Money Stuff: Predicting the Big Game

Yesterday was the Super Bowl of American football, which is also the Super Bowl of American advertising, and also increasingly the Super Bowl of American sports gambling, in which traditional sportsbooks take on newer prediction markets. Prediction markets have the momentum. Bloomberg’s Peyton Forte and Denitsa Tsekova reported on Saturday:

The bigger concern hanging over the [gambling] industry is the rise of prediction markets like Kalshi, which have come out of nowhere over the last year to offer a new way to bet on sports, bypassing the state-level gambling regulations that have restricted the spread of older gambling apps. …

Part of the allure of Kalshi is that it is using its novel structure to open up betting on everything from the length of the half time show to the likelihood of Jeff Bezos attending the event, while DraftKings and FanDuel are almost exclusively focused on the scoring and outcome of the game.

“Kalshi’s growth is fueled by ad campaigns, earned media, social virality and, above all, superior depth, breadth, and distribution compared to traditional online sportsbooks,” analyst Edwin Dorsey wrote in one of several recent posts on Substack about why he is bearish on DraftKings.

And Justina Lee and Ira Boudway wrote last week:

Traders on Kalshi and its chief rival, Polymarket, have swapped more than $800 million worth of contracts tied to the Super Bowl so far, compared with the $1.8 billion Americans are expected to wager on the game through regulated sportsbooks, according to the American Gaming Association.

We’ve talked about this a lot: Prediction markets have increasingly become sports gambling sites, and compared to traditional sportsbooks they have a range of regulatory advantages. Sports gambling is illegal in some US states and heavily regulated in most others, but Kalshi is in theory federally regulated, so it’s available everywhere and state regulators can’t do much to stop it. And because prediction markets are in theory financial markets, not gambling, they might be available in your brokerage app. “An emerging asset class like event contracts,” is how Robinhood described Super Bowl betting last year. 

In particular, Kalshi has better access to one important demographic of sports bettors, which is college students. Most states that allow sports betting have age requirements, and usually you have to be at least 21 years old to bet legally at a sportsbook. But Kalshi, which is federally lightly regulated, allows 18-year-olds to open accounts and bet on sports. Young men seem to be the most active and addicted gamblers, so getting them three years early seems like a big advantage for Kalshi.

Prediction markets have two other interesting features compared to regular sportsbooks:

1. They are much less focused on sports: They’ll let you gamble on all sorts of political and entertainment and financial topics, any sort of “event,” not just sports. If you want to construct a complex multi-leg parlay on the Super Bowl, you might prefer a sportsbook. But if you are a casual fan/gambler and you want to bet on the winner of the game and also the first song of the halftime performance, you might prefer a prediction market.
2. They are at least technically peer-to-peer: At a sportsbook, you are betting against the house, while at a prediction market you are betting against other bettors. There’s a winner for every loser.

One upshot of this is that bets at a sportsbook tend to be pretty easy to resolve. The sportsbook will tend to make clear rules about what counts as a win or a touchdown or whatever, and will resolve uncertainties swiftly and consistently and, you know, probably in its favor. But on a prediction market, with an infinite variety of possible events, and with any resolution being good for half the bettors and bad for the other half, sometimes you will get resolution disputes.

For instance, here is a Business Insider story about whether Cardi B performed at the Super Bowl halftime show:

That's the question that both Kalshi and Polymarket had to confront on Sunday night, after rapper Cardi B appeared at the Super Bowl halftime show as part of Bad Bunny's set.

Cardi B had no singing role, though she could clearly be seen dancing and mouthing words alongside singer Karol G and actors Pedro Pascal and Jessica Alba. …

Polymarket has decided that it did count as a performance, though that decision has been disputed and is set for a final review later this week. ...

With Kalshi, things were less clear. …

Due to the apparent lack of clarity, Kalshi invoked a rule that gives the platform sole discretion to settle contracts when there's an ambiguity not addressed by the rules.

Kalshi spokeswoman Elisabeth Diana told Business Insider in an email that the platform chose to settle the market at the last traded price, because Cardi B "was at the very least mouthing the words when dancing," but it's unclear "if she was singing," and it's "impossible to tell either way."

Will the US Commodity Futures Trading Commission weigh in? The Wall Street Journal recently profiled Polymarket founder Shayne Coplan, who calls his company a “global truth machine,” and I always think of claims like that when I read stories like this.

One set of things that you can bet on, in prediction markets, is events. You can bet on who will win the Super Bowl or who will win the election or when Timothée Chalamet will get engaged or, famously, whether Jesus Christ will return before 2027. We have talked about that one before, and I have explained it as largely a bet on interest rates, though I suppose it is also a bet on the Messiah. Conventionally the prices of these event contracts are described as “probabilities”: If a contract that pays $1 if Jesus Christ returns trades at $0.04, then that implies a 4% probability of His return, and people sometimes describe the price as “4%” rather than “$0.04.”

Another set of things that you can bet on, in prediction markets, is prices of financial assets. You can bet on whether the price of Bitcoin will exceed $70,000 this afternoon, or whether the S&P 500 index will close up or down today, or whether US housing prices will go up more than 2% this year. As we have discussed, these sorts of bets tend to be expressed as binary options, not continuous linear bets: The bet is not “I am long Bitcoin, and make a 10% return if Bitcoin goes up 10%,” but rather “I get $1 if Bitcoin goes above $70,000 and $0 if it doesn’t.” This seems to me like a weird artifact of prediction market architecture, but it might also be fun. You might like an all-or-nothing bet more than you’d like, you know, a regular investment.

Another set of things that you can bet on is other contracts. For instance: You can bet on whether the price of that Jesus Christ contract will exceed 5% for a majority of the one-hour period just after midnight on Feb. 17, 2026. Here, the underlying asset is the will-Jesus-Christ-return contract, and this is a derivative contract — a binary option — that pays out based on the trading price of the underlying contract at a particular arbitrary time (not maturity). Your bet here is not about whether Jesus Christ will return, or even exactly about the probability that He will return. It’s just about the price. The price of a contract on a different Polymarket market.

Here is a claim that a well-known prediction markets trader might have manipulated that contract, by (1) buying the derivative contract and then (2) buying the underlying contract to push its price up to 5%, thus causing the derivative to pay off. This strikes me as unlikely — the measuring period for the derivative contract is a week away, so manipulating it now doesn’t really help — but what do I know. 

On the other hand, I ask you, as a grown-up: What possible purpose could the “Jesus Christ return before 2027 Odds >5% February 17, 12-1 AM?” contract have other than as a plaything for market manipulation? I mean, we have joked around here about the underlying “Will Christ return” contract, and I suppose there are a few reasons you might trade that: genuine messianic beliefs, interest-rate and market-structure views, as a joke. But the derivative contract is pretty purely “I have gambled on Christ’s return and want a side bet that will resolve more quickly.” 

Obviously one way to play it is to buy the derivative and then buy enough of the underlying contract to push the price up above 5% during the relevant measuring period. The other way to play it is to sell the underlying contract to people who are buying it to manipulate it. Here is an X post from a trader suggesting that: “People think manipulating this market to YES will be the likely outcome, but that's actually the trap. … To win the DM [derivative market], a whale MUST manipulate the MM [main market] price up to 5.1¢+ and hold it there for 31 minutes. This forces them to become a ‘price insensitive buyer.’” So you sell them as much of the main Jesus contract as you can for a guaranteed [1]  5-cent return.

One point here is that derivative markets are often much smaller than the underlying markets, and that is the normal intuition: “The underlying market is real economic activity, while the derivative market is just some side bets.” But that is not always true; it is not true of, for instance, Indian stock options, or Libor. And there is no reason for it to be true in the market for the return of Jesus Christ in 2026. There’s no underlying real economic activity there; it’s just side bets on side bets. Though right now I see the underlying market at about $1.5 million of volume and the side-bet market at about $122,000.

Anyway the main Messiah contract is up 50%year to date (from under 2% to about 4%), presumably mostly on these sorts of shenanigans but, you know, it does feel like end times? 

Also the reader who sent me this proposed several section headers including “Messiah market manipulation” and “banging the Christ.”

The Super Bowl also featured a ton of artificial intelligence ads. Everyone already knows this, but those Claude ads are so good. Absolutely note-perfect acting and writing to nail the uncaniness of AI, the stereotyped AI chatbot mode of interaction, the way that chatbots respond to questions with a pause followed by obsequious flattery.

One thing that the ads made me think about is how cheap Super Bowl advertising is, for an AI company. A Super Bowl spot costs something like $10 million for airtime plus another few million to produce, for a total at the high end of maybe $20 or $30 million, or roughly the cost of paying one employee for one month at a leading AI lab. Mark Zuckerberg carries around $30 million in his wallet in case he runs into an OpenAI engineer at Starbucks. The cost of creating a cutting-edge AI model — in compute and researcher pay — is astronomical in a way that makes the cost of any advertising, even Super Bowl advertising, look like nothing.

Modern AI models are used for lots of things, and have huge business applications, but obviously they are also a consumer technology. The Super Bowl really drove that home, with ads for AI products that tended to emphasize personal use. I am not sure that the way to succeed in AI is by widespread consumer adoption; frankly it would be weird if that were true. (The way to succeed in AI is, like, create a superintelligent robot and use it to conquer the world, etc.) But clearly some AI companies want widespread consumer adoption; we have talked about OpenAI’s focus on chatbot user engagement over “original blue-sky research.” It is possible that, in practice, the way to AI dominance is wide adoption by consumers, not building the best model. 

Obviously building the best model is one way to consumer adoption, but at the levels we are currently at, I am not sure that consumers will be able to discern which models are better than others. And another way to consumer adoption might be … advertising? Branding? Coke vs. Pepsi? Which is, you know, hard and expensive, but easier and cheaper than building good AI models. What if the winner of the AI wars is not the company with the best model or the most data centers or the smartest researchers, but the one with the best ads?

Bitcoin has had a really bad couple of weeks, falling from about $90,000 at the end of January to about $70,000 today. We talked about this on Thursday, and I pointed out that what makes a crypto crash different from a stock-market crash is the lack of fundamentals. “Broadly speaking,” I wrote, “a stock’s price in the short term will reflect things like forced sales by its owners, but its price in the long term should reflect market expectations of the present value of its future cash flows.” With crypto, you have to look elsewhere for explanations.

Here’s a Wall Street Journal article titled “A New Crypto Winter Is Here and Even the Biggest Bulls Aren’t Certain Why,” though they wouldn’t be, would they? (The biggest bears are quite certain why.) One theory is boredom:

Market theories for the selloff ranged from investors’ pivot toward the prediction markets and other risky bets, to widespread profit-taking after a blistering bull run. …

There is no shortage of other markets for traders to make audacious bets, said [Anthony] Pompliano, the CEO of ProCap Financial. Prediction markets, gold, silver, artificial intelligence and so-called meme stocks are all vying for their attention of late, drawing eyes away from crypto.

“It used to be that bitcoin was the consensus view where asymmetry existed,” Pompliano said. “Now you have AI, prediction markets…many other areas where people can go and they can speculate.”

This sort of rhymes with my “Boredom Markets Hypothesis,” which says that the prices of speculative assets vary inversely with how many other fun things there are to do. If your theory of Bitcoin is “it is a thing to speculate on” then, you know, that has no scarcity at all. 

And here is a Bloomberg News story about the crypto crash in which everyone is similarly confused:

“It felt like we were due to puke down to these levels at some point,” said Jack Melnick, who works in digital assets and also trades his own funds. He says that the crypto market is searching for a story, “a bull case thesis,” but that it will find one eventually.

“Eventually we’ll figure out why we own this stuff,” okay.

Elsewhere, someone emailed me to point out that Bitcoin’s price seemed to be pinned near $69,420 at various points over the past few days. (69 and 420 are meme numbers.) “Who could 1) be that immature and 2) have that much capital,” my reader asked. I know the answer he was looking for, but the actual answer is “every single Bitcoin whale.”

I used to think that Tether was a silly place to keep your money. You could, I thought, keep your money in a bank or a US dollar money market fund; those are regulated by the US government, hold your money in relatively safe and transparent places, and will let you use it for pretty normal payments. Or there was Tether, which put its money in truly insane places back in the day. Why would you want that? Why was it so popular with crypto traders? And then a crypto trader pointed out to me that, for a lot of people, being unregulated and secretive was a feature of Tether. “This account is regulated by the Federal Reserve and the Federal Deposit Insurance Corp.” makes some customers feel safer, and makes some other customers feel much less safe.

That was a long time ago. Now Tether keeps its money mostly in reasonable safe places, has a certain amount of transparency, is buddies with regulators and does stuff like this:

On Jan. 30, Turkish authorities announced the freezing of more than half a billion dollars in assets owned by Veysel Sahin, who is accused of running illegal betting platforms and laundering criminal proceeds. An unidentified crypto company carried out the freeze at Turkey’s request, Istanbul’s chief prosecutor said.

That company was Tether Holdings SA, issuer of the $185 billion stablecoin USDT — part of its recent push to aid governments around the world in crackdowns against crypto’s use in a range of alleged crimes, from money laundering to drug trafficking to evading sanctions.

“Law enforcement came to us, they provided some information, we looked at the information and we acted in respect of the laws of the country,” Tether Chief Executive Officer Paolo Ardoino said in a recent interview with Bloomberg News. “And that’s what we do when we work with the DOJ, when we work with the FBI, you name it.”

Presumably some Tether customers, or at least some former Tether customers, wish it was otherwise. Elsewhere in Tether:

Tether has expanded its eccentric venture capital portfolio and launched a hiring spree as it seeks to move beyond its roots as a secretive provider of crypto-financial plumbing and create a global conglomerate built around “freedom”.

The world’s largest stablecoin issuer has long been run by a small circle of executives, who manage its $185bn token USDT, which serves as the main bridge between crypto and dollars.

The group, registered in El Salvador but with a base in Switzerland, has begun to deploy its large profits to build a sprawling portfolio that now spans 140 investments from a South American agricultural producer to a stake in Italian football club Juventus.

I guess they have to spend some of it on compliance now.

The big question in artificial intelligence economics is: If you are a company that sells some sort of knowledge-work service, will AI make you more efficient, or will it make you worthless? The bull case for, say, a consulting firm is that you can use AI to replace a lot of your expensive employees, cutting costs and providing better and faster consulting, increasing your sales and margins. The bear case is that your clients can use AI to replace you, cutting their costs and getting better and faster consulting, decreasing your sales and margins.

This creates a weird dynamic where knowledge-industry companies are simultaneously saying to their employees “you are all worthless and we will replace you with machines” and saying to their customers “oooh our employees are amazing and could never be replaced by machines.” I mean, not quite, not in so many words, but kind of. Anyway this seems like a weird move:

KPMG, one of the world’s largest auditors of public and private companies, negotiated lower fees from its own accountant by arguing that AI will make it cheaper to do the work, according to people familiar with the matter.

The Big Four firm told its auditor, Grant Thornton UK, it should pass on cost savings from the rollout of AI and threatened to find a new accountant if it did not agree to a significant fee reduction, the people said.

The discussions last year came amid an industry-wide debate about the impact of new technology on audit firms’ business and traditional pricing models. Firms have invested heavily in AI to speed up the planning of audits and automate routine tasks, but it is not yet clear if this will generate savings that are passed on to clients.

“Ehh auditing can basically be done by AI so why should we pay for it” is not a crazy thing for most companies to think, or to say to their auditors, but it is a crazy thing for an auditing firm to say to its auditor. KPMG should be paying Grant Thornton more. “In these crazy AI times, everyone needs to pay more for trusted human auditing; we’ll go first.”

CLO captive equity. VC blended tranches. ‘They want their pound of flesh’: bank bosses push strict return-to-office mandates. Easy-Money Loans Backfire on Rookies in the Home Flipping Market. The Chinese ‘Auntie’ Investors Behind the Gold and Silver Frenzy. China Urges Banks to Curb Exposure to US Treasuries. Goldman Says Hedge Funds Add Record Shorts on US Stocks in Rout. Blue Owl Founders Pledged $1.9 Billion Stake Before Stock Plunge. Alphabet Looks to Raise About $15 Billion From US Bond Sale. Banks push for speedy European IPOs to cut market risk. FOMC Insight Engine. KKR Sees Arctos as Linchpin for Building a $100 Billion Business. US embassy in London denies visas to executives over minor offences. Job Hunters Are So Desperate That They’re Paying to Get Recruited. The New Ways High Schools Are Teaching Teens About Money. Harvard Proposes Capping A’s to Curb Grade Inflation. Musk Shifts Focus to Moon as Mars Mission Remains Years Away. Pivot to AI.com.

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can subscribe to Money Stuff and other great Bloomberg newsletters here. Thanks!