On 12 August 2026 the Division of Market Oversight at the Commodity Futures Trading Commission published Staff Letter 26-23, an advisory on how exchanges should self certify the incentive programs they offer to traders. On its face it is a filing manual. It sets out what a submission must contain and why staff keep sending filings back.
Read it a second time and it is something else. The checklist it lays down cannot be filled in honestly by a program designed without those answers already in hand, which makes it a design test wearing procedural clothes. It also carries a working date that never made it into the press release.
What the advisory is, and what it is not
The letter goes to designated contract markets and is signed by Duncan Hennes, Acting Director of the Division of Market Oversight. Its subject is the self certification of market maker, liquidity, trading and incentive programs under Commission Regulation 40.6 and Regulation 40.5, the route by which an exchange files a rule and, absent objection, sees it take effect after a ten business day review.
Staff are careful about its status. The letter is, in their words, "informational and does not create new obligations," it does not supersede the Commodity Exchange Act or the Commission's regulations, and it represents the views of DMO staff rather than of the Commission. Nothing in it is enforceable on its own terms; all of it describes what the division will look for in your next filing, which is a different kind of binding.
The Commission's announcement of the advisory put the problem plainly. Incentive program filings under Regulation 40.6(a), particularly those tied to event contracts, have been arriving with procedural and substantive deficiencies.
The date that sits in a footnote
The announcement names no deadline. Footnote 15 of the letter does. DMO staff recommend that every exchange which has already self certified an incentive program under Rule 40.6(a) review those programs for compliance, and that amendments falling reasonably within the advisory's scope be submitted by 14 September 2026, either as a fresh 40.6(a) certification or under Rule 40.6(d) for non substantive revisions.
That is a retrospective ask, aimed not at the next program somebody dreams up but at the ones already running, certified and paying out. An exchange reading the advisory as guidance for future filings has read half of it.
Be precise about what the date is. Footnote 15 is a staff recommendation, not a rule, and the advisory disclaims any new obligation. But it names a date, a task and a filing route, and points all three at programs that already exist. That is the mirror image of the Commission's letter on pricing displays, where an acknowledgement date of 31 August 2026 was repeated as a compliance deadline it never was, as we set out in our piece on why a price is not an odds line.
A filing standard that is really a design test
The substance is in Appendix A, which lists what a submission has to cover: program type and objective, the contracts covered, participant obligations and volume thresholds, objective criteria for selecting participants, start and end dates, and for every incentive on offer, whether it is capped or unlimited, what it is worth, and how it is tied to trading activity in a way that furthers the stated purpose.
Two of the required elements are more demanding than they look. The first is that an exchange offering a range of incentives must analyse Core Principle compliance using, in the advisory's phrase, "the most generous application of the incentive structure." That is a stress test, not a description. It asks what the program pays the participant who works it hardest, under the reading most favourable to them. A program whose answer is unbounded has produced its finding.
The second is the requirement to summarise substantive opposing views from board members or market participants and explain why they were not adopted, or to state that none were received. An exchange can only answer that if somebody inside it was in a position to object before the program shipped. The filing asks, indirectly, whether it passed through a process at all.
The structures staff singled out
The advisory names specific designs DMO staff view as problematic. Volume based rewards with steep tiers or threshold bonuses can, in the letter's assessment, encourage participants to trade solely to reach volume targets, raising the risk of wash trading, pre arranged trading and other manipulative practices. Market maker programs that guarantee net profits, or cover a participant's losses through stipends and rebates, may incentivise artificial strategies.
Behind both examples is a single test. Staff write that compensation exceeding what is reasonably necessary to achieve the program's aim, or lacking reasonable limits, is more likely to produce the market effects the Core Principles exist to prevent. Disproportionate or unlimited payouts are named: unlimited rebates, risk free trades and market maker stipends. Staff tie those to the prohibition on guarantees in Commission Regulation 1.56 and warn that they undermine bona fide risk transaction activity.
The anchors are ordinary exchange law: the duties to prevent market disruption, run a competitive market and protect participants from abusive practices, at Core Principles 4, 9 and 12 of 7 U.S.C. section 7(d). When the hook is general market integrity law rather than the fight over whether event contracts are gaming, the obligation does not wait for the jurisdictional argument to settle, a point we made about operating while the courts contradict each other.
Volume bought at a threshold carries no information
Here the advisory stops being a compliance document and becomes an argument we have made since the manifest. A prediction market price is a claim about the world, and its only claim to accuracy is that people put money behind beliefs they actually hold. Every trade is supposed to be a piece of evidence. A trade placed because it moves a participant across a volume tier is evidence of nothing but the tier, and it enters the same book, prints on the same tape and counts toward the same headline figure as a trade someone made because they had a view.
So this is not only a question of fairness between traders. Manufactured volume degrades the product. The number an exchange publishes stops describing how many informed participants it has, and the first party misled by it is the exchange, which then plans and raises capital against a figure it paid for. We have argued before that volume is the loudest and least useful metric on a prediction market, and that depth, spread and executable liquidity describe a venue far better.
None of which makes incentives wrong. DMO staff say directly that properly designed programs support trading in new products, liquidity and depth, price discovery and orderly operation. We agree, and the distinction they draw is the one to build on. Paying a participant to quote both sides at a defined size for a defined time means paying someone to hold inventory and carry risk. They can lose. That is liquidity, and thin books are a real cost traders pay through the spread. Paying for volume regardless of whether risk was taken is paying for a number.
Directive 06 of our manifest commits signatories to revenue that comes from many informed participants over years rather than a few ruined ones over months. A book padded with incentive driven turnover cannot tell you which of those two you have. The measurement is the point, and the advisory is a regulator arriving at it from the direction of market integrity law.
The paragraph on chance is the one to read twice
Under Core Principle 2 an exchange must offer impartial access, and Regulation 38.151(b) requires comparable fee structures for members, persons with trading privileges and independent software vendors. From that principle staff derive a list that reads much less like fee policy: secret discount codes, non cash prizes, VIP or early access without formal disclosure, selectively offered retention bonuses, and unequal trading conditions delivered through faster market data or enhanced API access.
Then comes the sentence worth sitting with. Sweepstakes style or randomised rewards, and prizes based in whole or in part on pure chance rather than pre defined performance metrics, in staff's view likely run afoul of Core Principle 2. A footnote extends it to gamified, casino style or other chance based mechanisms, and names spin the wheel promotions as an example.
The stated reasoning is equal treatment: chance produces benefits that are unequal and non objective between participants in the same category. But look at where the line lands. It lands on the promotional layer, the surface where a prediction market most closely resembles a casino product, five days after a separate letter told the same firms not to display prices in bookmaker format. Two documents, two legal hooks, one message. An operator who treats promotional mechanics as marketing's business, downstream of compliance, has misread where the boundary sits.
Our position has never been that prediction markets are gambling. It is that they share mechanics which have ruined people before, and that the tools built to contain those mechanics belong in the product from the start. A randomised reward is one of those mechanics in its purest form, and it is hard to think of a justification for attaching one to a regulated derivatives market that survives being written down in a public filing.
The perimeter reaches your affiliates, again
Where incentives are delivered through third party affiliates or intermediaries, staff write that the exchange should hold sufficient oversight to ensure those parties comply with the program terms and actually pass the rewards to their customers. Three safeguards are named: limiting incentives to non discretionary customer orders, requiring affiliates to keep records substantiating qualified orders consistent with Regulation 1.31, and periodically reviewing affiliate access and permissions.
The first is the sharpest and the easiest to skip. A non discretionary order is one the customer chose. Tying an incentive to a discretionary order means paying an intermediary for trades it decided to place on someone else's behalf, which is a conflict with a fee attached. Limiting the program to non discretionary orders removes the incentive to churn a customer account, and does it at the level of design rather than through surveillance after the fact.
This is the second time in one week that the compliance perimeter reached past the exchange into the layer of partners and affiliates through which most users make first contact. Draw that map wider than your own product surface.
The conflict behind the advisory, and the rulemaking that is open
The advisory's shortest substantive section may be its most consequential. Core Principle 16 requires exchanges to minimise conflicts of interest in decision making. Staff note that this bites hardest when a market maker is a subsidiary or affiliate of the exchange itself, because the affiliation gives the exchange a direct financial reason to set program terms in its affiliate's favour.
For that proposition the letter cites something readers should follow up: not another advisory but a live proposed rule, the Commission's Conflicts and Affiliations rulemaking, published in the Federal Register on 6 August 2026 at 91 FR 50926. There the Commission states its preliminary view that an exchange administering matching priority, market data and connectivity arrangements, and market maker incentive programs for the benefit of an affiliate counterparty would compromise both its Core Principle 16 duty to minimise conflicts and its Core Principle 12 duty to protect participants from unfair treatment. The affiliate trades as principal, so any advantage converts into trading gains at the expense of unaffiliated participants on the other side.
Two things follow. Staff guidance has no binding force, but it is pointing at a Commission proposal that could acquire some. And the comment period on that proposal, per the Federal Register record for the document, runs to 5 October 2026. An operator with an affiliated market maker has a narrow window in which the useful response is a comment letter rather than a compliance memo, exactly as it was for the June 2026 event contracts proposal.
What we would do before 14 September 2026
- Inventory every live program, not only the new ones. Footnote 15 asks for a review of programs already certified, and past certification is not a finding about the present.
- Run the most generous application test on each. Model the maximum one participant can extract under the reading most favourable to them. If there is no ceiling, that is the answer, and it belongs in your filing before it arrives as a staff question.
- Separate liquidity from volume. For each reward, ask whether the participant carried inventory and could lose. Where the reward survives with no risk taken, the program is buying a number rather than a market.
- Remove chance entirely. Sweepstakes, randomised prizes and spin to win mechanics have no defensible place in a filing that has to describe objective performance metrics.
- Map incentives to surveillance. Staff describe the method: map each incentive to the behaviour it encourages, then build surveillance against that behaviour. In practice, alerts tuned for wash trading, closer review of trading clusters around threshold rewards, and monitoring for price moves linked to rebates.
- Walk the affiliate layer, then publish. Confirm that incentives reach customers, that qualifying orders are non discretionary, and that records prove both. Rule 40.6(a)(2) then requires the non confidential parts of the submission on your own website when you file.
- Cite the statute, not the advisory's numbering. Section V.D calls Core Principle 18 the prohibition on unreasonable restraints of trade. Under 7 U.S.C. section 7(d), paragraph 18 is recordkeeping and 19 is antitrust considerations, which is what the letter's own footnote 8 cites alongside Regulations 38.1000 and 38.1001.
The bottom line
Coverage by CoinDesk and InGame on 12 August 2026 got the headline right: the regulator is unhappy with the quality of incentive filings and worried about wash trading. That reading is correct, and smaller than the document.
What the letter supplies is a written test for something the category has been doing on instinct. An incentive program is not a marketing decision that happens to need paperwork. It is a decision about which trades an exchange will pay to bring into existence, and that shapes the price every other participant reads. Describing it in Appendix A terms, with caps, objective criteria, a stress tested maximum and a record of who objected, is not a filing exercise. It is the exercise, and the filing is only where it becomes visible.
An operator who cannot complete the description does not have a paperwork problem. Directive 06 is the same test from the other end: a business built on many informed participants over years can say where its volume comes from, because it knows. If your firm wants that written down and held to publicly, our commitment is open.