The Bizarre World of Prediction Markets

ELI5/TLDR

Prediction markets — Kalshi, Polymarket, and the like — have rebranded gambling as “event contracts” and slipped through a crack in federal commodities law. The pitch is that these are truth machines where putting money on the line produces accurate forecasts. The reality is thinly traded venues where a few million dollars can move the odds, hedge fund quants systematically extract money from retail punters, and insider trading is defended as a design feature. Boyle’s verdict: a financial product too sophisticated to be called gambling and too simple to be called investing, operating in a regulatory gray zone because no one in Washington can agree on what it is.

The Full Story

The onion clause and the CFTC’s drift

American commodities regulation started with wheat and cotton. The CFTC’s job was to make sure futures contracts served a real hedging or price discovery function — the “economic purpose test.” Over time, the exchanges noticed that interest rate and stock index futures generated far more revenue than pork bellies, and the definition of “commodity” quietly stretched to accommodate them. Bitcoin now counts. The lone exception is onions. A 1958 statute banned onion futures after two traders cornered the Chicago market, which means you can now legally bet on a geopolitical conflict, a meme token, or control of Congress — but not on the price of the thing in your French onion soup.

If you attempt to hedge your exposure to French onion soup, or really any onion soup, the federal government will step in to protect the public from you.

From election contracts to football

The CFTC had explicitly banned event contracts tied to war, terrorism, assassination, and gaming, and tried to block election contracts too — quite sensibly, since approving them would turn the commodities regulator into an election monitor. Kalshi sued. A federal judge sided with the platform. Having won, Kalshi then pushed further: if a presidential election is just an event, a Super Bowl is also just an event. Earlier this year they began self-certifying sports contracts on the NBA, the Masters, and the Super Bowl.

This was awkward for the states. Since the Supreme Court struck down the federal sports betting ban in 2018, nearly 40 states had built licensing regimes, compliance departments, and tax pipelines. Kalshi turned up claiming its functionally identical product was exempt from all of it because federal commodity swaps override state gambling law.

State pushback and an 18th-century lawsuit

Arizona filed criminal charges. Ohio produced something weirder: a private suit against Kalshi invoking the Statute of Anne, a 1710 British law that lets third parties recover other people’s gambling losses. Boyle notes this is the sort of legal instrument you’d expect to find in a museum rather than an active federal lawsuit against a Y Combinator startup.

The federal response has been stranger still. The CFTC and DOJ went to court to block Arizona from enforcing its own gambling laws against Kalshi. One detail worth registering: Donald Trump Jr. is a strategic adviser to both Kalshi and Polymarket.

I can’t think of why they hired him, but I suppose it’s still worth noting that the president’s son advises the companies that the federal government is currently shielding from state prosecutors. I’m sure that it’s all a coincidence.

The truth machine, in theory

The pitch: when people back opinions with money, biases strip away and the market price becomes an objective probability. The efficient markets hypothesis applied to elections, wars, weather, and Bad Bunny’s wardrobe.

The problem is scale. These markets are thin, so moving the odds is cheap. In 2012, a single trader dropped roughly $7 million on Intrade buying Mitt Romney contracts — not to win the bet, but to make the race look tighter and keep voter enthusiasm up. Cable news covered the Intrade odds nonstop. In 2021, YouTuber Brian Rose allegedly had associates place bets on his London mayoral bid so he could point journalists at the “momentum.” A truth machine that a few million dollars can rewrite is a PR tool with a chart.

Financial nihilism and the quants

Boyle borrows Demetri Kofinas’s term financial nihilism: young people who feel the traditional wealth-building path is closed and who substitute lottery tickets — meme tokens, bankrupt-company shares, now event contracts. Crypto has gone dull (Bitcoin up ~25% over five years; a 4% money market fund got most of the way there), so the energy has rotated to prediction markets, where at least you can watch the thing you bet on.

Where retail money pools, quants follow. Susquehanna and DRW are reportedly staffing dedicated prediction market desks, paying $200k base salaries to traders building algorithms that hunt mispriced contracts. The sharks and fish problem from mid-2000s online poker is the template — amateurs log on, pros (and then bots) show up, recreational survival time collapses, the fish stop logging in, liquidity dries up.

If you’re a retail trader betting on a geopolitical event based on a feeling and the person on the other side of your trade is a gamma neutral algorithm being run by a multi-billion dollar hedge fund, the odds are not in your favor. This is not a skill gap that can be closed by doing more research. It’s a structural disadvantage.

One actual advantage

Prediction markets are peer-to-peer exchanges. They match buyers and sellers and collect a fee either way, so they have no reason to throttle winners — which is exactly what traditional sportsbooks do to anyone who starts winning consistently. That makes prediction markets structurally fairer than DraftKings. Of course, if you’re winning, you’re probably a quantitative algorithm, so this is mostly good news for quantitative algorithms. Meanwhile DraftKings, FanDuel, and Fanatics have quietly launched their own prediction market products while also spending $48 million on a super PAC pushing sports betting legalization in Texas and Georgia — fighting and copying at the same time.

Insider trading as a “feature”

Last summer, a Polymarket user named Rico Suave 666 made precise, lucrative bets on the exact timing of Middle East military strikes. Israel later arrested two men, including an army reservist, for allegedly betting on classified operational intelligence. Similar pattern around the capture of Nicolás Maduro — large confident bets placed on Polymarket shortly after the US announced the operation.

Platform advocates argue this is fine, even good: the insider brings information, the price adjusts, society gets a better forecast.

If a military officer leaks classified operational plans so that his friend can win a few hundred thousand on a cryptobetting site, we should apparently all be grateful for the positive externality of slightly more accurate price discovery.

Boyle’s counter is straightforward. Insider trading bans in equities exist because markets only function when participants believe the game is at least roughly fair. Strong US institutions are why American companies can raise capital. Erode that trust and people log off — and the erosion doesn’t stay confined to prediction markets if the same regulators oversee both.

Since 2018, the US has been running a natural experiment in phone-based gambling. Recent academic work flagged by The Economist shows the introduction of online betting is associated with roughly a 12-point drop in average credit scores in affected states, plus higher bankruptcies and loan delinquencies. Adults can spend their money how they like, but bankruptcies eventually land on safety nets funded by the taxpayers who didn’t bet their rent.

Key Takeaways

Prediction markets exist in a regulatory gray zone built on the argument that a football bet is a “commodity swap” rather than a wager.
The CFTC’s economic purpose test eroded slowly over decades; event contracts are the endgame. Onions remain the only thing you legally can’t bet on.
Federal–state conflict is active: Arizona has filed criminal charges against Kalshi; DOJ is defending Kalshi against Arizona. An Ohio suit is using a 1710 British statute.
Thin markets make price manipulation cheap. $7M on Intrade in 2012 kept the Romney–Obama narrative close on cable news.
Quant desks at Susquehanna and DRW are industrializing the extraction side. Retail on the other side of those trades is structurally, not informationally, outmatched.
Insider trading on Polymarket is treated by advocates as a feature. Known cases: IDF reservist on Middle East strike timing; unknown actors on the Maduro capture.
Unlike sportsbooks, prediction markets don’t restrict winning accounts — because they’re exchanges, not house-books. Cold comfort when the winners are algorithms.
Online betting legalization correlates with ~12-point credit score drops and rising bankruptcies at the state level.
Donald Trump Jr. advises both Kalshi and Polymarket. The federal government is actively shielding those platforms from state prosecution.

Claude’s Take

Boyle is doing what Boyle does well — taking a fashionable financial product and methodically walking through the places where the marketing copy doesn’t match the plumbing. The truth-machine argument has always had a small-market problem that its advocates wave away, and he hits it cleanly: if a few million dollars can move the odds and the media treats the odds as signal, you have a PR laundering service, not price discovery.

The strongest section is the sharks-and-fish analogy. It’s not speculation — it’s pattern-matching to a case that already ran its full cycle (online poker, 2003-ish to collapse) with the same structural setup. That’s more useful than most prediction market critiques, which tend to get tangled in the philosophical case for or against gambling.

Where he’s a little too neat: the “wealth transfer mechanism” framing is rhetorically satisfying but applies to most zero-sum leveraged markets, including plenty that raise real capital elsewhere in the chain. And the credit-score study he cites is one data point — worth noting, not yet load-bearing. The Trump Jr. detour is fair game but does more work for tone than for argument.

Score 8/10. Tight, funny, well-sourced, genuinely useful if you’ve been hearing about Kalshi and wondering what the fuss is. Loses a point because the central critique (thin markets, quant extraction, insider leakage) isn’t new to anyone who followed Intrade, and another half-point for leaning on the “president’s son” detail harder than the evidence supports. Gains it back for the onion framing, which is the kind of regulatory absurdity that’s funnier the more you think about it.