> What is Kalshi? A plain-English guide to the CFTC-regulated event exchange, how contract prices become probabilities, and how it differs from Polymarket.
If you have watched the rise of prediction markets over the past few years, you have probably seen two names come up again and again: Polymarket and Kalshi. They look similar on the surface — both let people trade on the outcome of future events — but they are built on very different foundations. Kalshi is the one people reach for when they want to know what a *regulated* prediction market actually looks like.
This is a plain-English explainer of what Kalshi is, how its event contracts work, why a contract price behaves like a probability, and what "regulated" really changes. We will also look at how forecasters treat these prices as a benchmark. It is educational, not a trading guide.
What Is Kalshi?
Kalshi is a US-based exchange where people trade **event contracts** — legally, a type of financial contract tied to the yes-or-no outcome of a real-world question. It operates as a designated contract market under the oversight of the Commodity Futures Trading Commission (CFTC), the same federal regulator that supervises US futures and derivatives exchanges.
That regulatory status is the whole point. Rather than existing in a legal grey zone, Kalshi was authorized to list event contracts to US customers. In practice that means the questions it lists, the way trades clear, and the limits it enforces all sit inside a supervised framework, not on the honor system.
How an Event Contract Works
An event contract is about as simple as a financial instrument gets. The exchange poses a clearly defined question with a fixed resolution date — for example, "Will a given economic indicator come in above a stated level this month?" Each contract settles at **$1 if the answer turns out to be Yes, and $0 if it turns out to be No.**
You can buy the Yes side or the No side. If you hold a Yes contract and the event happens, it is worth $1; if it doesn't, it is worth nothing. Because the two sides always add up to $1, buying No is mathematically equivalent to betting against the Yes outcome. Every contract has an unambiguous rule for how it resolves, published in advance, so there is no argument after the fact about what counts as a win.
Why the Price Is a Probability
Here is the elegant part, and the reason forecasters pay attention. Because a Yes contract pays exactly $1 when it resolves true, its current trading price maps directly onto an implied probability.
If a Yes contract is trading at 63 cents, the market is effectively saying the event has roughly a **63% chance** of happening. A contract near 92 cents reflects near-certainty; one trading at 8 cents reflects a long shot. Divide the price by a dollar and you have read the crowd's probability estimate straight off the screen.
Why should that price track reality? Because traders have an incentive to correct it. If you think an outcome is genuinely 80% likely but the market prices it at 60%, buying underpriced Yes contracts is rational — and when enough informed participants do the same, the price drifts toward consensus. The market aggregates scattered knowledge — data, expertise, and money at risk — into a single number that updates continuously.
Kalshi vs Polymarket: The Key Difference
People often lump Kalshi and Polymarket together, but they differ in a way that matters.
**Polymarket** is a crypto-native platform. Trades settle on-chain, positions are denominated in stablecoins, and it covers a very wide, fast-moving range of world events. Its openness is its defining feature.
**Kalshi** is a regulated US exchange. You fund an account in dollars, it operates under CFTC oversight, and it is designed from the ground up to serve US customers inside that legal structure. The contracts are cleared and the platform answers to a federal regulator.
The mechanics of "price equals probability" are the same on both. The difference is the plumbing underneath: an on-chain, permissionless market on one side, and a supervised, dollar-denominated exchange on the other. Neither model is automatically "better" — they are optimized for different priorities.
What "Regulated" Actually Means
Regulation is an abstract word until you see what it changes day to day. For a market like Kalshi, it shows up in a few concrete places.
- **Clearing and settlement.** Trades are cleared through a regulated clearinghouse, which stands between buyers and sellers and enforces that winning contracts actually get paid the dollar they are owed. This reduces the risk that a counterparty simply vanishes. - **Position and eligibility limits.** A regulated venue can cap how large a position a single participant may hold and restrict who can trade, which limits the ability of one deep-pocketed player to distort a market. - **Defined, pre-approved contracts.** The questions themselves and their resolution rules operate inside a supervised process, so a contract's meaning is fixed and auditable rather than improvised. - **Oversight and reporting.** Operating as a designated contract market brings recordkeeping, transparency, and accountability obligations that an unregulated venue can skip.
None of this makes a market smarter or its prices more correct. What it does is change the *trust model*: participants rely less on the good faith of an anonymous operator and more on an enforced legal framework.
What Kinds of Markets Does Kalshi Offer?
The event contracts on a regulated exchange tend to cluster around questions with clean, verifiable outcomes. Common categories include:
- **Macroeconomics** — questions tied to published economic data such as inflation readings, interest-rate decisions, or employment figures. - **Weather and climate** — outcomes like whether a temperature or seasonal metric will cross a threshold by a set date, which resolve against official measurements. - **Politics and elections** — contracts on publicly reported outcomes, where the resolution source is a matter of record. - **Other public events** — a rotating set of clearly defined, date-stamped questions across culture, science, and current affairs.
The common thread is *resolvability*. A regulated exchange favors questions that settle cleanly against an authoritative source, because ambiguity is exactly what a supervised framework is built to avoid.
How AI Forecasters Use These Prices as a Benchmark
For anyone building forecasting models, regulated prediction-market prices are useful for a reason that has nothing to do with trading them: they are a **calibration benchmark**.
A well-built forecaster does not aim to be "right" on a single dramatic call. It aims to be *calibrated* — its 70% forecasts should come true about 70% of the time across many predictions. To know whether a model is calibrated, you need a trustworthy reference point to compare against, and a liquid market's implied probability is one of the toughest baselines available, because it already blends the judgment of many participants.
So forecasters line their own probability estimates up against the market's price and ask honest questions. Where does our model disagree with the crowd, and by how much? When we differ, who turns out closer once the event resolves? Over hundreds of questions, does our stated confidence match reality as well as the market's does? Scoring rules like the Brier score make that comparison precise, rewarding forecasts that are both confident and correct while penalizing overconfidence.
Used this way, a market price is not a tip to act on — it is a measuring stick that tells a forecaster whether their calibration is genuinely good or merely feels good.
The Bottom Line
Kalshi is worth understanding not because it promises anything, but because it shows what a prediction market looks like when it operates inside a regulated, cleared, supervised framework. Its event contracts turn yes-or-no questions into prices, and those prices read as probabilities the same way they do on Polymarket — only the machinery underneath is different. For forecasters, that transparent, continuously updating price is one of the best calibration benchmarks there is.
At NeuPortal Research, that benchmark idea is the whole experiment: we lock AI forecasts in advance, timestamp them, and score them in the open against market prices — losses included. You can follow the accountability experiment at neuportal.ai/experiment.
---
*Educational content — not financial advice, and not a betting tip.*