A Bet Is Still a Bet When One Side Already Knows the Answer

A football game ends. The scoreboard is final, the stadium is emptying, and the result is a fact in the world. Somewhere else, a contract written on that result is still accepting orders. Part of the market knows how it ends. Part of the market has not looked yet. Nothing about that situation is exotic, and nothing about it is new. It is the ordinary consequence of settling a contract against an official source rather than against reality, because the official source needs time to speak.

Two legal systems have been asked what to do about this. One of them answered in 2005 and the answer is still on the statute book. The other has the question split across contract terms, a rulebook definition and a federal regulation about insiders, and the piece that would actually cover it is not written down anywhere a user can read. This is a text about where that window sits, how long it really is, and what a category that wants to set its own floor should be publishing about it.

The British statute refuses to pretend the transaction is something else

Section 9 of the Gambling Act 2005 defines betting as making or accepting a bet on the outcome of a race, competition or other event or process, on the likelihood of anything occurring or not occurring, or on whether anything is or is not true. That much is unremarkable. The two subsections that follow are the interesting part.

Subsection (2) says that a transaction relating to the outcome of a race, competition or other event or process may be a bet despite the facts that the event has already occurred or been completed and that, in the statute's words, "one party to the transaction knows the outcome". Subsection (3) repeats the construction for transactions about the likelihood of something occurring, where the thing has already occurred or failed to occur and one party knows it. The section came into force on 1 October 2005 under S.I. 2005/2455, and the published text was current to 16 November 2026 when it was read for this article.

Read it twice, because the drafting choice is deliberate. Parliament did not write that such a transaction is void, or prohibited, or fraudulent. It wrote that the transaction is still inside the definition. Asymmetric knowledge of a finished outcome is not treated as a defect that makes the deal something other than a bet. It is treated as a case the regime has to be able to handle.

Britain answered with a licence rather than a prohibition

Once the transaction is inside the definition, the Act's architecture takes over. Section 33 of the same Act makes it an offence to provide facilities for gambling unless one of the listed exceptions applies, the most important being that the person acts in the course of a business carried on by a holder of an operating licence authorising the activity. The penalty under subsection (4) runs to imprisonment for a term not exceeding 51 weeks, a fine not exceeding level 5 on the standard scale, or both, with six months substituted for 51 weeks in Scotland.

So the British answer to "what if one side already knows" is structural. The law does not try to inspect each trade for who knew what. It decides who is allowed to run a venue where that can happen, and then supervises them. Whether that is the right answer for event contracts is a separate argument and not one this article needs to win. The point worth taking is narrower and harder to argue with: the question is twenty one years old, somebody already thought about it carefully, and the conclusion was that the window needs an operator who is accountable for it, not a pretence that it does not exist.

On an event contract the window has a name and a length

American event contracts do not leave this to inference. The contract terms say it plainly, and because CFTC Regulation 40.2(a)(3)(vi) requires an exchange to post its submission on its own website at the same time it files, the documents are public.

Take the election contract that KalshiEX LLC filed with the Commission on 14 October 2026 as a Regulation 40.6 notification, amending its "Which <party/person> will win <rank> place in <election>?" contract. The terms state that the Last Trading Date is the same as the Expiration Date and that the Last Trading Time is the same as the Expiration Time. They also describe, in the product's own words, that while trading is open members "are able to adjust their positions and trade freely", and that the Expiration Value and Market Outcome are determined at or after market close.

Then comes the number that matters. The outermost Expiration Date the terms allow is one year after the election date, with expiration moved to an earlier date and time under Rule 7.2 if certified results are published for the relevant geography. Expiration time is 10:00 AM ET. In other words, the market is open from the close of polls until certification catches up, and the outer backstop if certification never arrives is twelve months. The terms even contemplate the intermediate state directly: if a fully conducted election's results are nullified or not certified by the Expiration Date, the contract resolves on unofficial source results reflecting the counting of at least 95 per cent of votes cast. A document that resolves on a 95 per cent count is a document that knows there is a period when the outcome is effectively known and not yet official.

The same mechanism produces a window of minutes or a window of weeks

Compare that with a sports contract from the same exchange. In the Regulation 40.2(a) certification for its "How many <stat> will <entity> record in <time period> of <NFL game>?" contract, posted by KalshiEX LLC on its own regulatory page and dated 21 September 2026, the terms say the underlying becomes determinable when the relevant time period has ended and, for a period running through the end of the game, when the NFL has designated the game as final. When that happens, expiration is moved earlier under the same Rule 7.2. The backstop here is one week after the originally scheduled start, not one year.

That clause is the better piece of drafting, and it is worth saying so plainly. It names the trigger, ties it to an identifiable act by an identifiable body, and accepts the consequence: changes to official statistics published after expiration are not accounted for, in the terms' own words even where they "would have produced a different long payout". Somebody thought about the window and wrote down how it closes.

The two contracts use one rule and arrive at wildly different windows, and the reason is not inconsistency. It is that the trigger is the official source, not the event. The window is simply the distance between the moment reality settles and the moment the designated source says so. For a game marked final that is minutes. For a certified election it can be weeks. Any product whose settlement source is slow has a long window built into it by arithmetic, whatever anyone intends. That is the same structural gap we described in Settlement Risk, looked at from the clock rather than from the counterparty.

The insider rule that looks relevant binds the venue's own staff

The natural assumption is that trading on a known outcome is covered by insider rules. Read them and the assumption does not survive. The Polymarket US Rulebook, in the version dated 14 September 2026, carries Rule 2.10 on restrictions on certain persons who possess material, non public information. It reaches company personnel and members of any committee or panel established by the board, bars them from using or disclosing such information outside their duties, and in paragraph (c) bars company personnel from trading in any contract at all. Paragraph (e) sends the definitions to CFTC Regulation 1.59.

CFTC Regulation 1.59 defines material information as information which, if publicly known, would be considered important by a reasonable person in deciding whether to trade, and non public information as information which has not been disseminated in a manner that makes it generally available to the trading public. The persons it binds are the employees, governing board members, committee members and consultants of a self regulatory organisation, with a further prohibition on anyone who trades on material non public information they know was obtained in breach by one of those covered persons.

That is a coherent regime and it is aimed at a real risk, but it is aimed somewhere else. A spectator who watched the final whistle is not an employee of the exchange, did not receive anything from one, and holds information that is not non public in any meaningful sense. It was broadcast. The insider framework is not failing here. It was never pointed at this.

The one participant facing restriction points at the settlement source

There is a provision that does key a trading restriction to knowledge of the outcome. The same rulebook defines a Source Prohibition as the trading restrictions applicable to any person who has access to material non public information from a Source designated for any contract, or who is otherwise in a position to influence the determination or declaration of the Expiration Value. Both halves are sensible. The first covers the person inside the data provider. The second covers the person who can move the number, which is the problem our article on who counts as an insider when the secret is the outcome examined from the European side.

Two things follow. First, the restriction is keyed to the Source, not to the ordinary case of a market where some traders have seen the result and others have not. Second, and this is a detail worth flagging to the exchange rather than to a regulator, the definition says the restrictions are further specified in the applicable Product Specifications and Rule 2.9. In the 14 September 2026 version Rule 2.9 is headed Scheduled and Ad Hoc Maintenance Windows, and the same mismatch appears in the version dated 2 July 2026. The substance therefore lives only in the Product Specifications, and the rulebook's own pointer does not lead a reader to it. That is a cross reference worth correcting, not a compliance finding, and we say it in that order on purpose.

Real time monitoring is required, but this condition is not named

The federal duty that comes closest sits in CFTC Regulation 38.251, which requires a designated contract market to demonstrate an effective program for conducting real time monitoring of market conditions, price movements and volumes, and, where necessary, to make a good faith effort to resolve conditions that are, or threaten to be, disruptive to the market. Adjacent paragraphs require rules to prevent, detect and mitigate market disruptions associated with electronic trading, pre trade risk controls on all electronic orders, and prompt notification to Commission staff of significant market disruptions.

A market still accepting orders after its underlying event concluded is not named in any of that. Whether it counts as a condition that threatens to be disruptive is the exchange's judgement, made internally, and nothing obliges anyone to publish how that judgement went. The obligation to watch exists. The obligation to say what you saw does not.

Four things an operator can publish without waiting for anyone

This is where a voluntary standard does work that a rule cannot do quickly. None of the following needs a rulemaking, a court or a trade body. All four are inside the gift of any operator who wants Directive 01 and Directive 02 to mean something concrete.

  • State the window in words a user can read. "The Last Trading Time will be the same as the Expiration time" is precise and tells a retail reader nothing. A sentence saying that trading may continue after the event concludes and until the named source publishes, with the backstop date, tells them what they are buying.
  • Make the acceleration trigger mechanical and public. Which source, at what threshold, checked how often, and who is allowed to decide it has been met. The NFL clause described above is close to this already. A one year backstop on an election contract is the other end of the same spectrum and the difference should be visible to a user, not buried in a certification PDF.
  • Disclose at the point of the trade, not in a help article. If the window is open, the order ticket is where that belongs. A user who is told at the moment of committing money is a user who cannot later say nobody mentioned it.
  • Measure the realised gap and publish it. After settlement, the distance between the event's real conclusion and the expiration that followed is a number. Published per contract, it would let the category compare products and improve them. Unpublished, nobody can tell a well drafted window from a careless one.

None of this makes the window go away, and it should not. A market that stops the instant someone believes the result is known would be a market that halts on rumour, and the cure would be worse than the condition. The window is the honest cost of settling against a source that can be checked. What is not honest is leaving its length to be reconstructed from a certification filing.

The question is older than the category and the answer can be too

Three documents, read together, make an uncomfortable point. A British statute from 2005 says a transaction on a finished event is still a bet when one side knows. An American election contract filed in October 2026 leaves trading open until an official certification arrives, with a twelve month outer limit. A rulebook from September 2026 keys its only outcome based trading restriction to the settlement source and points the reader at a rule about maintenance windows. Nobody in that sequence is behaving badly. The window is simply a place where the category has not yet written down what it does, and it is easier to write it down now than to explain it later to someone who lost money inside it.

That is the whole argument of this initiative compressed into one mechanism. The floor we are describing is not a restriction on event contracts. It is the set of sentences an operator would have written anyway if the question had been put to them before the first complaint rather than after. Firms that want to put their name to that kind of floor can read the six directives and sign the commitment.

Share X LinkedIn Email