Polymarket, Kalshi, and the rise of prediction markets: what traders actually need to know

For most of the last decade, prediction markets sat in a strange corner of the financial-and-gambling crossover, taken seriously by a small community of academics, betting-market researchers, and crypto-native users, and ignored by almost e

For most of the last decade, prediction markets sat in a strange corner of the financial-and-gambling crossover, taken seriously by a small community of academics, betting-market researchers, and crypto-native users, and ignored by almost everyone else. In 2024 and 2025 that changed. Polymarket processed billions of dollars of volume on the US presidential election, the academic and journalistic citation rate of prediction-market prices on macro questions started to climb, and Kalshi won a regulatory battle in the United States that opened the door to election contracts on a CFTC-regulated venue. By 2026 the category has graduated from a curiosity into a question every careful retail trader has had to form an opinion about. Most of the opinions on offer are bad ones. Prediction markets are a real asset class with real liquidity properties, real regulatory complexity, and real educational value, and most retail content frames them as either a casino or a crystal ball — neither of which equips a serious trader to think about the category clearly.

What prediction markets actually are, and why the framing matters

A prediction market is a venue where you can buy a contract that pays out a fixed amount — typically one dollar — if a defined real-world event resolves in a specified way, and zero otherwise. The price you pay between zero and one is, by mathematical construction, the market's implied probability of the event occurring. If you believe the contract is mispriced relative to the true probability, you can take the side you think is mispriced and, if you are correct on average, the math works out in your favour over a large enough sample.

That is the textbook description. It is also where most retail content stops, and the stopping point is the source of most of the confusion downstream. The real story is more interesting and more constrained.

The contract is binary, but the market is not. A liquid prediction market on a major question — the result of an election, a Fed rate decision, a corporate-action outcome — has a continuous secondary market between the moment the contract is listed and the moment the event resolves. The price moves with information, with sentiment, with positioning, and with the same kinds of microstructural pressures that move any other liquid market. A trader on a prediction market is not just betting on the outcome; they are also taking a position on the path of probability between now and resolution, and on the liquidity dynamics that shape that path.

This is the first thing most retail content gets wrong. It frames prediction markets as a forecasting tool — "the market thinks there is a sixty percent chance of X" — without noticing that the path-dependent secondary trading has its own dynamics. The probability quoted at any moment is the price the marginal participant is willing to transact at; it is not necessarily a calibrated probability, and treating it as one without checking the liquidity context is the same category error as quoting a thinly-traded stock's last price as its fair value.

The second thing most retail content gets wrong is the assumption that prediction markets are fundamentally different from financial markets. Mechanically they have a great deal in common with binary options, with limit-order-book derivatives, and with index futures whose final settlement depends on a defined event. The differences that matter — and there are several — sit in regulation, in question universe, in counterparty structure, and in the social and academic role the prices play. The mechanics of price discovery, market-making, and adverse selection look very familiar to anyone who has spent time with traditional derivatives.

The Kalshi precedent and the US regulatory turn

Kalshi launched in 2021 as the first US prediction market explicitly licensed by the Commodity Futures Trading Commission as a Designated Contract Market. Its bet — pun intended — was that prediction-market contracts could be structured as legitimate event-driven derivatives under existing CFTC rules, rather than as gambling products under state-by-state rules. For two years the venue listed a deliberately tame universe of contracts: economic indicators, corporate decisions, weather outcomes, scientific milestones. The list was carefully curated to stay inside what the regulator was clearly comfortable with.

The interesting fight came when Kalshi sought to list contracts on US elections. The CFTC had historically taken the position that election contracts were not permissible event contracts under its rules, partly on the grounds that election outcomes are not the kind of "commercial activity" the CFTC is structured to oversee, and partly on a long-standing public-policy concern about creating financial incentives that could distort the political process. Kalshi's argument, distilled, was that the public-policy concern was contestable in a world where political-betting markets already existed offshore, where research had not validated the distortion claim with empirical evidence, and where the CFTC's mandate was to apply its rules consistently across event-contract types.

The case worked its way through the federal courts and, in late 2024, ended in a ruling that allowed Kalshi to list election contracts. The ruling was narrowly framed and continues to be appealed and refined, but the practical effect is that for the 2024 US election cycle a CFTC-regulated US venue offered retail-accessible election contracts for the first time in modern American history.

What the Kalshi precedent did, beyond its immediate effect, was establish that a regulated path for at least some prediction-market contracts exists in the United States. That is a structural change. It does not legalise prediction markets across all questions, all venues, or all jurisdictions; it does not affect EU retail access; and it does not eliminate the open public-policy debates. What it does do is create a legitimacy on-ramp that the category did not have before, and that on-ramp is part of why the conversation around prediction markets in 2026 is qualitatively different from the conversation in 2022.

Polymarket: scale, mechanics, and the question of access

Polymarket is the other half of the story and a very different animal. It is a non-custodial prediction market built on the Polygon blockchain, denominated in USDC stablecoin, operating without a central counterparty in the traditional sense. Anyone with a Polygon wallet and a USDC balance can, in principle, participate; in practice, the venue is geo-restricted in many jurisdictions, including most EU member states, the United Kingdom for retail, and the United States for retail since a 2022 settlement with the CFTC. The geo-restriction is enforced at the front-end web-app level, not at the underlying smart-contract level, which is a distinction that matters legally and matters less for the disciplined retail trader who should not be looking for technical workarounds in any case.

The 2024 US election cycle was Polymarket's coming-out moment in mainstream financial discourse. The election-outcome contracts attracted significant volume — billions of dollars by the end of the cycle, by widely-reported figures — and the prices were referenced extensively in mainstream and financial-media coverage of the campaign. The volume was concentrated heavily in non-US wallets, which is consistent with the geo-restriction, and which complicates any reading of the prices as an indicator of US public sentiment.

The mechanical difference between Polymarket and Kalshi is significant. Kalshi runs a centralised limit-order-book matching engine inside a CFTC-regulated venue. Polymarket runs an automated-market-maker model alongside a hybrid order-book layer on a public blockchain, with USDC as the unit of account. The two architectures produce different liquidity profiles, different fee structures, and different counterparty-risk surfaces. They are not interchangeable, and a trader thinking about either one needs to understand the architecture they are dealing with, not the cartoon version of "a prediction market" the press tends to reach for.

Polymarket's question universe is also broader than Kalshi's, by design. The non-custodial structure permits the listing of questions that a regulated US venue would not be permitted to list, and the platform has historically been comfortable with a wider range of cultural, political, and speculative contracts. The breadth is a feature for users who want exposure to questions traditional financial markets do not offer; it is also a regulatory and reputational liability that has shaped the platform's history with US and European regulators.

The math: liquidity, pricing, and the difference from traditional markets

For a working trader, the question that matters more than the regulatory framing is how the underlying math differs from the markets they already know. Three differences are worth dwelling on.

The first is the binary settlement. A prediction-market contract resolves at zero or one. There is no continuous payoff distribution, no Greek surface in the option-theoretic sense, no theta decay against a continuous underlying. There is a single resolution event and a price path leading up to it. This shapes everything about how the contract behaves. A trader who is used to thinking in dollars-per-point on a futures contract has to recalibrate to thinking in cents-per-contract against a fixed maximum payoff. The mental arithmetic is different. The risk decomposition is different.

The second is the liquidity profile. Prediction markets are typically deeply two-sided in the days before resolution and in the immediate aftermath of major information events; they are often thin in the long flat middle between listing and resolution. The bid-ask spread on a prediction-market contract can compress and widen by an order of magnitude over the contract's life. A trader who placed an order at quoted mid-price three weeks before resolution may not be able to exit at any reasonable price during the quiet stretch in the middle. The lesson — visible from the asset selection framework and from any honest treatment of the news fade — is that liquidity is not a property of the venue alone; it is a property of the question, the moment, and the participant mix.

The third is the adverse-selection profile. Because the question universe is small relative to the participant universe, prediction markets are particularly prone to participants with private information. The trader who shows up at a contract on a regulatory ruling without any insight beyond reading the news is, statistically, transacting against people who have spent the last six months building a domain-expertise position. This is true of any market, but it is more concentrated in prediction markets where the question is sometimes literally about a single decision by a single body. The retail trader who treats this as just another asset class will find themselves on the losing side of an adverse-selection trade more often than the headline win-rate math suggests.

The hardest thing for a retail trader to internalise about a prediction market is that the price you see is the price the most-informed participant is willing to deal at. Quoting "the market thinks there is a sixty percent chance" without asking who set the price and on what information is the same mistake retail traders make on thin equities, dressed up in different clothes.

— Internal note on prediction-market microstructure, Tradoki desk

Why prediction markets are not (just) sports betting

A reasonable reader will at this point ask: how is this different from sports betting? The mechanics share a great deal, the cultural overlap is real, and the platforms themselves often court that overlap. The honest answer is that meaningful differences exist, but the line is thinner than either industry likes to admit.

Three things distinguish the category in practice. First, the question universe — prediction markets list contracts on economic indicators, regulatory rulings, corporate actions, technological milestones, political outcomes, scientific results. A contract on whether a Fed decision will be a pause is structurally much more like a binary option on a macro indicator than like a parlay on a football game. Second, the regulatory framing — Kalshi's listings are CFTC-regulated event contracts, Polymarket's exist outside any single regulator's framework, sports betting is regulated separately from financial markets in most jurisdictions, and the classifications matter for tax, consumer protections, and what protections the user does and does not have. Third, the academic and price-discovery role — prediction markets have been studied by economists for decades, with a substantial literature on whether prediction-market prices are well-calibrated estimators of true probabilities; the same cannot be said of sports-betting lines, where the academic interest is mostly in the bookmaker's operations.

None of this makes prediction markets a benign retail product. Loss rates on retail participants are bad enough that the category sits well inside the family of products where most retail accounts lose money. But the public conversation does prediction markets a disservice by collapsing them into "sports betting in a different costume," because the regulatory and academic differences are real and they matter for how the category will evolve.

The European reader's situation: regulatory complexity, in detail

For the DACH and wider European retail audience, the most consequential thing about prediction markets in 2026 is that access is genuinely complicated and the legal exposure of using a non-permitted venue is non-trivial.

Polymarket has, throughout its history, restricted access from most EU jurisdictions. The restriction has been enforced inconsistently — the smart-contract layer is permissionless, the front-end is geo-blocked, third-party interfaces have come and gone — and the practical question of whether an individual EU resident can or should access the platform is one only a qualified adviser in their jurisdiction can answer. Kalshi is, at the time of writing, structured for US retail and is not generally available to non-US retail; the question of whether and how Kalshi might extend internationally is one for the future, not the present.

Inside the EU, prediction markets sit in a regulatory category that varies by member state. Some jurisdictions treat them as gambling products subject to gambling-specific licensing; others treat them as financial instruments subject to MiFID-like rules; others treat them as something in between, requiring case-by-case classification by the national regulator. The picture is heterogeneous and changing, and the only honest position to take in an educational piece is that an EU resident should not assume access is permitted, should consult the relevant national regulator, and should treat any platform's marketing claims about jurisdictional eligibility as one input rather than as a determination.

For the DACH-specific reader: BaFin in Germany has historically taken a cautious approach to prediction-market-adjacent products and would be the first port of call for a German resident considering the category. FMA in Austria operates within a similar EU framework. FINMA in Switzerland sits outside ESMA and applies its own rules, and the Swiss situation has historically been somewhat different from the EU one — but "different" does not mean "permissive" in any blanket sense.

The point of this section is not to provide regulatory advice, which we cannot. The point is that the regulatory question is the first-order issue for an EU reader and not an afterthought to bolt on to the end of an analysis. We have written about the broader regulatory horizon in the future of retail trading 2026 to 2030; the prediction-market piece of that horizon is one of the more uncertain elements, and the right posture for a retail trader in 2026 is to assume the regulatory perimeter will tighten before it loosens.

How to think about prediction markets as a category, not as a trade

The framing this essay is arguing for is that prediction markets are an asset class worth understanding for a serious retail trader, even if the trader never accesses them. Three reasons.

Prediction markets are now a price-discovery layer that mainstream financial coverage cites on macro questions. Whether or not you trade them, the prices quoted in financial media — implied probability of a Fed move, of a regulatory ruling, of a corporate-action outcome — are increasingly drawn from prediction-market venues. Reading those prices critically is part of the literacy of a 2026 macro-aware trader. The reading-the-macro framework is one part of that literacy; the prediction-market piece is another.

Prediction-market mechanics also teach a clean version of probability-as-price thinking. Most retail traders are weak at translating between probability and price. A binary contract priced at sixty cents is a clean way to think about an implied sixty percent probability, and the discipline of asking "if I think the true probability is X and the market is pricing Y, what is my edge?" is the same discipline a serious quant uses on options. The prediction market is the simplest possible instrument for practising it, even if you never trade it.

Finally, prediction-market regulatory developments will shape the broader retail-derivatives perimeter. Whether the Kalshi precedent extends, whether the EU clarifies its position, whether crypto-native venues retreat or expand — these events will affect the regulatory horizon for retail trading more broadly. The trader who is paying attention to the prediction-market perimeter is paying attention to a leading indicator.

What this article does not do is tell you whether to trade prediction markets, endorse Polymarket or Kalshi, or opine on whether any specific contract is mispriced. The omission is deliberate: the educational point is the category-level understanding, and the trade-level question is one we are not equipped, and not authorised, to answer for any individual reader in any specific jurisdiction. The broader pattern — a new asset class graduating from fringe into mainstream, regulators slowly defining their posture, retail content racing ahead of the regulation with bad analysis, serious traders quietly forming careful views — is one we have seen play out in crypto, in retail FX-via-CFD, in zero-day options. The category has its own specifics; the pattern is invariant. The work for a serious trader in 2026 is to ignore the noise, hold to the structural framing above, and let the regulatory and structural questions resolve before reaching for a position. Sometimes the right trade is to study the asset class without trading it. For most retail readers in DACH and the wider EU, prediction markets are exactly that asset class.

● FAQ

What is a prediction market?

A prediction market is a venue where contracts pay out based on the outcome of a defined real-world event — an election, an economic print, a sports result, a regulatory ruling. The contract price between zero and one is interpretable as the implied probability of the outcome. The mechanism is closer to a binary option than to a sports bet, even though the line between the two is thinner than either industry likes to admit.

What is the difference between Polymarket and Kalshi?

Kalshi is a CFTC-regulated US exchange that lists event contracts on a defined set of questions vetted by the regulator. Polymarket is a non-custodial market built on the Polygon blockchain, denominated in USDC, with a much wider universe of questions and structurally different access. The two are sometimes spoken of as competitors but they are operating under different regulatory regimes for different audiences.

Are prediction markets legal in the EU?

It depends on the jurisdiction and the specific platform, and the answer is changing. Polymarket has been restricted or unavailable to retail users in several EU member states for most of its history. Kalshi is US-only at retail. Anyone considering access from a DACH or wider EU jurisdiction needs to consult their national regulator and a qualified adviser before doing anything. This article does not, and cannot, provide that determination.

Are prediction markets just sports betting in different clothes?

Mechanically they share features, and culturally the line is thin. Where they differ is in regulatory treatment, the universe of questions, the price-discovery role they play in adjacent markets, and the academic interest. None of those differences turns prediction markets into a benign retail product, but they do make the category worth understanding rather than dismissing.

Should retail traders trade prediction markets?

This article does not answer that question. It argues that traders should understand what prediction markets are as a category, how the math differs from price-of-a-stock, and what the regulatory exposure is in their jurisdiction, before they have an opinion either way. Most retail content treats prediction markets as either a casino or an oracle. Neither framing helps you make a careful decision.