What Are the Odds? Are News and Media Really Betting Their Credibility on Kalshi?

Reading the Race Off a Betting Line

In September, CNN's chief data analyst Harry Enten opened the midterm stretch by pulling up Kalshi's odds, describing the market as the place "where people put their money where their mouths are," and reporting a 47% chance that Democrats win the Senate.

When did we start using betting odds to predict elections? It can't have been long, because regulated betting markets have only been allowed to offer election contracts to Americans since the fall of 2024, when a federal judge cleared Kalshi to launch them. Barely a year later, in December 2025, CNN made Kalshi its official prediction market partner, bringing its live odds to the network's television, digital and social coverage. CNBC signed a similar deal days later, and Kalshi has since added the AP and Fox News, though Fox's deal reportedly leaves politics out.

Money explains an advertising relationship, but not a replacement of a scientific method or evidence-based tools. Journalists are supposed to value evidence and verification. Betting markets are new and, in politics especially, untested. We've already seen how bad some of their estimates have been, and they are far more easily influenced than a poll.

This begs the question: why did news outlets and journalists turn a betting market into the political compass for where races stand in real time?

The Media's Bet on the Betting Lines

If money explains the partnership, the media's own reasons have to explain the replacement. First, the polls are clearly off, which has degraded the trust in the polling institutions. Pew Research Center acknowledges that in 2016, 2020 and 2024, many polls underestimated support for Trump, its own included, and this August, half of likely voters said they don't trust most political polls. This doesn't mean that polling doesn't work, or that it's not valuable, however. The likelier problem is how the variables are measured to build the polls, along with biases that interfere with a realistic representation of the electorate and how responsive it is. Pew found that Democrats were significantly more likely than Republicans to take part in 2020 preelection surveys, even after controlling for factors such as age, gender, race and urban versus rural residence.

The culmination of the above basically handed a win to the betting markets in 2024: a Vanderbilt study comparing Polymarket with polling found the market's odds tracked the result more closely, especially in contested states, and Kalshi declared that prediction markets can "decisively outperform polls and traditional media." A second Vanderbilt study of nearly 2,500 political markets from the final weeks of that same election found accuracy of just 78% on Kalshi and 67% on Polymarket. Kalshi's declaration was true in the moment, but not necessarily indicative of future results. After all, its track record was small, as it had only been cleared to offer election contracts about a month before the vote. As the markets became known to more and more of the public, and with them how easily they can be influenced, their odds drifted further from electoral results, or at least appear to have.

The rest of the case appears to be pure convenience and an emotional ploy. Few state-level races are polled regularly, so outlets like the Kansas Reflector turned to the markets for primaries with no traditional polls. A poll takes days to field, while a market reprices in minutes, and CNN pitched its partnership as giving its journalists a fresh, data-based angle. That makes the odds a ready tool to sell a story or a narrative, regardless of the intentions.

Broken trust in the polls, one anecdotal result in 2024, and a faster number that sells a story explain why the media made the switch. One result in the history of elections is anecdotal and not representative of performance in any other year but 2024. Betting odds earned their reputation in sports, where they are tested against a result recorded on the field.

The Bet: Reality Draws the Line

In sports, the bet is tied to objective data and one outcome: the payday. There are numerous ways to get paid out, from the spread to the total to a player prop, but every one of them settles on the same record of what happened in reality. Only the empirical results are supposed to shape that record. The event happens, the data is recorded, and it stays unchanged through to the outcome. Influencing that result is much harder and clearly illegal, both inside the sport and under the law, and when someone is caught, the leagues and prosecutors step in. The NBA banned Jontay Porter for life over a betting scandal, and in October 2025, federal prosecutors charged Terry Rozier in a scheme to bet on nonpublic information about NBA players and teams.

That is why sports odds sharpen as information comes in: a Harvard study of more than 88,000 Major League Baseball games found forecasts converged with outcomes as information improved in quality and accessibility, and that folding in current-season performance raised accuracy. The bettor's reward is direct: he uses Kalshi to make money on a result decided by objective accomplishment on the field. Rewards are directly influenced by reality, and the goalposts are the same for all betting parties.

The media sells the odds as if that standard carried over to elections, without explaining how the two are different, or even knowing why they are different. My bet is they didn't think that far about it. They sold it without considering the sciences between them, a large hint that the incentives are not honest journalism and align more with narratives and storytelling. Those incentives favor short-term wins over long-term success, which is obvious when we look at the trust in mass media: just 33% in Gallup's latest survey, roughly on par with the 28% of likely voters who trust most political polls.

Emotions Move the Odds

Politics, while often called a blood sport, doesn't share the same variable inputs that sports betting does. A voter's reasons to vote, and whom to vote for, are personal, removed from objective measure and reason, and slide with their subjective goalposts, which can be changed at any time, for any reason or no reason at all. Where the sports bettor's reward is direct, a voter's reward is personal.

Trump is a hyperbolic example of this, and the polling around him is the proof. Some people support or oppose him on data, on actual things he has done, his personality, or other provable things. But there is a strong, unusual correlation on both the pro- and anti-Trump sides that is all emotion, and those voters don't budge. In his first term, Gallup found his frequent controversies did little to move opinion: his approval stayed within an 11-point range over two and a half years, against a 30-point average for other elected presidents since World War II.

Voters could change their minds, or they could not, and the reasoning could be related to data, or it could not. That change isn't measurable, because the rule set behind it belongs to each person, if one exists at all. A betting market's live line can never hope to move with an electorate in general, even before intentional influence, let alone with specific regions or areas. Unless the market could get into people's minds in real time, it can't, and neither could anyone. That is why polling uses past behavior, like how people voted last time, to project future behavior. That isn't what betting lines do.

Money Buys the Emotions

The bad calls are already piling up. In California's 2026 primaries, prediction markets misidentified the second-place finishers in both the governor's race and the Los Angeles mayor's race, despite heavy trading. Across this year's primaries, an on-chain review of Polymarket found that in closer races, favorites won less often than their prices said they would.

Part of the reason is who sets the price. In 2024, Polymarket confirmed that four accounts holding a combined $28 million in bets on a Trump win belonged to one French trader. At one point, a single account placed enough bets to push Trump's odds of winning the popular vote from 26% to 39% in a few hours. Polymarket said it found no evidence of manipulation, and the trader said he was betting on his own read of the race. But that is exactly the concern: intentional or not, it influenced the odds, which means malfeasance is even easier to hide in plain sight, and with the benefit of the doubt. The Brennan Center notes that small-volume markets on down-ballot races are easier still to move, and the money behind the odds isn't disclosed.

Once the number moves, others follow it. The same Vanderbilt researchers who measured 2024's accuracy concluded that the platforms encourage herd behavior driven by visibility and hype rather than news.

Then the media puts the number on air, and the number becomes part of the story the media wants to tell. That matters, because the people telling it lean heavily one way: a Syracuse University survey of 1,600 journalists found that just 3.4% identify as Republicans, against 36.4% who identify as Democrats. And Harvard's Shorenstein Center found that Trump's first 100 days drew 80% negative coverage, reaching 93% at CNN and NBC, while Obama's first 100 days drew 59% positive. A newsroom that leans left, selling a number anyone with enough money can move, has every tool it needs to steer how voters see a race, whether it means to or not.

And voters do follow it. A Brennan Center study of 1,994 registered voters, released this month, found that when a candidate was favored by prediction markets, voters were 10 to 20 percentage points more likely to believe that candidate would win, no matter what the polls said. Warning voters about the markets made no difference. Even the Daily Wire's Ben Shapiro, whose show Kalshi sponsors, has said late-breaking movement in the markets reflects nothing "but the narrative power that's being presumed." Money influences emotions through the media's narrative, built on betting lines that aren't well correlated to results. Sell the story, and the story starts to create the reality.

Influence Runs Through the Newsroom

So what is the media's intention, and why did it pivot so quickly to Kalshi's markets? Why does Shapiro look at his own sponsor's odds and say, "these polls don't justify that," while CNN's Harry Enten presents them as the network's exclusive data partner, in segments brought to you by Kalshi? The media put a sports tool, built for results that reality decides, in charge of measuring voters whose reasons are personal, whose goalposts move, and whose emotions money can buy through the media's own narrative.

The next time a network puts the odds on screen, ask what that number is actually measuring, and who could have moved it.

What are the odds it's coincidence?

Editor's note (Oct 29, 2026): As I was finishing this piece, Polymarket halted trading on its international exchange after some users' funds were withdrawn without their authorization. Polymarket said a third-party provider was hacked. Two analytics firms told CNBC that one account pulled at least $2.7 million across 710 wallets. Polymarket says its U.S. platform wasn't affected and that it will refund the users who were hit. Trading resumed at 3:00 PM ET. A similar vendor breach happened in June.

Sources

Reading the Race Off a Betting Line 1. CNN News Central transcript, CNN, Sept. 8, 2026 2. Election Betting Takes Kalshi, Polymarket To Top of App Store Charts, Covers, Nov. 6, 2024 3. Exclusive: CNN strikes prediction data partnership with Kalshi, Axios, Dec. 2, 2025 4. First CNN, Now CNBC: Kalshi's Event Odds Go Prime Time, Finance Magnates, Dec. 2025 5. Fox to Integrate Kalshi Forecasts Across Platforms, Excluding Politics, DeFi Rate, 2026

The Media's Bet on the Betting Lines 1. Why and how we're weighting surveys for past presidential vote, Pew Research Center, July 23, 2025 2. Only 28% Trust Political Polls, Rasmussen Reports, Aug. 13, 2026 3. Weighting polls on party affiliation and past vote, Pew Research Center, 2025 4. Are Betting Markets Better than Polling in Predicting Political Elections?, Vanderbilt University, 2025 5. Polymarket predicted Trump's win. Now comes the hard part., Business Insider, Nov. 2024 6. Kalshi Claims it 'Beats Wall Street' Forecasts, Yahoo Finance, Dec. 2025 7. Election betting odds became less accurate as primaries neared, report finds, NBC News, Oct. 29, 2026 8. New Study Finds Prediction Markets Are Influencing Voters, Brennan Center, Oct. 20, 2026 9. Prediction markets offer insight into Kansas primary races, Kansas Reflector, Aug. 4, 2026 10. Kalshi Partners With CNN, Hits $11B Valuation, Legal Sports Report, Dec. 2025

The Bet: Reality Draws the Line 1. NBA coach Chauncey Billups, Heat guard Terry Rozier arrested in Mafia gambling probes, RNZ/AFP, Oct. 2025 2. FBI Arrests NBA Players, Coaches in Gambling Probes, Bloomberg Law, Oct. 27, 2025 3. Information and Market Efficiency: Evidence From the MLB Betting Market, Matthew Bouchard, Harvard College, 2019 4. Gallup: Americans' trust in news media rebounds from record low, The Desk, Oct. 2026

Emotions Move the Odds 1. Trump Job Approval Steady in Low 40s, Gallup, Aug. 2019

Money Buys the Emotions 1. Can prediction markets foresee elections? California's results offer clues, Straight Arrow News, 2026 2. How Accurate Is Polymarket? We Scored Every 2026 Primary, Bitquery, Oct. 2026 3. A handful of mystery Polymarket accounts…, Fortune, Oct. 24, 2024 4. Polymarket whale raises Trump odds, sparking manipulation concerns, Cointelegraph, Oct. 2024 5. Study Finds Prediction-Market Odds Shape Voters' Election Views, RG.org, Oct. 21, 2026 6. Are Polymarket and Kalshi as reliable as they say? Not quite, study warns, DL News, 2025 7. Just 3% of Journalists Identify as Republican, AllSides (Syracuse Newhouse data), Mar. 20, 2024 8. It's not your imagination: Study finds Trump coverage overwhelmingly negative, Fox News (Harvard Shorenstein data), May 22, 2017 9. A Pollster Got Caught Faking the Data, The Ben Shapiro Show, Oct. 21, 2026

Influence Runs Through the Newsroom 1. Republicans Are in Better Shape Than You've Been Told, The Ben Shapiro Show, Oct. 9, 2026 2. How Kalshi infects the news, Public Notice, July 2026