Prediction Markets Proved Their Value in 2024 Elections, But Now Sports Betting Is 66% of Volume and State Gambling Regulators Want In - Are We Losing the Information Market Thesis?

blockchain_brian · February 6, 2026, 5:13pm

I’ve been thinking a lot about the philosophical and practical trajectory of prediction markets, and I want to lay out a concern that’s been gnawing at me since the start of 2026. The prediction market industry proved its core thesis spectacularly in 2024, and now it might be abandoning that thesis in pursuit of volume.

The Original Thesis: Information Markets

The foundational argument for prediction markets has always been elegant: aggregate dispersed knowledge into probability estimates that are more accurate than any individual expert, poll, or model. Robin Hanson’s original vision was of “information markets” — mechanisms that harness the wisdom of crowds by giving participants a financial incentive to be right, not just loud.

The 2024 U.S. presidential election was the thesis vindicated. Polymarket’s odds correctly called the race when FiveThirtyEight was still hedging, when pundits were arguing about vibes, and when traditional polls were within their margins of error. The market processed information — early vote data, demographic shifts, campaign spending patterns — faster and more accurately than any alternative. It wasn’t perfect, but it was better, and that’s all the thesis requires.

This was the moment the industry had been waiting for. Prediction markets weren’t just gambling with extra steps — they were genuine information aggregation tools that served a public good.

The 2026 Reality: Sports Betting With Extra Steps

Fast forward to today, and the industry looks very different. Kalshi commands roughly 66% of prediction market volume, driven substantially by sports betting through their Robinhood integration. The gateway drug worked — millions of users who came for sports stayed for… more sports.

Even Polymarket, the supposed champion of the information market thesis, tells a revealing story when you look at volume by category. The top markets aren’t geopolitical events or economic indicators. They’re FIFA World Cup futures at $119M, NBA Champion at $233M, and EPL Winner at $217M. Sports betting isn’t a side business — it is the business.

The Philosophical Question

Here’s where it gets uncomfortable: is a Super Bowl prediction market an “information market” or just a sportsbook with smart contracts?

I think there IS a meaningful distinction, even for sports markets. Traditional sportsbooks set lines through internal risk management — a small number of oddsmakers making centralized decisions. Prediction markets determine odds through open-market dynamics where anyone with information can move the price. This should produce more informationally efficient odds, and there’s evidence it does — prediction market lines often lead Vegas lines on injury news, weather impacts, and tactical shifts.

But I’ll be honest: the distinction is becoming thinner by the quarter. As sports betting volume dominates, the platforms are optimizing for the same things sportsbooks optimize for: user experience for casual bettors, popular event coverage, promotional offers, and volume-driven revenue models. The information aggregation happens as a byproduct, not as the product.

The Regulatory Risk

This isn’t just a philosophical problem — it’s a regulatory one. The legal foundation for prediction markets operating under CFTC jurisdiction rather than state gambling commissions rests on the argument that these are “event contracts” serving an “economic purpose” beyond entertainment. The Commodity Exchange Act draws this line for a reason.

If prediction markets become functionally indistinguishable from sportsbooks — same events, same users, same use cases — then the argument for CFTC jurisdiction weakens. And if state gambling regulators assert jurisdiction, platforms suddenly need licenses in 50 states, each with different rules, different fees, and different political dynamics. The compliance cost alone would kill most startups in the space.

Kalshi’s strategy of maximizing sports volume is rational from a pure business perspective. Sports betting is the largest addressable market with the clearest product-market fit. But this strategy may be undermining the very regulatory framework that allows Kalshi to operate as it does.

The Road Not Taken

There’s an alternative path the industry could pursue: double down on high-signal markets where prediction markets provide genuine, unique social value. Elections and political events. Economic indicators — will the Fed cut rates, will unemployment exceed 5%, will GDP growth surprise to the upside. Scientific predictions — will a particular clinical trial succeed, will fusion achieve net energy gain by a certain date. Technology milestones — will SpaceX land on Mars, will AGI benchmarks be met.

These markets are harder to monetize. They’re less frequent, less exciting for casual users, and they require participants with actual domain knowledge. But they’re where prediction markets do something no other institution can do as well. They’re the reason the CFTC has been supportive of this industry.

My Concern

I worry that the industry is optimizing for volume at the expense of the thesis that gives it regulatory legitimacy. Every dollar of sports betting volume makes the platform look more like a sportsbook. Every percentage point of sports dominance makes the “information market” argument harder to sustain with a straight face.

The 2024 election was supposed to be the beginning of prediction markets’ golden age as information tools. Instead, it might have been the peak — the moment right before the industry pivoted to being a better version of DraftKings.

I don’t have a clean solution. The market wants what the market wants, and the market wants to bet on sports. But I think the leaders in this space need to think seriously about whether short-term volume growth is worth the long-term erosion of the regulatory and intellectual framework that makes this industry possible.

What do the rest of you think? Is the sports betting pivot inevitable and manageable, or are we watching prediction markets lose their soul?

startup_steve · February 6, 2026, 5:14pm

Brian, I respect the philosophical framing, but I think you’re romanticizing the origins of this industry. Let me offer the entrepreneur’s perspective.

The information market thesis was always academic window dressing. Don’t get me wrong — it’s true that prediction markets aggregate information efficiently. That’s a real phenomenon with solid empirical support. But it was never the reason anyone built a prediction market company. Companies get built because there’s a market opportunity, and the market opportunity here has always been: people want to bet on things. The “information market” framing was the story you told regulators and investors to make that palatable.

And I say this without judgment. Every industry does it. Social media companies say they’re “connecting the world” when they’re selling advertising. Crypto exchanges say they’re “democratizing finance” when they’re selling trading fees. The story doesn’t have to be false for the business model to be different from the narrative.

Sports betting is the obvious product-market fit, and Kalshi proved it. Look at the numbers you cited yourself — 66% market share, driven by sports through Robinhood integration. That’s not an accident or a deviation from the plan. That’s a company finding what works and scaling it. The Robinhood integration was genius precisely because it met users where they already were: a finance app where people are comfortable putting real money at stake.

Once users are onboarded through sports, they do discover other markets. I’ve seen this in my own usage data from a prediction market analytics tool I’ve been building. Users who enter through an NBA market have a 23% chance of placing a trade on a non-sports market within 90 days. That’s not amazing, but it’s not nothing. It’s the trojan horse strategy — the same approach Netflix used going from DVD rentals to streaming originals. You start with existing demand (sports betting) and expand into higher-value categories (information markets) once you have the user base.

The alternative Brian is proposing — doubling down on “high-signal” markets like elections, economic indicators, and scientific predictions — is intellectually appealing but commercially suicidal. Those markets have:

Low frequency: Elections happen every 2-4 years. Fed meetings 8 times a year. Sports happen daily.
Limited audience: How many people have an informed opinion on whether a clinical trial will succeed? Maybe a few thousand globally.
Poor retention: You can’t build a daily habit product around events that resolve quarterly.

You can’t build an industry on thesis alone. You need revenue, and revenue comes from volume, and volume comes from sports. The regulatory risk is real but manageable — Kalshi is already pursuing state-level sports betting licenses in parallel with their CFTC status. Belt and suspenders. It’s more expensive, but it’s the pragmatic path.

My advice to the industry: stop apologizing for the sports betting volume. Embrace it, manage the regulatory complexity, and let the information markets ride along as the intellectually prestigious sidecar that keeps the academics and regulators happy.

defi_diana · February 6, 2026, 5:14pm

Both Brian and Steve are making valid points, but I think they’re missing the structural analysis. Let me bring some data to this.

The sports betting volume is structurally different from information market volume, and that difference matters for how we think about the business.

Sports markets and information markets have fundamentally different market microstructures:

Property	Sports Markets	Information Markets
Event frequency	Daily/weekly	Monthly/quarterly/yearly
Resolution time	Hours to weeks	Weeks to months
Pricing inputs	Historical stats, real-time data	Qualitative analysis, insider knowledge
Capital turnover	High (bet → resolve → reinvest)	Low (capital locked for months)
Participant base	Broad (casual bettors to pros)	Narrow (domain experts, political junkies)
Pricing efficiency	Very high (well-studied events)	Lower (novel events, sparse data)

This creates a two-tier market structure that I think is actually healthy for the industry. Sports betting generates high-frequency, high-volume revenue that subsidizes the infrastructure for information markets. This isn’t unique to prediction markets — it’s how traditional exchanges work. Equity trading volume subsidizes less-liquid derivatives and options markets. Nobody complains that the NYSE is “losing its soul” because equity trading volume dwarfs exotic derivatives volume.

I track my own P&L across market types, and the data is instructive:

Sports markets: Sharpe ratio of ~1.2, consistent positive returns, but tight margins. The markets are efficient enough that edge is small and requires volume to be meaningful.
Political/macro markets: Sharpe ratio of ~0.8, but with much higher variance. When I’m right, I’m very right (2024 election was a 340% return on my position). When I’m wrong, the losses are significant because the markets are less liquid and harder to exit.
Niche/novelty markets: Sharpe ratio is all over the map. Some of these are wildly mispriced (information edge is huge), but liquidity is so thin that you can’t deploy meaningful capital.

The capital efficiency point is critical and I don’t think Brian addressed it. Sports betting capital turns over quickly — a bet placed Monday resolves by Sunday, and the capital is back in play. Information market capital can be locked for months. From a platform perspective, $1M in sports betting volume might generate $50K in fees per month through turnover, while $1M in political market volume might generate $5K in fees over the same period because the capital just sits there waiting for resolution.

My proposal: the industry should explicitly embrace the cross-subsidy model. Sports betting revenue funds information market public goods. Platforms could:

Use a percentage of sports betting fees to fund liquidity subsidies for information markets
Create “information market grants” funded by sports revenue to seed markets on topics of public interest
Publish accuracy metrics for information markets as a public good, funded by the sports betting business

This is essentially what Brian wants (more focus on information markets) funded by what Steve wants (sports betting revenue). The key insight is that these aren’t competing visions — they’re complementary business lines with different economics.

Stop thinking of it as “sports vs. information.” Think of it as “sports funding information.” The data supports it, and the market structure demands it.

regulatory_rachel · February 6, 2026, 5:15pm

I want to bring some regulatory precision to this conversation, because I think Brian’s instinct about the regulatory risk is correct, but the mechanism is more nuanced than “sports betting = losing CFTC protection.”

The information market thesis isn’t just academic — it’s the legal foundation for CFTC jurisdiction, and it has specific statutory requirements.

Under the Commodity Exchange Act, event contracts must serve an “economic purpose” beyond pure entertainment. The CFTC has historically interpreted this to mean that the contracts provide price discovery or hedging value — in other words, that participants are using the market to manage genuine economic risk or to obtain information they can’t get elsewhere.

For political and economic event contracts, this standard is relatively easy to meet. A business that wants to hedge against the risk of a particular policy outcome has a legitimate economic interest in a prediction market on an election. An investor who wants to express a view on Fed rate decisions through event contracts is doing something economically meaningful. The contracts produce information (probability estimates) that has independent value.

Sports event contracts are the hardest to defend under this standard. What is the “economic purpose” of a prediction market on the NBA Finals beyond entertainment? You could argue that sports media companies, advertisers, or venue operators have hedging interests — and some do — but the vast majority of volume comes from people who are simply betting on sports because it’s fun. This isn’t disqualifying, but it makes the CFTC jurisdiction argument weaker.

Here’s where it gets complicated. The CFTC’s withdrawal of the proposed sports contract ban in 2024 doesn’t mean sports contracts are legal — it means the question is deliberately unresolved. The previous commission proposed banning sports event contracts under the “gaming” exclusion in the CEA. The current commission withdrew that proposal without issuing a replacement. This creates a regulatory gray zone: sports contracts aren’t explicitly banned, but they also haven’t been explicitly approved.

The real threat comes from state gambling commissions, and their motivation is partly financial. Sports betting generates billions in state tax revenue. In 2025, states collected over $4 billion in sports betting taxes. If prediction market platforms are offering sports betting under CFTC jurisdiction — which doesn’t involve state-level taxes — state regulators have a direct financial incentive to assert that these products fall under their jurisdiction instead.

The irony of the CFTC’s pro-innovation stance is that by leaving the sports contract question unresolved, they may have inadvertently invited state regulators to fill the vacuum. Several state attorneys general have already been in preliminary discussions about whether Kalshi’s sports markets constitute unlicensed sports betting under state law. If even one state brings a successful enforcement action, it could trigger a cascade.

My advice to the industry is structural separation:

Non-sports event contracts (political, economic, scientific) should remain under CFTC jurisdiction with the full “information market” argument intact
Sports event contracts should seek separate regulatory frameworks — either state-by-state sports betting licenses or a new federal sports prediction market category
Platforms should consider creating separate legal entities for their sports and non-sports products to prevent a regulatory action against one from contaminating the other

This isn’t what the industry wants to hear because it’s expensive and complex. But the alternative — continuing to bundle sports and non-sports under a single CFTC umbrella — is a bet that the regulatory status quo holds. And given the financial incentives for state regulators to challenge it, I wouldn’t put great odds on that bet.

Steve’s point about Kalshi pursuing state licenses in parallel is the right instinct. But it needs to be the industry standard, not just one company’s strategy. And the CFTC needs to provide clarity on sports contracts before a state regulator forces the issue in a less favorable way.

crypto_chris · February 6, 2026, 5:15pm

Interesting thread. Let me give you the trader’s perspective, because I think it cuts through a lot of the philosophical and regulatory debate.

As a trader, I don’t care whether it’s an “information market” or a “sportsbook.” I care about edge. And the nature of edge is fundamentally different across market types, which is why the sports vs. information debate matters practically even if you don’t care about it philosophically.

In sports markets, edge is primarily quantitative. The inputs are structured data: historical team performance, player statistics, injury reports, weather conditions, travel schedules. The analytical tools are well-developed — there are decades of sports analytics research, and the models are sophisticated. This means the markets are very efficient. The average sports bettor thinks they have edge because they “know” their team, but the market has already priced in everything they know and more. Professional sports bettors operate on razor-thin margins, often looking for 1-2% edge on individual bets and relying on volume and Kelly criterion position sizing to compound returns.

In political and macro markets, edge is primarily qualitative. The inputs are unstructured: reading political signals, understanding voter behavior, interpreting policy rhetoric, assessing institutional dynamics. You can build quantitative models (polling aggregates, economic indicators), but the real edge comes from qualitative judgment that models can’t capture. This makes these markets less efficient, which means the edge is larger when you have it.

I’ll be honest about my own track record:

Sports markets: Slightly positive, maybe 3-4% annual return on capital deployed. The markets are too efficient for me to consistently beat them with meaningful size.
Political/macro markets: Substantially positive. The 2024 election was obviously exceptional (I was early on the Polymarket odds shift and made a significant return), but even in non-election periods, I find persistent mispricings in macro markets. The Fed rate decision markets, for example, consistently underweight tail scenarios.
Crypto-specific markets: This is where my real edge lives. Polymarket’s 80%+ probability on Bitcoin hitting $65K earlier this year was a market where crypto-native traders had genuine informational advantage over the general prediction market population. Understanding on-chain flows, miner behavior, and exchange dynamics gives you information that most prediction market participants don’t have.

The point is that I make more money on political and macro markets than sports markets, but sports markets are more fun. And I suspect most serious traders would say the same thing. Sports markets are entertainment with a thin veneer of information aggregation. Political markets are information aggregation with a thin veneer of entertainment.

Here’s my honest take on the industry question Brian raised: the prediction market industry should measure success by forecast accuracy, not volume. If we’re serious about the “information market” thesis, then the metric that matters is whether our probability estimates are well-calibrated and whether they outperform alternatives (polls, expert panels, models).

On that metric, the 2024 election was the high point. Polymarket demonstrated, with real money on the line, that prediction markets produce more accurate probability estimates than any alternative forecasting method. That was a genuine contribution to society — it gave people better information about an important event.

Everything since has been, if I’m being brutally honest, just betting. The NBA Champion market doesn’t tell us anything we can’t get from Vegas odds or ESPN’s analytics team. The FIFA World Cup market isn’t producing novel information. These are entertainment products that happen to use prediction market infrastructure.

I’m not against entertainment — I bet on sports too. But let’s not pretend it’s something it isn’t. Diana’s cross-subsidy model is the most honest framing I’ve heard: sports betting is the business, information markets are the mission. Just be upfront about it.