Cash-Settled Prediction Markets

What cash settlement is, and what it is not the opposite of

markets.settlement: cash means a winning contract pays US dollars. Twenty-one of the sixty cards here carry it, against ten on play_money and twenty-seven on crypto — so this page is the sibling of play-money platforms and not its complement. Money is real on the crypto venues too; what is different here is that the dollars arrive in a bank account, which means an institution with a banking relationship is holding them, which means somebody has to know who you are. Every consequence below follows from that chain.

Six venues, and every one of them answers to a regulator

Kalshi, Polymarket US, Crypto.com Prediction and ForecastEx are CFTC-designated contract markets. PredictIt operates under CFTC no-action relief with a hard cap of $3,500 per contract, which is a regulatory number and not a product decision. Robinhood Prediction Markets is not an exchange at all: Robinhood Derivatives is a futures commission merchant routing to KalshiEX, ForecastEx and its own Rothera joint venture, so a contract there settles under whichever exchange's rules it reached.

All six require identity verification — kyc_required: true on every card in the set, which is true of no other settlement value in this catalogue. Four of the six serve the United States and nowhere else; Kalshi lists the US, the EU, Asia and Latin America, and ForecastEx adds Canada. What you have to prove about yourself to trade is the page for what that check actually asks for, and Kalshi vs Polymarket US vs Robinhood for the three a reader most often confuses.

ForecastEx differs in shape: its rulebook bars individuals from direct membership, so access is always through a broker, and it lists no sports contracts. It also does something no other venue here does — every position is fully collateralised in cash at its clearing house, and Rule 612(c) requires the interest earned on that collateral to be passed to members as a monthly coupon. What your money does between the trade and the resolution is where that sits.

The fee is a different shape on every one

Each card carries the effective fee at fifty cents — this catalogue's own unit, because venues publish fees on incompatible bases and a published 3.00% can be a smaller charge than a published 1.75%.

Kalshi takes 1.75% from a taker and nothing from a maker. Crypto.com takes 1.75% and 0.44%. Polymarket US runs one symmetric formula exchange-wide that peaks at the midpoint and falls to almost nothing at the extremes: 1.74% from a taker at 50 cents, and a maker receives 0.31%, credited at the fill rather than accrued. ForecastEx charges 1% on both sides, Robinhood 2% on both.

PredictIt is the outlier and the reason both fee fields on that card read None rather than a number: nothing is charged on a fill at all. Ten percent of profit is taken whenever you sell above your cost or redeem at a dollar, and then five percent of whatever you withdraw, principal included. Losing sales earn nothing back, so a trader who is flat across two positions still pays on the winning one. What a trade actually costs is the page that puts the spread beside the fee, and how money gets back out the one for the second half of PredictIt's arithmetic.

Resolution is somebody's job here, not a vote

On a cash venue the outcome is determined by an institution with an obligation, which is the sharpest structural difference from the on-chain half of this catalogue. Kalshi names a settlement source per market and publishes it through its public API as a name and a URL. Polymarket US's rulebook names the exchange itself as the Source Agency, and a Contract Outcome Review under Rule 10.4 is final with no appeal from it. ForecastEx settles against the initial release of the named statistic, so a later revision does not change the outcome. PredictIt resolves against the Rules printed on each market page, which govern rather than the headline question does. Robinhood resolves nothing of its own, which is the one to hold onto — the venue you signed up to is not always the venue that decides. Who decides how a prediction market resolves is the general version.

Fifteen of the twenty-one settle nothing

The rest is APIs, clients, datasets and dashboards, where the field describes the venues behind the product. kalshi-api and polymarket-us-api are the exchanges' own developer surfaces; pykalshi, kalshi-python, the current Kalshi SDK, polymarket-us-python and polymarket-us-typescript are clients for them.

On a single-venue card that reading is unambiguous. On a cross-venue one it strains, and the cards say so: Adjacent and Predictefy both carry the value and both tell you in their bodies to read it as a property of the underlying venues. Brier.fyi carries it while scoring four platforms of which only one is cash. Artemis, Lychee, FinFeedAPI, the Apify actors and Prediction Market Analysis are the same shape — read the coverage on the card, not the enum. Tools that cover more than one venue is the page about what that flattening costs.

The identity check reaches the data too, on one of them. Polymarket US answers markets, books and price history with no key, but the developer key that opens either WebSocket requires the iOS app, an account and identity verification.

Nothing in this collection has been traded by this site with a funded account. Fee schedules, rulebook citations, jurisdiction lists and identity requirements above are read from each venue's own documents and dated on the cards.