I took a cultural anthropology class at my very large, sports-obsessed, state university (Hook ’Em!). For weeks, we discussed the role of sports and tribalism: the importance of feeling like you belong and an “us vs. them” mentality. This mentality was on full display every Saturday in the fall, while we watched our football team (mostly) win. Supporting university sports teams was an outward expression of our tribalism and one of the few things on campus that almost every student supported.
For sports fans, this still rings true. But there are now new ways to join a tribe. Prediction Markets let you pick a side on almost anything: Bad Bunny versus Drake for Spotify’s top artist or your candidate in the midterms. Prediction markets have turned picking a side into an everyday option, with odds that move in real time.
It’s the sportification of everyday life.
The role and growing importance of non-sports data in prediction markets is less about the data itself than about why it exists. As the loneliness epidemic grows, people begin to feel small and powerless. Rooting for something, even something like Drake streaming more on Spotify and climbing five points a week on a market, becomes tribal as people pick a side. That’s your guy making a comeback. There’s something intoxicating about a line that tells you how much hope you should have.
People want a team and need a way to track it. The problem is what they’re using to track. A prediction market tells you what a group of traders thinks will happen: it’s an opinion with money behind it. Unfortunately, more and more, people are treating the opinion like the fact and using it as a source for information. Polymarket told Reuters that more than 90% of visits to its site are to view markets, not to trade. CNN, CNBC and Fox have signed deals with Kalshi, Dow Jones has one with Polymarket, and the AP has licensed its election data to Kalshi. The everyday opinions with odds about your tribe are on your TV, in your feed, and in the newspaper acting as news. In sports, nobody confuses the point spread with the final score. Outside sports, that line is blurring, and once a number is treated as information, it has to be judged as information. That’s where non-sports data is weakest.
There is also the issue of whether the prediction market information you see before you make a trade is presented correctly. Is the bet talking about the right person? Is it asking the right question? My company has spent 20 years in sports data, so we deal with these data inconsistencies every day. For example, a rookie’s name might be spelled two ways by two different operators.
There’s something intoxicating about a line that tells you how much hope you should have
It’s messy, but at least there’s a fence around it because it’s a limited number of participants. There are only 32 NFL teams. Every player has an ID, and each provider’s ID can be matched to the others. In our company, one of our secret ingredients is our ability to match these IDs across operators and media. When something looks off, we use our “Rosetta Stone” to check against it.
Outside sports, there’s no fence, and nothing to check against, because the data is infinite and no one is giving Bad Bunny an ID number to carry with him through every line and trade. An example of this problem recently came up in the Texas Senate race. One platform listed a candidate as “J. Talarico” and another as “James Talarico (D).” Humans can assume that it's the same person, but a machine will see two. Platforms are listing markets in almost anything, and the number of markets in 2026 for midterm elections was more than 21,000, compared to 463 in the 2024 election cycle. That’s a lot of names, wording, and grammatical sources where consistency across data points is essential. When the list of possible “whos” is infinite, checking accuracy goes from a QA job to a behemoth of a challenge.
The responsibility to ensure this consistency falls squarely on the industry. If non-sports data is going to be part of our identity and our pitch to newsrooms, then the platforms, data providers and media partners have to make sure it’s consistent and correct. Focusing on honest numbers about scale, with handle and notional volume labelled for what they are, and a signal of how many people actually stand behind a number (like how a poll shows its sample size), might help to ensure that traders are seeing the real data and a more complete picture.
None of this is radical and it’s already done in sports data. It took leagues, providers, and operators years to build shared IDs and official feeds, but they did it because accuracy was the price of trust. We gather this data, check it, and make it digestible for partners who want to engage their audiences without misleading them. We’re not the only ones who should care about getting it right, and the whole industry will be better off when more of us do.