The Boundaries of Chance: A Game of Hide-and-Seek with the Definition of Gambling
From Polish district courts through pseudorandom-number generators to prediction markets on Wall Street: how the law of one country, working with a single word, tries to police an industry now worth a hundred billion dollars a year
The Philosopher and the Legislator
Randomness belongs to that class of concepts which philosophy has dissected for centuries and which the law must, nevertheless, define in a single paragraph and then apply on a Friday afternoon in a courtroom in the Polish town of Łomża. From this first tension all the others follow. A mathematician will tell you that a sequence is random when there exists no shorter description of it than the sequence itself. A physicist will add that even a coin toss is deterministic, given full knowledge of its initial conditions. A philosopher will ask whether randomness can exist at all if the universe is deterministic. The legislator, drafting a gambling-control statute, has no time for this. He writes that a game is one of chance if its outcome depends, in particular, upon chance, and goes to lunch.
From this comfortable imprecision flow several billion zlotys of the gray-market slot-machine economy, fifteen years of case law, and, on the global stage, a regulatory laboratory in which an industry that long ago ceased to be a gambling business has reinvented itself as the most ingenious experiment in legal arbitrage of our time.
With those caveats stated, we may turn to the substance.
The Anatomy of Gambling: Three Elements, One Pressure Point
A short detour through doctrine. In nearly every common-law and continental jurisdiction, gambling is constituted, in its classical definition, by three elements: a prize of value; consideration furnished by the participant; and chance determining the outcome. Negate or weaken any one of these three, and the product slips outside the gambling regime. Almost the entirety of industry ingenuity, for the past several decades, has concentrated on the third element, because chance is the softest of the three definitionally, the most vulnerable to technical reinterpretation, and the most dependent on the particular doctrinal test that local courts have adopted.
And the tests are several, and uncoordinated across jurisdictions. The dominant factor test (most U.S. states, India, Germany) asks whether chance predominates over skill. The material element test (federal law in the U.S. under the Illegal Gambling Business Act) asks whether chance is a material element in the outcome, even if skill matters too. The any chance test (rare, now nearly extinct) treats any chance element as sufficient. The broader-game test (Dutch Council of State, 2022) evaluates the chance mechanism within the context of the entire game in which it sits. This doctrinal fragmentation is not accidental: operators carefully select the jurisdictions that apply the most permissive test, and they design their products to land precisely on the border of whatever test happens to govern locally.
Poland, like Germany and India, applies the predominance test as a matter of principle, but, as we shall see, the Polish Supreme Court has gone considerably further than that formula alone would suggest.
The First Line of Defense: “An Algorithm Is Not Chance”
The first wave of defenders, in Poland as everywhere else, begins with the simplest possible argument: a slot machine runs on a complicated but deterministic algorithm, and therefore it does not contain chance. A pseudorandom-number generator, after all, is not random. It is pseudorandom, as the computer scientists themselves concede in the name. Therefore, the argument runs, the device contains no element of chance within the meaning of Article 2(3) of the Polish Gambling Act. Therefore it is not a slot machine within the meaning of the act. Therefore it requires no license. Therefore the charge under Article 107 § 1 of the Fiscal Penal Code is without basis. The argument is elegant, internally coherent, and drawn from the relevant chapter of a textbook on computer science.
The Polish Supreme Court, in its order of November 9, 2021 (I KZP 4/21, refusal to adopt a resolution), disposes of this in a single sentence. The concepts of “the character of chance” and “the element of chance,” as used in the statute, mean unpredictability of outcome to the person operating the device, regardless of how the machine, or the software controlling it, is constructed. In Polish law, randomness is not a mathematical concept, nor a philosophical one. It is a psychological concept. What matters is what the player sees and feels at the moment he presses the button, not what is happening in the registers of the processor.
The Court goes further. Even where the software offers a function by which the player can preview the outcome in advance, an option the player need not use but could, in theory, invoke, the state of unpredictability is not eliminated. The clause repays attention. The Supreme Court is in effect saying: it does not matter whether the player could have known. It matters whether, under typical conditions of play, he did know. The legislator was protecting against unpredictability as experienced, not unpredictability as objectively unavoidable. By that move, the cleverness of “but our customers can preview the result” becomes irrelevant, however carefully the manufacturer typesets the disclaimer.
Under the Hood: A Brief Technical Anatomy of the Generator
The argument from computer science deserves a closer look at what actually sits under the hood of a slot machine, an iGaming platform, or a loot-box system. The entire digital-gambling economy rests on three categories of generator, each with distinct mathematical properties and distinct evidentiary weight in court.
The first is the Pseudorandom Number Generator, or PRNG, the workhorse of the entire industry. A PRNG is a deterministic mathematical algorithm that, beginning from an initial integer called a seed, produces a long sequence of numbers with no discernible pattern. As John von Neumann once put it, with his characteristic economy, “anyone who uses arithmetic methods to produce random numbers is in a state of sin.” Three algorithms dominate in practice: the humble Linear Congruential Generator, used in the random-event tables of consoles such as the Nintendo 64; the Mersenne Twister, proposed in 1997 by Matsumoto and Nishimura, with a period of 2^19937 − 1, the default PRNG in Python, Ruby, and MATLAB; and the more modern Permuted Congruential Generator, which passes every known statistical test, including TestU01 and PractRand. All share a single weakness: if you know the seed and the algorithm, you can reconstruct every future number. This is why early online-poker sites that seeded their generators from the system clock, in milliseconds, fell prey to several documented frauds, in which players who knew the algorithm reconstructed the cards in advance.
The second category is the True Random Number Generator, or TRNG, which draws on physics: thermal noise in resistors (the Johnson-Nyquist effect), the timing of radioactive decay, the arrival times of photons in optical detectors, ring-oscillator jitter on chips, and, in modern Intel and AMD processors, the RDRAND instruction, which samples a hardware entropy source directly on the silicon. TRNGs are slow and expensive, which is why the iGaming industry uses a hybrid architecture: the TRNG produces a high-quality seed, which is then used to refresh a fast PRNG.
The third category is the Cryptographically Secure PRNG, or CSPRNG, occupying a middle position. It remains algorithmic, but must satisfy two additional properties not required of an ordinary PRNG: the next-bit test (given the first k bits of output, no polynomial-time algorithm can predict the k+1-th bit with better than fifty-percent accuracy) and resistance to state compromise (even if an attacker learns the generator’s current internal state, he cannot reconstruct earlier outputs). The two best-known implementations are AES-CTR DRBG, defined in NIST SP 800-90A, and ChaCha20, designed by Daniel J. Bernstein and used in the Linux kernel since version 4.8.
For a slot machine, the process unfolds roughly as follows. The PRNG maintains an internal state (for the Mersenne Twister, six hundred and twenty-four integers). When the player presses the spin button, the algorithm advances its state and produces a large integer, say 3,847,291. This number is then translated into a virtual reel position via modular arithmetic: 3,847,291 mod 20 = 11, mapping to symbol 11 on the first reel. The same procedure runs for each reel independently. But the trick lies not in the generator itself; it lies in the symbol weighting. The jackpot symbol corresponds to perhaps one of sixty-four virtual stops; the losing symbol corresponds to thirty of sixty-four. The generator is honest. The mapping is not. This distinction is the substance of the Telnaes Patent (U.S. 4,448,419), filed in February, 1982, and granted in 1984, which decoupled the virtual reel (with arbitrary stop weights) from the physical reel, thereby allowing precise control over probabilities while preserving the appearance of a mechanical device. Inge Telnaes was, in this sense, the patron saint of the modern slot machine; he changed gambling more in twenty pages than a century of cabinetmakers had changed it in their work before him.
From the legal point of view, only one thing matters: regardless of which generator the operator chooses, PRNG, TRNG, or CSPRNG, the outcome is unpredictable to the player. The Supreme Court’s position in I KZP 4/21 maps with surprising precision onto the technical architecture. The Court does not need to understand what a Mersenne Twister is. It understands that the player, before he presses the button, does not know what he will see after he presses it. That is enough.
There is, in jurisdictions that license such things, an entire industry built around the regulatory certification of random-number generators. In Malta, the United Kingdom, Gibraltar, Nevada, and most American states, an operator may not deploy a generator commercially without an independent certification from one of four major laboratories: Gaming Laboratories International, BMM Testlabs, iTech Labs, or eCOGRA. The dominant standard is GLI-19, which prescribes minimum seed complexity, minimum period length, statistical-test thresholds (NIST SP 800-22, which contains fifteen tests, plus Diehard and TestU01, run against samples of millions to billions of outputs), entropy requirements, and security architecture. All four laboratories operate under ISO/IEC 17025 accreditation. The Polish slot machines on which the Supreme Court has been ruling, more often than not, carry no such certification, and this is where the next layer of the argument begins.
The Second Line of Defense: Financial Instruments
When the first line falls, the industry climbs to the second floor of its ingenuity. If the essence of chance is unpredictability to the player, and the algorithm is unpredictable, then perhaps the trick is to replace the algorithm with something that is genuinely unpredictable but also objectively real. The choice falls on the financial markets. Devices appear, styled as slot machines, on which the payout depends not on the spin of the reels but on short-term fluctuations in foreign-exchange rates. The operator calls the device an investment platform, the instrument a binary option of the European type, itself a financial institution within the meaning of Article 4(1)(7)(h) of the Banking Act, and the machine installed on a gas station in the town of Z. a terminal for the conclusion of forward financial transactions. The legal argument runs: my company is governed by Article 7a of the Banking Act, which displaces the Gambling Act.
In May, 2022, the Supreme Court took up the question in V KK 173/22. The conditioning of payouts on short-term fluctuations in financial markets, the Court wrote, fluctuations so dynamic that they are unpredictable both to the user of the device and to the operator of the platform, must be evaluated in the light of randomness. That criterion, the Court continued, is not displaced by the mere fact that the payout is linked to actually occurring real-world events rather than to pure chance. A game is one of chance whenever its outcome is unpredictable to the player, and whenever there exists no strategy by which he can improve his result without breaking the rules.
Three lines, and the entire roadside-bar binary-options edifice collapses. But the Court then does something more: it reaches for an argument that, to any seasoned lawyer, sounds foreign, and is unanswerable. It writes that customers of devices ostensibly styled merely as gambling-themed would in fact be entering into knowing contracts for the transfer of funds to a provider of financial instruments, fully aware of the details and the mechanisms governing forward financial transactions, money-market instruments, and securities, under circumstances starkly at odds with the nature of such transactions. In roadside bars. At gas stations. In places where one is by definition only temporarily. The passage reads like a paragraph from a piece of reportage, and has the virtue of reportage: it describes the world as it is, not as the defense brief describes it.
The Court adds, almost as an afterthought, a procedural twist. For Article 7a of the Banking Act to apply, the defendant company would first have to document that it is in fact a financial institution within the meaning of the Banking Act, that it operates in compliance with Polish financial-markets regulation, that it maintains scrupulous records of its settlements, that it possesses copies of the transfer orders given by its customers, records of the bank accounts to which those transfers were directed, evidence that the financial transactions described were real and not merely simulated. The defendant has none of these documents. What he has are devices on which, as the customs and tax officers’ inspection report makes plain, the reels spin with the audible signature of a casino, and the symbols have been chosen to attract people interested in gambling. A strategy of conduct deliberately designed to circumvent the statutory rules on the organization of games of chance, the Court concludes, in a phrase that carries the particular weight of phrases a court does not deploy lightly.
The same logic now returns, on a global scale, in the dispute over prediction markets, Kalshi and Polymarket, platforms on which users bet on binary outcomes: results of sporting events, elections, macroeconomic indicators. The operators classify their products as event contracts and swaps within the jurisdiction of the Commodity Futures Trading Commission, rather than as gambling. According to Kalshi’s own filings, its users wager over a billion dollars a month, ninety per cent of it on sports. Massachusetts sued Kalshi in September, 2025, for operating sports betting without a license. In April of this year, thirty-eight state attorneys general, including those of New York, California, and Texas, filed an amicus brief contending that the C.F.T.C.’s authority under Dodd-Frank does not displace state gambling laws, and that Kalshi’s contracts amount to “bets dressed up as event contracts.” A federal appellate court, the Third Circuit, ruled by a divided 2-1 panel on April 6, 2026, that New Jersey may not enforce its gambling laws against Kalshi, on the ground that federal swaps law likely preempts them; the ruling, however, is a preliminary injunction rather than a final adjudication on the merits, and the dissenting judge found Kalshi’s products to resemble conventional gambling. The Ninth Circuit, hearing consolidated Nevada cases at the same time, appears poised to rule the other way, which would create the classical circuit split that invites Supreme Court review.
The Polish judgment in V KK 173/22, concerning slot-style binary-options machines in roadside establishments, and the American disputes over prediction markets concern, at root, the same question: does wrapping a gambling product in the language of a financial instrument change its nature, or is it merely a redactive maneuver? The Polish Supreme Court settled the question, on its side of the Atlantic, in 2022: it does not. If the U.S. Supreme Court accepts the matter on the anticipated circuit split, it will face the same question, with the difference that the answer there will carry implications measurable in the tens of billions of dollars.
The Third Line of Defense: “The Player Knew the Outcome”
The industry’s ingenuity does not stop. A third wave of defense reads I KZP 4/21 with care and discovers a crack in it. If chance is unpredictability to the player, then one can build a device in which the player has, before each round, the theoretical ability to preview what the outcome will be. A simple software-menu option, a single preview function, will do: the player could know, therefore the outcome is not unpredictable, therefore the game is not one of chance.
The Polish Supreme Court, in January, 2024 (II KK 596/23, an order dismissing a cassation as manifestly unfounded; the passages quoted have the character of obiter dictum), disposes of this argument with a sentence worth repeating. The possibility of displaying the outcome screens, and of deciding whether to continue the game on the basis of that information, does not stand in the way of finding the game to be one of chance, because each individual round has the character of chance. Knowledge of the outcomes of earlier rounds, and of the decisions made on the strength of that knowledge, does not alter the chance character of each individual outcome. So long as the outcome of any given round was unpredictable to the player at the moment the round began, the game was one of chance.
The reasoning closes the door from the other side as well. It does not suffice that the player could have looked; it does not suffice even that he actually did. What matters is the character of the individual event of the game, not the knowledge a player possesses across the set of possible events. The Polish definition of randomness, in other words, is microstructural, not aggregated.
The Fourth Argument: Common Sense
The most interesting moment in the line of authority comes, however, not where the courts dismantle clever constructions, but where they step back and articulate something the law does not bother to put on paper because it does not need to. In January, 2024, in IV KK 555/23 (also an order dismissing a cassation as manifestly unfounded, with the holding articulated in passing alongside the main ruling), the Polish Supreme Court formulated an argument from common sense that might serve as the motto for the entire field.
It is, the Court wrote, obvious that gambling is oriented toward the realization of substantial profits through the exploitation of the psychological mechanisms of gambling addiction. In situations of that sort, the game on any device must be configured in such a way that, in the aggregate, the operator always wins. The question of who wins or loses cannot therefore be made to depend on the dexterity of the player, on his knowledge, or on any other personal attribute of his, because to do so would create, for the operator, a serious risk not merely of failing to profit but of incurring losses outright, which, of course, he cannot afford.
The reasoning has the elegance of first-rate legal thinking precisely because it inverts the industry’s argument by a hundred and eighty degrees. Operators had claimed, for years, that their devices turned on the player’s dexterity, his knowledge, his skill, his powers of observation, and that for this reason the devices were not games of chance. The Supreme Court answers: if that were true, you would have gone bankrupt. Your business model assumes profit. No business model premised on the skill of the player generates a predictable profit. Therefore either your device contains chance, or you are not making money. And since you are making money, the material evidence of chance lies in your financial statements, whatever you choose to call the mechanism in your customer-facing terms of service.
This is a rare specimen of legal reasoning that is, at the same time, economic reasoning. And it has empirical company. Contemporary research on skill-based gaming machines shows that the addition of a skill element to a random mechanic does not change the statistical structure of payouts; it merely intensifies, in players, the illusion of control, which paradoxically increases their propensity to keep playing and to develop disordered patterns of use. The Polish Supreme Court’s economic intuition is confirmed in the experimental psychology: if the operator earns, the structure must be random; if the structure is random but appears to reward skill, the extraction mechanism is only more efficient.
A Global Map of the Escape Routes
The Polish Supreme Court dispatched these four lines of defense in four rulings. The global map of industry ingenuity is, of course, considerably broader: nine principal strategies are now in active use, and every one of them has either affected, or will soon affect, Polish practice.
Strategy One is the classical argument that the game depends on skill rather than chance. The paradigm case is poker. In United States v. DiCristina (E.D.N.Y. 2012), Judge Jack Weinstein, in a post-verdict judgment of acquittal entered on August 21, 2012, held that Texas Hold’em was not gambling under the Illegal Gambling Business Act, because the predominance test established the superiority of skill. The Second Circuit reversed on August 6, 2013, and the Supreme Court denied certiorari on February 24, 2014, leaving the deeper skill-versus-chance question unresolved while settling the federal-criminal question against the defendant.
Strategy Two is the technique known as “skill-stage injection,” the implantation of a superficial element of dexterity into a device that is, in substance, a slot machine. The most striking American example is Pace-O-Matic and its “Pennsylvania Skill” terminals, of which roughly sixty-seven thousand units are deployed across the Commonwealth of Pennsylvania, as against only twenty-five thousand licensed slot machines. The devices look like slot machines, but each round includes a second stage in which the player can manipulate the outcome and recover up to a hundred and five per cent of his stake. The Pennsylvania Commonwealth Court held in 2023 that they were not “slot machines” within the meaning of state law; the consequence was that they paid no gaming tax (whereas licensed casinos pay between fifty-four and fifty-five per cent). The Pennsylvania Supreme Court took up the case in late 2025, where the state attorney general argued, with the directness the situation deserves, that “a game that looks like a slot machine and plays like a slot machine is a slot machine.” A Philadelphia jury, in 2025, returned a verdict of fifteen-point-three million dollars against Pace-O-Matic on behalf of the estate of a convenience-store clerk killed during a 2020 robbery at a gas station. The theory of liability was not that the murderer was a losing gambler, the actual sequence was rather more prosaic: the perpetrator entered, shot the clerk within four seconds, and made off with a bag of cash that had been left for vendors. The plaintiffs contended, and the jury accepted, that the company had built an all-cash operation without adequate security safeguards against the foreseeable robbery risk such an operation generates. It was the first time in American history that a manufacturer of unregulated gambling machines was made to pay damages for the death of a third party caused by the theft of cash whose presence in the store was, itself, a consequence of the business model.
Strategy Three is daily fantasy sports, and the exploitation of a narrow statutory carve-out. Section 5362(1)(E)(ix) of the Unlawful Internet Gambling Enforcement Act of 2006 excludes from the definition of a “bet or wager” any fantasy contest whose outcome reflects the “relative knowledge and skill” of the participants. The clause was drafted with season-long fantasy leagues in mind, and was duly hijacked by FanDuel and DraftKings for the operation of single-day, high-frequency contests with entry fees from one dollar to several thousand. States have legalized the product as a skill game; D.F.S. is now legal in forty-four states. In March of this year, the Public Health Advocacy Institute at Northeastern University School of Law, led by Richard Daynard (the attorney who negotiated the two-hundred-and-six-billion-dollar tobacco settlement), filed suit against DraftKings, FanDuel, Genius Sports, and the National Football League, alleging that their A.I.-driven microbetting features constitute “a defectively designed product” engineered to maximize addiction. The argument will sound familiar: the same economic logic the Polish Supreme Court applied in IV KK 555/23 has now arrived in American product-liability litigation.
Strategy Four is the loot box, and its two-pillar legal defense. Pillar one: the loot box is part of a larger game of skill, and ought not to be assessed in isolation. Pillar two: the contents have no real-world monetary value, and so the prize element is missing. The Dutch Council of State accepted the first argument in 2022 with respect to FIFA Ultimate Team; the U.K. Gambling Commission has consistently maintained the second, holding that loot boxes are not gambling under the Gambling Act 2005, since their prizes have no real-world value. But where a secondary market exists, the analysis goes the other way: an Austrian court in Styria ordered Valve to refund fourteen thousand euros to a player who had bought CS:GO loot boxes, on the ground that the items were tradable on the Steam Marketplace. Belgium declared all paid loot boxes to be gambling in 2018, prompting Electronic Arts to withdraw FIFA Points; a study published in 2023 found that eighty-two per cent of the top-grossing iPhone apps in Belgium continued to feature randomized monetization mechanics, and that enforcement was “almost entirely absent.” The most aggressive regulatory regimes have been built in China (mandatory disclosure of probabilities, 2017) and South Korea (amendment to the Game Industry Promotion Act of March, 2024, with penalties up to two years’ imprisonment and twenty million won in fines for non-disclosure). Empirical compliance research on the Chinese rules has identified what it calls “suboptimal compliance”: disclosures buried in terms-of-service pages, partial, or structured so as to mislead.
The loudest American action in this area remains the Federal Trade Commission’s settlement with Cognosphere (HoYoverse), the developer of Genshin Impact, in January, 2025. The settlement carried a twenty-million-dollar penalty, ten years of compliance monitoring, and a prohibition on the sale of loot boxes to minors under sixteen without verified parental consent. The F.T.C. charged that the developer had built a multi-tiered virtual-currency system (real money to Genesis Crystals to Primogems to Wishes) for the express purpose of obscuring the real per-pull cost of a draw. The technical detail is worth pausing on. Beneath the multi-tiered currency sits an ordinary PRNG with a weighted probability table, augmented by a mechanism called the pity counter, which gradually raises the probability of a rare item after a string of failed draws. From a technical standpoint, this is still unambiguous gambling, with an added progressive feature. From a legal standpoint, so long as the player cannot calculate the actual dollar cost of a single draw, the mechanism constitutes a deceptive practice under Section 5 of the F.T.C. Act.
Strategy Five is the social casino: a slot-style game on which the player buys virtual chips with real money but cannot redeem his winnings for cash. No prize, the argument runs, therefore no gambling. The strategy worked for years, generating billions in revenue. The first breach came in Washington State, whose statute defines a “thing of value” broadly enough to encompass virtual chips that extend the duration of play. In June, 2023, a class settlement of four hundred and fifteen million dollars was approved against DoubleDown Interactive and International Game Technology. In February, 2025, a federal jury returned a verdict of twenty-four-point-nine million dollars against High 5 Games, the first class-action verdict in history against a social-casino operator. Evidence at trial showed that the company’s employees had systematically targeted players whom they referred to, internally, as “whales,” among them at least one self-identified gambling addict. Class actions are now expanding into Alabama, California, Illinois, Massachusetts, and New York, and into the platforms, Apple, Google, Meta, and Amazon, as entities profiting from the hosting of unlicensed gambling.
Strategy Six is the sweepstakes casino, a “dual-currency” model. Players buy “Gold Coins” for entertainment and receive, separately and ostensibly free of charge (by mail, or by e-mail), “Sweeps Coins,” which may then be redeemed for prizes. No consideration, the argument runs, therefore no gambling. The American Gaming Association calls the model “illegal gambling under another name,” and the regulatory response in 2025 has been swift: at least nine states (Washington, Michigan, Montana, California, New York, Connecticut, Nevada, Louisiana, and New Jersey) have banned the structure, with bills pending in roughly fifteen more.
Strategy Seven is skin gambling. Virtual game items, traded on active secondary markets (the CS:GO skin economy on the Steam Marketplace), serve as casino chips on third-party platforms. The 2016 scandal, in which the YouTube influencers “TmarTn” and “ProSyndicate” promoted CS:GO Lotto, a site they secretly owned, without disclosure, triggered the first wave of litigation. In February, 2026, the New York attorney general, Letitia James, filed suit against Valve Corporation, arguing that loot boxes in Counter-Strike 2, Team Fortress 2, and Dota 2 constitute illegal gambling under Article I, Section 9 of the New York State Constitution and under New York Penal Law §§ 225.05 and 225.10, and that the Steam Community Market amplifies the harm by enabling cash-out via the secondary market. Valve responded publicly in March, 2026, comparing its mystery boxes to physical trading cards (Pokémon, Magic: The Gathering) and noting that it had closed over a million Steam accounts engaged in gambling.
Strategy Eight is the prediction-market route we have already discussed. Strategy Nine is the most jurisprudentially elegant of all: the Indian experiment. In a country where the Constitution guarantees the freedom of trade (Article 19(1)(g)), and gambling is a state subject (Entry 34, Seventh Schedule), the Supreme Court has, in a series of rulings stretching from 1957 to 2024, built a precise doctrine: rummy (1968), horse racing (1996), and fantasy sports (2017 to 2021) are games of skill, in which skill “substantially and predominantly” determines the outcome. The Indian Parliament nonetheless adopted, in 2025, the Promotion and Regulation of Online Gaming Act, which bans all online games, including games of skill; the statute is now before the Supreme Court, which has signaled in November, 2025 that skill games may be exempt as constitutionally protected.
The Architecture of Evasion: Five Recurring Patterns
From this global map, five recurring architectural patterns emerge.
The first is multi-tiered virtual currency systems, designed to obscure the real cost of a single play and to sever the “consideration” from the moment of the draw (Genshin Impact, sweepstakes casinos).
The second is the free alternative entry, in which the chance element is distributed without charge alongside a paid path that is economically dominant (the sweepstakes-casino “no purchase necessary”).
The third is non-cashable rewards, in which virtual items are not officially redeemable on the platform but a secondary market exists in fact (Valve, the CS:GO skin economy).
The fourth is the injection of a skill layer, the addition of brief mini-games, second chances, or mechanisms by which the player can nudge the final random outcome (Pace-O-Matic, skill-based slot machines).
The fifth is regulatory reframing, the argument that the product is something other than gambling: a financial instrument, a promotional sweepstakes, a contest of skill (prediction markets, the Polish variant of binary options).
Each of these patterns has been tested. Each has worked, for a time. Each is now being recognized, by courts and regulators in turn, as an evasion mechanism. The cycle runs, in general, five to ten years: a new technology, a new product, several years of expansion in the definitional gray zone, the first class actions or regulatory enforcement, several years of litigation, ultimate categorization. After it, the next ingenious construction appears, and the cycle begins again.
Consequences for Current Designs
Anyone designing a product today on the borderline of gaming, gambling, investment platforms, play-to-earn schemes, blockchain-gaming mechanics, lottery-style social games, skill games with prize components, or, attentively, certain peer-to-peer models with a fee on the pot, should read the four Polish rulings in order, and then set them against the global background. From them emerges a test that may be condensed into four questions.
First: at the moment any individual round begins, is the outcome unpredictable to the player? Individual, not aggregated. If yes, the element of chance is satisfied, however deterministic the underlying algorithm may be.
Second: does the game admit of a strategy by which the player can improve his outcome without violating the rules? If no such strategy exists, the game is one of chance, however emphatically you label it a game of skill.
Third: does the business model assume the operator’s profit? If yes, the Supreme Court will infer chance from the simple fact of profitability.
Fourth: do the exterior, the terminology, the symbols, the sounds, the user experience suggest gambling? If yes, the rules of ordinary life experience (an idiom of art in Polish criminal procedure) will tell the court that the target audience consists of people seeking to gamble, and not of those engaged in capital investment under the conditions of a gas station.
These four questions render irrelevant the most sophisticated doctrinal architectures, because their logic does not yield to a change of label. You may call the device an investment terminal, you may wrap it in Article 7a of the Banking Act, you may call the game a mechanism for distributing rewards from a cumulative pool, you may deny the existence of a random-number generator: the result of the test will be the same, because the test measures not construction but function.
New Regulatory Horizons
There are new questions, of course, that the case law has not yet reached. What of play-to-earn games, in which rewards in blockchain tokens depend on a distribution algorithm verifiable in a smart contract? The industry has supplied its own answer: so-called provably fair systems, in which the server publishes the SHA-256 hash of a server seed before the round, the player contributes a client seed and a nonce, and after the round the server reveals the seed, so that the player may verify the hash and recompute the outcome. The mechanism eliminates the possibility of retroactive manipulation; it does not, as experts note, eliminate the possibility of generating a biased distribution in the first place, which is why external R.N.G. certification remains necessary. From the legal point of view, mathematical verifiability does not exclude chance within the meaning of I KZP 4/21, because the test remains psychological.
What of prediction markets offered to Polish residents? Here the question shifts from the definition of chance to questions of the territorial scope of the foreign license, Article 107 § 2 of the Fiscal Penal Code (participation, on Polish territory, in a foreign game of chance), and domain blocking. The Polish Supreme Court has already, in 2022, resolved the underlying question: wrapping a product in the language of a financial instrument does not exclude chance. If Kalshi or Polymarket begin actively to solicit Polish players without a Polish license, the test will apply to them as it applied to a binary-options device on a gas station in the town of Z.
The most interesting trap, however, is the one in which the regulator may himself become caught. The Dutch Council of State’s broader-game test could, in principle, influence Polish courts in the years ahead. If the loot box is not a game of chance because it forms part of a larger game of skill, then the same logic could shield cumulative mechanisms in play-to-earn games, or reward systems in mobile games more generally. The Polish Supreme Court rejected the analogous reasoning in I KZP 4/21, but it did so in the context of slot machines, not of video games. The first case in which this argument arrives in a Polish court will probably be decided in the next three to five years, and may set the Polish answer to challenges that the world’s regulators are only now beginning to face.
Conclusion
There is, in this history, something larger than the gambling industry itself. The Polish Supreme Court, in the course of a few years, accomplished a feat the significance of which exceeds the resolution of any individual case. From a concept that philosophy has been analyzing for centuries, the Court extracted an operational concept, legally useful, grounded in the perception of the recipient. It then demonstrated that this operational concept is resistant to the industry’s several attempts at escape: into mathematical determinism, into financial instruments, into preview functions, into the marketing of skill. And it finally observed that the concept’s correctness is confirmed by economics, since no profitable gambling model can escape chance, however subtle the form in which chance is incorporated.
The convergence of events in 2025 and 2026, the F.T.C. settlement with Genshin Impact, the New York suit against Valve, the Washington verdict against High 5 Games, the thirty-eight-state coalition against Kalshi, the Indian Online Gaming Act now before the Supreme Court, signals that the world’s tolerance for legal gray zones in this field is narrowing. The direction of travel is clear: the burden of proof is shifting from the regulator demonstrating violation to the operator demonstrating compliance. The Polish line of authority from 2021 to 2024 anticipated this direction, not by any awareness of the global trend, but through the simplicity and precision of a functional test.
There remains one question, which appears in none of the four rulings but which hangs over them all. Did the legislator, drafting the 2009 Gambling Act, know that he was writing a definition which, fifteen years later, the Supreme Court would defend against four ingenious waves of evasion, without need of amendment? Probably not. But that is precisely the rare moment in which a well-written statute proves to be wiser than its author, which is, as is well known, the highest compliment one can pay to the craft of legislation.
Games of Chance – Further Readings
The House Always Wins: How Caribbean Islands Became the World’s Casino Regulators
Event Contracts – The Casino That Pretends to Be a Stock Exchange
Robert Nogacki – licensed legal counsel (radca prawny, WA-9026), Founder of Kancelaria Prawna Skarbiec.
There are lawyers who practice law. And there are those who deal with problems for which the law has no ready answer. For over twenty years, Kancelaria Skarbiec has worked at the intersection of tax law, corporate structures, and the deeply human reluctance to give the state more than the state is owed. We advise entrepreneurs from over a dozen countries – from those on the Forbes list to those whose bank account was just seized by the tax authority and who do not know what to do tomorrow morning.
One of the most frequently cited experts on tax law in Polish media – he writes for Rzeczpospolita, Dziennik Gazeta Prawna, and Parkiet not because it looks good on a résumé, but because certain things cannot be explained in a court filing and someone needs to say them out loud. Author of AI Decoding Satoshi Nakamoto: Artificial Intelligence on the Trail of Bitcoin’s Creator. Co-author of the award-winning book Bezpieczeństwo współczesnej firmy (Security of a Modern Company).
Kancelaria Skarbiec holds top positions in the tax law firm rankings of Dziennik Gazeta Prawna. Four-time winner of the European Medal, recipient of the title International Tax Planning Law Firm of the Year in Poland.
He specializes in tax disputes with fiscal authorities, international tax planning, crypto-asset regulation, and asset protection. Since 2006, he has led the WGI case – one of the longest-running criminal proceedings in the history of the Polish financial market – because there are things you do not leave half-done, even if they take two decades. He believes the law is too serious to be treated only seriously – and that the best legal advice is the kind that ensures the client never has to stand before a court.
