State-by-State Legal Opinions for Skill-Gaming Platforms

Here is the pattern I see over and over: a founder builds a genuinely skill-based game, then takes 10 percent of the prize pool as the platform's cut. That one design choice can convert the platform from a contest organizer into something several states treat as raking wagers, selling pools, or holding stakes for a fee, and those offenses generally do not care that the game is pure skill.

The remediated structure that survives underwriting review: a separately stated, fixed, outcome-independent per-player fee, with the winner receiving 100 percent of the combined entries, and player funds held by a licensed provider rather than the operator. In a 51-jurisdiction engagement I completed in July 2026, that restructuring was the difference between zero processor-grade launch states and a defensible multi-state conditional footprint.

States test skill against chance differently: predominance tests ?, material-element tests ?, any-chance tests ?, and definitions that turn on structure rather than chance at all. But the test is only step one, and this is what generic articles miss.

The harder question for paid head-to-head play is the mutual-stake problem: when both contestants fund the prize with their own equal entries, some states recharacterize the match as wagering between the players regardless of skill. The classic entry-fee doctrine (an entry fee paid unconditionally to an operator who does not compete is not a bet) works cleanly when the operator funds the purse; it does not automatically carry a player-funded pot. What usually decides the state is whether it has an express exception for prizes awarded to the actual contestants in a bona fide skill contest: 22 jurisdictions have some version of that language on my current record, and a few are remarkably specific. Iowa's bona fide contest statute names chess by name and permits awards whether or not entry fees are charged; Arkansas enacted an esports safe harbor in 2023 under which compliant paid tournaments do not constitute gambling.

The least intuitive result in this field: a state can protect your players and still criminalize you. Illinois is the cleanest example. Its supreme court held that head-to-head contestants in a predominantly-skill contest are protected, yet a separate provision, 720 ILCS 5/28-1(a)(12), makes it an offense to operate an Internet site that permits a person to play a game of chance or skill for money, and its exceptions do not extend the players' defense to the operator.

That pattern repeats in different forms: bookmaking definitions with no chance element (a "bet" defined as hazarding value on any event), paid custodian-of-stakes provisions, pool-selling statutes, and 27 jurisdictions with some form of staking-on-skill-contest prohibition on my record. These operator-side structures, not the skill analysis, drive most exclusions in a serious footprint review. There is also a quiet civil tail: 33 jurisdictions have loss-recovery statutes letting losers, and sometimes any third party, sue to claw back money lost at games, several of them indifferent to skill.

If player deposits settle into your own account and you run balances on an internal ledger, two independent problems appear. First, gambling codes: paid custodian-of-stakes offenses fit an operator holding both players' entries. Second, money-transmission law: holding and transmitting third-party funds is the textbook trigger for state money-transmitter licensing, and "we are just a tech platform" is not an exemption analysis.

The architecture underwriting teams expect: custody with a licensed provider or sponsor bank, the platform transmitting only objectively determined result instructions, a published settlement matrix for draws, voids, disconnections, and cheating cases, and a documented chargeback allocation that does not automatically debit the innocent winner. This is also where the management factual certificate earns its place in the package: management certifies what the platform actually does, and the opinion applies the law to those certified facts.

Two federal statutes come up in every processor conversation. UIGEA (31 U.S.C. § 5362) does not contain a general safe harbor for paid skill contests; whether a transaction is "unlawful Internet gambling" turns substantially on state-law classification, which is why the state-by-state schedule is the heart of the package, not an afterthought.

The Wire Act, 18 U.S.C. § 1084, applies to a person "engaged in the business of betting or wagering" using interstate wires to transmit bets or wagers. The threshold argument for a properly structured skill platform is that a fixed entry fee paid by an actual contestant in a bona fide skill contest is not a bet or wager at all. The statute's outer scope is genuinely contested: the Department of Justice read it as limited to sports gambling in 2011, reversed itself in 2018, and the First Circuit then held it applies only to wagers on sporting events or contests (New Hampshire Lottery Commission v. Rosen, 986 F.3d 38 (1st Cir. 2021)). A careful opinion analyzes the threshold question expressly rather than waving at the circuit split.

A sample of what changed recently, from my research record (each of these has consequences for at least one platform model):

• 2024, October Michigan's Gaming Control Board issued cease-and-desist action against a major skill-gaming app operator, part of a broader pattern of states testing paid skill apps against their internet-gaming regimes.
• 2025 Utah rewrote and renumbered its gambling framework, adding fringe-gaming provisions drafted to reach electronic access to contests offering monetary prizes regardless of skill.
• 2025 Oregon's Attorney General issued Opinion 8297 on fantasy contests; New Mexico's Attorney General followed in December with Opinion 2025-17, declining to categorically bless paid online contests and requiring contest-specific skill-versus-chance analysis.
• 2025-26 The sweepstakes-casino ban wave: California's AB 831 and New York's S5935A took aim at dual-currency casino-style models, with language broad enough that hybrid promotional structures need re-checking.
• 2026 Illinois enacted a fantasy licensing regime (P.A. 104-0468) and Kentucky enacted HB 904. Minnesota went a different direction and banned prediction markets outright (§ 609.7615, effective August 2026, immediately challenged in federal court). Oklahoma's SB 1589 adds platform and vendor liability for specified online casino games effective November 2026.

These are illustrative mockups with fictional content, not client documents. The real deliverables run to dozens of pages; the point is that you receive artifacts your processor can file, not an email with opinions in it.

Will the processor approve me? No result is guaranteed. The opinion is prepared to address the processor's stated underwriting request, and it is the artifact underwriting teams ask for.

What if my current structure is not opinion-ready? I identify the blocking issues before issuing a favorable conclusion. Document remediation is separately scoped; you will know exactly what has to change and why.

Are rules and Terms of Service revisions included? Not in the opinion package unless expressly stated. The opinion analyzes your documents as written.

Are licenses, registrations, bonds, or local counsel included? No. The classification schedule identifies any such requirement per jurisdiction; implementation is separate.

What processor follow-up is included? One written clarification round tied to the issued opinion. New facts or material product changes are a new analysis.

When does the clock start? When I confirm your record is complete and opinion-ready, including the signed factual certificate. I confirm the delivery date before the engagement begins.