Event contracts are derivatives

So far this series has spent its time on what a prediction market does — it prices the future, it aggregates dispersed knowledge, it grades its own forecasts. This piece is about what an event contract is, and then the two questions that follow from the answer: what you actually do with the thing, and what it looks like when someone does. The answer to the first is older and more boring than the headlines suggest: an event contract is a derivative. The same family as a futures contract or a stock option. Once you see it that way, almost everything about the category stops being mysterious — why it's priced the way it is, why nobody can default on you, and why a derivatives regulator was the one that claimed it. And once you have the instrument in hand, its real job comes into focus: not betting, but hedging — and we'll finish with a regulated insurer that did exactly that, in the flesh, cheaper than the firm it would have called instead.

A derivative is just a contract whose value is derived from something else — the price of oil, the level of an index, whether a candidate wins. An event contract derives its value from a single fact about the world: did the event happen, yes or no. That puts it in a very specific corner of the derivatives universe, the one finance has known about for decades under a slightly unglamorous name.

A binary option, by its proper name

The instrument has a textbook name: a binary option, also called a digital option.¹ It pays a fixed amount if a condition is met at expiry, and nothing otherwise. On a prediction market the fixed amount is normalized to $1. Hold a YES contract to expiry and you receive exactly $1 if the event happened and exactly $0 if it didn't. There is no in-between, no "how much" — only "whether."

We can write the payoff with an indicator function — the mathematician's switch that equals 1 when its condition is true and 0 when it's false. If \(\mathbf{1}[\,\cdot\,]\) is that switch, the value of one YES contract at expiry is:

That indicator is the whole personality of the instrument. Plotted against the underlying outcome, the payoff is a step function — flat at $0, then a vertical jump to $1 the instant the condition is crossed. Engineers call that exact shape a Heaviside step; it is the cleanest payoff in finance, a single binary cliff.

Contrast it with a vanilla call option, the thing most people picture when they hear "option." A call pays nothing until the underlying clears a strike, and then pays more and more the further it travels — a sloped "hockey stick," continuous and unbounded on the upside. The binary doesn't care how far. One cent past the threshold or a mile past it, the payoff is the same $1. It has thrown away magnitude and kept only direction.

The price is the probability — again

If you've read What is a prediction market? or A price is a probability, the next step is familiar, but now it has a name attached. The fair value of a derivative is the expected value of its payoff. For a contract that pays $1 on the event and $0 otherwise, with the event assigned probability \(p\), that expectation collapses to almost nothing:

The price is the probability. A contract at 62¢ is the market quoting a 62% chance. That identity isn't unique to prediction markets — it's the standard way every digital option in the world is valued. The technical caveat is one word, and it's worth saying out loud because it's exactly the bridge to traditional finance: the relevant \(p\) is a risk-neutral probability, the probability implied by what people will actually pay, not necessarily their private gut belief.² In a market dominated by small, diversifiable bets the two are nearly identical, which is why "62¢ means 62%" is an honest reading. And to be fully correct you discount for time value — a dollar at expiry is worth a hair less than a dollar today — but on contracts that resolve in weeks at today's rates the discount rounds away:

Here \(r\) is the interest rate and \(\tau\) the time to expiry. Keep the discount factor in your back pocket; for the rest of this piece I'll quote the clean version, price \(=p\).

Why you can't get stiffed

Here is where an event contract stops resembling a casino bet and starts resembling a cleared derivative. When you place a bet with a bookmaker, you are extending the house credit: you hand over your stake, and you are trusting that if you win, the other side is good for the money. On a regulated event-contract exchange, nobody trusts anybody. Both sides post the full value of the contract up front.

Work the arithmetic on a single $1 contract. The YES buyer who pays 62¢ and the NO buyer who pays 38¢ are, between them, funding the entire dollar that will be paid out. The two stakes sum to exactly the maximum the contract can ever owe:

The exchange escrows that dollar — holds it in a segregated account — and at expiry it cash-settles: the side that was right collects the whole $1, the side that was wrong collects nothing, and the books are square to the penny.³ Four consequences fall straight out of that one design choice, and together they are the financial difference between this and a wager:

• Fully collateralized. Every dollar of potential payout is already sitting in escrow before expiry. The money to pay the winner isn't promised; it's posted.
• Defined maximum loss. The most you can lose is what you paid — 62¢ on a 62¢ contract. The payoff is bounded at both ends, $0 and $1, so your downside is known before you enter.
• No leverage. You can't lose more than you put in, because you funded your whole side of the dollar. There is no margin call, no liquidation cascade, no owing the house money you never had.
• No counterparty credit risk. You are not relying on the person on the other side to be solvent. Their stake is already in the pot. The exchange is a custodian of collateral, not a bookmaker taking the other side of your bet.

The atomic unit is composable

The reason finance bothers to name a primitive is that you build with it. A single binary is the atom; structured payoffs are molecules. The cleanest example is a range.

Suppose you don't want to bet that inflation comes in above 3% — you want to bet it lands between 3% and 4%. There's no special instrument for that. You just hold the difference of two threshold binaries: long the "above 3%" contract, short the "above 4%" contract. You collect $1 only in the band where the first pays and the second doesn't — that is, exactly when the outcome falls in \([3\%, 4\%)\). In symbols, a digital spread is the difference of two steps:

Two cliffs, subtracted, make a plateau: a payoff that is $1 inside the band and $0 outside it. Stack enough adjacent binaries and you can approximate any payoff shape you like over the outcome — the binaries are the pixels, and a structured position is the picture you draw with them. This is why an exchange that lists a clean strip of threshold contracts is quietly handing traders a construction kit, not just a menu of bets.

Why the word "derivative" carries the weight

None of this is pedantry about taxonomy. Calling an event contract a derivative is the move that decides who regulates it — and that is the whole ballgame for this category.

Futures and options are derivatives, and in the United States derivatives are the jurisdiction of one agency: the Commodity Futures Trading Commission. Because an event contract is structurally a binary option on a future fact, it falls under the CFTC's authority over derivatives, and a venue that lists them does so as a CFTC-regulated designated contract market — the same license category as a futures exchange.⁴ Kalshi was the first to win that designation. When the CFTC recognized event contracts as a regulated asset class in 2020, what it really did was confirm that this corner of the derivatives world was open for licensed business.⁵

That single classification is why the category looks the way it does today. A regulated, fully-collateralized derivative is an instrument that institutions can touch — which is the backdrop to the numbers that opened this series: monthly volume across the leading venues climbing from under $100M in early 2024 to past $13B by the end of 2025, the owner of the New York Stock Exchange committing up to $2B to distribute one venue's market data, and the odds now read out on CNN and CNBC next to the polls.⁶ You don't get there as a betting product. You get there as an asset class. How the 2020 decision rewrote the rules — turning a betting product into a licensed asset class — is a longer story in its own right.

What's hard: the binary option has a rap sheet

I owe you the uncomfortable half of this. "Binary option" is a phrase with a genuinely bad reputation, and pretending otherwise would be dishonest.

For years, offshore websites sold things they called binary options to retail customers, and many were close to outright fraud: unlicensed platforms, payouts rigged so the house won regardless, software that quietly moved the expiry against you, and customer funds that vanished when you tried to withdraw. Regulators on several continents warned about them and, in places, banned their sale to retail entirely.⁷ If your instinct on hearing "binary option" is to keep one hand on your wallet, that instinct was earned.

So the distinction has to be drawn precisely, because it is the entire point. The scam was never the instrument — a digital option is just a step-function payoff, a neutral piece of math. The scam was the venue: unregulated, uncollateralized, and self-resolving, where an opaque counterparty both set the terms and decided who won. The regulated version inverts every one of those failures. It is collateralized, so the money to pay you is already escrowed. It is exchange-listed under the CFTC, so a regulator stands behind the venue and customer funds are segregated. And it is transparently resolved against a settlement source named in the rulebook before the market opens, so no one gets to decide the answer after the fact — the subject of Who decides what's true?

Same payoff diagram; opposite trust model. That gap — between a rigged payout offshore and a collateralized, surveilled, rule-bound contract on a licensed exchange — is exactly the gap a license exists to certify. It's the reason I keep returning to the license in this work: at Seeker, the live MVP and demo are how we show the mechanism, but the license is the goal, because the license is what turns "trust us" into "trust the regulator standing behind us." The math was never the hard part. The trust was.

Strip away the politics and the sports — which, honestly, are still where most of the volume lives today⁸ — and what's left is unglamorous and a little reassuring: a binary option, fully funded on both sides, settled against a fact. A derivative. We've known how to price these, collateralize them, and clear them for a very long time. The only new thing is what they're written on — the future itself.

Which raises the question this piece turns to next. We've established what the instrument is. So what is it for? Because a derivative isn't interesting because of its payoff diagram; it's interesting because of the job it does. And the job an event contract does best is the one finance built derivatives for in the first place — older than any casino.

---

What it's for: hedging the uninsurable

Derivatives have a day job older than any trading floor. Long before anyone traded them for sport, farmers and millers used them to sleep at night — to lock in a price now so a bad harvest couldn't ruin them later. That job is called hedging, and it is the difference between a bet and an instrument. The full-collateral, $1-or-nothing binary we just dissected is, underneath, a tool for exactly this.

Here's the part most people miss. Insurance — the thing we usually reach for when we want to be made whole after something goes wrong — only works for a narrow slice of the risks that actually keep people up at night. A prediction market reaches a much wider slice. To see why, start with what an insurer can and cannot sell you.

What an actuary can price

An insurance premium is not a guess. It's a number an actuary computes from a mountain of past events: car crashes, house fires, mortality tables stretching back a century. With enough independent, repeatable incidents, the law of large numbers does the heavy lifting — the insurer can't predict your crash, but across a million drivers the rate is steady enough to price, pool, and profit on. That is the entire machine. No history, no pool, no premium.

So the machine breaks the moment a risk is one of a kind. What's the actuarial base rate for this election going a particular way? For this court striking down this rule? For a specific war breaking out this year, a specific supplier failing, your specific product launch slipping two quarters? These aren't rare in the insurable sense — a million independent draws from a stable urn. They're singular. There is no urn. An actuary has nothing to count, so there is no premium to quote. The risk is real, expensive, and uninsurable.

This is the cell insurance can never reach: real, costly risks with no usable history. A market reaches it anyway, because a market doesn't need a history. It needs two people who disagree about an outcome and are willing to put money on it. There's no insurer for "Candidate X wins" — but there can absolutely be a contract, and we've just spent the first half of this piece establishing exactly what that contract is.

The same instrument, opposite intent

So suppose the contract exists. How does buying it actually protect you? This is where the word hedge has to be earned, because the contract a speculator buys for profit is the very same contract you buy for protection. Nothing about the instrument changes. What changes is whether you were already carrying the risk before you bought it.⁹

Take a concrete case. You run a company whose economics get materially worse if a particular regulation passes — say it would force you to retire a product line and eat the cost. You can't insure that; no actuary has a base rate for this rule, this year, this committee. But there's a market on whether it passes, with a contract that pays 100¢ if it does. So you buy the YES contract. Now play it forward:

• The regulation passes. Your business takes the hit you feared — but every contract you hold pays out, and that payout is sized to land right on top of the loss. You're made roughly whole.
• The regulation fails. Your contracts expire worthless — you're out what you paid for them — but the thing you were afraid of never happened, so the business is fine. The premium was the cost of sleeping at night.

Look at what just happened. Whichever way the world breaks, your net outcome barely moves. You didn't take on a new gamble; you cancelled one you already had. That is the whole definition: a speculator takes risk for the chance of profit; a hedger sheds a risk already on the books. Same ticket, opposite purpose. The only honest way to tell them apart is to ask what the buyer's life looked like before the trade.

That flat NET line is the entire point of a hedge. The business is no longer holding a directional view on the regulation; it's neutral, free to get back to the work it's actually good at. You can write the idea in one line — your net payoff is the exposure plus the position you bought against it:

Here \(\omega\) is whether the event happens, \(E(\omega)\) is what it does to your business, \(p\) is the price you paid as a probability — the same \(p\) from the first half of this piece — and \(q\) is how many contracts you hold. Choose \(q\) so the hedge's slope offsets the exposure's slope, and the two move in opposite directions by the same amount — \(\text{Net}\) goes flat. That's not a trick of this market; it's the same arithmetic a wheat farmer has used for a hundred years, pointed at a risk no insurer would touch.

Why the institutions actually care

This is the unglamorous reason serious money pays attention — not the thrill, the risk transfer. A treasurer who can offset a political or regulatory exposure has a tool that simply didn't exist before. And the most valuable part isn't even the payout; it's the price. A liquid market on "this rule passes" prints a live, money-backed probability of exactly the kind of event a risk desk has always had to guess at.¹⁰

That's the cleanest read on the institutional money I pointed to earlier. When the owner of the New York Stock Exchange commits up to $2B to a prediction-market venue and structures the deal around data distribution — becoming the distributor of the platform's event-market prices⁶ — that is not a bet on casino entertainment. It is a bet on a new, real-time feed of prices on risks the rest of the financial system has never been able to mark. The hedge is the use case; the data is the asset.

Where the hedge gets imperfect

Now the honest part, because a hedge that's oversold is just a worse bet. Three limits keep this from being a magic wand, and any treasurer who's tried it has hit all three.

The first is liquidity. You can only hedge as much as the market can absorb. Push a large exposure through a thin order book and your own buying walks the price up — you end up paying far more than the fair odds, and past a point the depth simply isn't there. The category has grown enormously — combined monthly volume climbed past $13B by the end of 2025, from under $100M a month two years earlier⁶ — but that depth is concentrated in a handful of marquee markets. The specific, idiosyncratic risk you most want to hedge is often the one with the thinnest book.

The second is basis risk, and it's the subtle one. Your hedge pays off on the contract's exact wording, not on your actual loss — and the two rarely line up perfectly. The market resolves on "Rule §12 is struck down by December 31"; your business is hurt by a softer version that guts the rule without formally striking it, or by a delay that lands in January. The event you feared basically happened, your contract pays nothing, and the gap between "what hurt me" and "what the contract settles on" is the basis. The narrower you can write the contract to your real exposure, the smaller it gets — but a residual almost always remains.¹¹

The third is the quietest: a hedge has a cost, and that cost is the price itself. Pay 30¢ to insure against a 30% risk and, over many such years, you roughly break even before fees — exactly as you'd expect from a fair market. The hedge doesn't make the risk free; it converts a lumpy, catastrophic maybe into a smooth, survivable premium. For most of the singular risks in that top-left cell, that trade — turning a company-ending shock into a line item — is worth making.

So the line between a bet and a hedge was never about the instrument. It's about what you were already carrying. The speculator shows up empty-handed and takes on risk for the chance of a payout. The hedger shows up already exposed and pays to put it down. Insurance can only meet you for the risks an actuary has seen a thousand times before. A market will meet you for the ones no one has ever seen at all — the election, the ruling, the war, the launch that slips. That reach is exactly why this is an asset class, and not a game. But reach is a claim, and claims want evidence. So here is one.

---

Proof in the flesh: what the gambling built

A sports-insurance broker called Game Point Capital did something that reads like a contradiction. It went to Kalshi — a venue most people file under "bet on the game" — and used it to hedge. Not to wager. To lay off risk it was already carrying. And it did so at roughly half the price a traditional reinsurer would have quoted for the same protection.¹²

The headline version is "hedging, not betting" — and that's exactly the case the previous section just made in the abstract, now playing out with real money on a real book.¹² But the more interesting fact is hiding underneath it: the only reason an insurer could hedge here at all is that the order book was deep enough to absorb the trade — and that depth was paid for, ticket by ticket, by people betting on basketball. The gambling everyone loves to scold built a piece of real financial infrastructure.

The business nobody thinks about

Sports has an insurance industry, and it is not small. The market runs around $9B a year and is widely projected to roughly double by 2030.¹³ It covers the things that can blow a hole in a sports balance sheet: a sponsor's payout if a star is injured, a promoter's loss if a game is cancelled, disability on a guaranteed contract — and, most commonly, performance-bonus insurance.

Here's the mechanic. Teams routinely promise coaches and players large payouts triggered by milestones — making the playoffs, winning a championship, breaking a scoring record. Those bonuses are good for morale and brutal for planning, because they're lumpy: a great season can trigger a multi-million-dollar bill all at once, in the year you can least predict it. So teams hand that risk to a specialist. Game Point writes the policy, collects a premium, and takes on the obligation to pay if the milestone hits. It issues hundreds of millions of dollars of this coverage a year.¹⁴

But an insurer that simply holds every bonus it writes is one good season away from a very bad year. So it does what every insurer does: it offloads the risk somewhere else. That "somewhere else" is the whole story — and it's the exact move from the hedging section, run by a professional who does it for a living.

The two trades, and the gap

Game Point laid off two teams' bonus exposures on Kalshi. The contracts are the obvious ones — a market that pays out exactly when the bonus comes due:

• A bonus owed if a team makes the post-season. Kalshi priced that outcome at about 6%. The over-the-counter reinsurance quote was roughly 12–13%.
• A bonus owed if a team advances to the second round. Kalshi: about 2%. The OTC quote: roughly 7–8%.

Look at those pairs. For the same risk transfer, the exchange charged about half what the broker's traditional counterparty wanted — and on the second one, closer to a quarter.¹²

The mechanism is the one this piece keeps returning to: a market price is a probability. A contract that pays one dollar if a team makes the post-season, trading at six cents, is the market's live estimate that the team has a 6% chance — and it is also the actuarially fair premium to insure a one-dollar bonus on that exact outcome. To cover a bonus of size \(B\), you buy \(B\) of those contracts. The math is the same \(C = p\cdot B\) the hedging section set up, now with a price tag on each side:

The cost \(C\) to hedge is just the price \(p\) times the payout \(B\). If the team makes it, the contracts pay \(B\) — landing exactly on top of the bonus the team now owes. A lumpy, unpredictable liability becomes a fixed, known premium. The only question that matters is which \(p\) you pay: the market's, or the desk's. Game Point paid the market's, and the market's was half.

Why the exchange beats the broker

To see why the prices diverge so much, you have to look at where insurers normally send their risk. The traditional home for it is over-the-counter reinsurance — the world of Lloyd's of London and the desks like it.¹⁵ OTC means exactly what it says: you negotiate one-to-one. You call a reinsurer, describe the risk, and haggle over price and terms in private. There is no open book, no competing quote, no posted price.

That structure has a built-in tax. The reinsurer is a single counterparty with its own appetite, and it is conservative by design — it dislikes volatile, hard-to-model risk, and when it agrees to take some, it pads the price for two things at once: its margin, and its own uncertainty about an outcome it can't cleanly underwrite. You can't shop the quote, so you can't discipline it. The number you get is opaque and, for anything spiky, prohibitively high. It's the same basis risk and uncertainty premium from the previous section, except here the reinsurer charges you for bearing it.

An exchange inverts every part of that. Instead of one reluctant counterparty, you get many, competing in the open to take the other side. Instead of a private negotiation, a posted price anyone can see and undercut. Competition and transparency drag the price toward the true probability and squeeze out the padding. That is the entire reason Kalshi's 6% beats the desk's 12–13%: the OTC quote was carrying a spread the open market simply refuses to pay.

The depth the betting paid for

Here is the catch, and it's the one that usually kills this idea before it starts — the same liquidity limit the hedging section flagged. An exchange only beats a broker if it has a book deep enough that a real hedge can go on without the buyer's own order walking the price up. A thin market is worse than the OTC desk: try to lay off a serious exposure into it and you move it against yourself. For years that was exactly why "just hedge it on an exchange" stayed a thought experiment. There was no depth.

What changed is the part that makes purists wince. The depth got built by sports betting. Over the past stretch Kalshi's sports markets deepened enormously — to the point that during the Super Bowl the book could absorb a $22M trade without meaningfully moving the price.¹² That is not a casino statistic; that is an institutional-grade order book. And it exists because millions of people wanted to bet on a football game.

$9B

sports-insurance market · ~2× by 2030

$22M

single Super Bowl trade · no price impact

~½

the OTC price, on the exchange

This is the thesis cashing out. The volume that funds these books is mostly sports⁸ — and that's fine, because the volume is what funds the liquidity and the surveillance and the settlement that let the venue do something serious. Game Point is what "something serious" looks like in the flesh — a regulated insurer using a book the bettors paid for to move real risk off its balance sheet, cheaper than Lloyd's. The gambling was never the product. It was the capital that built the product.

Where it gets hard

The honest part again, because a story this clean has sharp edges — and they're the same two the hedging section warned about, now with teeth.

The first is the liquidity that made the trade possible: it's concentrated. This works beautifully for a marquee outcome with a thick book — will a popular team make the playoffs, win a title, reach a round. It does not work for the long tail of bonus structures — an obscure individual milestone, a third-string incentive, a clause on a team nobody trades. The contract you most want is often the one with no market at all, and there you're back at the OTC desk.

The second is basis risk, exactly as before. The hedge pays on the contract's exact wording, not on the bonus's. If the policy triggers on a specific seed or a date and the market resolves on "makes the post-season," the two can drift apart — the team backs in through a tiebreaker, or the bonus language has a wrinkle the contract doesn't. The closer the contract is written to the real liability, the smaller the gap, but a residual almost always remains.¹¹

And the third edge is the one that makes the whole thing legible: these are regulated event contracts — the derivatives we anatomized at the top, sitting under the CFTC — which is precisely why a licensed insurer can treat them as a real hedging instrument rather than an offshore wager. Strip away the regulation and Game Point can't touch them. Which closes the loop back to where we started: the license is what turns "a place to bet" into "a place a treasurer can clear risk."

---

So pull the three threads together. The instrument is a binary option, fully funded on both sides, settled against a fact — a derivative, which is why a derivatives regulator claimed it. What you do with it is hedge: shed a singular risk no actuary could ever price, the election or the ruling or the bonus, by buying a contract whose payoff lands on top of your loss. And the proof is a regulated insurer doing exactly that on a book deep enough to undercut Lloyd's by half — depth that the people betting on basketball paid for. The math was never the hard part; we've priced and collateralized and cleared these things for a century. The only new thing is what they're written on — the future itself — and, slowly, who has finally been given a clean way to lay it off.