Money Stuff: Sports Bets at the Stock Exchange

You can think about financial markets at different levels of abstraction. Most working traders and analysts spend most of their time thinking at the level of making correct predictions: You analyze some data about the world, form a hypothesis about what some asset prices will do, and test the hypothesis by trading the assets. Often this analysis is informed by experience and connoisseurship and intuitive pattern-matching, though increasingly these days the pattern-matching is done using machine learning techniques. 

Then there is the level of adaptive niches: Why do these patterns exist, and why should you be able to make money trading on them? A pattern with no explanation is a bit suspect; you want some psychological or structural explanation for why asset prices will move the way you expect and why they haven’t already. [1] “Long-only asset managers like to take some credit risk and get their duration via futures, so we can make a bit of money buying cash Treasuries and selling futures” is an explanation of this form. (While “futures usually pay a bit more than cash Treasuries” is a first-level explanation.) Or “index funds rebalance on particular dates so if we buy the stocks they need to buy ahead of time we can sell to them at a profit.” Or “retail investors are buying meme stocks at irrational prices so I can make money very carefully selling them those stocks.” Some sort of story about regulations or institutional constraints or mass psychology or something, some satisfying intuitive reason for the patterns.

And then there is the level of social purpose. Why is anyone doing any of this? How would you explain what you do to your grandparents? There seem to be two main social purposes to financial markets, or one purpose looked at from two ends: Financial markets allow people to save and participate in economic growth to meet their future material needs, and they help allocate resources to their most productive uses. When you run your machine-learning model that looks at piles of market data to figure out which 500 stocks to buy and which 500 stocks to short, you are doing your small part to make prices efficient and allocate capital to its best uses. I used to be a corporate equity derivatives investment banker, and I once watched with uncomfortable awe as my boss explained that a derivatives deal we did for a biotech company lowered its cost of capital and thus allowed it to develop a promising cancer treatment, so if you really thought about it we were basically curing cancer. Did I believe that? Did he believe that? I mean. Perhaps his grandparents did.

These things all loosely tie together: If you are a hedge fund analyst, you largely spot patterns that exist because of some structural inefficiency, and then you trade against those patterns to profit from that inefficiency and make money for yourself, but your trading pushes the markets in the direction of efficiency and moves capital a bit closer to its best uses.

But there is no reason to think they will fit neatly together all of the time. “Retail investors are buying meme stocks at irrational prices, and they’re going to keep doing that for the next month, so I’d better buy meme stocks” is a perfectly satisfactory first- and second-level explanation that tends to allocate capital to bad uses and make retirement savers worse off. But if it makes money you should do it. Your overall role in the market is to nudge capital toward its best uses, even if sometimes you trade the opposite way to make a quick buck. [2]

One way to think about crypto is that it created an enormous lucrative playground for traditional finance people (or their interns) to practice their first- and second-level skills — spotting patterns and understanding market niches — without worrying at all about the third level. That is quite rude and not the only way to think about crypto: There are lots of crypto believers who argue that crypto has created important innovations for the world, so allocating capital to crypto projects is a social good even if many individual projects are pretty bubbly.

But my impression of the glee with which a lot of traders jumped into crypto was that they were not particularly worried about the social good of crypto, about allocating capital to its best uses. They had all these tools for pattern recognition, and in traditional financial markets their application was difficult and constrained by economic reality, but crypto was a pure abstract casino for traders. You could spot patterns and turn them into money without anyone having to believe anything at all about capital allocation. As I once said to Sam Bankman-Fried: “You’re just like, well, I’m in the Ponzi business and it’s pretty good.”

That was practice, man. Sports gambling is exactly like that, and it is being colonized by finance. Quantitative trading firms are dipping their toes into sports betting: If you are good at finding patterns in data that predict future outcomes, modern sports have tons of data and tons of outcomes. There are structural inefficiencies in how people bet, inefficiencies that provide steady opportunities for people who are good at the first-level skills to exploit. By doing that, do you push the point spreads of National Football League games toward efficiency? Sure, probably! Is that a social good? I mean! The answer is not so much “no” as “who cares?” It might be! You could tell a story like “people want to be entertained, a little sports gambling is a more cost-effective form of entertainment for many people than various alternatives, and accurate point spreads improve their well-being.” Who am I to judge. You’re not curing cancer, or allocating capital to curing cancer. You’re making it a bit more fun for some people to watch the Jets game, and a bit less fun for other people.

I used to say this sporadically, but now I say it all the time: Financial markets and sports gambling are merging, and some big financial firms now go around acting like there is no difference between them. We have talked a lot about Robinhood’s move to become a sports betting app, and now here’s this:

Intercontinental Exchange Inc., owner of the New York Stock Exchange, plans to invest as much as $2 billion in cash in Polymarket, a crypto-based betting platform.

The transaction values the company, which lets traders wager on the outcome of real-world events such as elections and sports, at roughly $8 billion, ICE said in a statement Tuesday.

ICE will become a global distributor of Polymarket’s event-driven data, providing customers with sentiment indicators on topics in the market, according to the statement. The exchange operator and Polymarket have also agreed to partner on future tokenization initiatives. …

“Together, we’re expanding how individuals and institutions use probabilities to understand and price the future,” [Polymarket founder Shayne] Coplan said in the statement. “Realizing the potential of new technologies, such as tokenization, will require collaboration between established market leaders and next-generation innovators,” he said.

Perhaps I am way too cynical. Polymarket’s and ICE’s statement barely mentions sports, and Polymarket is not really explicitly a sports betting platform: It’s a general-purpose prediction market, and a lot of its volume and popular perception involve election betting. (It traded billions of dollars on the 2024 US presidential election, and there’s currently $130 million of volume on the New York mayor’s race.) But Polymarket is reportedly mostly a sports market these days, and it sure seems to me that the rapidly rising valuations of prediction markets (and the falling valuations of traditional sportsbooks) have something to do with the fact that they have moved into sports betting in a big (and government- favored) way. Predicting election outcomes is a niche business with a lot of seasonality; predicting sports outcomes is a large popular lucrative constant business.

And so the story here is kind of “the New York Stock Exchange is getting into the sports betting business.” [3] Because the thing that you do, on a first-order basis, at the New York Stock Exchange, is make bets on stocks. You use gut instinct or data or whimsy or computers to pick the stocks you think will go up, and you buy them. And at some level the purpose of the New York Stock Exchange is to allocate capital to its best uses, but at another level the purpose of the New York Stock Exchange is to give you the bets you want, and if you want to bet on sports then bet on sports. [4]  Bloomberg’s Joe Weisenthal writes today:

It seems like every company in the trading world is melding into some prediction market/sports betting/crypto conglomerate. All the lines between speculation and hedging or gambling and investing are going away.

Or I’m completely wrong! The alternative story here is what Coplan says: “We’re expanding how individuals and institutions use probabilities to understand and price the future,” where “the future” means something broader than this weekend’s NFL slate. Economists have long been fond of prediction markets on efficient-allocation-of-capital, complete-markets grounds: If you had market prices reflecting the probabilities of lots of important future events, you could better understand the world, be more prepared for the future, and hedge bad potential futures. Right now, the business of prediction markets involves a lot of sports betting, but that won’t necessarily be true forever. Perhaps the path to a robust prediction market for understanding and pricing the future runs through sports, because that’s what people already want, but it ends up in a more interesting and useful place. I wrote the other day:

One way to think about it is the Grossman-Stiglitz paradox: To get an efficient market (here, prediction markets that accurately predict important events), you need to create conditions where informed traders can make a lot of money making prices correct, which means that you need some noise traders willing to trade at incorrect prices to entice the informed traders. And you entice the noise traders with football.

ICE’s and Polymarket’s overall role is to nudge capital toward its best uses, even if sometimes they trade the opposite way to make a quick buck.

Everything is securities fraud, I often say, but this is the most securities fraud [5] :

1. I buy 1 million shares shares of Penny Stock X for $0.01 each.
2. I go on social media and tell my hundreds of thousands of followers, “I just got off the phone with the chief executive officer of Penny Stock X, she tells me that they used their proprietary quantum computers to find a cure for cancer, this is going to add $500 billion to their market cap, you’d better buy as much as you can right now.”
3. My hundreds of thousands of followers believe me, [6]  so they rush to buy the stock.
4. Their buying pushes the stock up to $0.20 per share.
5. I sell my million shares for $200,000, a $190,000 profit.
6. Obviously the stuff about quantum computers and a cure for cancer is a lie and the stock falls back to $0.01 per share.

This is called a “pump and dump.” In Step 2, I lie to pump up the price of the stock. In step 5, I dump the stock that I pumped. It is a very standard form of securities fraud, people have been doing it in some form or another for many decades, the US Securities and Exchange Commission and the Department of Justice regularly bring cases against people who try it, and they regularly win those cases. 

But is this securities fraud? I mean, yes, it absolutely is, come on. But could you make an argument that it isn’t? The argument that it isn’t might go like this:

1. “Fraud” means lying to someone to get money from them.
2. “Securities fraud” means lying to someone about securities to get money from them. [7]
3. Here, yes, sure, I lied to my social media followers about securities.
4. But I didn’t get money from them!
5. Oh, I got money. But I got money from the stock market. I got money from hitting the “sell” button at my brokerage account. The person who bought my stock from me gave me money, but there’s no particular reason to think that that person read my social media posts. Perhaps my broker sent my order to the stock exchange, and my stock was bought in an anonymous transaction and we’ll never know who bought it. More likely my broker sent my order to an electronic market maker like Virtu or Jane Street or Citadel Securities, who bought my stock from me, and none of those guys even follow me on social media.
6. Similarly, my social media followers spent money to buy the (worthless) stock. They sent money by hitting the “buy” button at their brokerage accounts. But, again, they don’t know whom they bought the stock from, and there’s no reason to think it was me. It was probably Virtu or Jane Street or Citadel Securities.
7. So whom did I defraud? I lied to my followers, but they didn’t give me money. The market makers gave me money, but I never lied to them. No fraud!

This is mostly stupid. The money in some obvious intuitive sense flows from my social media followers (who have been told lies, and believed them, and paid money) to me (who told the lies and got the money). It didn’t flow directly from them to me; there were a few steps in the middle. But those were all pretty mechanical steps. The market makers — and the market — respond to the information encoded in prices. When my followers saw my lies and started buying the stock, that mechanically pushed the stock price up. The higher stock price reflected demand from my followers, which in turn reflected my lies. The higher stock price was a lie: The higher stock price encoded the information “this company cured cancer with quantum computers,” which was false. In some sense I did trick Virtu or Jane Street or whoever was on the other side of the trade from me: They didn’t see or believe my lies directly, but their picture of the world was distorted by my lies, and I expected and intended and caused and profited from that distortion.

This is called the “fraud on the market theory,” and it is essential to modern securities fraud enforcement: Modern securities trading mostly occurs electronically and anonymously, so you’ll rarely be able to find cases where I lied about securities directly to the person who bought them from me. The US Supreme Court said in 1988:

The fraud on the market theory is based on the hypothesis that, in an open and developed securities market, the price of a company’s stock is determined by the available material information regarding the company and its business. Misleading statements will therefore defraud purchasers of stock even if the purchasers do not directly rely on the misstatements. The causal connection between the defendants’ fraud and the plaintiffs’ purchase of stock in such a case is no less significant than in a case of direct reliance on misrepresentations. 

This is all obvious stuff, this is the basic architecture of modern financial regulation, this is why people are not allowed to go around lying about stocks all the time. I apologize for explaining it at length, because, while many people probably haven’t thought about it, nobody is actually confused about it. Nobody goes around thinking “actually pump and dumps are legal because you’re not taking money directly from the people you lie to.”

Except this one judge? We talked in March of last year about what I called “an absolutely wild opinion” in a Texas federal district court, dismissing an indictment against some alleged pump-and-dumpers — who had nicknames like “MrZackMorris” and “Mystic Mac” and “The Stock Sniper” and who said stuff like “I’ll never get sick of pumping ... money into my followers bank accounts,” because the judge found that actually pump-and-dumps aren’t fraud. He wrote:

Unlike a traditional fraud case, in which the victim directly surrenders their property to the defendant (or an entity in the defendant’s control), the investors here surrendered their property to the stock market at market prices, and in return, received the benefit of the bargain in the form of securities. Thus, the scheme did not deprive investors of their money or property through any misrepresentation; the misrepresentations deprived them only of accurate information necessary to make discretionary economic decisions. ...

Even accepting as true that the alleged victims ultimately lost money on the stock market because the value of their shares went down, the Defendants did not obtain something of value from the entity to be deceived. The investors’ trading losses are one step too far removed from the Defendants’ alleged fraudulent misrepresentations.

That was weird! But maybe not that weird. The Supreme Court has issued a bunch of opinions over recent years narrowing the range of fraud and corruption laws, finding that various sorts of obvious fraud are not actually fraud because they do not “deprive people of traditional property interests.” One possibility was that the Texas judge looked at those precedents, thought “ehh they’re legalizing everything these days, probably pump-and-dumps are fine,” and got out ahead of things. One possibility was that he was right! Maybe that is where things are going. “Pump and Dumps are Legal Now,” was my headline at the time. That was not legal advice, and I didn’t really believe it, but I worried. If it’s right, I wrote, “the stock market, and social media, are going to get pretty weird.” I want to point out that things really have gotten a lot weirder since then (you can bet on football on Robinhood!), though mostly not because of that decision.

That said, last Thursday that opinion was overturned on appeal in a short, sensible, unanimous opinion of the US Court of Appeals for the Fifth Circuit:

While it is true that a defendant cannot be convicted of fraud for depriving an individual of potentially valuable economic information alone, the indictment here went further, properly alleging defendants’ scheme to defraud their followers of money. ...

The indictment does not mention the right-to-control theory. Rather, it alleges that defendants induced their followers, through misrepresentations, to purchase securities — that is, to part with their money or property. The indictment alleges a “fraudulent-inducement theory,” under which a defendant “uses a material misstatement to trick a victim into a contract that requires handing over her money or property.” …  Unlike the right-to-control theory rejected in Ciminelli, the fraudulent-inducement theory advanced here “protects money and property.” ...

As the indictment alleges, defendants’ object was to obtain money, and they did so by fraudulently inducing their followers to purchase securities. The indictment’s allegation that defendants pumped and dumped to gain a profit shows that they “had money in mind.” … “[A] fraud is complete when the defendant has induced the deprivation of money or property under materially false pretenses”—regardless of whether the victim suffered a net pecuniary loss. The indictment’s allegation of defendants’ fraudulently inducing their followers to purchase securities is sufficient to allege an injury.

So pump and dumps are illegal again. For now. I guess this could still go to the Supreme Court.

The two basic possible paths for generative artificial intelligence and the consulting industry are:

1. Consultants will master AI to produce better and more useful consulting engagements with fewer resources, making themselves more indispensable to clients while also improving their own profit margins; or
2. Clients will master AI well enough that, instead of paying a consulting firm for advice about critical business decisions, they can just ask ChatGPT and get a good enough answer, making the consultants completely dispensable and putting severe pressure on their margins.

Not to pick on consultants; really those are the two possible paths for most sorts of knowledge work. We’ve talked about the same conundrum for homeowners’ association management companies; also, please do not ask ChatGPT to write Money Stuff for you. But especially consultants.

So if I ran a consulting firm I would be very interested in simultaneously:

• Making sure that my consultants were really good and creative and cutting-edge about using AI to do consulting; and
• Making sure that their work product was highly differentiated from the raw output of large language models. If I was charging clients hundreds of thousands of dollars to answer a question like “how can we increase sales of our widgets,” I would be very very careful to give them answers that were not simply what you’d get by typing the same question into ChatGPT.

Deloitte apparently made different choices? The Financial Times reports:

Deloitte will partially refund payment for an Australian government report that contained multiple errors after admitting it was partly produced by artificial intelligence. ...

Australia’s Department of Employment and Workplace Relations ... had commissioned a A$439,000 “independent assurance review” from Deloitte in December last year to help assess problems with a welfare system for automatically penalising jobseekers. …

In late August the Australian Financial Review reported that the document contained multiple errors, including references and citations to non-existent reports by academics at the universities of Sydney and Lund in Sweden. ...

In the updated version of the report, Deloitte added references to the use of generative AI in its appendix. … While Deloitte did not state that AI caused the mistakes in its original report, it admitted that the updated version corrected errors with citations, references, and one summary of legal proceedings.

“Well if we just type our question into ChatGPT we’ll get a decent answer, but it might contain hallucinations, whereas if we pay hundreds of thousands of dollars to a big consulting firm we can at least avoid that risk,” you might think, incorrectly.

Buddy:

Glancy Prongay & Murray LLP, a leading national shareholder rights law firm, today announced that it has commenced an investigation on behalf of Firefly Aerospace Inc. (“Firefly” or the “Company”) (NASDAQ: FLY) investors concerning the Company’s possible violations of the federal securities laws. ...

On September 29, 2025, Firefly disclosed that during testing, the first stage of its Alpha Flight 7 rocket “experienced an event that resulted in a loss of the stage.”

On this news, Firefly’s stock price fell $10.29, or 27.8%, over two consecutive trading days to close at $26.67 per share on October 1, 2025, thereby injuring investors.

I should say that the answer to the question “if a rocket explodes, is that securities fraud” is not always “duh of course everything is securities fraud.” If a NASA rocket explodes, probably not securities fraud: NASA is a government agency and does not have securities. [8]  If a SpaceX rocket explodes, probably not securities fraud: SpaceX is a private company, its shares do not trade regularly in a public market, and so there is (probably) not some pool of investors who bought the stock last week and can say with a straight face “we overpaid for the stock because we thought your rockets wouldn’t explode but then your rocket exploded so, fraud.” But every bad thing that a public company does is securities fraud, and Firefly is public, and there you go.

OpenAI’s Golden Touch Spreads as Stocks Soar Off Mere Mentions. Investors Place Bets on Trump Team’s Next Investment Target. Elon Musk names former Morgan Stanley banker as new xAI CFO. Market for Synthetic Risk Transfers Is Riddled With Gaps in Disclosure, IMF Officials Say. Rising Fees Force Premium Credit-Card Holders to Choose Sides. Bitcoin Life Insurance Provider Meanwhile Raises $82 Million. CFA Level I Success Drops to 43% as AI Models Ace Final Exam. JPMorgan replaces European banking boss who was doing job from New York. JPMorgan paid $115 million for Charlie Javice’s lawyers. Fannie Mae and Freddie Mac have “perhaps the strangest IPO ‘bake-off’ ever.”

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!