The Rules We Choose

In the last week of September 1983, a line of semi-trailers drove north from El Paso, Texas, to the municipal landfill at Alamogordo, New Mexico, and tipped their loads into the ground. The cargo came from a factory Atari was closing: cartridges, consoles and parts, crushed by bulldozers and, from 29 September, covered with concrete. The Alamogordo Daily News counted between ten and twenty truckloads; the New York Times, which Atari confirmed the story to, said fourteen. An Atari spokesman described what was buried as “by-and-large inoperable stuff.”¹ Atari, which had supplied half of Warner Communications’ revenue in the first nine months of 1982, lost $538 million in 1983.²

The burial became a legend, and the legend outgrew the facts. It was said that millions of copies of E.T., often called the worst game ever made, lay under the desert. On 26 April 2014 a film crew dug the site up with the city’s permission. They found about 1,300 cartridges from 59 different games, E.T. among them but far from alone; a former Atari manager put the number originally buried at 728,000. The city sold 881 of the recovered cartridges on eBay for $107,930.15, and the Smithsonian took one copy of E.T. for its collection.³

The industry that buried its stock in 1983 is now among the largest forms of entertainment on earth. The analysts at Newzoo forecast that players will spend about $214 billion on games in 2026, more than half of it on phones.⁴ Nintendo had sold 157 million Switch consoles by the end of September 2026, and 26 million of its successor in sixteen months.⁵ Microsoft paid $68.7 billion for one publisher, Activision Blizzard, in 2023; a group led by Saudi Arabia’s Public Investment Fund paid about $55 billion for another, Electronic Arts, in 2026.⁶ Grand Theft Auto VI, whose release has been delayed twice, is due ten days after this essay is published. Regulators on four continents have ruled on whether a box of random prizes is gambling, and the World Health Organization classifies “gaming disorder.”

This essay follows how a demonstration built for a laboratory open day became that industry. It follows the decisions that shaped it: the cartridge slot left open to anyone, then closed by a lockout chip; the threat of federal regulation answered by a ratings board; the cartridge abandoned for the compact disc; the game delivered over the internet, then given away free and paid for in small amounts; the random prize sold for real money; the studio bought, the studio closed. Its deeper subject is the strangest thing about games, which a philosopher named in 1978: they are voluntary attempts to overcome obstacles nobody needs to overcome. People accept the rules of a game because they choose to. The history of video games is in large part a history of who writes those rules, and what happens when the people writing them are paid by the hour of attention or by the pull of a lever.

Before the screen: the game that was gambling

Arcade amusement machines were the first games to meet the law, and they met it on the question that would return in every decade after: is this skill or chance? In January 1942 New York City banned pinball. Mayor Fiorello La Guardia had the machines seized in all five boroughs and was photographed smashing them with a sledgehammer, on the reasoning that a game in which a ball’s path is governed by luck, and which rewarded players with free games, was a gambling device.⁷ The flipper, added to pinball machines from 1947, made the game more a matter of skill, but the ban stood for more than three decades. In April 1976 a writer named Roger Sharpe, appearing for the amusement trade association before a committee of the city council, played a machine to show that a skilled player could control the outcome; the committee voted six to nothing to license machines that rewarded good play with an extra ball rather than a free game, and the ban was lifted that spring.⁸

The distinction the city drew, between a game whose outcome the player shapes and one whose outcome is drawn by chance and paid for, never went away. It would be argued again in Brussels, The Hague and London about digital packs of football players, forty years later.

A game for visitors’ day

The first electronic games were built by scientists for no commercial reason at all. In October 1958 William Higinbotham, head of the instrumentation division at Brookhaven National Laboratory on Long Island and a physicist who had worked on radar at MIT during the war, wanted something to enliven the laboratory’s annual visitors’ day. In his own words, “it might liven up the place to have a game that people could play.” A technician, Robert Dvorak, spent about three weeks building it. Tennis for Two ran on an analog computer and showed a ball bouncing over a net on a five-inch oscilloscope screen. Hundreds of visitors queued to play on 18 October 1958.⁹

Three years later, at MIT, a group of young programmers around Steve Russell wrote Spacewar! for the university’s new PDP-1 computer. Playable by February 1962, it put two spaceships in combat around a star whose gravity pulled at them both. The code was shared freely, passed from one PDP-1 to another at universities and research firms, and the computer’s maker, Digital Equipment Corporation, featured it in a brochure in 1963.¹⁰

The box that plugged into a television

The first person to ask how to sell a video game was an engineer at a defence contractor. In August 1966 Ralph Baer, working at Sanders Associates in New Hampshire, imagined using the television sets already in tens of millions of homes as displays for games. Sanders approved a modest budget, and Baer’s team built a series of prototypes; the seventh, finished in January 1969 and known as the “Brown Box” for the wood-grain tape wrapped around it, played chase games, a ball-and-paddle game and a light-gun shooting game.¹¹ The television maker Magnavox signed a licence in January 1971 and unveiled the Magnavox Odyssey at Tavern on the Green in New York on 22 May 1972, at $99.95. It went on sale that September through Magnavox dealers in a limited number of cities. Lifetime sales were about 350,000, by Baer’s count.¹²

Among the people who saw the Odyssey demonstrated in May 1972 was Nolan Bushnell, who had already made Computer Space, a coin-operated game derived from Spacewar! and the first video arcade game sold commercially. On 27 June 1972 he and Ted Dabney incorporated Atari. Their first engineer, Allan Alcorn, built Pong as an arcade version of the Odyssey’s tennis game. A test cabinet of Pong was installed at Andy Capp’s Tavern in Sunnyvale, California, by August 1972. Atari built more than 8,000 cabinets by the end of production in 1974, and in 1975 sold a home version through Sears, which ordered 150,000 units for Christmas.¹³

Magnavox sued Atari and other arcade makers for infringing Baer’s patents in April 1974. In June 1976 Atari settled by taking a licence for $1.5 million. A federal judge in Chicago called Baer’s patent “the pioneering patent of the video game art,” and by the time the patents expired they had earned more than $100 million in licences and judgments.¹⁴ The first decision of the industry’s law was that the idea of playing a game on a television screen belonged to someone.

Atari’s open door

Atari needed capital to build a home console that could play many games rather than one, and in November 1976 it sold itself to Warner Communications for $28 million.¹⁵ The Video Computer System, later called the 2600, came out in September 1977. It took a new approach: the games were not wired into the machine but stored on cartridges that the player slotted in and swapped. The console became a platform, and the money moved to the cartridges.

Atari’s management under Ray Kassar, hired in 1978, treated the programmers who wrote those cartridges as anonymous staff. In May 1979 four of them, David Crane, Alan Miller, Bob Whitehead and Larry Kaplan, asked for royalties and credit on the box. By their own later account their games made up about 60 percent of Atari’s sales; Crane said his had brought in more than $20 million while he earned $20,000 a year. They were refused. That October they founded Activision with Jim Levy, a record-industry executive, as chief executive, and began selling their own cartridges for the 2600.¹⁶ Atari sued for theft of trade secrets. The case was settled in 1982 with Activision paying Atari royalties, and the settlement had a consequence neither side intended: it established that anyone could make games for Atari’s console without Atari’s permission, so long as they did not steal its secrets.

Hundreds of companies took the opportunity, including Johnson & Johnson, with a dental-hygiene game, and the pet-food maker Purina.¹⁷ Atari made two costly bets on licensed games. It produced about 12 million cartridges of its version of the arcade hit Pac-Man in 1982, more than there were consoles to play it, and shipped about 7 million; and it paid a reported $20 to $25 million for the rights to Steven Spielberg’s film E.T., then gave its programmer, Howard Scott Warshaw, about five weeks to write the game so it could be on shelves for Christmas. About 5 million copies of E.T. were made and about 1.5 million sold.¹⁸ Those figures rest on later reporting rather than Atari’s own records.

In December 1982 Warner announced that Atari’s earnings would grow far less than expected. Warner’s shares fell from 53⅝ to 30¾ within days, and the head of Atari’s home division was removed.¹⁹ Kassar left in July 1983. By then the market had collapsed. Unsold cartridges were dumped at a few dollars each, and the discounted games pulled down the price of everything else. Crane later described it bluntly: those “awful games flooded the market at huge discounts, and ruined the video game business.”²⁰ The most often quoted figures for the collapse, US home video game sales falling from about $3.2 billion at the peak to about $100 million in 1985, rest on weak sources and are best read as an order of magnitude.²¹ Warner sold Atari’s consumer business to Jack Tramiel in July 1984.

The crash was not caused by one bad game, whatever the legend in Alamogordo suggests. It followed from a structure. The console’s maker had no control over what was sold for its machine, so it could not stop a flood of poor and discounted cartridges; it sold its hardware cheaply and depended on cartridge sales it could not protect; and it overestimated demand for its own titles. The lesson the next company drew from this was the opposite of the lesson Activision had drawn.

Nintendo locks the door

Nintendo, a Kyoto company, released its Family Computer, the Famicom, in Japan in July 1983, just as the American market was collapsing. Bringing it to the United States meant persuading retailers who had been left holding unsold cartridges to try again. The company test-marketed the redesigned Nintendo Entertainment System in New York from October 1985 and released it nationally in 1986; accounts differ on how well the test went and on the month of the national launch.²²

The structural decision lay inside the machine. Every NES contained a chip, which Nintendo called the 10NES, that waited for a matching chip in the cartridge to send the right signal. Without it, as a federal appeals court later put it, the console “refuses to operate the cartridge.” The chip turned the console’s slot into a gate, and Nintendo kept the key. A company that wanted to publish a game for the NES had to sign Nintendo’s licence, and the terms, as the court described them, were strict: “Under the license, Nintendo would take Atari’s games, place them in cartridges containing the 10NES program, and resell them to Atari.” Nintendo manufactured every cartridge itself. “Nintendo limited all licensees, including Atari, to five new NES games per year,” and a licensee could not release the same game on a rival console for two years.²³ Reports that Nintendo also required large minimum orders paid in advance are widespread, but neither court opinion describes them.

The trade-off, typical of Markets With Two Sides, was explicit. Nintendo gave up the openness that had let Activision and hundreds of others flourish on the 2600, and in return it got control over the number and quality of games, a cut of every cartridge sold and protection from the flood that had sunk Atari. Developers lost freedom and margin; retailers and players got fewer bad games, at higher prices. The NES brought the home console back to North America.

The control reached beyond the cartridge. In 1991 Nintendo settled allegations by the states of New York and Maryland, with the Federal Trade Commission involved in the investigation, that from 1987 to 1990 it had pressured retailers to keep NES prices up by threatening to cut their supply. Nintendo admitted no wrongdoing and agreed to pay $29.75 million, including up to $25 million in $5 coupons redeemable against Nintendo games.²⁴ The remedy sent buyers back to the company accused of overcharging them.

The lock itself was tested in court twice in 1992, and the courts drew a line that still shapes the industry. Atari’s games arm, operating as Tengen, had become a Nintendo licensee in December 1987 while trying to break the lock. Its lawyer obtained a copy of the 10NES program from the US Copyright Office by stating, falsely, that Atari was already a defendant in a lawsuit over it. In September 1992 the Federal Circuit held that “reverse engineering object code to discern the unprotectable ideas in a computer program is a fair use,” but only when it was “untainted by the purloined copy.” Atari’s lie lost it the case.²³

Sega’s Genesis console had its own lock, which looked for the four bytes “S-E-G-A” in a cartridge and, when it found them, flashed a message that the game was “produced by or under license from” Sega. Accolade, a developer that had declined Sega’s licence because it would have made Sega the exclusive manufacturer of its games, took Sega’s code apart to learn how to unlock the console. Sega sued in October 1991. In October 1992 the Ninth Circuit held that where taking a program apart is the only way to reach its functional elements, and there is a legitimate reason, “disassembly is a fair use of the copyrighted work, as a matter of law.” It added that Sega itself was “primarily responsible” for any confusion caused by its trademark appearing on an unlicensed game, since its lock put it there, and that the device served “to limit competition in the market for Genesis-compatible games.”²⁵ Together the rulings meant a lockout chip could not legally stop a competitor who studied it honestly. Platform holders kept their gates, held closed by contracts, encryption and, after 1998, a federal law against circumventing copy protection.

A ratings board instead of a law

In the early 1990s video games began to show violence in a form realistic enough to alarm parents. Two titles became the focus: the fighting game Mortal Kombat on Sega’s Genesis and Night Trap, a game for Sega’s CD add-on. On 9 December 1993 Senators Joe Lieberman and Herb Kohl held joint hearings of two Senate committees on violence in video games, with a second session in March 1994. Howard Lincoln of Nintendo and Bill White of Sega testified.²⁶ The question before the industry was whether it would be regulated by Congress or regulate itself.

It chose to regulate itself. In 1994 the industry’s new trade association, the Interactive Digital Software Association, founded the Entertainment Software Rating Board, which announced a system of five age categories and seventeen content descriptors printed on every box.²⁷ Europe followed in April 2003 with PEGI, a common system that replaced most national ones. The ratings did what the industry hoped. They answered the charge that children were being sold violent games without warning, and they gave legislators a reason to hold back.

The question returned in a different form. California passed a law in 2005 that banned the sale or rental of “violent video games” to minors, with fines of up to $1,000. The Supreme Court struck it down on 27 June 2011, by seven votes to two, in Brown v. Entertainment Merchants Association. Justice Antonin Scalia, writing for the majority, held that video games “communicate ideas” through “many familiar literary devices” and so qualify for the protection of the First Amendment. He pointed out that children’s literature has never been gentle (“Grimm’s Fairy Tales, for example, are grim indeed”), that “disgust is not a valid basis for restricting expression,” and that the psychological studies California relied on “do not prove that violent video games cause minors to act aggressively,” showing “at best some correlation” with “minuscule real-world effects.” He also noted that the voluntary ratings already did much of what the law intended. Justices Samuel Alito and John Roberts agreed only that the law was too vague, and warned that the Court was “far too quick to dismiss the possibility” that playing games is different from reading or watching. Justice Stephen Breyer, dissenting, asked what sense it made to forbid selling a thirteen-year-old a magazine with a nude image while protecting a game in which he could torture and kill a woman.²⁸

The science that the case relied on was contested throughout. A 2010 meta-analysis by Craig Anderson and colleagues concluded that violent games increase aggression; Christopher Ferguson and others argued that its methods inflated the effect.²⁹ In March 2020 the APA revised its position to say that there is “insufficient scientific evidence to support a causal link between violent video games and violent behavior,” while maintaining a “small, reliable association” with aggressive outcomes such as yelling and pushing, and warning that its findings should not be used to blame mass shootings on games.³⁰ Several of the APA’s own divisions objected that even that association was overstated.

The disc

The next turning point came from a deal that collapsed. Nintendo had agreed to work with Sony on a compact-disc drive for its Super Nintendo; at the Consumer Electronics Show in June 1991 Sony showed its design, and Nintendo announced a partnership with Philips instead.³¹ Sony’s engineer Ken Kutaragi was put in charge of building a console of Sony’s own. Sony Computer Entertainment was founded in November 1993, and the PlayStation went on sale in Japan in December 1994.

The PlayStation’s decisive feature was its medium. Discs held far more data than cartridges, room for three-dimensional worlds, recorded music and film, and that changed the economics for publishers. Hironobu Sakaguchi of Square, whose Final Fantasy series had been one of Nintendo’s strongest assets, later said the company moved Final Fantasy VII to the PlayStation for two reasons: Nintendo’s cartridges kept getting more expensive, and discs could hold the 3D characters and video the game needed.³² Nintendo kept cartridges for the Nintendo 64. Sega, which had rushed its Saturn console onto shelves at $399 at the first E3 trade show in May 1995, watched a Sony executive walk on stage and announce the PlayStation’s American price in one word: “$299.”³³

The PlayStation sold more than 102 million units and its successor more than 160 million.³⁴ Sega’s Dreamcast, released in 1998, could not hold its ground. In January 2001 Sega announced that it would stop making consoles and become a publisher of games for others’ machines; Dreamcast production ended that March.³⁵ A company that had shared the market with Nintendo a decade earlier now wrote games for its former rivals.

Shareware, Live and Steam

On personal computers the industry developed along a different line, with no gatekeeper owning the machine. At midnight on 10 December 1993 id Software, a small company in Texas, released the first episode of Doom free of charge as shareware on a University of Wisconsin file server. Players who wanted the rest of the game paid for it. Demand crashed the server.³⁶ The free episode was the marketing, and the internet was the shop.

Microsoft entered the console business with the Xbox in November 2001 and, a year later on 15 November 2002, launched Xbox Live, an online service that required a broadband connection and a paid subscription.³⁷ Requiring broadband when few homes had it was a bet that infrastructure would catch up and that a smaller audience with a better experience was worth more. Online play became a service people paid for.

Valve, the developer of Half-Life, had built a program called Steam to deliver updates to its games. With Half-Life 2, released on 16 November 2004, Valve required every buyer to activate the game through Steam, even those who had bought it on disc in a shop. The login servers were overwhelmed on launch day, and many players saw the requirement as a form of copy protection imposed on them.³⁸ Valve used its most anticipated game to seed a store. Steam grew into the main shop for computer games, taking 30 percent of each sale. In 2018 it lowered its cut to 25 percent on a game’s revenue above $10 million and 20 percent above $50 million, a change that helped the largest publishers and drew complaints from small ones.³⁹ A class action brought by developers, led by Wolfire Games, alleging that Valve used its position to keep commissions high, was certified for about 30,000 publishers in 2024 and had not gone to trial by autumn 2026.⁴⁰

The audience nobody was serving

By the mid-2000s Nintendo had fallen behind Sony and Microsoft in the race for more powerful hardware. Its president, Satoru Iwata, chose not to compete in that race. “‘Expanding the gaming population’ is our globally common strategy,” he told investors in June 2006.⁴¹ Nintendo, he said, had been the first to think about how to expand the number of people who play video games. The Wii, released at the end of 2006, was built around a motion-sensing controller meant to be usable by people who had never played, and it reached them: Nintendo sold 101.63 million Wiis.⁵

The success was not as complete as the sales suggest. In April 2012 Iwata told investors that while the Wii had reached “a large number of new consumers,” the company “could not adequately create the situation that such new consumers played games frequently … As a result, we could not sustain a good level of profit.”⁴² Many of the new players bought the console, played the bundled games and stopped. Its successor, the Wii U, sold only 13.56 million units; in January 2014 Nintendo cut its forecast from a ¥100 billion operating profit to a ¥35 billion loss, and Iwata halved his own pay for five months.⁴³ The answer was the Switch, released in 2017, a single device that worked as both a home console and a handheld, consolidating two audiences Nintendo had served with separate machines. By 30 September 2026 it had sold 157.20 million units, within a few million of the more than 160 million Sony reports for the PlayStation 2. Its successor, the Switch 2, launched on 5 June 2025 at $449.99, sold 3.5 million in four days and 26.35 million by the end of September 2026.⁵

Free, and paid for in pieces

The largest change in how games are paid for began in South Korea. Around the turn of the century Nexon, a Seoul developer, gave away games such as QuizQuiz and later MapleStory and KartRider, and earned money by selling items inside them.⁴⁴ The model, known as free-to-play, reached Western audiences through Facebook. Zynga’s FarmVille, the most played game on Facebook from August 2009 to December 2010, was free to play and sold items inside the game. When Zynga prepared to go public in 2011 its filings showed that only about 6.7 million of its 230 million monthly users, less than 3 percent, had paid anything.⁴⁵

On 15 October 2009 Apple allowed free iPhone apps to sell items inside them; until then in-app purchases had been permitted only in apps people had already paid for.⁴⁶ Mobile developers adopted free-to-play almost at once. Supercell, a Finnish company, made $892 million in 2013, mostly from two free games, Clash of Clans and Hay Day. King, the maker of Candy Crush Saga, reported about 12 million paying users in the last quarter of 2013, out of some 300 million monthly players.⁴⁷ Epic Games made its new Fortnite Battle Royale free on 26 September 2017 and sold seasonal “battle passes,” which unlocked rewards for playing over the following weeks.⁴⁸ Its later fight with Apple over the payment rules inside Fortnite on the iPhone is told on this site in The Walled Garden.

The economics of free-to-play depend on a small number of players spending a great deal. The analytics firm Swrve, studying more than 20 million players of 40 mobile games in 2016, found that 1.9 percent made a purchase in a given month and that 0.19 percent of all players, the top tenth of payers, provided 48 percent of revenue; its earlier study had found 0.23 percent of players providing 64 percent.⁴⁹ The industry calls such players “whales,” a term borrowed from casinos. By 2025, on Newzoo’s figures, mobile games accounted for more than half of global games revenue.⁵⁰ A design that charges nothing at the door and depends on one player in several hundred has an obvious incentive: find the people most willing to spend and keep them spending.

The box of chance

The sharpest form of that incentive was the loot box: a purchase whose contents are revealed only after paying. Its earliest known form in a widely played game was a “Gachapon ticket” in the Japanese version of MapleStory in 2004. In Japan, mobile games built whole economies around “gacha” draws, and in May 2012 the Consumer Affairs Agency ruled one variant, “complete gacha,” in which players had to collect a full set of random items to win a rare prize, unlawful under the country’s rules on premiums; the major companies dropped it from 1 July 2012.⁵¹ The general form survived and spread to console and PC games, where randomised packs of footballers in Electronic Arts’ FIFA Ultimate Team became one of the most profitable products in the industry.

The backlash came with Star Wars Battlefront II in November 2017. Players found that the game’s most popular characters, such as Darth Vader, were locked behind long play or purchases, and that crates bought with real money contained upgrades that made players stronger. A player who had paid $80 complained on Reddit. A member of EA’s community team replied that the intent was “to provide players with a sense of pride and accomplishment for unlocking different heroes.” The reply became the most down-voted comment in Reddit’s history, with about 668,000 down-votes by Guinness World Records’ later count. A Hawaii legislator called it a “Star Wars-themed online casino.” On 16 November, hours before the launch, EA switched off all purchases in the game, and in March 2018 it rebuilt the progression system so that crates held only cosmetic items.⁵²

Regulators began to rule. In April 2018 the Belgian Gaming Commission found that the paid loot boxes in Overwatch, FIFA 18 and Counter-Strike: Global Offensive were illegal gambling under Belgian law. Blizzard switched off paid boxes for Belgian players. Five years later the researcher Leon Xiao found that 82 percent of the highest-grossing iPhone games in Belgium still sold randomised items, including 80 percent of those rated suitable for children of twelve and over: “The Belgian ban has not been effectively enforced.”⁵³ The Dutch gambling authority found in 2018 that loot boxes whose prizes could be traded for money broke its law, and in 2020 ordered EA to change its FIFA packs or pay up to €10 million. A district court upheld the order. On 9 March 2022 the Council of State, the highest administrative court, reversed it, holding that the packs were part of a game of skill, FIFA Ultimate Team, rather than a separate game of chance.⁵⁴ The argument that had decided pinball in New York in 1976 decided football packs in The Hague in 2022, the other way.

The United Kingdom chose not to legislate. A committee of the House of Commons recommended in September 2019 that paid loot boxes be regulated as gambling. In July 2022 the government concluded that research showed “an association between loot boxes and harms” but had not established “a causative link,” and asked the industry to act instead. The industry’s trade body published eleven voluntary principles in July 2023, including parental consent for under-18s and disclosure of odds. A study by Xiao and a colleague in 2025 found “widespread” non-compliance with them.⁵⁵ Australia, from 22 September 2024, required games with paid loot boxes to carry at least an M rating and games with simulated gambling an adults-only R18+.⁵⁶ Brazil went furthest: a child-protection law in force from 17 March 2026 bans loot boxes in games aimed at or likely to be used by minors.⁵⁷ The European Parliament called for common rules in January 2023, EU consumer authorities issued non-binding principles on in-game currencies in March 2025, and the European Commission’s Digital Fairness Act, expected to address both, had not been published when this essay was written.⁵⁸

In the United States the main instrument was consumer protection law. In December 2022 Epic Games agreed to pay $275 million for collecting children’s personal data in Fortnite without parental consent, the largest penalty ever for breaking an FTC rule, and $245 million in refunds to players the FTC said had been tricked by “dark patterns” into unintended purchases. In January 2025 the maker of Genshin Impact agreed to pay $20 million and to stop selling loot boxes to children under sixteen without a parent’s consent, after the FTC alleged it had misled players about their odds and costs.⁵⁹

The research on loot boxes and problem gambling has been consistent in one respect. In 2018 David Zendle and Paul Cairns surveyed 7,422 gamers and found that the amount people spent on loot boxes was linked to the severity of their problem gambling, far more strongly than spending on other in-game items, and they repeated the finding in 2019. Two meta-analyses published in 2021 and 2022 put the average correlation at about 0.26 to 0.27, small to moderate and repeatedly observed.⁶⁰ The authors and the British government’s own evidence review agreed on what this does not show: whether loot boxes lead people into problem gambling, or whether people already prone to problem gambling are drawn to loot boxes, remains unresolved.

Disorder, limits and the evidence

In May 2019 the World Health Assembly adopted the eleventh revision of the International Classification of Diseases, which came into force on 1 January 2022 and includes “gaming disorder”: a pattern of “impaired control over gaming,” “increasing priority given to gaming over other activities” and “continuation or escalation of gaming despite the occurrence of negative consequences,” severe enough to cause significant impairment and normally evident for at least twelve months.⁶¹ In 2017 scholars led by Espen Aarseth argued that the research base was weak, that the criteria borrowed too heavily from addiction and gambling, and that the label risked pathologising ordinary play; others replied that requiring real impairment guarded against that.⁶² A 2021 meta-analysis of 53 studies put the worldwide prevalence of gaming disorder at about 3 percent, revised to 2.96 percent in a later correction, with much lower estimates in the most rigorously sampled studies.⁶³

China acted more directly. On 30 August 2021 its National Press and Publication Administration limited children under eighteen to one hour of online games, from 8 to 9 p.m., on Fridays, weekends and public holidays, enforced through real-name registration; officials said many parents had found earlier limits too loose.⁶⁴ In December 2023 the same regulator published draft rules to curb spending incentives in games; nearly $80 billion was wiped off the market value of Tencent and NetEase, and within weeks the draft had been removed from the regulator’s website. By 2026 China was approving new games at a high rate while keeping the limit on children’s play.

The best evidence on what games do to wellbeing has improved in a way that cuts against both alarm and reassurance. In 2019 Amy Orben and Andrew Przybylski, analysing data on more than 350,000 adolescents, found that digital technology use explained at most 0.4 percent of the variation in their wellbeing.⁶⁵ Later studies at Oxford persuaded publishers to share records of actual play. A 2022 study of 38,935 players of seven games over six weeks found that time spent playing had “limited if any” effect on wellbeing, but that people who played because they enjoyed it felt better and people who played because they felt they had to felt worse.⁶⁶ The average effect of hours played is close to zero while a small minority are harmed; both can be true at once.

Consolidation and its costs

On 18 January 2022 Microsoft agreed to buy Activision Blizzard, the publisher of Call of Duty, for $68.7 billion, saying that “Mobile is the largest segment in gaming.” Britain’s Competition and Markets Authority blocked the deal in April 2023, finding that Microsoft already accounted for an estimated 60 to 70 percent of cloud gaming services. The US Federal Trade Commission lost its attempt to stop it in court in July 2023. Microsoft restructured the deal by selling Activision’s cloud-streaming rights outside Europe to Ubisoft, and the CMA cleared it on 13 October 2023. The Ninth Circuit upheld the ruling against the FTC in May 2025, and the FTC dropped its case.⁶⁷

Spending stopped growing. Newzoo estimated the global market at about $193 billion in 2021, falling in 2022 for the first time in its records and roughly flat through 2024.⁶⁸ Companies that had hired for continued growth cut. Trackers counted about 10,500 jobs lost in the industry in 2023 and between 14,600 and 15,600 in 2024, the peak; estimates for 2025 range from 5,300 to about 9,200, and by 20 October 2026 another 10,977 had been counted.⁶⁹ Microsoft cut 1,900 gaming jobs in January 2024, cancelled games in 2025, and in July 2026 said Xbox would lose about a fifth of its staff and part with four studios.⁷⁰ Embracer Group, a Swedish company that had grown by buying studios, lost an expected investment in May 2023, cut about 1,400 jobs in six months, sold Gearbox and Saber, and split itself into three companies.⁷¹

The tools developers depend on also showed their owners’ power. In September 2023 Unity, maker of one of the most widely used game engines, announced a fee of up to $0.20 for each installation of a game made with it, once a game passed thresholds of revenue and installs. Developers revolted, and Unity cancelled the fee a year later.⁷² Performers struck. Voice and motion-capture actors represented by SAG-AFTRA walked out on major game companies on 26 July 2024, chiefly over the use of artificial intelligence to replicate their voices and likenesses, and ratified a new agreement on 9 July 2025 by 95 percent, with requirements for informed consent and disclosure for digital replicas.⁷³ Developers themselves grew more hostile to generative AI as its use spread: in the Game Developers Conference’s annual survey the share who said it was harming the industry rose from 30 percent in 2025 to 52 percent in 2026, while by July 2025 nearly 8,000 games on Steam disclosed using it.⁷⁴

The industry’s largest buyout closed in 2026. On 29 September 2025 Electronic Arts agreed to be taken private by Saudi Arabia’s Public Investment Fund, the investment firm Silver Lake and Affinity Partners for about $55 billion, $210 a share, financed in part with $20 billion of debt. Shareholders approved it in December 2025, and it closed in August 2026; EA’s shares were delisted.⁷⁵ And the most anticipated game in years, Rockstar’s Grand Theft Auto VI, was delayed from 2025 to May 2026 and then to 19 November 2026, “to finish the game with the level of polish you have come to expect and deserve.”⁷⁶

What video games are today

In 2026 video games are less a product than several overlapping businesses that share a name. One is the console business, now run by Nintendo, Sony and Microsoft: closed platforms in Nintendo’s 1985 mould, where the maker approves every game, takes a share of every sale and sells hardware at prices that rising component costs have pushed up. Sony had shipped 97.9 million PlayStation 5s by September 2026.⁷⁷ A second is the computer business, centred on Steam, where anyone can publish and the store takes its cut. The third and largest is mobile, where games are free to download and paid for by a small minority of players through purchases inside them; on Newzoo’s forecast it accounts for more than half of the roughly $214 billion spent on games in 2026.⁴

Ownership has concentrated. Microsoft owns Activision Blizzard; a group led by a Saudi state fund now owns Electronic Arts. Regulation has fragmented. A loot box that is legal and unrated in one country carries an M rating in Australia, is unlawful for children in Brazil, requires parental consent for children under sixteen in one American publisher’s games by order of the FTC, and is technically illegal but widely sold in Belgium. Children in China may play online for three hours a week. Developers face a fourth straight year of heavy job losses, and a majority of those surveyed believe generative AI is harming their industry even as thousands of games use it.

The players are everyone; a Nintendo strategy built on that observation sold over a hundred million consoles. For most of them, the best evidence suggests, the hours spent playing do little harm or good on average, and why they play matters more than how long.

How video games changed play

From a game you bought to a game that kept selling to you

For most of their history games were sold like books: a fixed price for a finished thing. The cartridge and the disc made each game a product. Online connections, Apple’s decision of 2009 and the free-to-play model turned the game into a service, updated continuously and paid for in instalments, many of them small, many of them by a few players. The design question changed from “will people buy this?” to “will people keep playing, and will some of them keep paying?”

From an open slot to a gated platform

The 2600’s open cartridge slot let anyone sell a game, and the market drowned in them. Nintendo’s lockout chip made the console a gate and the platform owner a licensor, and every successful console since has followed that model, along with Apple’s and Google’s app stores. The courts in 1992 confirmed that studying a lock is lawful, but the gate survived through contracts and encryption. The fight over who controls the door runs from Atari v. Nintendo to Wolfire v. Valve.

From a moral panic to a health classification

The concern about games moved from what they showed to what they did. The 1993 hearings and the California law were about content, especially violence, and both the courts and the scientific bodies ended by rejecting the claim that violent games cause violent behaviour. The later concern was about compulsion and spending: gaming disorder in the WHO’s classification, loot boxes and problem gambling, China’s time limits. That concern rests on firmer evidence about a small group of people and weaker evidence about everyone else.

From a craft to an industry that consolidates and cuts

Early games were made by a few people, sometimes in weeks; large modern games take hundreds of people and years, funded by acquisitions and debt. When spending stopped growing after 2021 the same scale made the cuts large, and performers struck for control over digital copies of their voices.

What the history suggests

The history of games shows, in unusually sharp form, patterns seen elsewhere on this site. Who pays shapes what gets built: a game paid for once is designed to be worth buying, a game paid for by subscription to be worth returning to, and a game paid for by a fraction of a percent of its players to find and keep those players. The metric becomes the target: daily active users, session length and revenue per paying user became what free-to-play games were built to raise. A defining strength becomes the constraint: Atari’s open slot, which let the 2600 fill with games, drowned it; Nintendo’s attachment to cartridges, which gave it control, sent Square to Sony. Distribution beats elegance: Valve used its most anticipated game to put its store on millions of computers, and Apple’s decision in 2009 made it possible to give a mobile game away and charge inside it. Defences against abuse concentrate the market: the cure for Atari’s flood of poor games was a gatekeeper with the power to say no, and that power did not go away when the flood ended.

Games differ from most products in one respect: the product is a set of rules people accept voluntarily, for no purpose outside them. A word processor serves an end beyond itself; a game, for most players, is its own end. That makes games unusually sensitive to who writes the rules and why. When a designer adds an obstacle to make a game more absorbing, the player is served; when a designer adds the same obstacle to sell its removal, the player is being worked. From outside the two look identical, which is where the philosophy of play begins.

Understanding the philosophy

The connections drawn below between particular thinkers and the history of video games are the author’s interpretation. Most of the thinkers wrote before video games existed or before the developments described; the point is to see what their ideas reveal, not to claim they foresaw them.

1. A world apart: the magic circle

The idea in plain terms

When people play a game, they act as if a small region of the world has its own rules. Inside it, a bishop moves diagonally, a goal is worth one point, and a pawn captured is a pawn lost, though in ordinary life none of this means anything. Players know the frame is artificial and treat it as binding anyway. The idea of the magic circle is that games create such a bounded space, separate from ordinary life, in which new meanings hold.

The tradition

The Dutch historian Johan Huizinga, in Homo Ludens (1938), argued that play is older than culture, since animals play, and that law, war, poetry and ritual all grew from play-forms. He defined play as “a free activity standing quite consciously outside ‘ordinary’ life,” intensely absorbing, proceeding within its own boundaries of time and space according to fixed rules, and “connected with no material interest, and no profit can be gained by it.” Among the spaces set apart for play he listed “the arena, the card-table, the magic circle, the temple, the stage, the screen, the tennis court,” which he called “temporary worlds within the ordinary world, dedicated to the performance of an act apart.” He observed that societies are harsher on the spoil-sport, who refuses to take the game seriously, than on the cheat, who breaks the rules but “still acknowledges the magic circle.”⁷⁸

The phrase appears only a handful of times in Huizinga. It became a central concept through the game designers Katie Salen and Eric Zimmerman, whose textbook Rules of Play (2003) used it for the shared space of meaning a game’s rules create. Zimmerman later wrote that he and Salen “more or less invented the concept,” cobbling it together from Huizinga and Caillois: “Blame us for all the trouble, not Huizinga.”⁷⁹ The economist Edward Castronova found in 2001 that players of EverQuest traded virtual items for real money in such volume that the game’s world could be given a gross national product and an hourly wage; in Synthetic Worlds (2005) he called the boundary a porous membrane.⁸⁰ The media scholar Mia Consalvo argued in 2009 that the idea was too static and too formal: players bring their lives, money, friendships and cheating into games, and the boundary is constantly negotiated rather than given.⁸¹

Where it shows up

The history of video games is partly a history of the circle being made more porous on purpose. A cartridge game bought in 1983 had no link to the buyer’s bank account once paid for. A free-to-play game is built around the exchange between its economy and the real one. Loot boxes, the paid packs in FIFA Ultimate Team and the battle pass all put a price on what happens inside the game. The Dutch Council of State’s 2022 ruling turned on exactly this boundary: whether the packs were a separate game of chance or part of the game of skill inside which they were sold. Huizinga’s clause that play is connected with no material interest reads, in that light, less like a description than a test that much of the modern industry fails.

The strongest objection

A sceptic would say that there never was a sealed circle. People have bet on games, played for prizes and made friends and enemies over them for as long as there have been games. Zimmerman himself argued that critics were attacking a strict version nobody held. If the circle was always porous, then loot boxes did not breach anything; they made visible a traffic between game and life that had always existed, and a theory that judges them by the purity of the circle is judging against a fiction.

What it means today

So the circle was never sealed. What survives of the idea is consent and visibility. Players can choose to let the outside world into a game, as they do when they bet among friends. The question for a designer is whether the boundary is being crossed with the player’s understanding or behind it. Selling cosmetic items that do not change the game keeps the circle’s rules intact. Selling advantage, or prizes drawn by chance, imports the economy into the rules themselves. A product team should be able to say, of each thing it sells inside a game, whether it changes the game’s rules for those who pay.

2. Four kinds of play, and how they go bad

The idea in plain terms

Not all play is the same. Some is a contest, some is surrender to chance, some is pretending and some is the giddy pleasure of losing balance. Each kind satisfies something different, and each can turn into something harmful when it leaks into ordinary life.

The tradition

The French sociologist Roger Caillois, in Les Jeux et les hommes (1958, translated in 1961 as Man, Play and Games), criticised Huizinga for focusing on competitive play and for excluding games of chance, which his “no material interest” clause ruled out. Caillois sorted games into four kinds: agon, competition decided by skill or merit; alea, chance, in which the player gives up control to fate; mimicry, make-believe and role-play; and ilinx, vertigo. He set them along a line from paidia, spontaneous and unruly play, to ludus, play bound by deliberately difficult rules. And he described how each kind has a corrupt form when it escapes its frame: competition becomes violence and trickery, chance becomes superstition, make-believe becomes alienation, vertigo becomes intoxication.⁸²

Where it shows up

Caillois gives the clearest vocabulary for what changed in the 2010s. FIFA, Overwatch and Star Wars Battlefront II are games of agon, in which skill is supposed to decide the outcome. Paid loot boxes inserted alea into them, so that a player’s strength could depend on what they drew and how much they spent. The Belgian regulators treated the result as gambling; the Dutch Council of State treated it as a contest that happened to contain chance. Both were classifying a hybrid in Caillois’s terms. Battlefront II’s community saw it the same way: what enraged players was not chance but chance determining the outcome of competition.

The strongest objection

Caillois’s categories are blunt. Almost every game mixes them: a card game is chance and skill at once, and competition among players in a game of luck is still competition. Chance also has value in games; it lets weaker players sometimes beat stronger ones, keeps outcomes uncertain and creates moments of surprise. A theory that treats chance as a contaminant would condemn backgammon, poker and most board games, which people have played contentedly for centuries.

What it means today

Fair enough. Caillois’s useful contribution is the observation that chance is a different pleasure from skill, appealing to different motives and carrying different risks. The design question is which kind of play a game claims to offer and which it actually sells. Chance that varies the experience within the rules, such as a random map or a shuffled deck, is part of the game. Chance that is sold, and that decides who wins, turns a contest into a lottery while still presenting it as a contest. A team that cannot tell which of those it is building should not be surprised when regulators decide for it.

3. Unnecessary obstacles

The idea in plain terms

In ordinary life we take the easiest route to what we want. In a game we deliberately refuse the easiest route. A golfer could carry the ball to the hole and drop it in; the game is the decision not to. A game, on this view, is a voluntary acceptance of obstacles that exist only to be overcome.

The tradition

The Canadian philosopher Bernard Suits set out this definition in The Grasshopper: Games, Life and Utopia (1978), written as a dialogue in the manner of Plato, partly to refute Ludwig Wittgenstein’s claim in Philosophical Investigations (1953) that games share no common feature, only family resemblances. Suits’s short version is that “playing a game is the voluntary attempt to overcome unnecessary obstacles.” The full version has four parts: a goal that can be described before the game (getting the ball into the hole), rules that forbid the most efficient means to it, those rules being accepted, and the “lusory attitude” of accepting them simply because they make the game possible. Suits went further: in a utopia where all work was done for us, playing games would be what was left worth doing. Thomas Hurka accepted the definition but argued that game-playing is one good among many, not the highest.⁸³

The philosopher C. Thi Nguyen developed the idea in Games: Agency as Art (2020). Games, he argues, are the art form whose medium is agency: a designer specifies what the player wants, what they can do and what stands in their way, and the player temporarily takes on that way of acting. He distinguishes achievement play, where winning is what matters, from striving play, where the player takes up the goal of winning for the sake of the struggle. In a later paper, “Value Capture” (2024), he describes how simplified, usually quantified values supplied by institutions or technology, such as step counts, likes and rankings, can replace a person’s richer values without their noticing.⁸⁴

Where it shows up

Video games made Suits’s obstacles literally manufactured. Every jump a player fails and every boss that takes ten attempts was placed by a designer to be overcome. The trouble begins when the obstacle can be bought away. Battlefront II locked its most popular heroes behind dozens of hours of play and sold a way around the wait. In Suits’s terms, it sold the removal of the obstacle that made the activity a game. EA’s community reply, about “a sense of pride and accomplishment,” was a sincere Suitsian argument, that the struggle was the point, made in defence of a system that charged money for skipping the struggle. That is why it provoked such anger. Free-to-play games built on timers, energy limits and paid shortcuts have the same structure. Daily streaks, battle-pass tiers and leaderboards are, in Nguyen’s terms, metrics that can capture the player’s purpose, a case of When a Measure Becomes a Target: the player begins by playing for the struggle and ends by playing to keep a number from resetting.

The strongest objection

Defenders reply that players are adults who can judge for themselves. Some people have more money than time and would rather pay to skip a grind they find tedious; selling them that option respects their preferences. Nothing obliges anyone to buy. And games have always rewarded resources from outside the game: a child with a better racket or more hours of lessons plays better tennis.

What it means today

The answer depends on who designed the grind and why. If a game is built so that its obstacles are worth overcoming and players may also pay to skip some of them, the purchase is a convenience. If the obstacles were lengthened in order to sell their removal, the design works against the very thing a game is for. A practical test for a product team is to remove the shop and ask whether the obstacles would still be the same size. If they would shrink, the team is selling a remedy for a problem it created.

4. Flow and the machine zone

The idea in plain terms

People are often happiest not when relaxing but when absorbed in something difficult that they are able to do: a climber on a hard route, a surgeon mid-operation, a player on the last level. In that state they lose track of time and of themselves. Games are unusually good at producing it. But there is a second state that looks similar from outside, in which a person is absorbed not in mastering anything but in simply continuing, and it is the state that slot machines are designed to produce.

The tradition

The psychologist Mihaly Csikszentmihalyi described the first state in Beyond Boredom and Anxiety (1975) and Flow (1990). Flow arises when the challenge of an activity matches a person’s skill: too little challenge produces boredom, too much produces anxiety. The game designer Jenova Chen built a master’s thesis at the University of Southern California around the idea in 2006 and a game, flOw, that adjusted its difficulty to the player.⁸⁵ The challenge-skill balance is the part of Csikszentmihalyi’s model that has been hardest to confirm under careful testing, though the broader description of absorbed, enjoyable effort is widely accepted.

The anthropologist Natasha Dow Schüll described the second state in Addiction by Design (2012), based on years of fieldwork among machine gamblers in Las Vegas. Heavy players described a “machine zone” in which time, money and worry receded. Schüll argued that this state was produced by the interaction of the player with machines whose software, ergonomics and settings had been tuned by the industry to maximise “time on device.”⁸⁶

Where it shows up

The two states mark the line running through this history. Tennis for Two, Spacewar! and the early arcade games were built for mastery and short sessions; arcade machines took a coin and gave a challenge. The free-to-play business added an economic reason to prefer continuation over mastery, since revenue rose with time spent and with the number of sessions. Mechanics such as daily rewards that reset, energy that refills over hours and endless progression systems suit the machine zone better than flow. The Oxford studies point in the same direction from the player’s side: in the 2022 telemetry study, people who played because they wanted to reported higher wellbeing, and people who played because they felt compelled to reported lower.⁶⁶

The strongest objection

The distinction may be clearer on paper than in practice. A player deep in a long role-playing game, a puzzle player solving one more level and a gambler at a machine can all report losing track of time, and the same person can move between the states in an evening. Designers who aim for flow also aim for players to come back. And Schüll’s account is of slot machines, which pay out money; transferring it to games that do not pay out may overstate the resemblance.

What it means today

From a single session, then, the two states cannot be told apart. They can be told apart from a design’s purpose and its metrics. A game tuned to keep challenge matched to skill will let a player finish, master or put it down satisfied. A game tuned to maximise time on device has a reason to make stopping uncomfortable. Product teams can ask which they are measuring: whether players report that the time was well spent, or only how much time they spent.

5. Needs met on a screen

The idea in plain terms

People are motivated not only by rewards and punishments but by deeper needs: to feel capable, to feel that their actions are their own, and to feel connected to others. Activities that meet those needs feel intrinsically worth doing. Games meet all three with unusual efficiency, which explains both their appeal and some of their risks.

The tradition

Self-determination theory, developed from the 1970s by the psychologists Edward Deci and Richard Ryan, holds that human motivation depends on three basic needs: competence, autonomy and relatedness. When an activity satisfies them, people do it for its own sake and feel better for it; when they act under pressure or for external rewards alone, motivation becomes fragile and wellbeing suffers. In 2006 Ryan, Scott Rigby and Andrew Przybylski published “The Motivational Pull of Video Games” in Motivation and Emotion. Across four studies, from twenty minutes of Super Mario 64 in a laboratory to a survey of an online multiplayer community, they found that the extent to which a game made players feel competent and free predicted how much they enjoyed it and whether they wanted to keep playing, and that in online games feeling connected to others mattered independently.⁸⁷ Rigby worked for a company that sold a commercial version of the questionnaire used, a point worth noting when weighing the work; later studies using publishers’ records of actual play found that the reasons people played mattered in a similar way.

Where it shows up

The theory explains much of the history. Nintendo’s games, built around controls that are easy to learn and goals that are clearly within reach, are engines of competence. Open-world games offer autonomy. Xbox Live, World of Warcraft, Fortnite and Discord made relatedness the core of play for a generation. The theory also explains Iwata’s disappointment with the Wii: it gave new players an easy sense of competence, but many did not find enough in it to keep feeling it. And it explains the most important finding of the Oxford telemetry studies, that the hours people played predicted little about their wellbeing, while the reasons they played predicted a good deal: intrinsic motivation went with feeling better, the sense of having to play with feeling worse.⁶⁶

The strongest objection

Against this, satisfying needs in a game may displace satisfying them elsewhere. A person who feels competent only in a game, free only in a game and connected only to an online guild may be meeting the needs in a form that does not carry over to work, family or friendship, and the very efficiency of games at meeting them could make the rest of life feel thin by comparison. A further objection is that the theory is so broad it can explain almost any behaviour after the fact.

What it means today

The displacement concern is real for some people, and it is the strongest reason to take gaming disorder seriously as a clinical category for a minority. But the evidence suggests that for most players games add to need satisfaction rather than replacing it, and that the warning sign is not how long someone plays but whether they still play because they want to. Designers who understand the theory can see the difference between building competence, autonomy and connection, and building their counterfeits: the feeling of progress without skill, the feeling of choice inside a funnel, the feeling of obligation to a team that the game uses to pull players back.

6. Chance, control and the near miss

The idea in plain terms

People are bad at telling luck from skill. When a game of chance contains features that look like skill, such as choosing, practising or competing, people believe they can influence the outcome. And when they almost win, they feel they are getting closer, even when each attempt is independent of the last. Both errors make games of chance more compelling than their odds justify.

The tradition

The psychologist Ellen Langer, in “The Illusion of Control” (1975), reported experiments in which people behaved as though cues that normally signal skill, such as choice, familiarity and competition, gave them influence over outcomes that were purely a matter of chance. Direct replications of her original studies are scarce; a related study of hers was replicated with preregistration in 2023 and the effect appeared, smaller than first reported.⁸⁸ R. L. Reid, in “The Psychology of the Near Miss” (1986), noted that near misses are widely believed to encourage future play “even in games of chance where the probability of winning remains constant,” and that some commercial gambling systems were “contrived to ensure a higher frequency of near misses than would be expected by chance alone.”⁸⁹ Luke Clark and colleagues reported in Neuron in 2009 that near misses on a simulated slot machine, though rated less pleasant than outright losses, increased the desire to play and activated brain regions that respond to wins, but only when the player had chosen the outcome: “near-misses invigorate gambling through the anomalous recruitment of reward circuitry.”⁹⁰ A review and new experiments by J. M. Pisklak and colleagues in 2020 failed to find that near misses make people keep playing longer, so the effect on persistence is disputed even if the effect on feeling is not.⁹¹

Where it shows up

Loot boxes combine the ingredients. The player chooses when to open them, often watches an animation that builds suspense, frequently sees rare items flash past before a common one lands, and opens them within a game that rewards skill in other respects. Langer’s work explains why chance inside a game of skill may feel controllable; Reid’s explains why near misses would be worth engineering; Clark’s suggests the feelings are real at the level of the brain. Zendle and Cairns’s finding, repeated in two meta-analyses, that loot box spending correlates with problem gambling at about 0.26, fits this picture without proving it.⁶⁰ The Japanese ban on “complete gacha,” in which players had to draw a full set of random items to win, targeted a mechanism that exploits these biases directly: each draw feels like progress towards the set.

The strongest objection

Critics counter that the evidence for causation is weak, and the replication record of the underlying findings is mixed. The illusion of control is smaller than first reported; the near-miss effect on persistence has failed to replicate; and the correlation between loot boxes and problem gambling could run the other way, with people already prone to gambling drawn to loot boxes. On this evidence, banning loot boxes would be regulating on the basis of plausible stories. The British government’s decision in 2022 rested on exactly this gap.

What it means today

That is a real gap in the evidence, and it should make anyone careful about claiming that loot boxes cause gambling problems. It does not show that their design is neutral. A company that times animations, shows near misses and ties random purchases to competitive advantage is applying techniques whose purpose in casinos is well understood. The responsible position for a designer does not wait for proof of harm: if a mechanism would be unremarkable in a casino, it deserves a casino’s safeguards, including disclosed odds, spending limits and no sales to children, whatever the eventual verdict of the research.

7. Panic and evidence

The idea in plain terms

Each new medium attracts a fear that it corrupts the young. Sometimes the fear is justified; often it is out of proportion to the evidence and says more about the anxieties of the time than about the medium. Telling the two apart requires a standard of evidence that holds whether the conclusion is comfortable or not.

The tradition

The sociologist Stanley Cohen, in Folk Devils and Moral Panics (1972), studied the British reaction to clashes between youth groups, the Mods and Rockers, at seaside towns in the 1960s. He described a moral panic as a process in which a condition or a group of people comes to be defined as a threat to a society’s values and interests, amplified by the press, experts and authorities out of proportion to the actual threat.⁹² Later scholars specified criteria including disproportion. The closest precedent for games is the American comics scare of the 1950s, when the psychiatrist Fredric Wertham’s Seduction of the Innocent (1954) blamed comics for juvenile delinquency, a Senate subcommittee held hearings, and the industry adopted a self-censoring code. In 2012 the information scientist Carol Tilley compared Wertham’s published cases with his clinical notes and documented that he had manipulated, overstated and in places fabricated the evidence.⁹³

Where it shows up

The 1993 Senate hearings followed the pattern closely: graphic examples, worried witnesses, a threat to an industry and a self-regulatory code in the form of the ESRB. California’s 2005 law, and the research cited to support it, faced the test Cohen’s framework implies. The Supreme Court in 2011 found the studies showed “at best some correlation” with “minuscule real-world effects.” The American Psychological Association, which had treated the link to aggression as established in 2015, said in 2020 that there was “insufficient scientific evidence” of a causal link to violent behaviour and warned against blaming mass shootings on games.³⁰ The later Oxford telemetry studies, built on records of actual play rather than self-report, found little effect of playing time on wellbeing either way.

The strongest objection

One answer is that “moral panic” can become a way of dismissing every concern about a popular product. Some worries about new products prove correct, and calling a concern a panic can delay recognising a real harm. The APA still finds a small, reliable association between violent games and aggressive behaviour, and a small effect spread across hundreds of millions of players is not trivial. The language of panic can protect an industry as effectively as the language of danger can attack it.

What it means today

Both errors are real, so the useful lesson is about method rather than conclusion. The violent-games debate was settled not by argument but by better data: larger samples, preregistered studies and behaviour measured directly. The same standard now applies to loot boxes, gaming disorder and screen time, and it cuts in both directions. Product leaders should welcome independent research with access to their data, as several publishers did for the Oxford studies, because it is the only thing that can distinguish a real harm from a panic, and a company that refuses it will be judged by the panic.

8. Protecting players from themselves

The idea in plain terms

Adults are generally entitled to spend their time and money as they choose, even badly. Children are not, because they cannot yet weigh consequences fully. Between those positions lies the hard question for games: when, if ever, should the state or a company limit what an adult may do to themselves, and how should children be protected without making every adult prove their age?

The tradition

John Stuart Mill, in On Liberty (1859), set out the harm principle: “the only purpose for which power can be rightfully exercised over any member of a civilised community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant.” “Over himself, over his own body and mind, the individual is sovereign.” Mill excluded children explicitly: the doctrine applies “only to human beings in the maturity of their faculties,” and those still in need of care “must be protected against their own actions as well as against external injury.”⁹⁴ The legal philosopher Joel Feinberg, in Harm to Self (1986), distinguished hard paternalism, which overrides a person’s voluntary choice for their own good, from soft paternalism, which intervenes only to establish whether a choice is truly voluntary or when it is not. He rejected the first and argued that the second is not really paternalism at all but a refinement of respecting autonomy.⁹⁵

Where it shows up

The regulatory responses to games map onto these positions. China’s limit of three hours a week for minors is protection of children taken to the edge of state control over family life. The rules on loot boxes in Brazil, Australia and the FTC’s order against the maker of Genshin Impact all focus on children, as Mill would. Odds disclosure, required by Apple from 2017 and Google from 2019, is soft paternalism in Feinberg’s sense and a piece of The Architecture of Choice: it does not stop anyone buying, but tries to ensure that the choice is informed. The FTC’s case against Epic was about choices that were not voluntary at all: purchases the FTC said were made by children without their parents’ knowledge and by players misled by the interface. Belgium’s ban on paid loot boxes for everyone, adults included, is the closest to hard paternalism, and its weak enforcement shows the cost of a rule that much of the market does not accept.

The strongest objection

Perhaps the line between adult and child, and between voluntary and not, does not hold in games. Children play on adult accounts; age checks are easily evaded or require collecting the personal data that privacy law tries to protect; and the biases that make loot boxes compelling affect adults as well. If a mechanism exploits predictable flaws in human judgement, a choice made under its influence is not fully voluntary even for an adult, and soft paternalism collapses into hard. Conversely, a rule strict enough to protect children reliably may require surveillance of everyone.

What it means today

There is no clean resolution, but there is an order of priorities that most of the history supports. First, protect children, and accept the friction that requires. Second, make adult choices informed: show odds, show real prices rather than prices in invented currencies, and make spending visible. Third, hold the design itself to account where it is built to exploit, rather than relying only on warnings. For companies the lesson of the Epic and Battlefront II episodes is that waiting for the regulator is expensive.

Bringing the threads together

Taken together, the eight ideas circle a single tension. Huizinga and Suits say what a game is: voluntarily accepted rules whose obstacles exist to be overcome. Caillois and Nguyen show how that space changes from inside, when chance enters a contest or a score captures a player’s purpose. Csikszentmihalyi, Deci and Ryan explain why games satisfy; Schüll, Langer, Reid and Clark explain how the same satisfactions can be engineered toward continuation and spending. Cohen warns against judging all this by fear, Mill and Feinberg against judging it by nothing.

What the theory cannot supply is the decisions. Atari’s open slot, Nintendo’s lock, the ESRB, the CD, Steam, the Wii, Apple’s free apps, the loot box and the leveraged buyout each moved the rules of play toward or away from the player’s own purposes. What decided the direction, again and again, was how the game was paid for.

The rules we choose

A game is one of the few things people do in which the rules are accepted freely and serve no end beyond themselves. That freedom is what makes play valuable, and it is fragile in a particular way: a person who has agreed to follow rules for their own sake is easy to steer by changing the rules. Nobody notices a new obstacle in a game as an imposition. They notice it as part of the game.

Several working principles follow from the history.

First, decide what the game is for before deciding how it earns money, and check that the second does not quietly rewrite the first. The test is whether the game’s obstacles would be the same if nothing were for sale.

Second, keep chance and contest separate when money is involved. Random rewards earned by play are part of a game; random advantages sold for money turn a test of skill into a lottery dressed as one, and regulators on four continents have learned to see through the dress.

Third, measure what players value, not only what they do. Time played and revenue per user are easy to count and easy to raise in ways that make the game worse. Whether players feel the time was well spent is harder to measure, and it is the number that predicts whether the game deserves to last.

Fourth, treat children as Mill did, as people who must be protected against their own actions as well as others’, and accept the cost. The two largest American penalties in this history, against Epic and against the maker of Genshin Impact, both turned on what children could do or buy without their parents’ consent.

Fifth, when a decision is controversial, open the data. The violent-games argument ran for two decades on weak evidence and ended when researchers could measure real play. Companies that share their data with independent researchers will be judged on evidence; those that do not will be judged on suspicion.

The cartridges in Alamogordo were buried because a company had flooded its own market with games nobody wanted. Forty years later the industry’s problems are different, almost the opposite: games so good at holding attention and so carefully priced that governments ask whether they are gambling, and players ask whether they are still playing or being played. The rules people choose to accept are the oldest thing in games. Who writes them, and for whose benefit, is the question every decision in this history has answered.

Selected bibliography

• Mill, John Stuart. On Liberty. London: John W. Parker and Son, 1859.
• Huizinga, Johan. Homo Ludens: A Study of the Play-Element in Culture. Dutch original 1938; English edition London: Routledge & Kegan Paul, 1949.
• Wittgenstein, Ludwig. Philosophical Investigations. Oxford: Blackwell, 1953.
• Caillois, Roger. Les Jeux et les hommes. Paris: Gallimard, 1958. English: Man, Play and Games, translated by Meyer Barash. New York: Free Press of Glencoe, 1961.
• Cohen, Stanley. Folk Devils and Moral Panics: The Creation of the Mods and Rockers. London: MacGibbon & Kee, 1972.
• Csikszentmihalyi, Mihaly. Beyond Boredom and Anxiety. 1975; and Flow: The Psychology of Optimal Experience. 1990.
• Langer, Ellen J. “The Illusion of Control.” Journal of Personality and Social Psychology 32, no. 2 (1975): 311–328.
• Suits, Bernard. The Grasshopper: Games, Life and Utopia. 1978; 3rd edition, with an introduction by Thomas Hurka, Peterborough, Ontario: Broadview Press, 2014.
• Reid, R. L. “The Psychology of the Near Miss.” Journal of Gambling Behavior 2, no. 1 (1986): 32–39.
• Feinberg, Joel. Harm to Self. Vol. 3 of The Moral Limits of the Criminal Law. New York: Oxford University Press, 1986.
• Atari Games Corp. v. Nintendo of America Inc., 975 F.2d 832 (Fed. Cir. 1992).
• Sega Enterprises Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992).
• Salen, Katie, and Eric Zimmerman. Rules of Play: Game Design Fundamentals. Cambridge, MA: MIT Press, 2003.
• Ryan, Richard M., C. S. Rigby, and Andrew K. Przybylski. “The Motivational Pull of Video Games: A Self-Determination Theory Approach.” Motivation and Emotion 30, no. 4 (2006): 347–363.
• Clark, L., A. J. Lawrence, F. Astley-Jones, and N. Gray. “Gambling Near-Misses Enhance Motivation to Gamble and Recruit Win-Related Brain Circuitry.” Neuron 61, no. 3 (2009): 481–490.
• Consalvo, Mia. “There Is No Magic Circle.” Games and Culture 4, no. 4 (2009): 408–417.
• Anderson, Craig A., et al. “Violent Video Game Effects on Aggression, Empathy, and Prosocial Behavior in Eastern and Western Countries: A Meta-Analytic Review.” Psychological Bulletin 136, no. 2 (2010): 151–173; and Ferguson, Christopher J., and John Kilburn, comment, ibid., 174–178.
• Brown v. Entertainment Merchants Association, 564 U.S. 786 (2011).
• Schüll, Natasha Dow. Addiction by Design: Machine Gambling in Las Vegas. Princeton: Princeton University Press, 2012.
• Zendle, David, and Paul Cairns. “Video Game Loot Boxes Are Linked to Problem Gambling: Results of a Large-Scale Survey.” PLOS ONE 13, no. 11 (2018): e0206767; and “Loot Boxes Are Again Linked to Problem Gambling: Results of a Replication Study.” PLOS ONE 14, no. 3 (2019): e0213194.
• Orben, Amy, and Andrew K. Przybylski. “The Association between Adolescent Well-Being and Digital Technology Use.” Nature Human Behaviour 3 (2019): 173–182.
• Nguyen, C. Thi. Games: Agency as Art. New York: Oxford University Press, 2020; and “Value Capture.” Journal of Ethics and Social Philosophy 27, no. 3 (2024): 469–504.
• Vuorre, M., N. Johannes, K. Magnusson, and A. K. Przybylski. “Time Spent Playing Video Games Is Unlikely to Impact Well-Being.” Royal Society Open Science 9 (2022): 220411.

Research limitations

Much of the early history rests on later histories (Donovan’s Replay, Smith’s They Create Worlds, Goldberg and Vendel’s history of Atari) as summarised in reference works, and on participants’ recollections rather than company records. The often-quoted fall in US home video game sales from about $3.2 billion to about $100 million is poorly sourced and is given as an order of magnitude. Cartridge figures for Pac-Man, E.T. and the Alamogordo burial (variously 700,000 to 800,000 buried) rest on later reporting. Sharpe’s “called shot” in 1976 is a popular account; the trade press confirms only the demonstration and the vote. Accounts differ on the NES test launch, on why the Sony–Nintendo CD project collapsed and on Kassar’s departure. Reported minimum-order terms in Nintendo’s licences are not confirmed by the court opinions. Nintendo’s and Sony’s unit figures are shipments. The status of the antitrust case against Valve rests on limited reporting. Newzoo’s revenue estimates are revised repeatedly (its 2024 total has been reported as $182.7 billion, $184.3 billion and $187.7 billion), and layoff counts differ between trackers, especially for 2025. Free-to-play spending figures come from one analytics firm’s samples. Research on loot boxes shows a consistent association with problem gambling but not causation, and the near-miss and illusion-of-control findings have mixed replication records. The 2026 survey of developers on generative AI was reported second-hand. The EU’s Digital Fairness Act had not been published, and Grand Theft Auto VI not released, at the time of writing. The connections to Huizinga, Salen, Zimmerman, Castronova, Consalvo, Caillois, Suits, Wittgenstein, Hurka, Nguyen, Csikszentmihalyi, Chen, Schüll, Deci, Ryan, Rigby, Przybylski, Langer, Reid, Clark, Pisklak, Cohen, Wertham, Tilley, Mill and Feinberg are the author’s interpretation; their works are paraphrased apart from short, attributed phrases. The decisions in the decisions box are judged by the author, separating what could be known at the time (process) from what happened (call).

Footnotes

1. “Atari video game burial,” Wikipedia, summarising the Alamogordo Daily News (27 September 1983) and the New York Times (28 September 1983). ↩

2. “Atari, Inc.,” Wikipedia, citing Goldberg and Vendel, Atari Inc.: Business Is Fun (2012); “Video game crash of 1983,” Wikipedia. ↩

3. Deseret News / AP, 26 April 2014; CNN Money, 1 September 2015. ↩

4. Newzoo’s 2026 forecast as reported in “Newzoo 2026 global games market report,” Wccftech, and Game Wisdom. ↩ ↩²

5. Nintendo, “Dedicated Video Game Sales Units,” to 30 September 2026; on the Switch 2 launch, VGC and Nintendo of America, 2025. ↩ ↩² ↩³

6. Microsoft, “Microsoft to acquire Activision Blizzard,” 18 January 2022; Electronic Arts, “EA Announces Completion of Acquisition by PIF, Silver Lake, and Affinity Partners,” SEC Form 8-K exhibit, 2026. ↩

7. Christopher Klein, “That time America outlawed pinball,” History.com; “The War on Pinball,” Atlas Obscura. ↩

8. Play Meter, May 1976, p. 15, via the Arcade Museum eLibrary; Music & Amusement Association, commemoration; “Roger Sharpe,” Wikipedia. ↩

9. Brookhaven National Laboratory, “The First Video Game?”. ↩

10. “Spacewar!,” Wikipedia, drawing on J. M. Graetz, “The origin of Spacewar,” Creative Computing, August 1981; Computer History Museum, “Steve Russell”. ↩

11. “Magnavox Odyssey,” Wikipedia, drawing on Smith, Donovan and Ralph Baer, Videogames: In the Beginning (2005). ↩

12. “Magnavox Odyssey,” Wikipedia; Smithsonian Institution, Odyssey unit in the collection. ↩

13. “Atari, Inc.,” Wikipedia. ↩

14. “Magnavox Odyssey,” Wikipedia, citing Donovan and Goldberg and Vendel on the Atari licence, and NPR, 8 December 2014, on total earnings. ↩

15. “Atari, Inc.,” Wikipedia. ↩

16. “Activision,” Wikipedia. ↩

17. “Tooth Protectors,” Wikipedia; “Chase the Chuck Wagon,” Bad Game Hall of Fame. ↩

18. “Atari video game burial,” Wikipedia, citing NPR (2017), Next Generation (1998), Snopes and IGN. ↩

19. Christian Science Monitor, 10 December 1982; “The Game Turns Serious at Atari,” New York Times, 19 December 1982, transcribed on AtariAge. ↩

20. David Crane, interviewed by Arcade Attack, 9 May 2016, as quoted in “Video game crash of 1983,” Wikipedia. ↩

21. “Video game crash of 1983,” Wikipedia, on the $3.2 billion and $100 million figures and the June 1983 home computer price war. ↩

22. “Nintendo Entertainment System,” Wikipedia. ↩

23. Atari Games Corp. v. Nintendo of America Inc., 975 F.2d 832 (Fed. Cir. 1992). ↩ ↩²

24. National Association of Attorneys General, “New York v. Nintendo of America, Inc., 775 F. Supp. 671 (S.D.N.Y. 1991)”; “Nintendo price-fixing case settled,” Seattle Times, 11 April 1991. ↩

25. Sega Enterprises Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992). ↩

26. “1993–94 United States Senate hearings on video games,” Wikipedia; Sega Retro. ↩

27. Entertainment Software Rating Board, “History”; on PEGI, “PEGI,” Wikipedia. ↩

28. Brown v. Entertainment Merchants Association, 564 U.S. 786 (2011): syllabus, opinion of the Court, Alito concurrence, Thomas dissent and Breyer dissent, Legal Information Institute. ↩

29. Anderson et al., Psychological Bulletin 136, no. 2 (2010): 151–173, and Ferguson and Kilburn, ibid., 174–178, bibliographic record; on the 2013 letter, Stetson University. ↩

30. American Psychological Association, “APA reaffirms position on violent video games and violent behavior,” March 2020; on objections within the APA, Division 46 Amplifier. ↩ ↩²

31. “Super NES CD-ROM,” Wikipedia; “Sony Interactive Entertainment,” Wikipedia. ↩

32. “Hironobu Sakaguchi on why Final Fantasy VII moved to PlayStation,” Automaton, 15 February 2024. ↩

33. “Flashback: Reliving the drama of E3 1995,” Time Extension; “Sony and Sega surprises light games market,” Supermarket News, 1995. ↩

34. Sony Interactive Entertainment, “Business Data and Sales”. ↩

35. “Sega to halt production of Dreamcast,” The Daily Record, 30 January 2001; “Game over for Dreamcast,” Deseret News, 31 January 2001; Sega, corporate history. ↩

36. “A quick look back at the creation of Doom,” Neowin; Doomworld, “Page of Doom”. ↩

37. Microsoft, “Xbox Live arrives in stores,” 15 November 2002; on the criticism of broadband-only, “Xbox Live is now 20 years old,” VGC. ↩

38. “Half-Life 2 at 20,” Inverse; “Half-Life 2 pushed Steam on the gaming masses,” Ars Technica, mirrored. ↩

39. “Valve’s new revenue sharing favours big-budget games,” PC Gamer; GeekWire, 2018. ↩

40. In re Valve Antitrust Litigation, No. 2:21-cv-00563 (W.D. Wash.), docket and class certification order; on its status in 2026, Tech Insider. ↩

41. Nintendo, Corporate Management Policy Briefing, Q&A, 7 June 2006. ↩

42. Nintendo, Financial Results Briefing, Q&A, 27 April 2012. ↩

43. Engadget, 29 January 2014; Pocket-lint. ↩

44. “QuizQuiz,” Wikipedia; “Nexon,” Wikipedia. ↩

45. Zynga, Form S-1, 2011; “Zynga’s new S-1 reveals slowing growth,” CNBC, 4 November 2011. ↩

46. Apple, “In App Purchase Now Available for Free Apps,” 15 October 2009. ↩

47. “Supercell’s 2013 revenue was $892 million,” PocketGamer.biz; “Few users actually pay for Candy Crush,” Statista; The Register, 18 February 2014. ↩

48. Epic Games, “Fortnite Battle Royale goes free,” September 2017. ↩

49. “0.19% of players account for 48% of free-to-play game revenue,” Adweek, 2016; Swrve, 2015 report release, 9 April 2015. ↩

50. Newzoo’s 2025 figures as summarised by PocketGamer.biz and Game Dev Reports. ↩

51. “Loot box,” Wikipedia; “Japan consumer agency says complete gacha illegal,” Bloomberg, 18 May 2012; “Why kompu gacha was banned,” Game Developer. ↩

52. GameRevolution; TheGamer on the Guinness count; Inverse; TechCrunch, 16 November 2017; Game Informer, 16 March 2018. ↩

53. “Belgium defines video game loot boxes as illegal gambling,” Public Gaming, 25 April 2018; Leon Y. Xiao, “Breaking Ban,” Collabra: Psychology (2023). ↩

54. Kansspelautoriteit, “Study into loot boxes,” 2018; Raad van State, ruling 202005769/1/A3, 9 March 2022; CMS. ↩

55. UK government, “Government response to the call for evidence on loot boxes,” 18 July 2022; SCL on the 2019 report; Game Developer on the principles; 5Rights Foundation on compliance. ↩

56. Australian Classification, “New mandatory minimum classifications for gambling content in games”. ↩

57. “Brazil’s ECA Digital,” Mattos Filho; Pocket Gamer.biz. ↩

58. European Parliament, resolution 2022/2014(INI), 18 January 2023; CPC Network, Key Principles on In-game Virtual Currencies, March 2025. ↩

59. Federal Trade Commission, Epic Games settlements, December 2022; FTC business blog on HoYoverse, January 2025. ↩

60. Zendle and Cairns, PLOS ONE (2018) and PLOS ONE (2019); Garea et al., International Gambling Studies (2021); Spicer et al., New Media & Society (2022). ↩ ↩²

61. World Health Organization, “Addictive behaviours: Gaming disorder” and “ICD-11 2022 release,” 11 February 2022. ↩

62. Aarseth et al., Journal of Behavioral Addictions 6, no. 3 (2017), repository record. ↩

63. Stevens et al. (2021), PubMed, and corrigendum. ↩

64. Xinhua, 30 August 2021; Reuters via Business Standard, 23 January 2024. ↩

65. Oxford Internet Institute, “Technology use explains at most 0.4% of adolescent wellbeing”; Orben and Przybylski (2019), RePEc. ↩

66. Vuorre et al., “Time spent playing video games is unlikely to impact well-being,” Royal Society Open Science (2022). ↩ ↩² ↩³

67. Competition and Markets Authority, “Microsoft / Activision deal prevented,” 26 April 2023; Advanced Television, 13 October 2023; A&O Shearman on the Ninth Circuit. ↩

68. “Global video game market to decline in 2022,” Axios, 15 November 2022; Game World Observer, 26 June 2025. ↩

69. “Gaming layoffs slowed from 15,631 in 2024 to 9,053 in 2025,” GamesBeat, 3 January 2026; TweakTown, October 2026. ↩

70. NBC News, January 2024; Game Informer, 2 July 2025; CNBC, 6 July 2026. ↩

71. “Embracer Group to split into three companies,” Game Informer, 22 April 2024. ↩

72. Nintendo Life, September 2023; Unity, “Unity is canceling the Runtime Fee,” 12 September 2024. ↩

73. “The video game voice actors’ strike officially ends,” PC Gamer; TheWrap, 12 June 2025. ↩

74. Game Developers Conference survey via Business Wire, 21 January 2025, and WinBuzzer, 2026; VGC on Steam. ↩

75. Electronic Arts, announcement exhibit, September 2025, and completion exhibit, 2026; on the shareholder vote, KitGuru. ↩

76. “Grand Theft Auto VI is delayed to November 19, 2026,” Engadget; AP via KSAT, 3 May 2025; “Take-Two confirms November debut,” Bloomberg, 21 May 2026. ↩

77. Sony Group, Form 6-K, November 2026; Gamereactor. ↩

78. Johan Huizinga, Homo Ludens (Roy Publishers printing, 1950), Internet Archive. ↩

79. Eric Zimmerman, “Jerked Around by the Magic Circle: Clearing the Air Ten Years Later” (2012). ↩

80. Edward Castronova, CESifo Working Paper 618, RePEc; Synthetic Worlds (2005), library record. ↩

81. Mia Consalvo, “There Is No Magic Circle,” Games and Culture 4, no. 4 (2009), record. ↩

82. “Man, Play and Games,” Wikipedia; MIT OpenCourseWare, CMS.300 session notes. ↩

83. Broadview Press, The Grasshopper, 3rd edition; “Lusory attitude,” Wikipedia. ↩

84. Review of Games: Agency as Art, Philosophy; C. Thi Nguyen, “Value Capture,” Journal of Ethics and Social Philosophy 27, no. 3 (2024). ↩

85. “Flow in psychology: a review,” Frontiers in Psychology (2020); Jenova Chen, Flow in Games, thesis abstract. ↩

86. Natasha Dow Schüll, Addiction by Design (Princeton University Press, 2012), publisher’s page. ↩

87. Ryan, Rigby and Przybylski, “The Motivational Pull of Video Games”, Motivation and Emotion 30 (2006). ↩

88. Langer (1975), cited in Springer reference list; on the 2023 replication of Langer and Roth, RePEc. ↩

89. R. L. Reid, “The Psychology of the Near Miss”, Journal of Gambling Behavior 2 (1986). ↩

90. Clark et al., “Gambling Near-Misses Enhance Motivation to Gamble,” Neuron 61 (2009). ↩

91. Pisklak, Yong and Spetch, “The Near-Miss Effect in Slot Machines,” Journal of Gambling Studies 36 (2020). ↩

92. “Moral panic,” Wikipedia; Maynooth University paper, 2005. ↩

93. “Researcher proves Wertham fabricated evidence against comics,” Comic Book Legal Defense Fund, 2013. ↩

94. John Stuart Mill, On Liberty (1859), chapter 1, Project Gutenberg. ↩

95. Joel Feinberg, Harm to Self (1986), Oxford Academic; review by Andrew Bailey. ↩