For the past week, I’ve found myself playing the same 23-second CNN clip on repeat. I’ve watched it in bed, during my commute to work, at the office, midway through making carrot soup, and while brushing my teeth. In the video, Harry Enten, the network’s chief data analyst, stares into the camera and breathlessly tells his audience about the gambling odds that Donald Trump will buy any of Greenland. “The people who are putting their money where their mouth is—they are absolutely taking this seriously,” Enten says. He taps the giant touch screen behind him and pulls up a made-for-TV graphic: Based on how people were betting online at the time, there was a 36 percent chance that the president would annex Greenland. “Whoa, way up there!” Enten yells, slapping his hands together. “My goodness gracious!” The ticker at the bottom of the screen speeds through other odds: Will Gavin Newsom win the next presidential election? 19 percent chance. Will Viktor Orbán be out as the leader of Hungary before the end of the year? 48 percent chance.
These odds were pulled from Kalshi, which hilariously claims not to be a gambling platform: It’s a “prediction market.” People go to sites such as Kalshi and Polymarket—another big prediction market—in order to put money down on a given news event. Nobody would bet on something that they didn’t believe would happen, the thinking goes, and so the markets are meant to forecast the likelihood of a given outcome.
Prediction markets let you wager on basically anything. Will Elon Musk father another baby by June 30? Will Jesus return this year? Will Israel strike Gaza tomorrow? Will the longevity guru Bryan Johnson’s next functional sperm count be greater than “20.0 M/ejac”? These sites have recently boomed in popularity—particularly among terminally online young men who trade meme stocks and siphon from their 401(k)s to buy up bitcoin. But now prediction markets are creeping into the mainstream. CNN announced a deal with Kalshi last month to integrate the site’s data into its broadcasts, which has led to betting odds showing up in segments about Democrats possibly retaking the House, credit-card interest rates, and Federal Reserve Chair Jerome Powell. At least twice in the past two weeks, Enten has told viewers about the value of data from people who are “putting their money where their mouth is.”
Only a minuscule amount of critical thinking is required to reject the hypothesis that cumulative betting odds on news items are predictive of real outcomes. He mentioned sports betting as a similar case, do sports betters think increased betting odds (say on a specific horse to win) increase the likelihood that that horse wins? I kind of feel like they understand what the numbers represent but I’m not super steeped in gambling culture so I’m genuinely curious.
News outlets framing betting odds as useful stats for predicting events is worrisome though. Readers expect journalists to provide only relevant information so mixing in stuff you’re paid to include breaks the expectation between writer and reader since the expected intention of the article (information on a topic) is different than the actual one (fulfilling an advertising obligation to a gambling company).
The thinking isnt that that you putting money on a horse to win increases it’s odds of winning, its that by signalling your belief that you think it will happen by a costly signal (you lose the money if you are wrong) you are updating the overall odds to be closer to the true probability by the power of crowds. if 200 people are betting something will happen and only 10 are betting it wont then that is evidence that the thing is more likely than not to happen.
There are flaws in this thinking, it doesnt take into account manipulation of events to win bets is a particularly big one, and it also gets worse the more removed the thing being bet on is from everyday life as people make less informed choices.