITIONAL.

The claim under test is the one the screen is built to sell: that the information on it, the insider’s scoop, the odds in the broadcast, the ticker beside the news, gives a bettor an edge. The ruling is that the claim mistakes who the information is for. By the time a scoop reaches a follower’s phone, it has reached the book’s traders, and on the most famous night of the insider era the market lurched toward a pick that never came. The ticker’s price is set by whoever trades, and a few people with money and a motive can move a thin market.

The case against

The case against this verdict is real. Traders who hold information before it is public stand to make real money: the House Oversight Committee has opened an investigation into insider trading on prediction markets, citing a soldier accused of turning classified knowledge into more than $409,000 in profit. On that view, a bettor who watches the best-informed screens closely enough should share in it.

Why the verdict holds

The verdict survives because every case documented so far, proven or alleged, rests on information held before it was published, which is what a viewer does not do. Speed decides who gets the pre-scoop price, and article 2.3 showed the bettor loses that race. Then there is the source. The scoop comes from a reporter partnered with a sportsbook, or employed by a network that is, or from no reporter at all; the odds carry a sportsbook’s logo; the ticker, by its supplier’s account, comes to the network without a license fee. Each is worth what it makes people bet, not what it gets right. The edge is real, and it is not the viewer’s.

ILLUSION Information on the Screen

Conclusion: The Billboard and the Price Feed

Go back to November 6, 2025. The broadcaster lost its bet on being the bookie, and what it kept afterward suited it better. It keeps the advertising revenue and carries none of the book’s losses. It keeps the betting shows, the odds graphics and the insiders, and none of it has to show up on a balance sheet as a stake in its viewers’ losses.

That is why the bust should not be read as a retreat. The equity era made the conflict visible. A warrant that pays off when the audience loses can be pointed at, and critics pointed at it. A rate card cannot, and neither can a scoop, a bet tracker in the corner of the screen, or a data feed CNN does not pay to license, according to Kalshi, that turns a news ticker into a price board. The complex survived by shedding the one part of itself that anyone could see.

The bettor inside the House of Mirrors should draw one practical conclusion. Everything on the screen that looks like information, the odds graphic, the insider’s tweet, the probability beside the headline, should be read as if it were put there by, or for, someone who profits when he bets. That does not make it false. It makes it an advertisement until shown otherwise, and an advertisement is never an edge.

Part 2 has now followed the machine from the bettor’s brain to the app, the bonus, the live market and the broadcast. Part 3 follows the same machine into the bettor’s own feed, beginning with article 3.1, The Algorithm’s Goal.

Sources

This is 2.4 of House of Mirrors: A Bettor’s Guide to the Illusions of Modern Sports Gambling, a serialized investigation of the American sports betting industry. Next up is article 3.1, The Algorithm’s Goal.