Introduction

623.486 billion RMB. This is the total lottery sales for 2024 announced by the Ministry of Finance of China in February 2025, hitting a record high.

850 billion USD. This is the industry estimate for the total annual gambling turnover of citizens from the Greater China region, and it is a conservative figure.

While there are exchange rate differences and technical variations in statistical methods between these two numbers, the core fact requires no complex calculation: the legal market is just the tip of the iceberg. China possesses the world's most stringent anti-gambling legal system, yet it simultaneously supports the world's largest gambling consumption market. In 2024, public security organs investigated 73,000 cross-border gambling cases and dismantled over 4,500 online gambling platforms. However, the total volume of capital flow has not decreased as a result; like mercury spreading across the floor, while its form has changed, the total volume remains constant—it has simply become more fragmented, concealed, and difficult to track.

Policy has not failed. The rules are still operating as designed.

Whose Expected Value?

To understand the operational logic of this trillion-level empire, one must first accept a counter-intuitive mathematical fact.

The payout ratio of the national lottery is approximately 50%. Spending 100 RMB on a lottery ticket yields an average return of 50 RMB. In contrast, the house edge in Baccarat is only 1.06%. Spending 100 RMB at a gambling table yields an average return of 98.94 RMB. The former is operated by the Ministry of Finance and is legal; the latter is illegal, and practitioners face criminal prosecution. If measured by the expected value of "who takes a deeper cut from the gambler," the answer might be uncomfortable: the "rake" of legal gambling is 47 times that of illegal casinos.

The real difference lies not in probability theory, but in the ultimate destination of the profits. When funds flow into the Ministry of Finance's public welfare fund account (which raised a total of 161.031 billion RMB in 2024), the behavior is called "purchasing lottery tickets"; when the same funds flow into the safes of Macau junket operators or the crypto wallets of northern Myanmar warlords, the behavior is called "gambling." The scale of the legality spectrum is not determined by the degree of harm, but by the ownership of the profits.

This spectrum is the core analytical framework of this report. From the national lottery on the far left to the telecom fraud parks on the far right, each chapter moves in a more clandestine direction. What readers should gain upon completion is not a moral judgment, but a complete map of capital flows and power distribution.

The Cost of Closing the Valve

The function of prohibition is not to eliminate demand, but to change the form of supply.

In 2021, Beijing launched the most severe crackdown in history on Macau's VIP rooms. Suncity Group collapsed, Alvin Chau was imprisoned, and the VIP room model essentially ended. From the perspectives of anti-money laundering and preventing capital flight, the goal was achieved. But from the perspective of capital flow, the result was a dam-breaking overflow.

Macau's total gaming revenue in 2024 reached 28.4 billion USD, a year-on-year increase of 24%, primarily relying on the recovery of the mass and premium mass markets. Meanwhile, VIP business revenue, which once accounted for half the market, has only recovered to about one-third of 2019 levels. Analysis by JP Morgan confirms this profound rupture. The vanished VIP funds did not "return"; they flooded into regulatory vacuums. Crypto casinos recorded 26 billion USD in betting volume in the first quarter of 2025, doubling year-on-year. Daily active user data for Telegram gambling bots is breaking records on a monthly basis. Nearly 60% of crypto bets are settled in USDT—anonymous, cross-border, and instantly cleared—each characteristic precisely evading traditional regulatory means.

Every time a gray channel is closed, capital flows in a more hidden direction. This is the secondary effect of prohibition and a pattern presented repeatedly throughout this report.

Public Security Incident or Proxy War?

The destination of capital overflow is Southeast Asia. It is not just a tourist destination but has become a new type of industrial infrastructure.

The latest reports from Interpol and the UNODC estimate that between 300,000 and 500,000 people are held in various gambling and scam parks across Southeast Asia, with annual profits of approximately 40 billion USD. The UNODC calls this a phenomenon at the level of a "humanitarian and human rights crisis." In the face of this threat, China's response is not singular but consists of three parallel lines of geopolitical operation.

In Myanmar: Tacitly allowing the Myanmar National Democratic Alliance Army (MNDAA) to launch Operation 1027 under the banner of "anti-fraud" to seize Laukkai by force. On the surface, it is a Myanmar internal affair; in substance, it is a China-backed proxy war using armed force to solve public security issues at the cost of reshaping border order. In Cambodia: Applying economic leverage to the Hun Sen family, with two chartered flights in 2024 extraditing over 1,200 suspects from Cambodia. In the Philippines: The existence or abolition of POGOs (Philippine Offshore Gaming Operators) has become a diplomatic bargaining chip in the South China Sea standoff.

The same public security issue is dismantled into three operations: armed proxies, economic pressure, and geopolitical exchange. The "anti-fraud operation" label in the news masks the true logic: this is a classic case of a major power projecting influence in zones of weak sovereignty.

Reading Path

This report does not require linear reading.

Policy researchers can focus on "Why prohibition is ineffective": Chapter 2 (Who Defines "Gambling"?) → Chapter 5 (The Lottery of 600 Million People) → Chapter 9 (The Singapore Model) → Chapter 11 (The House Always Wins).

Readers focused on China issues can track the undercurrents of capital: Chapter 3 (The Economics of Prohibition) → Chapter 4 (Macau: The Legal Valve) → Chapter 7 (Digital Casinos).

Southeast Asia and geopolitics observers can go directly to Chapter 6 (Northern Myanmar: The Gambling and Scam Empire), which details the park economic model and the logic of China's differentiated responses.

Readers interested in addiction mechanisms and behavioral economics can start from Chapter 1 (A Game of Negative Expected Value) to understand how casinos and lottery agencies precisely exploit cognitive loopholes.

Boundary Statement

This report does not provide moral judgments. "Gambling corrupts social morals" is a platitude that explains no mechanism.

It does not provide policy panaceas. "Complete liberalization" and "strict prevention" are not two options but two different modes of failure: the former expands the base of addiction (after legalization in 38 US states, betting volume reached 149 billion USD in 2024, with a simultaneous rise in problem gambling hotline calls), while the latter creates underground windfall profits (the current situation in China).

It does not pretend to be neutral. The selection of material is itself a judgment: dismantling 4,500 platforms is a tactical victory, but the strategic opponent—a system driven by neuroscience for demand, provided with infrastructure by crypto technology, and given physical sanctuary by weak sovereign states—is becoming more decentralized and difficult to locate.

When 600 billion in legal funds and trillions in illegal funds surge in parallel, the real question to ask is no longer "how to eliminate gambling," but what society is prepared to exchange for what when this behavior, as old as human history, dons the shell of blockchain and AI.

Games with Negative Expected Value

In 2023, Chinese public security organs solved 71,000 cases of cross-border gambling, seizing and freezing over 300 billion yuan in gambling-related funds. In the same year, the Ministry of Finance announced national lottery sales of 579.696 billion yuan, a year-on-year increase of 36.5%, setting a new record. In a country where gambling is strictly prohibited, legal lottery sales approach 600 billion yuan annually, while the capital involved in illegal gambling is measured in the trillions.

The contradiction between these two sets of figures cannot be explained by "greed" or "weak willpower."

Behind every deposit is a neurochemical reaction. The finger pressing "Recharge another 500" is driven by a force far more primitive than greed. This force existed before humans appeared, etched into the neural circuits of all organisms that rely on uncertainty for foraging. The brains of addicted gamblers are not constrained by the ability to calculate wages; instead, they are firmly locked in by an industrialized stimulation protocol precise to the millisecond. The design principles of this protocol share the same neural pathways as the mechanism by which heroin acts on opioid receptors.

The Second Bought with 500 Yuan

For a bricklayer earning a daily wage of 300 yuan, 500 yuan is nearly equivalent to two days of labor. This same person might haggle over two cents at a wet market or fuss over the placement of a single brick on a construction site. Yet, once they enter a mobile Baccarat interface, this economic rationality seems to vanish. This phenomenon has long been misunderstood as "greed" or "weak willpower."

Both explanations are untenable. If the motivation were simply "wanting to win money," continuous losses should prompt one to stop quickly. The reality is the opposite: problem gamblers increase their bets during long-term losses, sometimes even continuing to bet after "breaking even" until they hit zero again. Clinical interviews with deeply addicted individuals show striking consistency: many state that money is merely the entry ticket; what they truly seek is a mental state known as "The Zone."

That state is similar to the self-dissolution found in meditation, where the sense of time disappears and external pressures temporarily recede. Debts, foremen, family conflicts, and anxiety about the future are all muted during the few seconds the slot machine spins. Winning money actually becomes a distraction, as the result interrupts the numbing sensation brought by continuous betting, forcing the gambler to make a choice: "continue or leave." Many addicts subconsciously even hope for their balance to hit zero: only hitting zero can end the exhausting cycle.

This indicates that rational persuasion—telling gamblers they will lose money—is almost ineffective for addicts. Addicts are not buying the probability of winning, but a temporary shield from the real world. A chemical sedative without a prescription, injected through an audiovisual interface, lasting from a few seconds to several minutes each time, at a cost ranging from dozens of yuan to total bankruptcy.

So, where does this shielding effect come from? The answer lies in a set of pigeon experiments from the 1950s.

Skinner's Pigeons

In a Harvard basement, B.F. Skinner revealed the most unsettling law of behavioral psychology.

A pigeon that receives food every time it pecks a button (fixed-ratio reinforcement) will work methodically and stop once it is full. Switch to another scheme: food drops become random—sometimes it appears after one peck, sometimes not even after fifty. Under conditions of variable-ratio reinforcement, the pigeon will peck frantically and continuously until its physical strength is exhausted, even if it is no longer hungry.

The pigeon does not have a psychological defect. It is the sophisticated programming of evolution at work.

In the uncertain environments of antiquity, if hunters gave up entirely after three failed hunts, the species would have long been extinct. The brain evolved a reward system specifically to handle this situation: dopamine neurons in the Ventral Tegmental Area (VTA) do not release dopamine only when a reward is obtained, but release the most when the reward is most unpredictable. Neuroscience calls this "Reward Prediction Error" (RPE). When the outcome is better than expected, or completely unpredictable, dopamine release reaches its peak.

Uncertainty itself is the reward. This is the core parameter of the reward system.

The modern gambling industry has industrialized this parameter. Whether it is a VIP room at Wynn Macau or a "Sands Entertainment" app on a phone in a construction site dormitory, the underlying logic is a commercial variant of the Skinner box. Slot machines compress the frequency of "variable-ratio reinforcement" to the millisecond level: every spin, every deal of the cards, is a precise trigger for the brain's primitive foraging circuits. Every trigger releases dopamine. Every release drives "one more time." The cycle does not need a winning result to sustain itself, as long as the uncertainty remains.

PET scans provide direct evidence. The dopamine synthesis capacity of pathological gamblers is 16-17% higher than that of healthy controls, concentrated in the caudate nucleus, putamen, and ventral striatum. This does not mean these gamblers are happier, but rather that their reward system is in a state of chronic hunger. Tolerance to dopamine forces addicts to seek higher intensity and higher frequency stimulation to achieve the same "Zone" experience. This is almost identical to the tolerance curve of cocaine addicts.

Multiple neuroscientific reviews have reached a clear conclusion: the way variable-ratio reinforcement activates the brain's reward centers is "similar to cocaine and heroin."

If a white powder could reshape dopamine pathways in this way, the seller would face the death penalty in China. When the same neurochemical consequences are produced by an audiovisual interactive program, that program is not only legal but also contributes hundreds of billions of dollars in taxes and gray profits globally every year. Slot machines are dopamine syringes. The difference lies not in the effect, but in the medium.

The Business of "Almost Winning"

Skinner discovered the loophole. Modern gambling engineers turned this loophole into a design science precise to the first decimal point.

The reels of a slot machine stop between two jackpot symbols, with the third just one position away. The conclusion of probability theory is clear: this is a loss. A complete loss. Mathematically, it is no different from being ten positions away.

The brain, however, does not process it this way.

In 2009, an fMRI study published by Clark et al. in Neuron confirmed for the first time at the imaging level: a "near-miss" activates the ventral striatum and the anterior insula. These two brain regions are also activated during an actual win. In other words, an objective loss is encoded by the brain's reward system as a "partial victory." Subjects in subjective ratings described near-misses as "more unpleasant than total failure," while simultaneously reporting a stronger desire to continue playing. The emotional evaluation (negative) and the behavioral drive (positive) of an event are completely decoupled. The cognitive system says "this is bad," while the reward system says "do it again." The latter prevails.

A follow-up experiment by Sescousse et al. in 2016 in Neuropsychopharmacology further deconstructed the mechanism: the striatal response to near-misses in pathological gamblers is significantly amplified compared to healthy controls. This explains why the same slot machine has different "stickiness" for different people. The brains of addicts suffer from a more severe miscoding of near-misses; "almost winning" has a stronger behavioral drive for addicts.

Beyond near-misses, gambling engineering has developed more sophisticated tools of deception: "Losses Disguised as Wins" (LDW). In a multi-line slot machine, a player bets 10 yuan, the machine flashes lights, plays winning sound effects, and spits out 2 yuan. The net loss is 8 yuan. Yet, the signal received by the brain is "victory." Skin Conductance Response (SCR) measurements show that the physiological arousal produced by LDW is almost indistinguishable from an actual win.

The Dixon team at the University of Waterloo discovered that there is a precise optimal deception frequency for LDW: 19.6%. This is the peak of an inverted U-shaped curve. If the LDW ratio is too low, the deception is insufficient, and players leave out of frustration; if it is too high, players begin to notice the contradiction between their continuously declining balance and the frequent "winning" experiences. 19.6% is the point of maximized cognitive dissonance; at this frequency, players' overestimation of the number of wins is most severe, and their stay time is longest.

The precise tuning of cognitive deception parameters is not a matter of luck in product design. 19.6% is an engineering specification, belonging to the same category of precise numbers as the load-bearing coefficient of a bridge or the effective dose of a drug. Casino designers are calibrating the failure thresholds of the human perceptual system and then locking the product to run at that threshold.

A 2024 ERP study added key evidence at the electrophysiological level: when an LDW occurs, the Reward Positivity generated by the brain is statistically indistinguishable from the signal during an actual profit. It is not that the player "feels" they have won; it is that the brain is truly deceived at the level of neural signaling.

When DSM-5 Calls It Addiction

In 2013, the psychiatric community finally acknowledged the conclusion toward which all evidence pointed.

The DSM-5 moved "Gambling Disorder" out of the category of impulse control disorders and into "Substance-Related and Addictive Disorders." This was the first non-substance behavior in medical history to be formally classified as an "addiction."

The significance of this classification change extends far beyond psychiatry itself. This decision confirmed that gambling addiction shares a pathophysiological basis with opioid addiction. Both exhibit tolerance: requiring larger bets or doses to achieve the same level of excitement. Both have withdrawal symptoms: the irritability, insomnia, and anxiety when stopping gambling overlap significantly with reactions to stopping benzodiazepines. Both meet the core diagnostic criterion of "inability to stop despite serious negative consequences."

fMRI studies provided a direct comparison: when problem gamblers and cocaine addicts watch videos related to their respective addictions, the activation patterns in the ventral striatum show changes in the same direction—both are weakened relative to healthy controls. This supports the "Reward Deficiency Syndrome" hypothesis: those predisposed to addiction are born with an underactive reward system and are therefore driven toward extreme stimuli that can produce intense dopamine release. Gambling and drugs are simply different paths to the same neurochemical destination.

However, the logic of the legal system remains stuck in a 19th-century view of substances. Heroin is contraband, and sellers face the death penalty. A Baccarat app is an "entertainment product," and the harshest penalty for operators is administrative closure. The law punishes the act of "ingesting a certain chemical substance" while turning a blind eye to the "design that manipulates neural circuits." No regulations prohibit "designing a game that makes people addicted." Regulatory logic draws lines based on the medium, without questioning the effect at all.

In other words, a 19.6% LDW frequency is completely legal. The visual design of near-misses is completely legal. Algorithms that compress variable-ratio reinforcement to the millisecond level are completely legal. Every link in the addiction-manufacturing assembly line does not violate existing laws; the only thing constrained by law is "who is allowed to operate this assembly line."

For gamblers who are already addicted, these fine legal distinctions are meaningless. The ability of the addict's prefrontal cortex to inhibit impulses has already been damaged, and dopamine receptor density has already changed. What caused this damage was an app on a mobile screen, not liquid in a syringe. The damage is the same, the paths are different, but the legal treatment is worlds apart.

The Treatment Vacuum for 60 Million People

In China, these 60 million people whose brains have been altered are almost invisible.

Multiple epidemiological surveys estimate that the gambling addiction rate in mainland China is approximately 2.5% to 4% of the adult population. The global average for problem gambling is about 1.29% (according to a 2024 systematic review in The Lancet Public Health, covering 23 studies in 14 countries). In a country where gambling is completely banned, the addiction rate is two to three times higher than the global average.

This data needs to be deconstructed rather than simply attributed. There are three mechanisms for the high incidence of addiction in a prohibited environment. First: illegal gambling products are not subject to any consumer protection constraints; there are no deposit limits, cooling-off periods, or self-exclusion mechanisms. The LDW frequencies, near-miss designs, and betting caps that are regulated in legal casinos are all removed in underground apps. The addictiveness of the product is pushed to the extreme. Second: legal consequences prevent people from seeking help. If problem gamblers go to a hospital, they must admit to participating in illegal activities. Most choose silence. Third: a severe lack of a treatment system.

A survey published in Addiction Biology in 2019 provided a jarring figure: out of 110 psychiatrists in China, only 1 had received systematic training in treating gambling addiction. Not 1%. 1 out of 110.

71.8% of surveyed doctors believe that gambling addiction is an important public health issue and expressed a desire to be more involved in related treatment work. There is no lack of willingness. What is missing is the entire support system: training courses, clinical guidelines, insurance coverage, and referral channels. The National Health Commission has yet to include "Gambling Disorder" in the standard catalog of mental health services. While the DSM-5 completed the scientific characterization in 2013, China's public health system has still not made an institutional response 12 years later.

The root of the problem is not a lack of resources, but the logical consequences of the power of definition. If gambling "should not exist" (as the law has already declared), then there is no institutional prerequisite for the existence of a treatment system for gambling addiction. To admit that addicts need treatment is to admit the widespread existence of gambling behavior and to admit that the prohibition has limited effectiveness. In policy discourse, this is unacceptable.

Thus, 60 million people are trapped in a double denial: the law denies that their behavior is a disease, and the health system denies that this disease needs treatment. A survey by the China Lottery Research Center at Beijing Normal University further refined this picture: among 200 to 400 million lottery buyers, about 7 million meet the criteria for "problem gamblers," of whom at least 430,000 belong to the category of severe addiction. Even considering only the "legal" lottery-buying group, the scale of the problem is already visible, let alone those involved in underground gambling.

A 2024 survey covering 1,920 sports lottery buyers in six provinces showed that addiction risk is concentrated among young, low-income, male groups, particularly in central and western economic regions. The profile of these people overlaps significantly with typical underground gambling participants: no alternative upward mobility, no accessibility to mental health services, only a recharge button on a mobile screen that costs half a month's income.

A bricklayer who has lost 500 yuan has no hotline to call and will not tell anyone. In the Chinese social context, a gambler is a lawbreaker, a person with weak willpower, and a morally corrupt "gambling ghost." They are not seen as a patient whose brain reward system has been reshaped. These two identity definitions lead to completely different social responses: the former deserves punishment, while the latter needs treatment. China has chosen the former.

Prohibition has cut off legal supply but failed to touch the painful demand. It has created addicts while denying their existence. The United States has gone in the opposite direction: after legalization in 38 states, the volume of calls to problem gambling help hotlines has climbed simultaneously. Legalization has turned gambling from a hidden vice into a normalized consumer activity in one's pocket, comprehensively increasing accessibility. The problem with prohibition is not that it is "too strict," but that it manages supply while ignoring demand. The problem with legalization is not that it is "too loose," but that it treats accessibility as a harmless variable.

Both paths are avoiding the same question: when the brains of 60 million people have already been hijacked, who will provide a way out? The answer is not in legal statutes, nor in moral preaching. It lies in the direction represented by that 1 out of 110 psychiatrists who received training.

Who Defines "Gambling"?

Three in the morning. Two men stare at glowing smartphone screens, doing exactly the same thing: betting their entire balances on extremely uncertain outcomes.

One has opened 125x leverage on Binance, shorting the 5-minute K-line of Ethereum. According to academic simulation data, the survival probability at this leverage ratio is 1.68%, with an expected loss of $65.67 for every $100 invested. Average survival time: 22.5 minutes. His occupation is listed as "Crypto Trader."

The other has bet on the "Banker" in a mobile Baccarat app. The win rate is 48.94%, with an expected loss of $1.06 for every $100 invested. His label in the public security system is "Online Gambler."

The law's attitude toward the two is diametrically opposed. The former is protected by the exchange's user agreement, while the latter faces administrative detention or even criminal penalties. Mathematics' verdict on both is identical: in the long run, both have negative expected values. The difference lies not in probability, nor in risk, nor in addictiveness. It lies in who defines "gambling."

125x Leverage vs. Baccarat

If the definition of "gambling" were based on the degree of risk, 125x leverage contracts should be the most strictly prohibited products in human history.

The mechanism of perpetual contracts: a 0.8% fluctuation in the underlying asset's price triggers liquidation. Simulation studies based on Binance's historical data show that the average survival time for a 125x leverage account is 22.5 minutes, with a liquidation rate as high as 98.32%. For reference: the single-round survival probability of Russian Roulette is approximately 83%. From a purely probabilistic perspective, the survival rate of putting a revolver to one's temple for a single bet is 50 times higher than opening a 125x leverage contract.

This is called "financial innovation."

In 2024, the total trading volume of perpetual contracts across the world's top 10 centralized exchanges reached $58.5 trillion, doubling from $28 trillion in 2023. Internal research from Binance (2020) showed that 79% of users use more than 20x leverage. Capital evaporates in the millisecond jumps of K-line charts, flowing into the exchanges' insurance funds and the accounts of quantitative market makers. On the day in December 2024 when BTC flash-crashed by 7%, over $400 million in long positions were liquidated within hours. In an even larger crash in October 2025, the Hyperliquid platform alone generated $10.3 billion in liquidations.

The house edge for Baccarat is 1.06%. Mathematically, it is one of the gambling games closest to being fair for the player. If the person using 125x leverage were forcibly moved to a Baccarat table with the same principal, their expected loss would drop from $65.67 to $1.06. In other words: Baccarat's mathematical benevolence toward participants is 62 times that of a 125x contract.

An action with an EV (Expected Value) of -65% is called "investment" or "trading," while an action with an EV of -1.06% is called a "crime." The correlation coefficient between "gambling" in the legal sense and probability theory is zero.

The Expected Value Spectrum

If we arrange common "uncertainty betting behaviors" on the market by mathematical expected value from worst to best:

At the far left of the spectrum is the most exploitative product: the China Sports Lottery/Welfare Lottery. The payout ratio is about 50%, meaning an EV of -50%. For every 100 yuan spent on a lottery ticket, the expectation is to recover 50 yuan. This is legal. Not only is it legal, it is operated directly by the Ministry of Finance and sold publicly in supermarkets, subway stations, and convenience stores. In 2024, national sales reached a record high of 623.486 billion yuan.

Moving to the right are high-leverage crypto derivatives. Considering slippage, transaction fees, and funding rates, the comprehensive EV of a 125x contract is roughly between -65% and -40% (depending on holding time and market volatility). In most countries, these are legal financial products, governed by the exchange's own rules rather than gambling laws.

Continuing right, American Roulette. The EV is -5.26% (due to the double zero). European Roulette is -2.7%. These are legal in Las Vegas and Macau, but illegal in mainland China.

At the far right of the spectrum are the games that are mathematically most gentle to participants: Baccarat with an EV of -1.06%; Blackjack with an EV of about -0.5% when using perfect strategy. Under the Chinese legal framework, these two games are classified as "gambling," and operators face fixed-term imprisonment of 5 to 10 years (Crime of Opening a Casino under Article 303 of the Criminal Law).

When legal status is overlaid on this spectrum, a perfect inversion emerges: the products that are mathematically most exploitative are operated by the state, while the products that are mathematically fairest to participants are prosecuted by the state. The relationship between legality and expected value is not non-existent; it is negative.

So, what is the variable that determines legality?

If we mark the "profit destination" for each product on the spectrum: Lottery profits → Treasury public welfare fund accounts; Contract fees → Exchanges (Binance's average daily income from fees alone was about $3.9 million in 2024); Casino profits → Private gambling companies or individual bookmakers.

A pattern emerges: products where profits flow into the national treasury are legal; those where profits flow into licensed institutions are gray; those where profits flow into unlicensed individuals are illegal. The relationship between the level of expected value and legality is zero, while the relationship between profit attribution and legality is nearly one.

Whoever Issues the License Defines It

"Whether a product is gambling" is not a factual judgment. It is an administrative decision.

On December 31, 2017, trading Binary Options in the UK—a simple contract betting on the short-term direction of an asset's price—was considered a gambling activity under the jurisdiction of the Gambling Commission.

On January 3, 2018, the same British person doing the same thing, trading the same product, using the same interface, saw the legal nature flip to "investment." Regulatory authority shifted from the Gambling Commission to the Financial Conduct Authority (FCA).

The product did not change. The odds did not change. The user did not change. The only change: the header of an administrative document changed from "gambling" to "financial instrument." The gambler became an investor on the spot without changing any behavior.

Malta took the opposite path during the same period: moving binary options from its national gambling law into the EU's MiFID financial instrument framework. Under Malta's previous classification, binary options providers needed a license from the Lotteries and Gaming Authority. After the MiFID integration, the same business required an investment services license. The product remained the same, the legal definition flipped, and the required license moved from one office to another.

Across the Atlantic, the arbitrariness of the power to define has sparked even more intense jurisdictional wars.

The prediction market platform Kalshi obtained a Designated Contract Market (DCM) license from the U.S. Commodity Futures Trading Commission (CFTC) in 2020. This license granted Kalshi the legal status of a "derivatives exchange." During the 2024 U.S. election, Kalshi launched event contracts on "who will win the presidential election." The CFTC attempted to block it on the grounds that "this is similar to gambling and contrary to the public interest." Kalshi sued the CFTC. The federal court ruled in favor of Kalshi: the court determined that the CFTC had no right to use its own "public interest review," which was not authorized by Congress, to prohibit contracts already self-certified within the DCM framework. The court's wording was blunt: the ruling had nothing to do with whether the court liked the product.

Meanwhile, Polymarket was fined by the CFTC in 2022 for "operating an unlicensed derivatives trading platform" and was banned from accessing U.S. users. During the 2024 election, Polymarket processed a cumulative $3.3 billion in election bets, all from non-U.S. users. It wasn't until September 2025 that Polymarket obtained DCM status through a $112 million acquisition of the licensed exchange QCX, officially returning to the U.S. market in December.

The logical loop: Has a license → Derivative → Legal. No license → Gambling → Illegal. The license is issued by the same institutions that claim to "crack down on gambling." The product structures can be identical.

This loop was immediately attacked by state-level regulators. Between 2024 and 2025, Attorneys General and gambling regulatory boards in over 20 states launched lawsuits or injunctions against event contracts on Kalshi, Robinhood, and Crypto.com, alleging they were "unlicensed sports betting" or "illegal gambling." Federal district courts in Nevada and New Jersey held that the federal Commodity Exchange Act (CEA) took precedence over state gambling laws, and Kalshi won. The federal district court in Maryland held the opposite view, denying Kalshi's injunction request and ruling that the federal preemptive effect of the CEA does not necessarily extend to the realm of state gambling laws.

The same app is a legal hedging tool in Washington D.C. but may be illegal gambling software in Maryland. The answer to "is this gambling" depends on GPS coordinates. The power to define has itself become a weapon for federal and state governments to compete for jurisdiction and fine revenue.

China's Supreme People's Court drew the legal boundary of gambling from another perspective. Guiding Case No. 146 confirmed: trading platforms operating under the name of "binary options," if the profit and loss results are not linked to the actual magnitude of the underlying asset's price fluctuation (i.e., not "earn as much as it rises" but "double if you guess the direction right"), are essentially "betting on big or small, win or lose" and constitute the crime of opening a casino. The key criterion is "whether profit and loss are proportional to the magnitude of price change." How much you earn in a futures contract depends on how far the price moves; binary options only care if the direction is correct. This is currently the only operational standard in Chinese law to distinguish gambling from speculation.

This standard immediately raises a question: the profits and losses of 125x leverage contracts are indeed linked to the magnitude of price fluctuations (satisfying the "non-gambling" criterion), but a price movement of only 0.8% leads to a 100% loss of principal (the actual effect is equal to gambling). The crack between legal form and economic substance is torn wide open by high-leverage contracts.

China's Three Types of Gambling

The "gambling" targeted by Article 303 of the Chinese Criminal Law has an implicit prerequisite: the profit destination is not the state. When profits flow to the state, the same behavior receives a completely different name in the discourse system.

The first type of gambling is called "Public Welfare." In 2024, total lottery sales in China were 623.486 billion yuan. The public welfare funds raised, as announced by the Ministry of Finance, were 161.031 billion yuan. This indicates that lottery buyers collectively transferred a net loss of over 160 billion to the national treasury. The mathematical implication of a ~50% payout ratio is that it is the most exploitative of all "betting products" on the market, legal or illegal. If a private company designed a game where customers got back an average of 50 yuan for every 100 yuan spent, it would be classified as "fraud" or "unfair business practices" in most countries. When the operator is the Ministry of Finance, the classification becomes "social public welfare."

The second type of gambling is called "Investment." Retail investors account for about 70% of trading volume in the A-share market (CSDC data), the highest retail participation among major global stock markets. Research published by the Harvard Law School Forum on Corporate Governance in 2024 showed that in the 2014-2015 bull market cycle, 85% of retail investors lost money. The top 0.5% of profitable accounts earned 254 billion yuan, while the bottom 85% of losing accounts lost a total of 250 billion yuan. The two figures are almost equal.

Structurally, this is identical to a casino's profit model: the vast majority of participants transfer wealth to a tiny minority, and the platform (exchange stamp duty + brokerage commissions) takes a risk-free cut from every transaction. The difference is that the casino's house ensures a win through probabilistic advantage, while the "winners" in the stock market achieve the same effect through information advantage and capital volume advantage. The information gap between retail investors and institutions (asymmetric access to high-frequency data, corporate insiders, and policy trends) is equivalent in effect to the probability gap between gamblers and the casino.

The China Securities Regulatory Commission (CSRC) will not admit that "the stock market is statistically equivalent to a casino for retail investors." This is because capital markets carry the macro functions of financing and price discovery. These functions do exist; it's just that retail investors are almost never the beneficiaries of them. Free-market advocates also do not attack such structures; "willingness to bet and accept loss" is the price of market efficiency.

The third type of gambling is called "Economic Development." Chinese real estate has been the largest speculative market encouraged by the government for 20 years. The pre-sale system allows developers to collect full payment before houses are built, which is structurally equivalent to a futures contract: the buyer pays cash to purchase an underlying asset that will only exist in the future, bearing the delivery risk. A 30% down payment means 3.3x leverage. 70% of homebuyers entered the market with investment appreciation as their primary goal, rather than housing needs. Economists call the developer model of "using new home pre-sale funds to build old houses" a "variant of a Ponzi scheme," where the maintenance of the capital chain depends on a constant stream of new buyers.

In 2016, "houses are for living in, not for speculation" was written into national policy. By this time, the profit model of the entire industry was already highly dependent on speculative demand. After demand was compressed by policy, a chain reaction began: in 2024, new home prices recorded their largest annual decline in 9 years, and 39.1 billion square meters of completed housing sat vacant (a 72% increase from 2021). Evergrande was found to have inflated its revenue by $78 billion. Throughout the entire cycle, no developer or local official was prosecuted for "opening a casino." The reason is clear: land transfer fees accounted for 40-50% of local fiscal revenue at their peak. When the government is the largest "house" (selling land, collecting pre-sale funds, and obtaining taxes from property appreciation), the conditions for applying gambling laws are not met.

Only when funds flow out of these three "legal" pipelines—toward private bookmakers, overseas casinos, or underground networks—is the enforcement power of Article 303 of the Criminal Law activated. What is prosecuted is not the act of "betting on an uncertain outcome" itself. It is the fact that "the profit did not pass through the state."

The Universal Blur Zone

The aforementioned legal boundaries are being compressed from both ends by technological and product innovation.

The Meme coin launch platform Pump.fun on the Solana chain is a textbook case. In 2024, over 11 million types of tokens were launched on the platform. 98% went to zero within 24 hours. The platform collected over $500 million in fees. The product structure (betting → waiting for a random outcome → the vast majority losing everything → the platform making a certain profit) is identical to a casino. Pump.fun faces federal lawsuits alleging it "extracted nearly $500 million in fees" while 98% of its products collapsed in a day. To date, no major regulator has successfully classified it as "gambling." Because at the code level, Pump.fun is just a "decentralized token exchange protocol." Functional labels have replaced the substance of the behavior.

The broader Meme coin market reached a peak of $137 billion in 2024 before plunging 62% to $490 billion in 2025. An analysis report by Chainplay classified 55% of the studied meme coins as "malicious." Data from Merkle Science showed that $500 million was lost to meme coin scams in 2024, with 75% occurring on the X platform. The behavioral patterns of participants (chasing narrative hype, ignoring fundamentals, frequent entry and exit during short-term surges and crashes) are psychologically indistinguishable from gamblers chasing "hot tables" in a casino.

The loot box/gacha mechanism in the gaming industry is another sample of the failure of the power to define. In 2018, Belgium became the first country in the world to define paid loot boxes as illegal gambling. The Belgian Gaming Commission's definition criteria have the broadest coverage: even if virtual items are non-transferable and have no monetary value, as long as they are "valuable to the player" (such as rare skins), it constitutes gambling. Theoretically, this is the strictest loot box regulation in the world.

In actual practice: a study published by the University of California in 2023 showed that among the 100 highest-grossing iPhone games in Belgium, 82% still offered some form of paid randomized monetization mechanism. Game companies bypass legal definitions through visual redesigns (using "previews" instead of "unboxing" animations) or payment process adjustments (buying virtual currency first and then using it to draw). The power to define does not automatically equal the power to enforce.

The Netherlands adopted a different legal standard: items must be "transferable" or "have real monetary value" to be considered gambling bets. The Dutch Gaming Authority fined EA 5 million euros based on this (FIFA Ultimate Team packs were deemed gambling), but EA appealed to the Administrative Jurisdiction Division of the Council of State and won. The court determined that FIFA packs did not meet the "transferable" threshold of Dutch law.

Japan banned "Complete Gacha" in 2012, a model requiring players to collect a full set to obtain an ultimate reward. Single draws remain legal. The ban only blocked one variant with the strongest addictive properties; the main business remains unaffected. Japan's Pachinko industry has an annual output of over 20 trillion yen, bypassing gambling laws technically through the "three-shop system" (the casino exchanges balls for tokens, which are then exchanged for cash at a separate shop): because the balls are not directly exchanged for money, it is "not gambling."

The legal concept of "gambling" is losing its coverage capacity. When tokens that 98% go to zero are called "DeFi," when contracts with an EV of -65% are called "derivatives," and when 82% of unenforced bans are written in the statutes, the power to define is being hollowed out from within by product innovation and enforcement inertia.

The Profit Destination

Recapping the threads of this chapter, three hidden rules governing the operation of the global "power to define gambling" can be distilled:

Rule One: Profits flow to the national treasury = Legal. No matter how low the expected value (Lottery -50%), no matter how vulnerable the target population (low-income lottery buyers), the product is named "public welfare" rather than "gambling."

Rule Two: Profits flow to licensed financial institutions = Legal or Gray. Even if the liquidation rate is 98% (125x contracts), even if 85% of participants lose money (A-share retail investors), the product is protected by financial law because it nominally carries the discursive functions of "price discovery" or "risk management."

Rule Three: Profits flow to unlicensed private individuals = Crime. Even if the game rules are the fairest to participants (Baccarat EV -1.06%), even if the amounts are insignificant, as long as the profit destination is not the state or a licensed institution, the behavior is categorized under Article 303 of the Criminal Law.

These three rules explain the position of every product on the "gambling legality spectrum." These three rules are not a conspiracy theory; they are an honest description of the governance logic shared by all modern states. The only difference lies in the degree of compression of the spectrum: 38 U.S. states have legalized sports betting (a wider spectrum), while China retains only the lottery as a legal outlet (an extremely narrow spectrum). A narrow spectrum means more funds are pushed underground and offshore. This is the theme of the next chapter.

Based on these three rules, a judgment can be made on recent trends: before 2027, the "guessing" products of the China Sports Lottery (Jingcai Football, Jingcai Basketball) will evolve into regulated sports betting. The driving force is not a change in policy philosophy, but the competitive pressure from underground markets. When underground sports betting offers a payout ratio of 95-98% while the official Jingcai offers only 69%, the flow of funds underground is difficult to reverse. The response of the Sports Lottery system will be to increase payout ratios (approaching 75-80%), expand event coverage, and shorten settlement cycles. At that time, the definition of "gambling" will narrow again: as long as sports result guessing is conducted through Sports Lottery terminals, it will not be called "gambling" but "Jingcai experience upgrade." The formal ban remains unchanged. The actual gambling channels expand. The profit destination remains the fiscal accounts—this is the perpetual motion machine of the power to define self-exemption.

The Balloon Effect

It is a Friday night in a village in Shaoyang, Hunan. The lights in the three chess and card rooms are brighter than those in the village committee office.

Old Liu sits by the window, in front of him are plastic mahjong tiles and a red plastic basket containing a few scattered ten and twenty-yuan bills. It took less than three months for the stakes at this table to rise from "one yuan per point" to "fifty yuan per point." The professional composition of the four people at the table is typical: two masons (Old Liu and Old Zhou from the neighboring village), a middle-aged woman who runs a grocery store in town, and a young man who just returned from a construction site in Guangdong for the New Year. The young man plays the largest hands, losing two hundred yuan in a single round without batting an eye.

The next table is even more intense. Another young man back from Shenzhen has set up a phone stand and is teaching four middle-aged men how to download an app called "Sands Entertainment." The minimum deposit is 500 yuan, betting on "Banker" or "Player" in Baccarat. The young man sends WeChat voice messages while teaching the operations. Every time he gets someone to top up, he immediately receives a "referral reward" of several dozen yuan on WeChat. This young man does not gamble. The four middle-aged men sitting next to him are the ones gambling. The young man's role is called an "agent," and his monthly income is three to five times higher than hauling bricks at a construction site.

This village has no library. No cinema. No KTV. The nearest internet cafe is in town, a 25-minute ride by electric scooter. After dark, the only places still open and lit are these three chess and card rooms. Old Zhang, an auxiliary police officer from the local station, occasionally rides by on his scooter, his headlights sweeping across the rolling shutters of the card rooms. Every time he passes, someone chases him out to hand him a pack of Hard Chunghwa cigarettes. Old Zhang takes the cigarettes, does not enter, asks no questions, and turns the corner to leave.

This is a standard Friday night in hundreds of thousands of similar villages across the country. The law declares that gambling "should not exist." No one answers the prerequisite question: other than gambling, where do the villagers go at night?

The Balloon Effect

In economics, there is a concept called the "balloon effect": squeeze one end of a balloon, and the other end expands. If the legal supply for a demand is prohibited, illegal supply will fill the void with higher profit margins. The demand remains unchanged; only the channel changes. The channel shifts from transparent to black-box operations, and regulation subsequently fails.

The most complete historical sample is the American Prohibition (1920-1933). Before the ban, New York City had about 15,000 legal bars that paid taxes, underwent health inspections, and held licenses. After the ban was implemented, the number of illegal bars (speakeasies) ballooned to between 30,000 and 100,000, with no licenses, no health standards, and no consumer protection. Al Capone's criminal empire had an annual income of $100 million (in 1920s currency), paying $500,000 a month in bribes to Chicago police and politicians. During the 13 years of Prohibition, American per capita alcohol consumption returned to 60-70% of pre-ban levels after a brief decline. The ban failed to erase the desire for alcohol; instead, it stifled legal supply and gave birth to a massive criminal empire, turning what would have been tax revenue into bribes.

China's ban on gambling has produced a structurally identical outcome, only on a scale two orders of magnitude larger.

Since the early days of the People's Republic of China, gambling has been banned nationwide, with Article 303 of the Criminal Law explicitly stipulating the "crime of opening a casino" and the "crime of gambling." Under more than seventy years of high-pressure crackdowns, legal lottery sales amount to approximately 600 billion RMB annually, with a payout rate of 50% (the house takes half). Estimates of the size of the underground gambling market vary widely: the most conservative is 1 trillion RMB, the most aggressive is 18 trillion RMB, with most industry analyses centering on the 6-10 trillion RMB range. The International Center for Sport Security (ICSS) estimates that approximately $600 billion in cross-border gambling funds flow out of China annually. Taking a conservative median, the ratio of legal lottery to illegal gambling is about 1:10. In a country with a total ban on gambling, the scale of illegal gambling is ten times that of legal gaming.

The gap in profit margins is even more pronounced. The house edge in legal casinos (such as those in Macau) is between 1-5%, with operating profit margins around 15-25%. The payout rates of Chinese underground gambling websites are usually set at 95-97% (appearing "fairer" to gamblers), but after adding the money laundering premium (capital channel fees of about 3-8%), protection umbrella costs (daily payoffs to auxiliary police/village officials), and technical infrastructure costs (servers, app development, customer service), the comprehensive profit margin can still reach 40-60%. This is far higher than the legal gaming industry. Enormous profits attract supply, and supply continues to emerge.

When taking down a gambling website requires six months of investigation by public security, three months to close the net, and several months of court trials, while setting up a new platform takes only 72 hours and 5,000 USDT (about 36,000 RMB), the cycle of law enforcement will always be longer than the cycle of replication. In 2024, Chinese public security shut down more than 4,500 online gambling platforms. In the same year, it was estimated that tens of thousands of similar platforms were still in operation. This is entirely consistent with the logic during Prohibition: for every speakeasy closed, three new ones opened on the next block.

Level One: Village Gambling Tables

In the three-tier structure of underground gaming, the lowest, most dispersed, and hardest to combat is village-level gambling.

Some villages have more than 30 mahjong parlors, which locals call the "largest tertiary industry." Stakes range from 1 yuan to 500 yuan, with high frequency (3-5 times a week) and broad participation (mainly middle-aged and elderly men, with increasing female participation in recent years). A common escalation path: starting with "friendship matches" at 1 yuan per point, where wins and losses are a few dozen yuan and everyone is just looking for fun. Within three months, it rises to 50 yuan per point, with wins and losses of three to five hundred yuan a night. Six months later, someone suggests "the stuff online is more exciting," and mobile app gambling begins to penetrate: Baccarat, Dragon Tiger, slot machines, open 24 hours a day with no cap on stakes. Village gambling tables serve as an entrance to the abyss of online gambling. The transition between the two has almost no threshold. In 2024, the Ministry of Public Security held a special "On-site Meeting on Combating Rural Gambling," rarely listing rural gambling as a key focus of special governance.

The effectiveness of crackdowns is cyclical: arrests are followed by releases, releases by reopenings, and reopenings by more arrests. Mobile casinos have become the primary means of evasion, changing locations every day (today in Old Wang's living room, tomorrow in a construction shed at the village entrance, the day after in the woods by the river). Times also change, shifting from fixed evenings to the early morning or afternoon to avoid the police station's shift schedule. Some mobile casinos hire dedicated lookouts at the village entrance; as soon as an unfamiliar vehicle enters the village, everyone disperses.

Rural gambling is difficult to eradicate, and the lack of law enforcement resources is not the only reason. Chess and card rooms fulfill social functions beyond gambling: information exchange (who has work, who is hiring, whose daughter is getting married), social maintenance (daily lubrication of neighborhood relations), and even dispute mediation (settling matters over a few rounds of mahjong is faster than finding the village committee). For the left-behind elderly and middle-aged, the card room is the only public gathering space, the only social network node, and the only place that still provides light and human voices after dark. Abolishing card rooms, while banning gambling, also destroys the village's only remaining social infrastructure.

Grassroots law enforcement is well aware of this. "Arresting without ending" is not entirely due to incompetence; it is partly a tacit tolerance of social functions. The police station chief knows: if all three card rooms are closed, the village elders will have nowhere to go, and there will be far more petition letters than gambling cases.

Level Two: The Agent Pyramid

One level up from the village gambling tables is the industrially operated online gambling agent system.

The structure is a five-tier pyramid: Major Shareholders (platform owners, usually in Cambodia/Philippines/Dubai) → Shareholders (technical operations + fund management, a core team of 2-5 people) → General Agents (regional heads, divided by province or city) → Agents (front-line promoters, numbering in the thousands to tens of thousands) → Members (gamblers, numbering in the hundreds of thousands to millions). Each level takes a 10-15% commission, layered progressively. The work of a bottom-level agent is simple: post advertisements in WeChat groups, Tieba, and short video comment sections to recruit people to register. Every time a recruit tops up, the agent immediately receives 10-15% of the deposit amount as commission. Some platforms also have a "daily active reward": as long as the recruited member places a bet that day, the agent receives an additional 1-3% of that member's total daily wager.

The young man teaching people to download apps in the card room at the beginning of this article is at the second-to-last tier of this pyramid. His monthly income of ten to thirty thousand yuan is three to five times higher than hauling bricks at a construction site. His risk awareness is close to zero: "It's not like I'm running a casino; I'm just recommending an app." Court verdicts usually tell such people: the act of recommendation constitutes the "crime of aiding in opening a casino," with a prison sentence of one to three years.

The scale of this system has surpassed most people's imagination. Alvin Chau's Suncity Group (before its collapse in 2022) had a monthly betting volume exceeding 100 billion HKD, with an agent network spanning all provinces in mainland China and more than 20,000 subordinate agents. The Sensen Group online gambling case involved 2.93 million registered users and a gambling turnover of 82.3 billion RMB. These figures are close to the user scale of a medium-sized internet company. The difference is: the monetization rate of an internet company is usually 1-3%, while the rake of a gambling platform is 5-15%; the economic output of every active user is far higher than that of any legal app.

The distribution of legal crackdowns is highly uneven. In 2024, public security solved 73,000 gambling cases and prosecuted 19,000 people. Among those arrested, agents accounted for 68%. Those bottom-level agents earn a few thousand to tens of thousands of yuan a month, register with personal phone numbers, and collect payments via personal WeChat accounts, possessing almost no counter-reconnaissance capabilities. The major shareholders at the top of the pyramid use multi-layer VPNs, virtual identities, and encrypted communication tools (Telegram/Signal), staying overseas with assets scattered under the names of shell companies in multiple countries. The "results" of 73,000 cases sound impressive. Divided by the trillion-level underground market scale, the reach rate is approximately one in a thousand. Of that one-thousandth reached, 68% are at the bottom of the pyramid.

Level Three: USDT Cross-border Channels

The financial arteries of the pyramid used to be bank transfers and traditional underground banks. After 2020, USDT (Tether) mirror-accounting underground banks became the mainstream channel, drastically changing the traceability of gambling funds.

The mechanism is not complex. The gambler gives RMB to a domestic recipient A (usually via WeChat/Alipay/bank transfer). A does not remit this money abroad but instead transfers an equivalent amount of USDT (1 USDT ≈ 7.2 RMB) from their own crypto wallet to a wallet designated by the overseas gambling platform. The gambler's account is immediately credited, and they can start betting. Throughout the process, there are no cross-border bank transfers and no foreign exchange purchase records. What the bank monitoring system sees is just an ordinary domestic RMB transfer, possibly labeled as "payment for goods" or a "loan." Traditional anti-money laundering tools (large transaction reports, suspicious transaction reports) are almost useless against this.

The scale of this system exceeds the level of individual cases. In the Hubei Shayang case reported by the Ministry of Public Security in 2024, a single USDT underground bank involved a turnover of 400 billion RMB, serving multiple gray industries including gambling, telecommunications fraud, and smuggling. The first case defined as the "crime of aiding information network criminal activities" involving USDT gambling payments involved 120 million RMB. Another virtual currency underground bank case involved 17 billion RMB. These cases are just the tip of the iceberg that has been detected and publicized.

Compared to traditional underground banks, the difficulty of law enforcement for USDT channels has grown exponentially. Traditional underground banks rely on corporate bank accounts and personal bank cards to form a capital chain; every transfer leaves a trail within the banking system, and one can trace upstream and downstream by following the bank statements. USDT mirror-accounting achieves "funds not crossing the border": domestic RMB and overseas gambling funds each form an independent closed loop, and the only connection point is a transfer between two anonymous wallet addresses on the blockchain. In the Hubei Shayang 400-billion-turnover case, it took investigative organs 18 months to map out the capital topology using on-chain analysis tools.

This is entirely consistent with the logic of rum runners during the Prohibition era. In the 1920s, rum runners transferred alcohol on the high seas to bypass Coast Guard inspections. In the 2020s, USDT mirror-accounting moves funds on the blockchain to bypass the monitoring of the banking system. The ban remains the same, the technology has changed, and the result is the same: funds reach places the law does not want them to go. The only difference is speed; a smuggling ship takes days, while a USDT transfer takes seconds.

Protection Umbrella Economics

The scene at the beginning, where a cigarette is handed to auxiliary police officer Old Zhang, is not a literary imagination.

In cases reported by the Discipline Inspection Commission, the fee for auxiliary police providing "risk control services" to underground casinos is 1,500-3,000 RMB per day. The service content is clear: notifying the casino of inspection times in advance (ensuring the casino is cleared 15-30 minutes before the patrol arrives), providing information on other casinos being raided in the jurisdiction (used to judge current risk levels), and occasionally stepping in to "coordinate" disputes caused by gambling debts (to prevent police reports). The profile of officials corrupted by gambling shows a regular distribution: 62.5% are aged 29-49, and their positions are primarily grassroots police station officers, auxiliary police, and village officials.

This corruption income is by no means a so-called "extra" gain; it is actually a fixed item in the casino's operating budget, alongside rent, utilities, and staff wages. Casino owners view it as a natural security cost rather than a bribe. Even after removing the "protection umbrella fees," the casino's profit margin remains at 30-40%. Without a protection umbrella, the expected survival time of a casino shortens from months to days. What the protection fee buys is "time"—the time to receive a warning before law enforcement arrives.

Al Capone paid $500,000 a month in bribes to Chicago police and judges in the 1920s. A hundred years later, an auxiliary police officer in a Chinese village receives a pack of Hard Chunghwa or 1,500 RMB in cash every day. The amounts differ by several orders of magnitude, but the economic logic is identical: prohibition creates enormous profits → enormous profits require protection → protection requires bribes → bribery funds come from the enormous profits. It is a self-sustaining closed loop. As long as the ban exists and demand is not extinguished, corruption transcends the moral scope of individual officials and becomes an inherent output of the system.

Disciplinary education cannot solve the deep-seated problem. An auxiliary police officer earns a monthly salary of 3,000-5,000 RMB, while a casino owner is willing to pay an amount equivalent to the officer's monthly salary every day in exchange for safety. Where does the economic rationality for refusal come from? If you replace one auxiliary officer, the new one faces the same interest structure, and the probability of making the same choice is extremely high. The crux lies in the incentive mechanism, not individual character. The ban creates enormous profits, the enormous profits create the economic foundation for corruption, and the economic foundation determines the probability distribution of individual choices. Catching one auxiliary officer does not solve the overall dilemma, just as catching one agent does not solve the pyramid problem.

Why It Never Ends

Back to the data.

In 2024: 73,000 gambling cases, 19,000 people prosecuted, 4,500 platforms shut down. These numbers grow every year, and every year-end report emphasizes "significant results." Meanwhile, the estimated size of the underground gambling market also grows every year. The simultaneous rise of both sets of numbers indicates that while the intensity of crackdowns is increasing, the market size is growing even faster.

Three core reasons keep law enforcement perpetually behind the market:

Demand cannot be eliminated. The first chapter already argued: China has about 60 million gambling addicts, plus hundreds of millions of occasional participants. Addiction is a change at the neurochemical level; the reshaping of dopamine pathways is not reversed by legal statutes. Demand is rigid and unaffected by supply-side crackdowns. If one platform is taken down, addicts will find a replacement within 24 hours.

Supply is infinitely replicable. An entry cost of 5,000 USDT means that anyone with basic technical skills can set up a new platform within 72 hours (buying ready-made source code, renting overseas servers, integrating USDT payment channels). Take down one, and ten replacements appear. In a market where entry barriers are near zero, it is impossible to clear it by attacking supply. This is a basic law of economics that does not fail because of increased law enforcement intensity.

Capital channels are unilaterally upgrading. From bank transfers to traditional underground banks to USDT mirror-accounting, every technological upgrade has increased the difficulty and cost of tracking. The evolution of law enforcement tools is destined to be slower than that of money laundering tools. The speed of solving a 400-billion-RMB case in 18 months can never catch up with the speed of launching a new platform in 72 hours. The time gap exists objectively and cannot be bridged simply by deploying more police force.

Dividing 73,000 cases by a trillion-level market scale yields a reach rate of about one in a thousand. 68% of those caught are bottom-level agents. The ban has failed to eliminate gambling; instead, it has pushed it into darker, more hidden, more out-of-control, and harder-to-intervene places. The larger the underground market, the harder it is for addicts to get help (seeking help is equivalent to admitting to a crime; the first chapter already demonstrated the severe lack of a gambling addiction treatment system in China), the harder it is to monitor capital outflows (the anonymity of USDT channels will only continue to enhance with technological iterations), and the firmer the economic foundation for corruption becomes (the closed loop of enormous profits and protection needs repeatedly reinforces itself).

The ban has created a self-reinforcing cycle: the more it is banned, the larger it grows; the larger it grows, the more it is banned; the more it is banned, the more underground it becomes; the more underground it becomes, the more uncontrollable it is. This cycle has no inherent termination mechanism. The only variables that could break the loop are demand-side intervention (addiction treatment systems) or legal alternative supply (regulated legal gaming), neither of which exists within China's current policy framework.

The result is a paradox: the major economy with the world's strictest gambling ban has the world's largest underground gambling market. The world's most intensive anti-gambling enforcement actions (73,000 cases/year) have the lowest reach rate (one in a thousand). One of the countries with the world's highest gambling addiction rates (2.5-4%) has the lowest treatment coverage (only 1 in 110 psychiatrists has received training). These three "mosts" point to the same conclusion: as a policy tool, the ban's inhibitory effect on gambling demand is near zero, while its distortion effect on the supply structure is near maximum.

When this system needs a legal pressure relief valve, there is one place that undertakes this function. That place is a 15-minute walk from the Zhuhai Gongbei Port, absorbing that small portion of gambling demand from mainland China that is willing and able to go abroad. It is called Macau.

04. Macau: The Legal Valve

In 1849, on a cobblestone street behind the Senado Square on the Macau Peninsula, a Chinese merchant received a piece of paper from a Portuguese colonial official. Written on it in Portuguese was a "Fan-tan Operation License." The stakes were small, just a few copper coins per round. The significance of this paper lay not in the gambling itself, but in establishing an institutional logic that would endure for 177 years: gambling could exist legally as long as its profits flowed into the colonial treasury.

The colonial government's motive was simple: tax revenue. Macau, which had just declared gambling legal in 1847, was undergoing a severe fiscal crisis. As the opium trade withered, this 33-square-kilometer peninsula desperately needed new sources of income, and gambling filled the gap. From the beginning, gambling for Macau was not just an entertainment industry, but the cornerstone of its finances.

For over a century, from 1847 to 1961, Macau's gambling industry underwent multiple management adjustments. Initially, Fan-tan houses were scattered throughout the streets and alleys, small in scale and primarily serving local residents and passing merchants. By the 1930s, the Hou Heng Company and Tai Heng Entertainment Company successively obtained exclusive rights, and casinos began to centralize operations. However, the true turning point came in 1961: Stanley Ho won the monopoly with a record bid of 3.167 million Hong Kong dollars. Over the next 40 years, he transformed Macau from a dilapidated colonial town into the gaming center of Asia.

On a Tuesday afternoon in 2019, in the corridor of a VIP room at the Venetian Macao, Ah Jie leaned against a marble pillar waiting for guests. He was 32 years old, from Putian, Fujian, and had been in the industry for three years. Known in the circle as a "Bajai" (tout), he was at the bottom of the junket system (Die Ma Tsai). His job was to bring mainland guests to the gambling tables; he earned a commission based on the "rolling chip turnover" of whatever the guests wagered. The days of earning 200,000 RMB a month lasted for three years, and Ah Jie thought this life would continue forever.

Two years later, this industry completely vanished.

Three Eras

The 177-year history of Macau's gaming industry can be divided into three policy eras, each corresponding to a different degree of the valve's opening.

The first era was Stanley Ho's exclusive dynasty. In 1961, Ho founded the Sociedade de Turismo e Diversões de Macau (STDM) and obtained a 40-year gaming monopoly. One licensee, one government, one narrow pipe. Ho's business genius lay in upgrading casinos from low-end Fan-tan houses to modern integrated resorts, introducing Baccarat, Roulette, and slot machines. VIP rooms already existed during this period, but the junket population numbered only a few hundred, handling funds in the billions of HKD annually. The scale was manageable, and the government had a clear understanding of the industry's operations. A valve opened only slightly: there was capital outflow, but the volume was limited.

The second era was the explosion of gaming liberalization. In 2002, the Macau SAR government decided to break STDM's monopoly, issuing three main licenses (Galaxy, Wynn, Sands China), each of which could further split off a sub-license (Melco Resorts, MGM China, SJM Holdings), eventually forming a competitive landscape of six licensed companies. The Las Vegas giants brought not only capital but also industrial operational experience. The Venetian Macao, opened by Sands China in 2007, cost $2.4 billion and featured 3,000 suites, becoming the largest single building in Asia. Revenue curves soared accordingly: from $2.36 billion in 2002 to $10.4 billion in 2007, $33.9 billion in 2011, and reaching a historical peak of $45.2 billion in 2013. It grew 19-fold in 11 years. In 2006, Macau's gaming revenue officially surpassed that of Las Vegas, making it the world's number one gambling city. At its peak, VIP room revenue accounted for over 70% of total revenue. The number of junket operators expanded from a few hundred to thousands, forming a massive recruitment network covering every province.

This was not a story of tourism success. The true core was that the capital outflow channel had been expanded from a village road into a highway.

Behind the 19-fold growth was a key question: where did the money come from? The answer was not "tourism consumption." Macau receives about 30 million tourists annually, with an average stay of 1.2 days. If revenue relied primarily on tourist spending, per capita spending would need to reach $15,000 to support an annual revenue of $45.2 billion. Such per capita expenditure is unimaginable in any tourist destination. What truly supported this revenue curve was the astronomical betting volume in VIP rooms: a small number of ultra-high-net-worth individuals (mostly mainland Chinese businessmen and officials), with single bets in the millions and cumulative rolling chip turnover in the hundreds of billions. The betting volume of one VIP guest in a single night could exceed the total consumption of ten thousand ordinary tourists.

The third era began on November 27, 2021, the day Alvin Chau was arrested.

Dead Chip Economics

To understand Macau's true function, one must first understand a special type of chip: the "dead chip" (also known as rolling chips).

The biggest difference between dead chips and ordinary casino chips is that they cannot be exchanged for cash. Dead chips can only be used for wagering. If you win, the casino pays out in "cash chips"; if you lose, the dead chips return to the casino. Cash chips can be exchanged for Hong Kong dollars or US dollars at the casino counter at any time.

This seemingly simple distinction constitutes a sophisticated currency conversion system.

The complete process is as follows: A mainland guest carries RMB (or remits it through an underground bank) to find a junket operator. The junket operator converts the RMB into dead chips at an agreed exchange rate and gives them to the guest. The guest takes the dead chips to the gambling table. If they win, they get cash chips; if they lose, the dead chips go to the casino. Regardless of winning or losing, as long as they participate in the table cycle, a portion of the RMB completes its conversion into HKD or USD. The entire process leaves no bank transfer records, requires no foreign exchange purchase approval, and involves no capital control filings. The gambling table is the intermediary carrier, and the dead chip is the conversion medium.

The commission extracted by the junket operator is 0.7% to 1.3% of the rolling chip turnover. The numbers might seem small, but Suncity Group (under Alvin Chau) had a monthly rolling chip turnover exceeding 100 billion HKD. 0.7% of 100 billion equals 700 million HKD per month. This was the monthly income of just one junket company.

A report by the Congressional-Executive Commission on China estimated that annual capital outflow through Macau was approximately $202 billion. Global Financial Integrity (GFI) estimated that total illegal capital flight from China between 2005 and 2011 amounted to $2.83 trillion. Suncity Group's online betting volume alone exceeded one trillion RMB per year.

These figures describe not the scale of gambling, but the scale of banking.

A typical VIP client scenario: A real estate developer from a certain province carries 20 million RMB in "gambling funds" (pre-remitted to the junket operator's Macau account via an underground bank). The junket operator converts the 20 million RMB into approximately 24 million HKD worth of dead chips at the agreed rate. The developer "rolls" the dead chips at a Baccarat table: betting 1 million, winning gets them 1 million in cash chips plus their 1 million in dead chips back; losing means losing 1 million in dead chips. After several rounds of betting, assuming a total loss of 3 million HKD (the house edge in Baccarat is only 1.06%, so in the long run, this loss is essentially a "service fee"), the remaining 2100 million HKD is exchanged from cash chips into Hong Kong dollars and deposited into a Hong Kong bank account. The entire process: 20 million RMB enters, 21 million HKD (or equivalent USD) flows out. Capital controls are completely bypassed. The "loss" at the table is the cost of currency exchange.

The role of the junket operator goes far beyond "intermediary." In actual operation, the junket operator simultaneously plays three roles. The first is as a lender: providing credit lines to guests (ranging from hundreds of thousands to tens of millions), so guests do not need to carry cash across the border and can settle after gambling. This is equivalent to an unlicensed private bank. The second is as a money changer: completing the conversion from RMB to HKD and USD, with the 0.7-1.3% commission acting as the fee. The third is as a debt collector: when guests fail to repay, the junket operator employs a violent collection network. Debt collection gangs in the coastal areas of Fujian and Guangdong have long collaborated with the junket system.

Combined, these three functions make the junket operator a complete financial institution: deposit-taking, lending, settlement, and collection. It just so happens that the place of business is next to a gambling table, and the financial license does not exist.

After the Door Closes

On November 27, 2021, Alvin Chau was arrested in Macau. The 289 charges included illegal gambling operations, organized crime, and money laundering. In January 2023, the sentence was handed down: 18 years in prison and a fine of 24.865 billion MOP.

The Alvin Chau case was not just the fall of one man. Suncity Group's agent network spanned every province in mainland China, with monthly betting volumes exceeding 100 billion HKD and annual revenues in the tens of billions. Suncity's online platform also operated internet gambling businesses with even more staggering betting volumes. This was the largest tree in the era of Macau's VIP rooms. After cutting down this tree, the junket operators understood that winter had arrived for the entire forest.

The chain reaction from the Alvin Chau case spread rapidly. Levo Chan of Tak Chun Group was investigated during the same period. Many small and medium-sized junket companies voluntarily deregistered or ceased operations in early 2022. The number of VIP rooms plummeted from over 200 at the peak to almost zero. The number of practitioners in the entire industry (junket operators and their subordinate touts, customer service, and debt collectors) was estimated to exceed ten thousand; within less than half a year, they collectively became unemployed or changed careers. Some returned to their hometowns to do business, some became online agents, and others went to casinos in Cambodia and the Philippines to continue similar work.

In June 2022, the Macau Legislative Assembly passed a new gaming law. There were four core changes: prohibiting junket intermediaries from directly sharing gaming revenue; gaming enterprises must take direct responsibility for VIP business; intermediaries must be locally registered Macau companies and are not allowed to lend to clients; and strengthening anti-money laundering and source-of-funds scrutiny. In January 2023, new contracts for the six gaming enterprises took effect, with the license term changed to 10 years. As a profession, the junket operator was legally eliminated.

The effect was immediate in the numbers. The share of VIP revenue dropped from over 70% at its peak to about 25% by 2025. The mass market now contributes 75% of revenue. Total gaming revenue for the full year of 2025 was 247.4 billion MOP (approximately $30.9 billion), recovering to 84.7% of 2019 levels. While the surface numbers are recovering, the internal structure has changed drastically.

Ah Jie completed his last deal before the Spring Festival in 2022. Only two guests came that month, both old regulars, with a combined betting amount of less than 2 million HKD. News of the new gaming law had already spread through the circle. Ah Jie left Macau before the law was officially implemented. Returning to Putian, Ah Jie used his savings and remaining connections to become an online agent, helping mainland guests open accounts and top up on Southeast Asian gambling websites. His monthly income is only 30,000 to 50,000, less than a quarter of what it used to be. The upside is not having to worry about being stopped and questioned by the Judiciary Police on the streets of Macau.

The Balloon Re-inflates

Ah Jie's career transition is not an isolated case; it is a microcosm of the entire industry.

In 2025, Macau's total gaming revenue recovered to 84.7% of 2019 levels. On the surface, the recovery momentum seems acceptable. But the structure has completely shifted: VIP contributes only 25%, while the mass market contributes 75%. During the peak, VIP contributed an absolute value of over $30 billion annually; in 2025, it was less than $8 billion. Over $20 billion in annual VIP funds evaporated from Macau.

These funds went in three directions.

The first direction is Southeast Asia. Sihanoukville, Cambodia, saw an influx of Chinese-invested casinos between 2016 and 2019, with over 90 in operation at the peak. The Philippines' POGO (Philippine Offshore Gaming Operators) system employed over 100,000 Chinese workers in online gaming in 2019. Casinos in the special regions of Myanmar's Shan and Kachin States operate under the protection of warlords. Industry analysis estimates that after Macau's crackdown, over 340 casinos are operating in Southeast Asia, absorbing the VIP clients overflowing from Macau. Regulation in these casinos is far less stringent than in Macau: no FATF assessments, no junket registration systems, and no source-of-funds scrutiny. Anti-money laundering compliance is virtually non-existent.

The second direction is the USDT channel. The previous chapter demonstrated how USDT "mirroring" underground banks achieve "funds not crossing the border." When Macau, this semi-legal capital outflow valve, was closed, cryptocurrency provided a decentralized alternative. No gambling table needed, no junket operator needed, no need to buy a plane ticket to go anywhere. Just a mobile phone, a wallet address, and funds arrive in seconds. The cost is also lower: the fee for USDT mirroring is about 2-5%, while the comprehensive cost of the Macau dead chip cycle (table losses + junket commission + travel and accommodation) is typically 5-10%. It is cheaper, faster, and more covert.

The third direction is a return to the underground. Some VIP clients simply stopped going abroad and started betting directly on online gambling platforms from within the mainland. The servers for these platforms are overseas, but the clients and funds are domestic. The agent pyramids discussed in Chapter 3 absorbed this demand.

Among the people Ah Jie knows in Putian, about one-third went to Southeast Asia, one-third moved online, and the remaining third quit entirely. Among those who quit, several were sought out by the public security bureau—not for their past in Macau, but because they violated mainland gambling laws after becoming online agents. The irony is: the three years spent as a junket operator in Macau were completely legal (Macau law permits intermediary activities), but becoming an online agent back in the mainland became a crime. The same person, the same skill, the boundary between legal and illegal depends only on which side of the border one stands.

With these three directions combined, the total volume of capital outflow has not decreased. What has changed is the path: from a centralized, monitorable channel (Macau gambling tables, where at least the Gaming Inspection and Coordination Bureau exists) to countless decentralized, unmonitorable channels. This is entirely consistent with the "balloon effect" mentioned in Chapter 3: squeeze the Macau end, and all other ports inflate simultaneously. From the regulator's perspective, the net loss is certain: at least in the Macau era, funds flowed through a system with the Gaming Inspection and Coordination Bureau, FATF assessments, and real-name casino registration, allowing for some tracking of data flows. Once funds move into Southeast Asian jungle casinos and blockchain wallets, even the starting point for tracking cannot be found.

Catch-22

The Macau government faces an unsolvable dilemma.

Gaming tax accounts for 70% to 80% of Macau's fiscal revenue. Direct and indirect employment in the gaming industry accounts for about 60% of Macau's workforce. Forty-one casinos are distributed across 33 square kilometers of land, averaging one per 0.8 square kilometers. This city's economic heart has only one chamber, and the blood flowing through it is gambling money.

Beijing says: economic diversification. Macau says: with what money shall we diversify? There is only one answer: gaming tax.

Using the profits of gambling to eliminate the dependence on gambling. This logical loop has no exit.

For a decade, Macau's "economic diversification" policies have included developing the MICE (Meetings, Incentives, Conventions, and Exhibitions) industry, building a world center of tourism and leisure, constructing a commercial and trade cooperation service platform between China and Portuguese-speaking countries, and introducing non-gaming enterprises to Hengqin. The annual revenue of the MICE industry is less than one percent of the gaming industry's, and the results of diversification have been limited. The hotels and shopping malls of the tourism and leisure center still depend on casino complexes; tourists' primary motivation for coming to Macau is the casino, with shopping being incidental. The development plan for the "Cooperation Zone" in Hengqin has progressed slowly, lacking an independent reason to attract non-gaming talent to work and live there. Non-gaming enterprises cannot compete with casinos for labor: the monthly salary of a casino dealer already exceeds most white-collar positions, so what is the motivation to switch careers?

Employment dependence has an even deeper dimension. Macau's higher education system has designed multiple related majors (gaming management, hotel management, tourism services) around the gaming industry, and the career paths of graduates naturally point toward casino complexes. The city's human capital stock is highly specialized. Asking these people to find "diversified" employment is like asking a city that only has a medical school to suddenly transform into a technology hub. It is not a question of will, but a question of skill mismatch.

"Removing gambling" is equivalent to creating mass unemployment and fiscal collapse. "Not removing gambling" is equivalent to tacitly allowing the capital outflow channel to continue existing (though it has now shrunk to a level where VIP accounts for 25%).

This is Macau's version of Catch-22: to solve the problem, you first need money; to have money, you must rely on the problem itself.

This dilemma does not belong to Macau alone. It is a miniature version of China's entire gambling policy. The state needs gambling money (lottery sales reach 600 billion annually) yet needs to maintain a stance of banning gambling; it once needed Macau's tax contributions and capital valve function, yet demanded that Macau "diversify its economy." When contradictions are irreconcilable, the policy choice is always the same: maintain the surface ban while tolerating the actual existence. For Macau, "economic diversification" is the direction written in policy documents; "gaming tax continues to support everything" is the reality written in the fiscal balance sheets. The two documents coexist, without interfering with each other.

And the ultimate embodiment of this "surface ban + actual tolerance" is the protagonist of the next chapter: the National Lottery. With annual sales of 600 billion, covering 600 million buyers, and an expected value of -50%, it is nonetheless called "public welfare." The sole reason: the final destination for the profits is the treasury account.

05. The Lottery of 600 Million People

If you rank all gambling products in China in 2024 by their expected value (EV) from lowest to highest, the one at the bottom is not an underground casino, nor online baccarat, nor a VIP room in Macau—it is the national lottery. With a payout rate of 50%, for every 2 yuan invested, you can only expect to get back 1 yuan on average. EV = -50%.

The top-ranked is Baccarat, with a payout rate of 98.94%. EV = -1.06%.

The former is operated by the Ministry of Finance and sold openly at convenience stores and mall entrances. Scratch-offs are placed right next to the cash register, requiring no identity verification and having no purchase limits; even purchases by minors are rarely effectively blocked. The latter is listed as a key target for crackdown by the Ministry of Public Security, and participants may face criminal prosecution.

In 2024, 200 million Chinese people participated in this EV = -50% game, contributing 623.486 billion yuan in sales. This was a 7.6% year-on-year increase, setting a new record. No one calls this gambling.

A Mathematical Fact

What does a 50% payout rate mean? For every 100 yuan spent on lottery tickets, you will only get back 50 yuan in the long run. The remaining 50 yuan goes into government accounts (35% for the Public Welfare Fund + 15% for issuance fees).

Compare this with other products on the "gambling legality spectrum" mentioned in Chapter 2. Baccarat has an EV of -1.06%, meaning a long-term loss of 1.06 yuan for every 100 yuan wagered. Roulette has an EV of -5.26%, a loss of 5.26 yuan. Slot machines have an EV between -5% and -15%. Underground baccarat websites have an EV roughly between -1.5% and -3% (slightly higher than physical casinos due to the "rake"). 125x leverage contracts have an EV of about -40% to -65% (due to the combination of forced liquidations and transaction fees).

It is clear at a glance. On this spectrum, the lottery is at the far left, the most unfavorable to players. Baccarat is at the far right, the most favorable to players.

Yet, the legal attitude is exactly the opposite: the most exploitative is completely legal, while the fairest is completely illegal.

If "protecting consumers" is the reason for banning gambling, logically, the first thing to be banned should be the lottery. A product with a 50% payout rate exploits players 47 times more than Baccarat (50% ÷ 1.06% ≈ 47). A product 47 times more predatory than a casino is sold next to convenience store registers without identity verification or purchase limits, and buyers often don't even realize what they are doing.

The probability of winning the first prize in the Double Color Ball: 1/17,721,088.

One in 17 million. This probability far exceeds the intuitive understanding of the human brain. Behavioral economics has a term for this called "probability neglect": humans cannot distinguish the actual difference between one in a million and one in ten million. The brain categorizes all extremely small probabilities into the same class—"unlikely but possible." It’s not that lottery buyers don't know the probability is low; it's that the brain itself cannot process such orders of magnitude.

For comparison: the probability of being struck by lightning is about 1/1,200,000. The probability of a person being struck by lightning 15 times is still higher than winning the first prize in the Double Color Ball. Nationwide, about 2,000 people are killed or injured by lightning strikes every year. No one gets excited about the possibility of being struck by lightning. The difference is that the lottery adds a layer of positive narrative to the negative expected value: "What if I win?" Lightning strikes don't come with a packaged story.

This is the cognitive foundation upon which the lottery has been sold for years. Lottery buyers are not stupid; rather, the human nervous system did not evolve to handle probabilities of one in ten million. Lottery products precisely exploit this cognitive shortcoming.

The Silence of 7 Million Addicts

A survey by the China Lottery Research Center at Beijing Normal University shows that there are about 7 million "problem gamblers" nationwide, including 430,000 severe addicts. The standard for a problem gambler is spending more than 20% of monthly income on the lottery.

Typical characteristics of the 430,000 severe addicts: 18-45 years old, high school or vocational college education, monthly income of 1,500-3,000 yuan, and self-identifying as "below middle class" in social status.

With a monthly income of 2,000 yuan, spending over 400 yuan a month on the lottery has far exceeded leisure and entertainment and entered the realm of pathological gambling. DSM-5 classifies pathological gambling as a "behavioral addiction," sharing the same neural circuitry (the dopamine reward pathway) as drug addiction. Addicts cannot stop even when they know the expected value is negative because the brain's response to a "near-miss" is almost indistinguishable from a "real win."

Contrast this with consumer protection measures in commercial casinos. Singaporean casinos have a mandatory self-exclusion system (problem gamblers can apply for a lifetime ban), and citizens must pay an entry fee of 100 SGD per visit. UK gambling has purchase limits, and the system automatically locks accounts once exceeded. All gambling outlets in Australia are equipped with addiction screening questionnaires. Las Vegas has statutory self-exclusion registries and free problem gambling hotlines.

The Chinese lottery: no self-exclusion system, no purchase limits, no addiction screening, and no standardized risk warnings.

In the Changqiao area of Xuhui District, Shanghai, where residents relocated from shantytowns and low-income households are concentrated, four lottery stations opened within just a few hundred meters. The interiors are decorated in red and gold, with walls covered in "trend analysis" charts and past winning information. Statues of the God of Wealth are placed in the most prominent positions. Not a single poster warns of gambling risks. The decoration of lottery stations and the psychological design of casinos are identical: targeting vulnerable populations and using environmental cues to reinforce the illusion that "the next one will be the winner." The difference is that Las Vegas casinos, under regulatory pressure, at least provide formal protection mechanisms. The national lottery doesn't even need the formality.

A deeper issue: there are no "lottery addiction" clinics in China's public healthcare system. In official classifications, buying lottery tickets is not considered "gambling behavior," and lottery addiction is not included in the diagnostic and treatment scope of "pathological gambling." The 7 million problem gamblers have no formal treatment channel. There is no diagnostic classification, no outpatient clinics, and no epidemiological statistics.

Making the behavior that causes a certain disease "not count as gambling" is the most effective way to ensure the disease is not recognized.

The Poor Man's Tax

Who is buying the lottery?

Online surveys show that 90.9% of lottery buyers have a monthly income below 3,000 yuan. Mobtech's lottery buyer profile report shows that lottery buyers' incomes are mainly concentrated in the 3,000-10,000 yuan range, and the TGI index for the worker group is as high as 354 (meaning the penetration rate of lottery buyers among workers is 3.5 times the national average). For laid-off, underemployed, and unemployed people, lottery spending accounts for as much as 20.86% of their income. Geographically, more than 60% of lottery buyers are concentrated in second, third, and fourth-tier cities. The provinces where the Public Welfare Fund accounts for the highest proportion of local fiscal revenue are mostly in the economically backward regions of the West and Northeast.

An analysis by The Economist quantified this pattern: for every 10% decrease in median household income, lottery spending increases by 4%. The poorer people are, the more they buy.

Data from the United States is similar: the lowest-income third of households purchase more than half of all lottery tickets in the country. The class effect of lottery products is universal and not unique to China. In economics, this is called a regressive tax. The poor contribute a far greater proportion than the rich. The difference is that most US states have problem gambling research funds and free treatment hotlines; China does not.

Now let's calculate the scale of this "poor man's tax."

In 2024, national lottery sales were 623.4 billion yuan. After deducting payouts, about 300 billion yuan entered public accounts in the name of Public Welfare Funds and issuance fees. Low-income groups account for over 70% of lottery buyers. Proportionally, about 2100 billion yuan comes from low-income groups.

Where does this money go? The allocation structure of the Public Welfare Fund: 60% to the National Social Security Fund, 30% to Central Special Projects, 5% to the Ministry of Civil Affairs, and 5% to the General Administration of Sport.

60% goes into the Social Security Fund. The Social Security Fund subsidizes pensions for the entire society. "Entire society" includes internet professionals earning 50,000 yuan a month and finance professionals earning 100,000 yuan a month.

In other words: a laid-off worker with a monthly income of 2,000 yuan spends 400 yuan a month on the lottery (20% of income), of which 200 yuan becomes Public Welfare Fund, and 120 yuan enters the Social Security Fund. This money subsidizes everyone's pensions, including white-collar workers whose income is 25 times that of the lottery buyer. This laid-off worker is contributing to the social security of the entire society with an "investment return" of -50%.

This is a reverse transfer of payment from the bottom to the rest of society. The gambling stakes of the poor, under the name of "Public Welfare Fund," become a source of funding for national pension security. No policy document will ever admit to this logic. The official line is: "The lottery is a voluntary act of public welfare."

The prerequisite for "voluntary" is being fully informed. How many of the 7 million problem gamblers can accurately state that the probability of winning the first prize in the Double Color Ball is 1/17,721,088? How many understand what EV = -50% means? How many know that 60% of the Public Welfare Fund subsidizes the Social Security Fund rather than "helping poor children"? If this information is not transparent, "voluntary" is merely a legal assumption, not a fact.

Where Did the Money Go?

In 2024, the central lottery public welfare fund collected 79.26 billion yuan into the treasury, which, combined with balances from previous years, resulted in a total disposable amount of 101.882 billion yuan. These figures come from Announcement No. 13 of 2024 by the Ministry of Finance.

The allocation of this money has clear figures: 49.651 billion yuan (60%) for the Social Security Fund, 21.436 billion yuan (30%) for Central Special Projects, 4.138 billion yuan (5%) for the Ministry of Civil Affairs, and 4.138 billion yuan (5%) for the General Administration of Sport. It looks standardized, with every penny accounted for.

Now look at the results of the special audit by the National Audit Office in 2015.

The National Audit Office conducted spot checks on lottery funds in 18 provinces. They discovered 16.932 billion yuan in illegal or irregular funds, accounting for 25.73% of the total funds audited. One out of every four yuan had a problem.

Specific methods of violation: the civil affairs system used Welfare Lottery Public Welfare Funds to build office buildings (illegal construction of government buildings), distributed public welfare funds as welfare expenses (illegal distribution of allowances and subsidies), and used lottery public welfare funds for commercial investments (misappropriation to non-public welfare projects). The violations within the civil affairs system alone amounted to 4.27 billion yuan.

The audit conclusion was written very clearly: the information on the use of public welfare funds was "too general for the public to supervise."

That was 2015. Ten years ago. What happened afterward is: no public rectification report. No announcement of subsequent accountability results. No explanation of whether "these 16.9 billion yuan in irregular funds have been recovered." For a system that handles hundreds of billions annually, the last major audit found that a quarter of the funds were in violation, and the follow-up process is completely invisible to the public.

200 million lottery buyers hand over 300 billion yuan to this system every year. Regarding the final destination of this money, the contributors know almost nothing. The transparency of the Public Welfare Fund is far lower than that of general taxation. The use of tax revenue must at least undergo review by the People's Congress and budget disclosure. The Public Welfare Fund does not require these procedures because the official designation is "voluntary participation by lottery buyers" rather than "taxation."

To make an analogy: if a listed company's annual report was found by an audit to have 25.73% of its funds used irregularly, and it failed to release any progress on rectification for the next ten years, the company would be delisted immediately. But the "National Lottery Company" has no shareholders who can vote with their feet, no delisting mechanism, and no independent audit. 200 million "shareholders" are still putting their spare change into this machine at convenience store entrances every day, not knowing and having no way to know where the money went.

The Dangerous Evolution of Guessing-type Lotteries

The change in the structure of lottery sales in 2024 deserves separate analysis.

Guessing-type lottery sales for the year were 286.810 billion yuan, a year-on-year increase of 16.4%, accounting for 46% of total lottery sales. Nearly half. Among them, sports guessing (football lottery, basketball lottery, sports betting, etc.) is the main force.

The gameplay of guessing-type lotteries: guess the outcome of a match and win a prize. Guess the score of a football match, the winner of a basketball game, the total number of goals in various sports events, or the half-time/full-time results. Win and take the money; lose and lose the money.

How is this different from "sports betting"? There is no difference. It is betting on the outcome of sports events.

The payout rate for guessing-type lotteries is about 69%, much higher than the 50% for lotto-type lotteries (like Double Color Ball). A 69% payout rate means an EV of -31%. Comparing this with underground sports betting markets: those markets are usually between 95-98%. The state-run version and the underground version are doing the exact same thing, with only two differences: first, the state version has a lower payout rate (earns more); second, the state version is legal.

With 286.8 billion yuan in annual sales and a 16.4% growth rate, this speed indicates that guessing-type lotteries are rapidly cannibalizing the market share of traditional lotto-type lotteries. In 2019, guessing-type lotteries accounted for less than 35%; by 2024, they reached 46%. Based on this trend, guessing-type lotteries will surpass lotto-type lotteries to become the main force of lottery sales around 2027. By then, more than half of China's lottery revenue will come from sports betting.

The analysis in Chapter 2 predicted: in order to compete with underground markets for users, state-run gambling must gradually increase payout rates, expand event coverage, and increase the variety of gameplay. The increase in the payout rate of guessing-type lotteries from 50% to 69% is a result of competitive pressure. The logical endpoint is an all-encompassing sports betting platform with a payout rate close to 85-90%, which is almost indistinguishable from underground markets at the product level, with the only competition being the convenience premium brought by legality.

The state is expanding its "legal gambling" boundaries at an annual growth rate of 16.4%, while continuing to use Article 303 of the Criminal Law to crack down on private bookmakers doing the same thing. The posture of banning gambling hasn't changed at all, yet the actual scope of gambling expands every year.

One final observation. Take down the sign of the "Lottery Management Center" and replace it with any private company. Without changing any game rules: a 50% payout rate, no purchase limits, no addiction screening, dense placement in low-income communities, and no risk disclosure to buyers. This company would immediately violate Article 303 of the Criminal Law for "opening a casino," and the operators would face three to ten years in prison.

The only legal difference between the lottery and gambling lies in who gets the profit. If it flows into the fiscal account, it's called "public welfare." If it flows into a private account, it's called "crime." 623.4 billion yuan, 200 million lottery buyers—the rules have never changed. The only thing that changes is the person collecting the money.

06. Northern Myanmar: The Gambling and Fraud Empire

On the night of October 20, 2023, a bus departed from a telecommunications fraud park in Kokang. More than 600 Chinese nationals on board were being forcibly transferred under armed escort by the Ming Xuechang family. During the journey, some attempted to escape, and the guards opened fire. Four people died, and four were injured. Chinese officials formally confirmed the casualty figures in December 2024.

This was not a skirmish in a war zone. It was a criminal syndicate transferring assets, except these "assets" were living human beings.

Data from the Ministry of Public Security in 2024 shows that in the direction of Northern Myanmar alone, more than 120,000 Chinese nationals were lured and trafficked into telecommunications fraud parks. Behind these 120,000 individuals are 120,000 families: mortgaging properties to scrape together ransom money, not knowing if payment would truly secure a release, not knowing which country their loved ones were in, and uncertain if they were even still alive.

From Casinos to Telecommunications Fraud

Telecommunications fraud parks did not appear out of thin air. Tracing the upstream of the industry chain, the starting point is the casino.

Chapter 4 discussed the economic logic of Macau's VIP rooms: the gambling table is the carrier for capital conversion, and "dead chips" are the medium for money laundering. In 2021, Macau rectified and closed this channel. However, the industrial infrastructure of gambling—enclosed parks, armed security, cross-border capital channels, and customer recruitment networks—did not disappear. This infrastructure found a new use in Southeast Asia.

The timeline is as follows: Starting in 2016, Chinese capital poured into Sihanoukville, Cambodia, to invest in casinos. At its peak, Sihanoukville had over 90 Chinese-run casinos, with hundreds of thousands of ethnic Chinese engaged in gambling-related work. In 2019, the Cambodian government announced a ban on online gambling, and in 2020, the COVID-19 pandemic led to the closure of physical casinos. Casino revenue dropped to zero. Park owners faced a choice: disband their teams and accept the loss, or pivot using existing infrastructure.

Most chose to pivot. Enclosed parks required no modification. Security teams did not need to be re-recruited. Cross-border capital channels did not need to be rebuilt (USDT underground banks serve both gambling and fraud). The only change was the business content: shifting from operating gambling websites to operating fraud scripts. Turning a casino into a fraud factory only required changing the software on the computer screens.

The economic logic of the pivot was equally clear. Even the highest house edge in a casino is only 1-5% (1.06% for Baccarat, 5.26% for Roulette). To earn 100 yuan, the house needs the guest to wager 2,000 to 10,000 yuan first. The "house edge" of telecommunications fraud is 100%: not a single cent transferred by the victim ever returns. For the same 50-person team, operating a casino might yield monthly profits in the hundreds of thousands, while pivoting to fraud can yield monthly profits in the millions or even tens of millions. A profit gap of over 10 times drove this "industrial upgrade."

There is another overlooked factor. Gambling requires guests to come voluntarily (passive waiting), while fraud can actively seek out guests (active strike). Gambling is constrained by physical location (guests must fly to Cambodia), while fraud is not limited by geography (a single mobile phone and a WeChat account can reach the entire country). Gambling requires real capital to build tables and buy chips, while fraud only requires a phone, a script, and a trained "customer service representative." In every commercial dimension, telecommunications fraud became the "upgraded version" of gambling.

The criminals did not simply "fall." Rational, profit-driven business decisions were at work.

Park Economics

The income structure of a medium-sized telecommunications fraud park (200-300 people) is far more complex than the outside world imagines. Fraud income is only the first of five layers of extraction.

The first layer is the fraud income itself. "Pig butchering scams" (romance-investment fraud) can reach millions of yuan per transaction, aiming to keep the victim investing until they are bankrupt. Stock chat group scams can involve amounts as high as 800 million yuan in a single case. Employee commissions do not exceed 12%. Most bottom-level employees have no salary because they have no "performance"; they are pure forced labor. The organizational structure goes from Group → Branch Office (200-300 people) → Regiment → Group (about 10 people), squeezing every level. Each month, a company can successfully defraud 10-20 cases, each worth at least $5,000, with some reaching over a million dollars.

The second layer is site rent. Park property management charges site fees to the resident fraud teams. The Liu family group (the "richest" in Kokang) operated four gambling and fraud parks, with site rents dropping from 1.5 million yuan to 1 million yuan per month (a price war caused by oversupply). Site rent is "guaranteed" fixed income: whether the team succeeds in fraud or not, the property manager collects the fee. This is perfectly consistent with the logic of commercial real estate: the landlord does not need to care about the tenant's business performance.

The third layer is the trade in human heads. This is the most disturbing part of the park economy.

For "employees" who enter the park voluntarily, the park pays the recruiter an "introduction fee" of about 100,000 yuan. For those lured in by deception, the payment is 66,000 yuan. For "skilled workers" who have already worked in a park for some time, the resale price can reach 300,000 to 500,000 yuan. One inmate, Gong Yu, who was sold to the Sunshine Park in the Philippines, was priced at 500,000 yuan. The resale of people between parks is described in the industry as "similar to player transfers." Those with good performance appreciate in value, while those with poor performance depreciate and are resold to parks with harsher conditions. The price of a human being has been market-defined. This is not a metaphor, but a literal price tag: 100,000, 300,000, 500,000.

The Liu family group specifically established a "guarantee company" responsible for labor supply, forming a complete chain from recruitment, transportation, and smuggling to control. This is the "logistics link" in the human trafficking industry chain.

The fourth layer is internal circulation consumption. Parks are equipped with supermarkets, restaurants, and KTVs, with prices 4-5 times higher than outside. A bottle of mineral water costs 10 yuan, a bowl of instant noodles 30 yuan, and a decent meal 80-100 yuan. The salaries and commissions "distributed" by the park are recycled back into the boss's pocket through high-priced internal consumption. This is identical to the company store system of 19th-century American mining companies: paying wages in company scrip and reclaiming it through a monopoly store. Employees remain in a state of net debt forever.

The fifth layer is ransom. To leave the park, one must pay a "breach of contract fee," usually ranging from 100,000 to 500,000 yuan. After the contract expires, the ransom is halved (from 120,000 to about 60,000). Most people cannot raise the ransom due to family hardship and are trapped in an infinite loop: no performance → no salary → cannot afford ransom → continue working → no performance. Third-party civilian "rescue teams" operate by helping families pay ransoms and taking a cut of the negotiations; some teams "can earn tens of millions of yuan a year." Rescue itself has become an industry.

With these five layers stacked, a 200-person park can generate annual revenue ranging from hundreds of millions to over a billion yuan. The Liu family group alone had over 2.6 billion yuan in fraud-related funds and over 8 billion yuan in gambling-related funds. The Bai family group controlled Kokang for 15 years, building 41 large-scale gambling and fraud parks with cases involving tens of billions of yuan, and controlled an armed force of about 2,000 people. The Bai family's "punishment system" included: cage confinement, cutting off fingers, dark rooms, and forced prostitution. Underperforming team members faced threats of being "buried alive."

The 2024 indictment of 39 members of the Ming family by the Wenzhou Procuratorate outlines a typical promotion path within the parks: a bottom-level employee lured into the park is promoted to group leader after three months due to "good performance." A group leader's monthly salary is about 20,000 yuan (10 times that of a bottom-level employee, but one-thousandth of the park owner's income). A group leader manages about 10 people, most of whom were deceived, a few who entered "voluntarily" (mostly due to domestic debt), and others resold from different parks. The group leader's duties include supervising script execution, tallying performance, and reporting anomalies. When a member tries to escape, the group leader must report to security immediately.

Multiple former park personnel interviewed by the media after returning to China described the same psychological state: they did not feel like victims, nor did they feel like criminals. Those trapped inside did things to avoid being beaten. This sentence accurately describes the power structure of the park: everyone is both a victim and a perpetrator, and there is no clear moral boundary between levels.

Three Domains

The geographical distribution of fraud parks follows a clear pattern: they are located in "sovereignty weakness zones"—civil war zones, high-corruption areas, and regions carved up by local factions. The criminal industry is the filler for the absence of sovereignty.

The first domain is Kokang, Northern Myanmar. Kokang is a Chinese autonomous region in Shan State, Myanmar, covering about 2,000 square kilometers with a population of about 300,000. It uses the Chinese language and the Renminbi. After the MNDAA was expelled by government forces in 2009, Bai Suocheng's forces controlled Kokang, forming a co-governance relationship with the military junta. This co-governance gave the four major families (Bai, Wei, Liu, Ming) ample gray space: the junta provided political asylum and armed endorsement, while the families provided economic benefits and "stability maintenance" functions. The 41 large-scale gambling and fraud parks operated under this protection for over a decade. The four families were essentially the junta's "franchisees" in Kokang.

The second domain is Cambodia. According to a 2025 Amnesty International report, Cambodia has at least 53 confirmed active fraud parks, mainly distributed in Sihanoukville, Poipet, and Bavet. The annual net inflow from telecommunications fraud is approximately $19 billion. Cambodia's GDP is only $46 billion. Fraud accounts for 41% of the national economic output.

41%. This number requires a moment of reflection. Crime has become an economic infrastructure; it is not just a matter of government penetration. Government officials and members of parliament acting as umbrellas is not incidental corruption, but deep dependence. If all fraud parks were eliminated tomorrow, the Cambodian economy would lose its largest source of foreign exchange. No government is willing to bear that consequence. This is why Cambodia repeatedly refused to attend the quadrilateral anti-fraud meetings between China, Laos, Myanmar, and Thailand (missing all three in August 2024, February 2025, and July 2025).

The Prince Group is a landmark case of Cambodian telecommunications fraud. The actual controller, Chen Zhi, established fraud parks in countries like Cambodia and Myanmar, systematically trafficking and detaining laborers. In 2024, the US Department of Justice seized 127,271 Bitcoins controlled by Chen Zhi, valued at approximately $15 billion. An asset scale of $15 billion exceeds the sovereign wealth funds of most small nations. The volume of a "fraud company" is equivalent to the treasury of a medium-sized country.

The third domain is the Philippines. The POGO (Philippine Offshore Gaming Operators) system had about 300 operators and over 60,000 practitioners (mostly Chinese) at its peak in 2019. POGOs were initially legitimate offshore gambling businesses licensed by the Philippine gaming regulator, PAGCOR. Among the approximately 300 operators, most operated telecommunications fraud under a legal shell.

The 2024 Bamban raid became a turning point. Philippine authorities raided Zun Yuan Technology Inc. (ZYTI) in the small town of Bamban, arresting 800 people, including 427 undocumented Chinese citizens. The umbrella for this fraud park was the local mayor, Alice Guo (Guo Huaping), a person accused of being a Chinese citizen who forged a Philippine identity to run for public office. Guo's business partners were also linked to a 3 billion SGD money laundering case in Singapore. On July 22, 2024, President Marcos announced a total ban on POGOs in his State of the Nation Address. In November 2025, Alice Guo was sentenced to life imprisonment for human trafficking.

Three domains, three types of sovereign relations. China's crackdown strategies are therefore completely different: using armed proxies for Northern Myanmar, diplomatic pressure for Cambodia, and internal political tools for the Philippines. The intensity of the crackdown depends on geopolitical relations, not the number of victims.

Operation 1027

The bus mentioned at the beginning of this chapter, with over 600 people being transferred under armed escort and four shot dead, was the "October 20 Incident." Seven days later, the war began.

On October 27, 2023, the Myanmar National Democratic Alliance Army (MNDAA), joined by the Ta'ang National Liberation Army (TNLA) and the Arakan Army (AA), formed the "Three Brotherhood Alliance" and announced a coordinated attack on multiple locations in northeastern Shan State, Myanmar. The publicly stated goals were twofold: to maintain ethnic territorial rights and to combat border telecommunications fraud. They deployed over 20,000 troops, and with allies like the Kachin Independence Army, the total force was even larger.

Progress of the war within 60 days: On the first day, they controlled Chinshwehaw and blocked the road from Lashio to Muse; in the second half of November, they besieged Laukkaing; on December 1, they attacked the Laukkaing urban area; on December 26, the Myanmar Defense Services and Border Guard Forces surrendered in Laukkaing. The MNDAA gained full control of the entire Kokang region. Over 8,000 Myanmar army and fraud militia members were killed or captured.

Military operations proceeded in sync with Chinese law enforcement actions. On November 12, the Ministry of Public Security issued reward notices for four people, including Ming Xuechang and Ming Guoping. On November 15, Ming Xuechang committed suicide out of fear of punishment. On November 16, three people, including Ming Guoping, were captured and handed over to the Chinese side. On December 10, reward notices were issued for 10 people, including Bai Suocheng. In December 2024, the Wenzhou Procuratorate filed a public prosecution against 39 members of the Ming family on charges including fraud and intentional homicide.

On January 12, 2024, under Chinese mediation, the Myanmar military junta and the Three Brotherhood Alliance signed the "Haigeng Peace Agreement" in Kunming. The MNDAA obtained actual governance rights over the Kokang region. This armed force returned to its homeland after 14 years in exile. On August 5, 2024, the MNDAA further captured Lashio, the seat of the Northeastern Military Command, marking the greatest military defeat the Myanmar army has suffered in the 76 years since the founding of Myanmar.

The logic of the event needs to be reassembled. The MNDAA had been in exile for 14 years after losing Kokang in 2009. During those 14 years, there was no sufficient political window to launch a counteroffensive. In 2023, the telecommunications fraud issue provided this window: Chinese public anger toward Northern Myanmar fraud reached a peak (multiple accounts from returnees about park atrocities triggered a public opinion storm), and the Chinese government faced intense domestic pressure. The MNDAA acted under the banner of "anti-fraud," which highly aligned with China's domestic political needs. China mediated the ceasefire, received suspects, and issued reward notices—the timeline of these actions was highly synchronized with the MNDAA's military advance.

How is this defined in international law? It was not a peacekeeping operation (no UN mandate). It was not a counter-terrorism operation (telecommunications fraud is not classified as terrorism). It was not extradition or law enforcement cooperation (it involved the armed occupation of territory). Ultimately, a sovereign state tacitly allowed an overseas armed organization to launch military operations and occupy part of a neighboring country's territory in the name of "combating crime." This event has no precedent, no international law classification, and no formal name.

The effects can be quantified: a cumulative 55,000 Chinese suspects were captured, the four major families were annihilated, and large-scale border fraud parks were completely eradicated. The cost is equally clear: Myanmar's sovereignty in the Kokang region has been effectively hollowed out. The principle of "non-interference in internal affairs" was overridden in this case by the needs of "protecting overseas citizens" and "combating cross-border crime." The long-term impact of this precedent is currently not being discussed publicly.

The Immortal Hydra

The phrasing at the press conference during the 2025 Two Sessions was "telecommunications fraud parks in Northern Myanmar near the border have been completely cleared." The wording deserves word-by-word analysis. "Near the border" excludes deep inland areas. "Large-scale" excludes small gangs. "Northern Myanmar" excludes Eastern Myanmar and other countries. Precise qualifiers point to a precise reality: the large parks in Kokang are gone, but fraud activities continue.

After Kokang was cleared, fraud activities shifted in three directions.

The first direction is Myawaddy. Myawaddy, in Kayin State, Myanmar, is located on the Thai-Myanmar border, only 500 kilometers from Bangkok, and is politically a fragmented zone. Controlling this area is Colonel Saw Chi Thu of the Kayin Border Guard Force, an armed force of 6,000 men who officially broke away from the Myanmar government army in 2024 to establish themselves independently as the "Karen National Army." The KK Park is the most notorious fraud complex in Saw Chi Thu's controlled area: co-founded by Macau Triad figure Wan Kuok-koi ("Broken Tooth Koi"), it features 4-meter-high electrified wire fences and 635 buildings (Phases I to IV), including casinos, hotels, and fraud centers. Compared to the Northern Myanmar parks, Myawaddy is larger, better equipped, uses more diverse criminal methods, and involves more complex factions.

In early 2025, the Thai government decided to implement the "Three Cuts" on Myawaddy: cutting off electricity, internet, and fuel supplies. The effect was extremely direct: the parks could not operate without power. In February 2025, senior officials from China, Myanmar, and Thailand arrived in Myawaddy to form a joint working group. 2,876 Chinese suspects were escorted back to China. In October 2025, the Myanmar military captured and closed the KK Park, releasing over 2,000 trapped individuals and confiscating 30 Starlink terminals. In January 2026, all 635 buildings were completely demolished using controlled explosions and heavy machinery.

Physically annihilating a "city." This is the most extreme means of crackdown. The operators had already relocated before the demolition.

The second direction is deeper into the interior. Small towns in eastern Myanmar like Wan Hai and Tangyan received the overflow of teams from Myawaddy. The Golden Triangle Special Economic Zone in Laos (controlled by the Zhao Wei Group) became a new gathering point. These places are more remote, have poorer infrastructure, greater regulatory vacuums, and higher logistical costs for crackdowns.

The third direction is "park-less operation." In recent years, under pressure, some teams have pivoted to "Western markets" (shifting fraud targets from Chinese to Westerners), reducing the jurisdictional coverage of Chinese law enforcement agencies. A more critical technical change is the introduction of Starlink. The 30 Starlink terminals seized at KK Park demonstrate that satellite internet has become standard equipment for telecommunications fraud. The "internet cut" part of the "Three Cuts" strategy is technically bypassed; fiber optics can be cut, but satellite signals cannot be intercepted. Coupled with diesel generators for power and trucks for fuel, the effectiveness of the "Three Cuts" is shortening.

What is the final form? A mobile phone, a Starlink terminal, and a USDT wallet address. No fixed park required, no armed security needed, no physical walls necessary. Fraud teams can operate anywhere with a view of the sky. There are no buildings to demolish, no electricity to cut, and no addresses to raid. The "balloon effect" argued in Chapter 3 reaches its extreme in the field of telecommunications fraud: as physical space is compressed, activity moves into information space. From Kokang to Myawaddy to Wan Hai to online, every "successful sweep" pushes criminal activity into forms that are harder to track. In 2024, 294,000 cases of telecommunications network fraud were solved nationwide, and 315.1 billion yuan in case-related funds were urgently intercepted. If border parks are "completely eradicated," why is it still necessary to intercept 315.1 billion yuan in a year? Because the growth rate of decentralized cybercrime exceeds the speed of physical crackdowns. Eradicating parks destroys buildings, not criminal networks.

In July 2024, under the unified coordination of the Ministry of Public Security, the Myanmar side handed over a batch of Chinese telecommunications fraud personnel to China. A cumulative 55,000 people were captured and repatriated. Those who returned had lost dozens of pounds, and some were missing fingers—the punishment for "failing to meet performance targets." Ransoms ranged from 100,000 to 500,000 yuan, with middlemen taking a 30,000 yuan "handling fee."

Most returnees found jobs in their hometowns with monthly salaries of three to four thousand yuan. Much less than a "customer service rep with a 20,000 yuan salary." Much cheaper than a finger. At least they can answer the phone.

07. Digital Casinos

$81.4 billion.

This is the global Gross Gaming Revenue (GGR) of crypto casinos in 2024. It represents a fivefold increase year-on-year. For comparison: Macau's total gaming revenue for the full year of 2024 was approximately $30 billion. The scale of crypto casinos is already 2.7 times that of Macau.

The aforementioned casinos have no buildings, no dealers, no chips, no security guards, and no cameras. Their form of existence is: a bot program within an instant messaging app, a string of cryptocurrency receiving addresses, and an overseas cloud server with a monthly rent of $50. They require no physical facilities, are not subject to the gambling laws of any country, and do not need to apply for licenses from any regulatory body. In cyberspace, casinos have evolved into an entity that cannot be eradicated by bulldozers.

Ajie stands at a middle node of this evolutionary chain. 32 years old, a native of Putian, Fujian. In 2019, he was a "pazhai" (a ground-level junket agent) in a Macau VIP room. His responsibility was to help the boss recruit clients to gamble and earn commissions from the "rake" on chips. Chapter 4 described the economic logic of this trade: junket agents are the terminal nerves of the gambling industry chain, relying on personal networks for a living, with unstable income and the constant risk of being replaced by younger competitors. In 2021, Macau launched a massive crackdown on the VIP room system; Suncity collapsed, and the junket agent community faced collective unemployment. Ajie returned to Putian and stayed for six months, living off his savings. Then, a peer he knew from Macau pulled him into a group. The group was recruiting for "online promotion."

Now, Ajie operates three gambling groups. His clients include the working class from Guangdong, Fujian, and Zhejiang, as well as Southeast Asian Chinese communities. He earns about 80,000 RMB per month—twice what he made in Macau. There are no physical venues, no entry-exit risks, and no worries about gamblers defaulting on debts. Payments are made in cryptocurrency, with funds arriving overnight. He has never seen the face of any client.

Gambling in 30 Seconds

The operational process of gambling groups has been compressed to under 30 seconds: open the instant messaging software, search for a specific bot name, click "Start," select the game type (Baccarat, dice, slots, or sports betting), enter the bet amount, and confirm. The result is instantly displayed in the dialog box; winnings are automatically credited, and losses are automatically deducted. All steps combined take no more than 30 seconds.

30 seconds. This duration is worth measuring carefully.

Traditional physical casinos require leaving the house, taking a flight or driving for hours, exchanging chips at a counter, and finding a table to sit down. From the moment a gambling impulse arises until the actual bet is placed, the friction time is measured in hours or even days. This time difference is a natural "cooling-off period": long enough for most people to abandon the impulse while on the way. Traditional online casinos that emerged in the 2010s lowered this threshold: registering an account, filling in identity information, binding a bank card, waiting for verification, and depositing funds—the entire process takes about 5 to 10 minutes. Although 5 minutes is short, there are still several nodes where the impulse might be interrupted: hesitation when entering an ID number, alertness when binding a bank card, or second thoughts upon seeing the account balance.

Crypto gambling groups compress all of this into 30 seconds. There is no registration process (no window for hesitation), no identity verification (no fear of being discovered), no bank card binding (no reminder of visible balances), and no app downloads (no risk of the phone being checked by family). The only physical barrier between impulse and betting is the distance a finger slides across the screen.

Behavioral economics research has proven that inserting a time interval between an impulsive decision and its actual execution is the most effective mechanism for blocking irrational behavior. Based on this, gambling regulations in many countries require online casinos to set mandatory waiting functions: the UK requires a 24-hour self-exclusion option, Australia requires preset deposit limits to trigger cooling-off periods, and Singapore casinos charge a 100 SGD entry fee for their own citizens (economic friction). The common logic of these mechanisms is to create an interval between impulse and action.

Crypto gambling groups have eliminated this interval. The disappearance of the interval is by no means a side effect of "technological progress"; it is a precise engineering of the gambling addiction mechanism. Chapter 1 analyzed the neurological mechanism of the dopamine reward circuit's response to "near-misses." Crypto group casinos have ramped up the frequency of triggering this circuit from several times per hour (physical casinos, requiring waiting for the dealer to deal) to several times per minute (instant results plus one-click re-betting). The higher the stimulation frequency, the faster the dopamine tolerance threshold rises, and the more rapidly addiction forms.

The underlying infrastructure supporting this system is the blockchain. In 2024, the number of users on the world's largest instant messaging platform surpassed 1 billion. The number of accounts on the blockchain network deeply integrated with it exploded from 4 million to 128 million (a 3100% increase in one year). In April 2024, a stablecoin issuer deployed 60 million USD-pegged tokens on this network; by the end of 2025, this number grew to 1.43 billion—an increase of nearly 24 times. There are over 650 decentralized applications on the network, with gambling applications accounting for a significant proportion. Users do not need to register a new account; the messaging software username is the casino account. Conversion between stablecoins and native tokens can be done with one click. Zero identity verification, complete anonymity.

What does the 1-billion-user base of the instant messaging app signify? It means that 1 out of every 8 people globally already has the entrance to a crypto casino installed on their phone. No additional downloads are required. The casino is already in their pocket.

Cryptocurrency: The Blood of Gambling

Why did underground gambling choose USD-pegged stablecoins instead of Bitcoin or bank transfers?

Bitcoin prices fluctuate wildly (a 15% single-day drop in October 2025, and a 35% crash in November); gamblers do not want to bear the additional risk of "currency fluctuation" on top of "gambling wins and losses." Bank transfers have real-name records; every transaction can be tracked by regulators, frozen, or used as judicial evidence. Cash requires physical transport, which is particularly difficult across borders, and large amounts of cash face customs inspections and anti-money laundering reporting obligations.

Stablecoins solve these three pain points. Price stability: 1 token is always pegged to 1 USD, eliminating currency fluctuation risks. Anonymity: Issuers do not perform identity verification on on-chain end-users; anyone can generate a receiving address to receive tokens. Borderless: Transaction fees on specific public chains are less than 1 cent, with 24-hour arrival between any two points globally, bypassing the banking system and avoiding foreign exchange control rules. China's $50,000 annual individual foreign exchange limit simply does not exist here.

The United Nations Office on Drugs and Crime (UNODC) 2024 report on transnational organized crime in Southeast Asia identified this stablecoin as the "currency of choice" for regional cybercrime. Within one year (ending September 2023), $17 billion worth of such tokens were closely linked to underground exchanges and criminal activities. The report specifically noted that a certain public chain is the main channel for criminal fund flows due to its extremely low transfer costs (less than 1 cent) and fast speed (approximately 3-second confirmation), perfectly matching underground gambling's demand for "fast, low-cost, and hard-to-trace" funding channels.

Law enforcement cases by the Chinese police validate this assessment. In 2024, police dismantled a money laundering gang covering 26 provinces, involving approximately $2 billion and the arrest of 193 suspects. The gang's operational model was described by law enforcement as "motorcycle squad" style rapid exchange: gamblers hand over RMB to offline "squad" contacts, who then transfer an equivalent amount of crypto tokens into the gambler's platform account within minutes. Once the exchange is complete, the "squad" members immediately move locations, leaving no fixed base and retaining no transaction records. The entire exchange process, from contact to completion, takes no more than 20 minutes.

Estimates from 2020 show that the annual scale of China's black-market gambling industry is approximately $145 billion. Stablecoins serve as the payment and clearing layer for this massive shadow economy. Without such tools (or similar crypto payment solutions), the cost of fund flows for cross-border online gambling would increase by 10 to 50 times due to the anti-money laundering compliance requirements of the traditional banking system.

All three of Ajie's gambling groups settle in crypto tokens. Clients transfer funds through daily payment tools to "exchangers" (nodes in an exchange network spread across cities nationwide), who then deposit equivalent tokens into the client's built-in messaging app wallet within 30 minutes. Ajie never touches fiat currency. Commission rakes (about 3% to 5%) remain in on-chain addresses in token form. If an investigation occurs one day, the on-chain assets can be moved to a brand-new address within 10 minutes. There are no bank accounts to freeze, no corporate entities to dissolve, and no physical servers to search and seize.

125x Stakes

The world's largest cryptocurrency exchange offers up to 125x leverage for Bitcoin perpetual contracts. Second-tier exchanges offer 200x. Some small offshore platforms even offer 1000x.

The mathematical implication of 125x leverage is simple and brutal: invest $1,000 to control a position of $125,000. If the Bitcoin price moves 0.8% in an unfavorable direction, the margin is exhausted, the entire principal is wiped out, and the position is forcibly liquidated by the exchange's system. In 2025, Bitcoin's average daily volatility was between 2% and 5%. Statistically, the probability of a trader with a 125x long position (a bet on a price increase) surviving until the day's close is extremely low.

Place this figure within the gambling probability framework established in Chapters 1 and 2. The house edge in Baccarat is 1.06% (an expected value of negative 1.06%); a gambler bringing 10,000 RMB to a casino would, on average, need about 940 hands to lose all their principal (assuming 100 RMB per hand). The equivalent expected value of a 125x leverage contract is approximately negative 40% to negative 65% (due to the combination of forced liquidation mechanisms and transaction fees). A trader opening a position with the same 10,000 RMB would have an average survival time measured in hours, certainly not days. The capital survival rate for a Baccarat player is far higher than that of a high-leverage contract holder.

In 2025, the total amount of forced liquidations of leveraged positions in the global crypto market reached $150 billion. The story of October 10th illustrates the scale of the problem: after news broke that the U.S. would impose 100% tariffs on Chinese goods, Bitcoin plummeted from $125,700. Within 24 hours, $19.37 billion in leveraged positions were forcibly liquidated by exchange systems, and 1.6 million traders lost their entire margins on the same day. 85% to 90% of the liquidated positions were longs—bets on an upward direction. From November 20th to 21st, Bitcoin crashed again by 35% (falling from $126,080 to $81,600), resulting in 396,000 traders' positions going to zero in a single day, with total liquidations exceeding $2 billion.

A trader who lost over $100 million cumulatively said candidly in a post-event interview: "It was basically just gambling. I got greedy and stopped seriously analyzing the numbers."

The evaluation by a co-founder of an exchange provides a more industry-oriented perspective: "Professional institutions use leverage to optimize capital efficiency, never to place a big bet. Retail investors use the same tools to gamble and are usually wiped out."

If the product "Bitcoin Perpetual Contract 125x Leverage" were renamed "Bitcoin Up/Down Bet, up to 125x Odds," the probability structure would remain unchanged, but the product would legally shift from a "financial derivative" to a "gambling tool," and the operator would be in violation of gambling laws. The only thing that changes is the name. The core question from Chapter 2 resurfaces here: Is gambling defined by its probability structure or its product name? Is regulation targeting the substance of the behavior or the industry classification?

The Unstoppable Platforms

In 2024, Chinese public security organs dismantled more than 4,500 online gambling platforms and investigated 73,000 cases of cross-border gambling. During the "Summer Action," cluster campaigns were launched against 45 major cases, resulting in the arrest of over 11,000 suspects. The "Golden Eagle 2024" operation repatriated or persuaded 3,700 gambling-related criminal suspects to return from abroad. Chongqing police destroyed the "DC" gambling group that had been entrenched in the Philippines for years.

4,500 platforms. This number should be encouraging. However, its actual effect needs to be evaluated against another set of figures.

What is required to build an online gambling platform based on cryptocurrency? Open-source gambling program code (available for free on code-hosting platforms), an overseas cloud server ($50 to $100 per month), a cryptocurrency receiving address (anyone can generate countless addresses for free), and a domain name (registrable for $10 a year). The total startup cost does not exceed 10,000 RMB. A skilled programmer can have a platform live and operational within three days.

Is the domain name blocked? Register a new one; DNS configuration is completed in 30 minutes. Is the mobile app removed from the app store? Move to a bot program on instant messaging software with zero additional development cost. Is the server shut down by overseas law enforcement cooperation? Rent a new server; migrating the database takes about 2 hours. The entire "reconstruction cycle" of a platform does not exceed half a working day.

Chapter 3 argued in detail the "balloon effect" in the economics of prohibition: pressure does not eliminate activity; it only changes its geographic distribution and form. In digital space, this effect is amplified to an almost irreversible degree. The reconstruction of a physical casino (like the KK Park in Chapter 6) requires billions in investment and years of construction. The reconstruction of a digital gambling platform requires 10,000 RMB and half a day. The cost gap between the two is approximately 100,000 times. This proves that the speed of physical strikes can mathematically never catch up with the speed of digital reconstruction. If 4,500 are taken down, 45,000 can be rebuilt. Each costs less than 10,000 RMB.

Ajie has been an online gambling agent since 2022. In three years, he has never been contacted by any law enforcement agency. Contrast this with his life as a junket agent in Macau: he had at least two tense experiences of being followed by plainclothes police every month, client information in his phone contacts could be used as evidence at any time, and abnormal flows in his bank account would trigger a suspicious transaction report every six months. The online world provides a completely different sense of security: there is no physical location to be raided (he is at home in Putian), no face to be identified by surveillance (he never meets clients), and no bank accounts to be frozen (all funds flow on-chain). The only theoretical risk comes from client reports, and in anonymous groups, clients do not even know the agent's true identity.

Borderless Spectrum

The world's largest crypto casino recorded $2.6 billion in Gross Gaming Revenue in 2023, ranking it as the seventh-largest gambling group globally. The platform holds a gambling license from the Caribbean island nation of Curaçao and is considered a legal operation. A famous Canadian rapper promotes the platform with an annual endorsement fee of $100 million and has personally wagered over $1 billion on it. Starting in 2024, the platform became the title sponsor of a Formula 1 racing team, with a three-year contract worth $100 million. The casino's logo is printed on the cars as they race across every track in the world.

The same platform: is banned from access throughout the European Union. In China, it is a target of Article 303 of the Criminal Law. In the UK and Australia, it is considered an unlicensed operation. Its legal status in Brazil and Japan is ambiguous. A significant proportion of its 600,000 regular users log in via Virtual Private Networks (VPNs) to bypass geographic blocks.

The Curaçao government says it is legal. The EU says it is illegal. China says it is criminal. The spokesperson says it is "inevitable." Formula 1 says it can be a title sponsor. The same platform simultaneously holds at least five different legal identities at the same point in time.

The experience of another prediction market platform is even more absurd. In November 2024, the FBI raided the founder's residence to investigate whether the platform allowed U.S. users to participate in "gambling." In July 2025, the investigation was closed without charges. In September 2025, the Commodity Futures Trading Commission (CFTC) officially approved the platform as a regulated contract market. In October 2025, the parent company of the New York Stock Exchange invested $2 billion in the platform, valuing it at approximately $8 billion. In January 2026, the Nevada Gaming Control Board filed a lawsuit against the platform, citing "gambling activities without a gaming license."

The same product saw its legal identity flip four times within 12 months. It went from "being raided" to "being approved," to "receiving investment from the parent company of the world's largest exchange," and finally to "being sued as illegal gambling." Singapore bans the platform. Thailand bans it. Romania bans it. Australia bans it. The U.S. federal government officially approves it. The gambling regulator in Arizona publicly calls on federal regulators to "take action."

The question posed in Chapter 2, "Who defines gambling?" finds its most absurd real-world answer here: everyone is defining it, but no one's definition is effective for global users. Each country can only govern within its own borders. Borders are ineffective against digital products. For a user to bypass geographic blocks via a virtual network takes only 10 seconds and a $5 monthly subscription fee.

Ajie does not care about the legal arguments of any country. Since switching from being a junket agent to an online agent, his sole criterion over the past three years has always been "will I get caught." In Macau, the answer to that question became increasingly clear after 2021: "Yes." In virtual space, the answer remains "No." Moral judgment has never entered Ajie's decision-making framework; the risk-reward ratio is the only basis for calculation. Just like the moment of hesitation when Lao Liu's finger hovered over "Deposit another 500" in Chapter 1, every one of Ajie's business decisions can be measured by expected value. The difference is: Lao Liu's expected value is negative (the gambler always loses), while Ajie's expected value is positive (the house always wins).

08. The Transformation of Las Vegas

On December 26, 1946, a pink building lit up its neon lights along the desert highway.

Construction cost: $6 million. What did that money mean in America in 1946? An average house sold for about $8,000 back then. The cost of the Flamingo was equivalent to 750 houses. The investor was Benjamin "Bugsy" Siegel, a member of the New York Mafia, with funds coming from the criminal syndicate's cash reserves.

It lost $300,000 in its opening week. Driving here from Los Angeles took four hours through the Mojave Desert, with no gas stations or air conditioning along the way. Gamblers were unwilling to come to this desolate area. Hollywood stars were invited to the opening ceremony; most played a few rounds and flew back to Los Angeles. The air conditioning system frequently failed in the desert heat, construction was unfinished, and the walls of some guest rooms were still bare plasterboard.

Six months later, on June 20, 1947, Siegel was struck in the right eye by a bullet that came through a window at his girlfriend's residence in Beverly Hills. Before his body had even cooled, his mob associates had taken over the operations of the Flamingo.

Subsequent development: Under new management, the Flamingo turned a profit. The business model of a casino, luxury hotel, and resident shows proved effective. Within a decade, the Sands, Sahara, Riviera, and Tropicana appeared on the Strip. Each replicated the Flamingo model: criminal capital investment, the casino as the core traffic driver, and the hotel and entertainment as auxiliary profit points.

A mobster validated a business model with his life. Siegel's innovation was this: previously, Nevada casinos were just tin shacks in the desert, offering gambling tables and cheap whiskey, serving passing truck drivers and miners. The Flamingo was the first to package gambling as a "vacation experience": swimming pools, gardens, resident bands, and steak dinners. The gambling tables remained the source of profit, but gamblers no longer felt they were "gambling," but rather "vacationing." This cognitive shift became the core logic of every evolution of Las Vegas for the next eighty years.

Three Waves of Whitewashing

First: In 1966, billionaire Howard Hughes moved into the top floor of the Desert Inn. After staying too long and being asked to check out, Hughes's response was to buy the entire hotel. Over the next two years, Hughes acquired the Sands, Frontier, Landmark, and other casinos, with a total investment exceeding $300 million. Wall Street and banks began to notice that casinos could be a legitimate business.

Second: In 1969, Nevada passed the Corporate Gaming Act, allowing publicly traded companies to hold gaming licenses. Previously, only individuals or partnerships could be licensed. The new act opened the doors for Wall Street capital to enter the gambling industry. Mob funds were gradually squeezed out, as the compliance requirements and auditing processes of public companies significantly raised the cost of backroom operations. In the 1970s, mainstream hotel groups like Hilton Hotels and Holiday Inn entered Las Vegas. The ownership structure of casinos shifted from "a few Italian-American families" to "tens of thousands of anonymous shareholders." The criminal tint was diluted within the figures of securities regulatory filings and quarterly dividends.

Third: In 1989, Steve Wynn's The Mirage opened. It cost over $1 billion. An artificial volcano erupted every 15 minutes in front of it. In the back, there were dolphin pools and white tiger exhibits. This was the first true "mega-resort." Wynn's core philosophy was: the casino is just a tool to attract people to the resort for other types of consumption. The gambling table was no longer the destination, but a traffic-generating device.

In 1998, Wynn's Bellagio opened. The musical fountains in the artificial lake out front remain the Strip's most famous free attraction. Inside, it featured art galleries with original works by Picasso and Monet, Michelin-starred restaurants, and Cirque du Soleil residencies. The casino still existed, but it was no longer the protagonist.

The capital forms of the three whitewashings: Criminal cash (1946–1966) → Individual billionaire wealth (1966–1989) → Wall Street public capital (1989–present). Each upgrade pushed the boundaries of legitimacy further out. What Bugsy Siegel did with mob money, MGM Resorts now does with pension funds and mutual funds. The product structure has not undergone profound changes. The tables still run, the house edge still takes its cut, and the addiction mechanisms still operate. What has changed is the identity of the holders and the transparency of the capital sources. Legitimacy is not an inherent attribute, but a function of capital form.

The Invisibilization of Gambling

In 2024, gaming revenue on the Las Vegas Strip was $8.62 billion. Non-gaming revenue accounted for approximately 64% of the Strip's total revenue. Hotel accommodations accounted for 34%, food and beverage for 18.5%, drinks for 6.5%, and entertainment shows for 5%.

Contrast this with 1984: Gaming revenue accounted for 58% of the Strip's total revenue.

Over forty years, this proportion dropped from 58% to 26%. However, absolute gaming revenue did not decrease; $8.62 billion is a historical high. The Strip's strategy is not to abandon gambling, but to layer more revenue sources on top of it. The casino has transformed from a profit center into infrastructure.

A visitor survey showed: Only 2% of visitors stated that gambling was their primary reason for coming to Las Vegas. Top-ranking reasons were dining, shows, shopping, and pool parties. The casino has become a place to "stop by." The typical Strip visitor profile: spends $300 on champagne at a pool party during the day, $200 on a Cirque du Soleil ticket in the evening, $500 on a nightclub table late at night, and finally loses $100 playing a few hands of blackjack at the tables. Gambling is the smallest amount in the entire consumption chain.

MGM Resorts' total revenue in 2024 was $17.2 billion, with non-gaming revenue accounting for about 65%. Caesars Entertainment's 2024 revenue was approximately $11 billion, with the digital business (sports betting apps) being the fastest-growing segment. These companies' annual reports look more like those of hotel groups and tech companies than casino operators.

The gambling license remains the legal foundation of the entire business empire. Without a casino license, one cannot obtain integrated development rights and tax incentives in Nevada. Gambling is the legal foundation, not the commercial ceiling. Understanding this structure explains why Strip hotel room rates can be sold below cost (subsidized by the casino), why buffets can be priced below the cost of ingredients (subsidized by the casino), and why parking was once free (subsidized by the casino). The pricing strategy of the entire resort revolves around one core assumption: as long as people stay inside the building long enough, gambling revenue can cover all losses. The casino acts like an invisible tax, levied on every visitor who stays for more than three hours.

Symbols of the Post-Gambling Era

In September 2023, the Sphere opened. Costing $2.3 billion, it is the most expensive pure entertainment venue in history. Its exterior is covered by 580,000 square feet of LED screens, making it the world's largest LED display structure. Inside, it has 17,600 seats equipped with an immersive audio-visual system.

The Sphere has no gambling tables, no slot machines, and no sports betting terminals. Zero gambling functions. Annual revenue exceeds $1 billion. It is expected to become profitable starting in fiscal year 2025. Abu Dhabi has already signed a license to build a second venue of its kind. Las Vegas is no longer the only gambling city, but it is becoming the only experimental ground for super-entertainment technology.

Placing the Sphere on the Flamingo's timeline: In 1946, Bugsy Siegel's innovation was "casino plus hotel." In 2023, the largest single investment in Las Vegas is "pure entertainment, no gambling." In seventy-seven years, the functional positioning of gambling in this city has undergone a 180-degree shift, from the sole selling point to an optional configuration.

The same logic is reflected in the intensive introduction of sporting events. In 2020, the NFL's Raiders moved from Oakland to Las Vegas, with the newly built Allegiant Stadium costing $1.9 billion. In 2023, the F1 Las Vegas Grand Prix was held on the Strip for the first time, with the track running past the front doors of casinos; a single race brought over $1.5 billion in economic benefits to the city. In 2024, the Super Bowl was held in Las Vegas for the first time. NBA and MLB expansion teams are also in site negotiations. Professional sports are gradually replacing gambling as the new core of attraction for Las Vegas.

The Strip's net income fell 40.4% in 2024. Total visitor numbers fell 8%. However, per capita spending rose, with average hotel rates at $275/night (up 12% year-on-year). Las Vegas is shifting from a "high volume, low transaction value" model to a "low volume, high transaction value" model.

Blind Spots of Legalization

In 1931, Nevada became the first state in the U.S. to legalize gambling. The motivation was simple: tax revenue was desperately needed during the Great Depression, and underground gambling could no longer be suppressed. The "prohibition economics" discussed in Chapter 3 found its earliest validation here: when the cost of prohibition exceeds the cost of tolerance, legalization becomes the rational choice.

Ninety-four years later, Nevada possesses the most mature gaming regulatory system in the U.S.: the Gaming Control Board (established in 1955), the Gaming Commission (1959), rigorous license reviews (applicants must disclose financial and criminal records), and a "Black Book" system (banning specific individuals from casinos).

However, in terms of problem gambling protection, Nevada's investment is among the lowest in the nation. Per capita spending on problem gambling is only $0.40. There are no mandatory active exclusion laws for physical casinos (only for online gambling). Casino exclusion systems are voluntary and not legally binding. A gambling addict can walk into any casino, even if they signed a "self-exclusion" form at another casino the day before. Casinos have no obligation to verify, nor is there a shared database of excluded individuals.

This contradiction points to an overlooked fact: legalization solves supply-side issues (turning underground casinos into tax-paying enterprises) but does not solve demand-side issues (how to protect addicts). Las Vegas casinos pay billions of dollars in gaming taxes to the state government every year. Funding for the treatment and prevention of problem gambling comes from that same tax pool. The interests of the regulators and the regulated are highly aligned: the more the casinos earn, the more the state government receives. Problem gambling protection is forever a secondary agenda in this interest structure. Nevada's problem gambling hotline receives over 10,000 help calls annually. Yet, the state government's annual budget for the Advisory Council on Problem Gambling is less than $3 million—not enough to carpet a single casino corridor. Every tile in a casino lobby is meticulously designed to prolong a gambler's stay, but not a single signpost marks a hotline for addiction help.

The next chapter will showcase Singapore's alternative model: also legalized, but with a more complete protection mechanism established through entry taxes, restrictions on domestic citizens, and mandatory exclusion systems. The comparison between the two models answers one question: after legalization, what remains unresolved?

Sports Betting: Re-Gamblingization

In 2024, revenue for legal sports betting operators across the U.S. was $13.71 billion, a year-on-year increase of 25.4%. Mobile sports betting revenue in Nevada grew by 96.8%.

Thirty years after "de-gamblingization," sports betting has brought gambling back to the core business of Las Vegas under the guise of "legal entertainment."

In 2018, the U.S. Supreme Court overturned PASPA (the Professional and Amateur Sports Protection Act), allowing states to decide on the legality of sports betting for themselves. In the six years since, 38 states plus Washington D.C. have legalized sports betting. This wave of legalization meets the core question of Chapter 2 once again: "The same act, given a different name, changes its legal characterization."

In Las Vegas, sports betting is redefining the spatial layout of casinos. The core areas of traditional casinos were baccarat tables and arrays of slot machines. Now, the core areas are giant screen walls, sports bars, and real-time odds display terminals. Casinos have transformed from "silent concentration" to "boisterous revelry." Gambling behavior has not decreased; it has simply taken on a more socially acceptable form.

The total revenue of the U.S. gaming industry in 2024 was $72 billion, a record high for the fourth consecutive year. Sports betting plus iGaming (online casinos) combined for over $21 billion, accounting for nearly 30% of total revenue. The share of traditional table games is being eroded by odds buttons on mobile screens. A tourist sitting by a Strip hotel pool uses their phone to bet on the outcome of that night's NFL game; if they win, they continue drinking in celebration; if they lose, it doesn't dampen their vacation mood. This type of "embedded gambling" has less ritual than walking into a casino and sitting at a card table, lower psychological barriers, and higher frequency.

The transformation story of Las Vegas, from Bugsy Siegel's Flamingo to the Sphere, is on the surface an evolutionary narrative of "from gambling city to entertainment city." The underlying logic remains unchanged: every generation of capital has found a way to package gambling more acceptably. The mob added a hotel to the gambling tables. Hughes added compliance to the hotel. Wynn added art and luxury to compliance. Sports betting added "sportsmanship" to gambling. The packaging changes. The stakes do not. In 2024, total revenue for the U.S. gaming industry was $72 billion, breaking historical records for the fourth year in a row. The "successful transformation" of Las Vegas is not evidence of gambling's disappearance, but a specimen of gambling's evolution. The most successful casinos are those that make gamblers forget they are gambling.

09. The Singapore Model

Nevada: Ninety-four years of legalized gambling. Per capita investment in problem gambling protection: $0.40. No mandatory physical casino exclusion laws. Problem gambling rate: approximately 2-3%.

Singapore: Fifteen years of legalized gambling. 356,000 active exclusion orders. Entry levy: 150 SGD per day. Problem gambling rate: 1.1%.

The gap between these two sets of figures is not merely a difference in policy details, but a fundamental disagreement regarding the assumption of citizens' rights to gamble. Nevada assumes that every adult who walks into a casino possesses fully rational autonomous judgment, and the government has no right to intervene. Singapore assumes that citizens may be unable to resist the addictive mechanisms of gambling, and the government has an obligation to proactively intervene and protect them.

One is the logic of liberalism. The other is the logic of paternalism. Gambling policy is a litmus test for how a government views the rational capacity of its citizens. Singapore chose the latter. Fifteen years of data can now answer one question: Is proactive protection effective? And at what cost?

A Casino Not Called a Casino

In 2001, Singapore faced an economic recession. The Economic Review Committee proposed legalizing gambling to stimulate tourism. Then-Deputy Prime Minister and Minister for Finance Lee Hsien Loong rejected the proposal, citing that "social harms outweigh potential revenue."

Three years later, on August 12, 2004, Lee Hsien Loong was inaugurated as Singapore's third Prime Minister. Ten days later, in his National Day Rally speech, the same man announced that the government was considering legalizing casinos. The economic situation had not substantially improved. What had changed was the position of the decision-maker: from a commentator to the person in charge.

On April 18, 2005, during a special session of Parliament, Lee Hsien Loong announced the Cabinet's decision: to build two "Integrated Resorts" (IR) at Marina Bay and Sentosa. This naming was not a random choice. The word "casino" triggered intense moral fear in Singaporean society (especially within the Malay Muslim community). "Integrated Resort" wrapped the gambling tables within a narrative framework of hotels, convention centers, shopping malls, and theaters. Linguistic choice became an administrative tool to lower the public's threshold of fear. The question of "who defines gambling" from Chapter 2 finds a precise administrative case here: the government redefined the social perception of gambling through its power of naming.

The ruling People's Action Party (PAP) took the unprecedented step of lifting the party whip, allowing Members of Parliament to vote freely according to their conscience. This was a rare acknowledgment of the issue's moral complexity. During the six-month public consultation period, tens of thousands of people signed petitions in opposition. Religious groups, social workers, and psychologists collectively expressed their dissent.

Lee Hsien Loong's strategy was not to ignore the opposition but to respond to it through institutional design. Simultaneously with the announcement of legalization, the government unveiled a social safeguard framework: the establishment of the National Council on Problem Gambling (NCPG), the design of an entry levy system, and the creation of an exclusion order mechanism. Casino license approvals were bundled with social protection measures, making them inseparable. On May 26, 2006, Las Vegas Sands won the Marina Bay site with the highest development investment of 3.85 billion SGD. On February 14 of the same year, the Casino Control Act was passed in Parliament. On April 2, 2008, the Casino Regulatory Authority was established. In 2010, the two integrated resorts opened one after another. From the parliamentary debate to the opening of the doors, the process took five years. Every step was a meticulous balance between "economic gain" and "social risk."

The 150 SGD Cooling-Off Period

Singapore citizens or permanent residents must pay a daily entry levy of 150 SGD (approximately 810 RMB) or an annual levy of 3,000 SGD to enter a casino. Foreign tourists enter for free.

What does 150 SGD mean for a low-income Singaporean family? For a bottom-tier worker earning 2,000 SGD a month, a single entry fee equals 7.5% of their monthly income. For a middle-class family (monthly income of 8,000 SGD), the entry fee is about 1.9% of their monthly income. This pricing was not arbitrary; it was deliberately set at a psychological threshold between "making the middle class hesitate" and "making the low-income give up."

Chapter 7 discussed the harms of the "zero-friction" design of Telegram casinos: there is no buffer between impulse and betting. The entry levy is reverse engineering, artificially creating a monetary barrier between the impulse and the gambling table. The 150 SGD is not a ban, but a threshold that requires clear consciousness to cross. Every act of paying the entry levy is itself a "confirmation": yes, I have decided to gamble today, and I am willing to pay the price of 150 SGD for it.

Contrast this with the Telegram casinos in Chapter 7: the entire operation is completed in 30 seconds, with zero monetary cost and zero identity verification. The Singapore entry levy stands at the exact opposite end of that spectrum: every act of gambling is attached to a clear, unavoidable economic cost. This cost is not high enough to constitute a "prohibition," but it is sufficient to make everyone who decides to enter clearly answer the question: "How much do I plan to spend on gambling today?" before they even reach the table.

Verification of effectiveness: The gambling participation rate among Singapore residents dropped from 52% in 2017 to 40% in 2023. During the same period, the problem gambling rate remained steady at 1.1%. The entry levy filtered out "casual gamblers" who just wanted to play a quick round, leaving behind participants who made a conscious choice to gamble. 92% of gamblers participate through legal channels, with Singapore Pools (the state-owned lottery operator) being the primary channel. Casinos are not the first choice for most Singaporeans; the entry levy has turned them into a consumer product for foreign tourists.

In April 2024, an administrative error occurred: the Ministry of Home Affairs forgot to renew the 2019 price increase order (the original order was valid for five years), and the entry levy automatically reverted to the old standard (100 SGD per day). This oversight lasted for 34 days, during which approximately 4.4 million SGD was collected above the legal standard. Parliament amended the law in September 2024 to rectify this issue. Even in a city-state known for administrative precision, a calendar-reminder-level mistake was made regarding a core social protection measure. The vulnerability of a system often lies not in its design, but in whether execution can maintain focus.

350,000 Exclusion Orders

As of December 31, 2023, there are 356,659 active casino exclusion orders in Singapore.

Singapore's total population is approximately 5.92 million, with about 4.04 million citizens and permanent residents, and approximately 3.4 million aged 18 and above. 356,000 exclusion orders mean that about 10% of adult citizens and permanent residents are banned from entering casinos.

Exclusion orders are divided into three tiers:

The first tier: Self-Exclusion (184,054 orders). Gamblers apply themselves, acknowledging an addiction risk and requesting that casinos refuse them entry. This is a proactive admission of one's own rational deficiencies.

The second tier: Family Exclusion (approximately 2,200 orders). Spouses or immediate family members apply to the NCPG, determining that a family member's gambling behavior has harmed the family's well-being. This is a negation of an individual's judgment by family members. In 2024, a "Family Visit Limit" was added: violators can face fines of up to 10,000 SGD and one year in prison.

The third tier: Automatic Exclusion (Statutory Trigger). Bankrupts, recipients of government financial assistance, legal aid recipients, and tenants with HDB (public housing) rental arrears are automatically excluded from casinos by law. No application or approval is required; the exclusion takes effect as soon as the condition is triggered. In 2024, the scope was expanded to include those receiving government-funded criminal legal defense.

These three tiers of exclusion orders answer the question raised in Chapter 2: "Who has the right to define a problem gambler?" Three entities—the individual, the family, and the state—simultaneously hold the power of definition. This is a structure almost unique in global gaming regulation. Contrast this with Nevada in Chapter 8: no mandatory exclusion, non-binding self-exclusion, and casinos have no obligation to enforce it.

Casinos that violate an exclusion order (by allowing an excluded person to enter) can be fined 20,000 SGD. The figure is not massive, but the enforcement records are made public long-term, and the threat to brand reputation far outweighs the fine amount.

The Economics of Two Casinos

In 2010, Marina Bay Sands (MBS) and Resorts World Sentosa (RWS) opened in succession. The initial combined investment for the two integrated resorts exceeded 10 billion SGD.

Over fifteen years, the cumulative EBITDA has exceeded 30 billion USD. MBS is expected to reach an annual revenue of 2.5 billion USD by 2025. Hotel occupancy is at 95.6%. In the third quarter of 2025, RWS saw a 22% growth in gaming revenue and a 7% growth in non-gaming revenue. Both operators have announced future expansion investments totaling 10.6 billion USD.

Why are only two casinos allowed?

Supply control is one of the core mechanisms of the Singapore model. Limiting the number of casinos to two (a duopoly confirmed to continue until 2030) means the market size for each casino is sufficient to support high investment and high tax burdens, while regulators only need to monitor two entities.

Casino tax rates: 18%-22% for mass gaming (tiered by GGR) and 8%-12% for premium gaming. This is significantly higher than Nevada's 6.75%. The government's annual gaming tax revenue is approximately 1.9 billion SGD. The ten-year license fee is 150 million SGD.

High tax rates usually compress corporate profit margins. Singapore's solution is to trade supply restriction for individual profitability: with only two casinos competing for a market of over 20 million international tourists annually, the revenue density of each casino is much higher than the landscape in Las Vegas, where dozens of casinos on the Strip compete with each other. High taxes and high profits can coexist in a market with restricted supply. This is classic monopoly economics: restrict supply quantity → increase individual pricing power → even with a heavier tax burden, profit margins remain ample. The government and operators have formed a community of shared interests: operators gain monopoly premiums, and the government gains high tax revenue; neither side has the incentive to break the existing status quo.

Contrast this with Nevada: dozens of casinos on the Strip compete for the same group of tourists on the same road. Price wars have driven down room rates and food and beverage premiums, and casino EBITDA margins are far lower than in Singapore. Competition has diluted the profitability of each casino and the taxes the government collects from any single entity. Singapore bypassed this predicament through supply restriction. MBS and RWS have now announced expansion plans totaling over 10.6 billion USD. MBS plans to build a new tower and entertainment facilities with an investment of about 4.5 billion USD. RWS's second-phase development budget is approximately 6.8 billion SGD, including new attractions like an aquarium, luxury hotels, and a retail complex. These additional investments fifteen years later prove one thing: the return on investment in the Singapore casino market remains high enough to attract billions of dollars in incremental capital.

The Cost of 1.1%

The problem gambling rate remains at 1.1%. The gambling participation rate has dropped from 52% to 40%. These numbers look like evidence of policy success.

Looking closer at the other side of the data: while the number of gambling participants has decreased, the average amount wagered per participant has risen. The entry levy filtered out "casual gamblers" who would "spend 100 SGD to play for two hours and leave," leaving behind "committed participants" who "prepare to gamble seriously all night after paying the 150 SGD entry fee." While the policy reduced the base population of gamblers, it may have deepened the level of involvement among the remaining participants.

A larger crack exists in the digital space. A 2023 survey showed that the participation rate in illegal online gambling rose from 0.3% in 2020 to 1%. The absolute figure seems small, but the increase is more than threefold. The Telegram casinos and USDT settlement systems described in Chapter 7 are not subject to entry levies, are not governed by exclusion orders, and do not rely on any physical space. Singapore's meticulously constructed protective walls are effective against physical casinos but powerless against gambling buttons on smartphone screens.

This is a variation of the "balloon effect" from Chapter 3 in the context of legalization: pressure does not eliminate demand; it only rewrites the channels through which demand is expressed. Singapore blocked the entrance to physical casinos, and part of the gambling demand flowed into the regulatory vacuum of online spaces. The entry levy has zero binding force on a gambling button on a smartphone. Exclusion orders have zero jurisdiction over a crypto casino registered in Curacao. The premise of maintaining a 1.1% problem gambling rate is that the statistical scope only covers known, monitorable gambling channels. How many participants are there in underground channels? Survey questionnaires rarely reach respondents who are unwilling to admit their participation in illegal gambling.

Two Types of Legalization

This chapter and Chapter 8 together constitute a comparative experiment: even with the same legalization of gambling, differences in institutional design lead to starkly different social consequences.

The Nevada Model: Unrestricted supply (dozens of casinos in free competition), extremely low tax rates (6.75%), thin protection (0.40 USD per capita investment), and trust in citizen rationality (no mandatory exclusion). Result: Problem gambling rate of 2-3%, casino profitability facing downward pressure under competition, and government gaming tax revenue diluted by competition.

The Singapore Model: Strictly restricted supply (duopoly monopoly), higher tax rates (18-22%), multi-layered protection (350,000 exclusion orders + entry levy + NCPG), and the assumption that citizens need protection. Result: Problem gambling rate of 1.1%, extremely high casino profitability (due to restricted supply), and concentrated, efficient government tax revenue.

A simpler comparison: The same addictive gambler in Nevada can freely walk into any casino to continue betting until bankruptcy. In Singapore, if they go bankrupt, they are automatically excluded from casinos; if their family reports them, they are forcibly excluded; if they receive government subsidies, they are banned from entry. The boundaries of the same person's right to gamble are completely different under different systems.

There is no simple "correct answer" between the two models. Nevada's liberal model requires every adult to bear the full consequences of their gambling decisions, including addiction and bankruptcy. Singapore's controlled model replaces individual judgment with state power, the cost of which is the surrender of civil liberties (350,000 people deprived of the right to enter a casino without a judicial process).

What happens if the logic of the entry levy is pushed to its extreme? If 150 SGD is effective, is 300 SGD more effective? If exclusion orders covering 10% of the population are effective, is 20% more effective? Where exactly is the line drawn between protection and control?

These questions have no end point. An entry levy can restrict the entrance to a physical casino, and an exclusion order can ban specific groups from stepping through a casino's doors. But when the medium of gambling shifts from a table to a phone screen, from physical chips to crypto tokens, and migrates from a casino building into instant messaging software, from whom should the entry levy be collected? To which server should the exclusion order be sent? The regulatory logic of physical space cannot find a corresponding enforcement fulcrum in virtual space.

The next chapter will approach this from another angle: when the definition of gambling itself expands infinitely, when blind boxes, prediction markets, and leveraged contracts all sit on the edge of gambling, any regulatory model faces the same dilemma: what are we protecting people from?

10. Everything is Gambleable

In 2024, Pop Mart's revenue reached 6.28 billion RMB, a year-on-year increase of 14.3%. Its market capitalization exceeded 148.8 billion HKD, with overseas revenue accounting for nearly 50%. In the same year, the annual scale of China's blind box market was approximately 58 billion RMB, accounting for 65% of the global share. A country where gambling is illegal across the board has nurtured the world's largest random consumption market.

Each box costs 69 RMB. The probability of obtaining a hidden LABUBU is 1/144. This probability is lower than the house edge in Baccarat (1.06%). Buyers do not need to show ID, pay an entry tax, face exclusion orders, or wait for license approval from a gambling regulatory authority.

What is displayed on the shelves are "designer toys." What the cash register prints are "commodity sales invoices." Payment, randomness, and the possibility of obtaining a rare item worth far more than the purchase price—these three elements are called "consideration, randomness, and prize" in gambling law. In business registration, it is called "retail." In the core business districts of Beijing, Shanghai, and Chengdu, Pop Mart's vending machines (which the company calls "smart retail terminals") contribute nearly 20% of sales. Every machine is a small slot machine that does not require a gambling license.

The 69-Yuan Gamble

The business model of blind boxes can be decomposed into three components: fixed consideration (69 RMB/box), a known probability distribution (1/12 for regular versions, 1/144 for hidden versions), and an uncertain return (the secondary market price of hidden versions can reach 10-50 times the original price).

These three components are completely isomorphic with the structure of a slot machine: fixed input (the cost of each pull), known odds (RTP usually 85-95%), and an uncertain jackpot (the top prize can be thousands of times the input). The difference lies in one key variable: the "prize" of a blind box is a physical toy, not cash. Based on this, the UK Gambling Commission determined that loot boxes do not constitute gambling, on the grounds that prizes "cannot be directly converted into monetary value."

This legal loophole ignores one fact: the existence of a secondary market gives every blind box prize a clear monetary price. The transaction price of a Pop Mart LABUBU hidden version on Xianyu often exceeds 2,000 RMB, a 29-fold return on the original 69 RMB price. The law says "it's not gambling." Market behavior says "this is gambling."

In 2023, China's State Administration for Market Regulation issued the "Guidelines for the Operation of Blind Boxes (Trial)." In 2024, the "Code of Conduct for Blind Box Management" was officially implemented: requiring businesses to disclose drawing probabilities, prohibiting sales to children under 8 years old, and setting a single consumption limit of 500 RMB. 30% of small and medium-sized blind box brands exited the market due to rising compliance costs.

The logic of these regulatory measures acknowledges the gambling attributes of blind boxes: if it's not gambling, why is probability disclosure necessary? Why is there a need to protect children? Why is a consumption limit required? Every regulation is a response to gambling risks, yet every regulation carefully avoids using the word "gambling." The 58-billion-RMB market size explains this linguistic caution.

Slot Machines in Games

In November 2017, EA's release of Star Wars: Battlefront II ignited global controversy. The game's "loot boxes" allowed players to pay for random equipment, while unlocking rare characters required dozens of hours of gameplay or direct payment for draws. One player calculated and posted on Reddit: unlocking Darth Vader required 40 hours of gameplay or approximately $260 in repeated paid draws. The game's payment structure was designed as a long-term probabilistic temptation: every box had a tiny chance of yielding a rare character, while failure resulted in low-value items, with sunk costs driving the next payment.

The post became the most downvoted comment in Reddit history (over 680,000 downvotes), and EA was forced to remove the paid loot box feature 48 hours before launch.

By 2025, the global game loot box market size is approximately $15 billion to $20 billion. Regulatory stances across countries are extremely polarized: Belgium determined that loot boxes violate gambling laws and banned them entirely; South Korea requires all games to disclose item probabilities (mandatory since March 2024); the UK believes that because prizes cannot be "cashed out," they are not gambling; the US has no federal law regulating them. In January 2025, the FTC penalized MiHoYo, the developer of Genshin Impact, alleging it "deceived children regarding the true cost of in-game transactions and the probability of obtaining rare prizes." In Austria, contradictory rulings appeared between courts: a Vienna court ruled that FIFA loot boxes were not gambling (players do not aim for profit), while a Styria court ruled that Steam loot boxes were gambling (items have monetary value in the secondary market).

The same behavior led two courts in the same country to opposite conclusions. The boundary of legal definition has a clear crack here. A global market of $15 billion to $20 billion operates in a gray space without a unified legal definition. Developers exploit this space to repeatedly test the limits of regulatory tolerance: lower probabilities, higher prices, and more extreme rarities.

Six Million Casinos

In January 2024, a platform named Pump.fun launched on the Solana blockchain. Its functionality is minimalist: anyone can create a new crypto token (Meme coin) in under 30 seconds, with no coding knowledge or qualification approval required, and a creation fee of less than $1.

Eighteen months later, the platform has seen the creation of over 6 million Meme coins. The platform itself has earned approximately $800 million in revenue through transaction fees.

What do 6 million tokens signify? The total number of licensed casinos in the United States is about 1,000. The number of "speculative tools" created by Pump.fun in a year and a half is thousands of times the number of all licensed casinos globally. Every Meme coin is a miniature gamble: early buyers can get hundreds of times in returns if they sell during a price surge, while latecomers bear all the losses. According to a 2024 Chainplay report, 55% of Meme coins are classified as "malicious." Throughout 2025, over 11.6 million crypto projects failed, setting an annual record.

The CTO of a16z compared Meme coin trading to a "high-stakes casino." A more accurate metaphor is: a casino without a license, without a regulator, without exclusion orders, and without problem gambling protections, open to anyone in the world with a smartphone. In the single month of January 2025, 1.7 million new Meme tokens were added. Every minute, a new "gambling table" is created, and every minute, an old "gambling table" goes to zero. 91% of new Meme coins on the Base chain have at least one security vulnerability, and one in every six tokens is a carefully designed scam.

Presidents have also joined this gamble. The $TRUMP token issued by Trump surged after launch, then plummeted 88%. Argentine President Milei endorsed the LIBRA token, which surged and crashed within hours, leaving a wasteland of retail investors. The gambling nature of Meme coins is so obvious that even political figures cannot pretend it is "investment," directly turning their names into chips.

Prediction Markets: Redefining the Bet

In 2022, the U.S. Commodity Futures Trading Commission (CFTC) identified Polymarket as an "illegal derivatives trading platform," imposing a fine and banning U.S. users.

In July 2025, the same Polymarket acquired QCEX, a derivatives exchange under CFTC regulation. In September, it received a No-Action Letter from the CFTC. In December, it officially reopened to U.S. users. In October, it received a $2 billion investment from Intercontinental Exchange (ICE), with a valuation of $8 billion.

The product has not changed: users are still "betting" on event outcomes—elections, interest rate decisions, sports matches. What has changed is the regulatory classification: what was previously seen as "illegal gambling" is now classified as "regulated event contracts." The "legitimacy spectrum" discussed in Chapter 2 finds a real-time case here: the same product, the same platform, transitioned from illegal to legal within three years.

In 2025, Polymarket and Kalshi together executed over $12 billion in transactions. After Robinhood launched its prediction market feature, it quickly accumulated 9 billion event contracts. FanDuel and DraftKings (the two largest sports betting companies) proactively gave up their Nevada gaming licenses to pivot toward prediction market businesses.

The motivation for this transition is worth scrutinizing: why would sports betting companies give up their gaming licenses? Because prediction markets are classified as "financial derivatives" rather than "gambling," they are not subject to state gaming laws and can theoretically operate in all 50 U.S. states. The same behavior (betting on the outcome of a sporting event), under a different legal framework (shifting from gaming law to commodity futures law), immediately changes the constraints. State regulators have realized this loophole: Ohio warned licensed sports betting operators that they might lose their licenses if they offer event contracts. Michigan, Arizona, and Massachusetts have issued bans one after another. Kalshi, Robinhood, and Crypto.com are dealing with over 20 lawsuits and cease-and-desist orders. This debate is likely to eventually head to the Supreme Court: is predicting the outcome of a sporting event gambling or a financial instrument? The answer will reshape the legal boundaries of a multi-billion-dollar market.

The Firework-Launching Brokerages

On January 18, 2024, Massachusetts securities regulators fined Robinhood $7.5 million. The reason: the platform used "gamification" design to induce users into irrational trading. Specific violations included: digital confetti and fireworks falling on the screen after a trade was completed, a milestone badge system, and push notifications urging users to check stock price fluctuations.

What does $7.5 million mean to Robinhood? In April 2025, the platform had 25.9 million funded accounts. Based on average daily trading volume, this fine is roughly equivalent to one day's worth of transaction fee revenue. The fine is a cost of doing business, not a behavioral constraint.

Reports from Gamblers Anonymous show a long-term increase in cases of gambling addiction among young men caused by trading apps. The behavioral patterns of these help-seekers are identical to those of traditional gambling addicts: chasing gains and selling on dips, using leverage to increase positions, hiding losses, and repeatedly swearing "this is the last time." In 2021, the Pennsylvania gambling hotline received more crypto and stock-related help calls than in the previous six years combined. The SEC proposed a draft rule in 2023 requiring brokerages to identify and eliminate conflicts of interest in gamified designs. This rule was officially withdrawn in June 2025 under industry lobbying pressure.

Robinhood's business logic is completely isomorphic with that of a casino. Casinos provide free drinks, free parking, and low-cost buffets to keep gamblers staying longer. Robinhood provides zero commissions, instant deposits, and social features to keep users trading more frequently. The casino's revenue source is the house edge. Robinhood's revenue source is payment for order flow (PFOF), where every user trade is sold to market makers. The more frequent the trading, the higher the platform's revenue. The price of "free" is that the user becomes the product itself.

The End of Definitions

Returning to the question raised in Chapter 2: What is gambling?

The law usually requires three elements to be met simultaneously: consideration (paying money or something of value), randomness (the outcome is not entirely determined by skill), and a prize (the possibility of obtaining money or a valuable return).

Pop Mart Blind Boxes: There is consideration (69 RMB), there is randomness (1/144 probability), but the prize is a physical toy rather than cash, thus exempting it from being defined as gambling. Pump.fun Meme Coins: There is consideration (purchase cost), there is randomness (price determined by market sentiment), and the prize is a crypto asset, thus it is packaged as "investment" to avoid suspicion of gambling. Polymarket Prediction Contracts: There is consideration (bet amount), there is randomness (event outcome unknown), and there is a prize (cash return), yet it avoids the gambling label under the name of "financial derivatives." Genshin Impact Loot Boxes: There is consideration (648 RMB for a ten-pull), there is randomness (0.6% probability for a limited character), but the prize is an in-game character that cannot be cashed out, so it is likewise not viewed as gambling.

Every case steps precisely on the boundary of legal definitions, using a technical reason to evade the "gambling" label. Blind boxes do not pay out cash. Meme coins brandish "investment" rather than "betting." Prediction markets call themselves "derivatives" to distinguish from "betting." Loot box prizes are "locked within the game."

These distinctions are meaningless to a brain experiencing a dopamine-hijacking cycle. Neuroscience does not distinguish between the "excitement of opening a blind box" and the "excitement of pulling a slot machine." Addiction mechanisms do not care about legal labels. The dopamine response triggered by variable ratio reinforcement is consistent in intensity across all scenarios.

When the definition of gambling is infinitely expanded, and when every product falls just outside the boundary, the "definition" itself loses its protective function. The entry tax in Singapore mentioned in Chapter 9 can stop citizens from walking through the casino doors. An entry tax cannot stop any consumer from spending 690 RMB on ten consecutive blind boxes in a store. An exclusion order can ban a bankrupt person from entering Sentosa. An exclusion order cannot ban anyone from opening Pump.fun to create a new Meme coin.

The next chapter will return to the final question: no matter how gambling disguises itself, migrates, or renames itself, one rule never changes—the expected value is negative. The house always wins.

11. The House Always Wins

It is Old Liu's 37th day of quitting gambling.

The Baccarat app on his phone has been uninstalled, but the green button on the recharge page still surfaces when he closes his eyes. This is not imagination, but a visual residue at the level of a conditioned reflex, much like the light spots seen after staring at a bright light for a long time and then closing one's eyes. The brain has recorded the position and color of that button, as well as the vibration feedback when his finger touched the screen, using a similar mechanism.

Psychology textbooks state that 21 days is the minimum threshold for habit formation. 37 days slightly exceeds this number. What Old Liu does not know is that statistical data shows 50% to 90% of problem gamblers will relapse. Among those who participate in Gamblers Anonymous, fewer than 10% remain abstinent after one year. Data from the World Health Organization is even colder: only 0.14% of people with gambling disorders worldwide have ever sought formal help.

0.14%. In China, this proportion may be even lower. Gambling is entirely illegal, and admitting to having a gambling addiction is almost equivalent to admitting to long-term participation in illegal activities. The psychological threshold for seeking help is not just shame, but also fear of the law. Currently, China has neither unified treatment standards for gambling disorders nor approved medications for addiction treatment. A few private gambling cessation institutions adopt a "5-day intensive treatment plus 5-year follow-up consolidation" model; 90% of the patients they see are online gambling addicts, mobile gamblers like Old Liu. The medications used in clinical treatment (such as Naloxone, mood stabilizers, and antidepressants) are all used off-label, as none have undergone clinical trials or regulatory approval specifically for gambling addiction. Therapists can only borrow treatment frameworks from alcohol and drug addiction because a dedicated treatment plan for gambling addiction has not yet emerged in China.

Day 37 is not the end, but a statistically highly unstable point in time. The brain during the withdrawal period is still undergoing neuroplastic remodeling. Addiction researchers using fMRI scans have found that the activity in the prefrontal cortex of individuals with gambling disorders only begins to approach normal levels six months after cessation. Old Liu's brain is currently in the early stages of this remodeling process. Every day is a dual battlefield of resistance and repair.

An Eternal Equation

The first chapter used pigeon experiments and dopamine circuits to explain the neural mechanisms of gambling addiction. The subsequent nine chapters displayed the full picture of the gambling empire: legal definitions, the effects of bans, Macau's legal valves, the institutional design of lotteries, the criminal industry chain in Northern Myanmar, the technical architecture of digital casinos, the capital evolution of Las Vegas, Singapore's protection experiments, and the expansion of definitions for blind boxes and Meme coins.

All these discussions point to a common conclusion at the mathematical level: the expected value is negative.

The house edge for betting on the "Banker" in Baccarat is 1.06%. This means for every 100 yuan wagered, the mathematical expectation is a loss of 1.06 yuan. The house edge for European Roulette is 2.7%, and for American Roulette, it is 5.26%. For slot machines, the house edge ranges from 2% to 10% depending on the settings. The payout rate for Chinese Sports Lottery and Welfare Lottery is approximately 50%, meaning the house edge is as high as 50%.

The meaning of these numbers is very straightforward: in a sufficient number of repeated bets, the gambler will inevitably lose. "Sufficient" is not a philosophical concept but a calculable number. Old Liu's daily wage is 300 yuan. If he plays Baccarat with a cumulative daily wager of 5,000 yuan (assuming 100 yuan per bet, playing 50 hands), a 1.06% house edge means a mathematical expectation of losing 53 yuan per day. Over a year, the loss would be 19,345 yuan, equivalent to 64 days of wages. No strategy can change this number. There is no "skill" in Baccarat; the result of each hand is independent of the previous one, and past wins or losses do not affect the future probability distribution.

Compare the house edge of other forms of "gambling": the probability of a "hidden" figure in a Pop Mart blind box is 1/144. A regular figure sells for 69 yuan, while a hidden figure averages about 1,500 yuan on the secondary market. After calculating the expected value, the mathematical expectation for each blind box is approximately negative 10 yuan (assuming a residual value of about 50 yuan for non-hidden figures). On Pump.fun, 95% of Meme coins eventually go to zero; the expected value is not "negative a few percent," but "negative 95%." Polymarket prediction contracts charge a fee for each transaction; after deducting the fee, the zero-sum game becomes a negative-sum game.

The names have changed. Casinos have become retail stores, blockchain platforms, and financial derivative exchanges. The math has not changed. Every product that "exchanges payment for a random result" has an implicit or explicit house edge. The commission rates differ, but the outcome is consistent.

The Flow of $450 Billion

The annual revenue of the global gambling industry from 2024 to 2025 is approximately $450 billion to $540 billion. The US market alone contributes about $120 billion, online gambling about $80 billion to $95 billion, and sports betting about $100 billion. These numbers are still growing. Forecasting agencies estimate that by 2030, the number of global gambling users will reach 1.9 billion, with a compound annual growth rate maintained between 3% and 5%. The growth rate in the Asia-Pacific region is faster, as mobile betting is replacing traditional casinos. Every new user is a potential source of net loss. The growth curve of the industry is the loss curve of global gamblers.

Understand these numbers in another way: the "revenue" of the global gambling industry equals the "total losses" of global gamblers. $450 billion is not value created out of thin air, but cash flowing out of the pockets of hundreds of millions of gamblers. This money is redistributed through the industrial structure: casino shareholders receive dividends and stock price appreciation, governments receive gambling taxes, employees receive salaries, regulatory agencies receive funding, and problem gambling protection projects receive a small portion of the funds.

In this flow chain, the gambler is the only participant with a net negative value. Behind every gambler who wins money, there are many more losing gamblers supporting the system. The existence of "winners" does not change the overall direction of the system: money flows from participants to operators. This flow direction is the defining characteristic of gambling. If the flow of money were reversed, casinos could not survive, lotteries could not operate, and blind boxes could not be profitable. The entire industry is built on the mathematical foundation that gamblers will inevitably lose.

At the industrial level, "the house always wins" is more absolute than at the individual level. An individual gambler might be a winner once, for one day, or even for one year. But the industry as a whole cannot lose. If casinos consistently lost money, they would close. The fact that the global gambling industry has grown for consecutive years proves that the house has not only won but is winning more and more. In 2024, the total revenue of the US gaming industry was $72 billion, breaking records for the fourth consecutive year. Every "record-breaking" moment means record-breaking losses for gamblers. This is an accelerating wealth transfer system.

Winning is More Dangerous than Losing

Gambling addiction shares the same dopamine pathway as drug addiction. The difference lies in one key detail: after withdrawal, drug addicts no longer come into contact with the harmful substance, whereas the positive stimulation received by a gambling addict when they "win" reinforces the neural circuits that drive the addiction.

In other words, a counterintuitive fact is: winning is a trigger for relapse. If a person who has quit gambling for 37 days wins 200 yuan through "casual gambling for fun," this "success experience" will reactivate the neural pathways that had been suppressed. The dopamine release confirms the prediction model deep in the brain: gambling can bring rewards. The probability of returning to gambling within the following days will rise sharply.

This is why the core principle of Gamblers Anonymous is "lifelong abstinence" rather than "moderate participation." There is no safe zone for "moderate gambling." For an addict, any single act of gambling can restart the entire addiction cycle. The time scale of neuroplasticity is much longer than 21 or 37 days. Some addiction research suggests that changes in the dopamine systems of gambling addicts may take years to partially recover. fMRI studies show that when gambling addicts recall experiences of winning money, the activity in the brain's nucleus accumbens (the reward center) is almost the same as when they actually receive a reward. Memory itself can trigger a virtual dopamine release. This explains why many relapses occur without any actual contact with gambling. A memory, an association, or a push notification related to gambling can trigger the reactivation of neural circuits.

The green button Old Liu sees when he closes his eyes is not a psychological problem, but a signal from the brain's reward circuit still saying "press that button." 37 days is not enough time for these signals to fall completely silent. Clinical data shows that about two-thirds of untreated individuals with gambling disorders will worsen, while one-third will recover on their own. Which group Old Liu belongs to cannot be determined on day 37. Statistics only provide group trends, not predictions of individual fate.

The Boundaries of Policy

Entry taxes can reduce the number of people walking into casinos, exclusion orders can prevent specific groups from accessing gambling tables, and probability disclosure can let consumers know the winning rate of opening boxes. These measures change "who is at the table" and "how many people are at the table." None of these measures can change the "math faced by the person sitting at the table."

From Baccarat's 1.06% to the lottery's 50%, from the house edge of over 20% in Northern Myanmar's gambling fraud parks to the 95% probability of Meme coins going to zero, every form covered in the first ten chapters points to the same mathematical structure: a negative expected value. The names have changed, the carriers have changed, and the absolute value of the house edge differs, but the direction is always the same.

Institutional design can optimize the boundaries of protection, allowing fewer people to be harmed and making the losses of victims smaller. Institutional design cannot change the mathematical essence of gambling. As long as the house edge exists, and as long as a person sits at the table and begins to bet, time will complete the harvest for the house. This common premise is the starting point for all gambling policies.

Old Liu's 38th Day

After the 37th day comes the 38th day.

This chapter cannot provide an "ending" for Old Liu. This is because addiction narratives do not have neat endpoints. Relapse could happen on the 38th day, or it could happen on the 380th day. Spontaneous recovery might quietly occur at some unpredictable moment. Clinical data provides probability distributions, not individual destinies. Old Liu has not read this data. Most gambling addicts do not know they are experiencing a mental disorder formally classified by the World Health Organization.

What is certain is this: if Old Liu relapses and reopens a Baccarat app to recharge 500 yuan, the math he faces will be exactly the same as on the first day. The 1.06% house edge will not change because he quit gambling for 37 days. If he turns to blind boxes, the 1/144 probability of a hidden figure will not increase because the buyer was once a gambling addict. If he buys a Meme coin, the 95% probability of it going to zero will not decrease due to any personal factors.

The coldness of the Law of Large Numbers lies in the fact that mathematical laws operate regardless of what an individual believes. A person can believe on the 38th day that they have "successfully quit," can believe "this is just a try," or can believe "luck has turned for the better." These beliefs do not affect the probability distribution.

The house always wins. The meaning of this sentence is not fatalism. It does not mean that every gambler will inevitably lose everything, but rather it is a precise mathematical statement: in a game with a negative expected value, the longer the participation time, the closer the cumulative loss will be to the mathematical expectation. Individual fluctuations in luck are just noise. The signal always points in the same direction: from the gambler's pocket to the house's vault.

For Old Liu, the choice faced on the 38th day is the same as on the 1st day: to press that green button, or not. The only difference is: 37 days ago, the signal sent by the brain was "this time I might win." 37 days later, if the data in this report has ever reached the cognition of a bricklayer, there is at least one more fact: in a game with a negative expected value, time itself is the house's most loyal employee.