Pay to Win, Episode 1

Aug. 13, 2021

This Video Contains Spoilers for Slot Machines


Diablo: Immortal is a Diablo game for phones which was later ported to PC. Diablo is a notable series for lots of reasons, but its primary gameplay loop is a cycle where you kill enemies, collect their loot, and then upgrade your gear to kill harder enemies and collect betterer loot. It's incredibly addictive.

Immortal is the collision of Diablo's addictive gameplay with the state of the art in free-to-play game design; it features premium currencies, shops, and gacha mechanics and it's making the developer-slash-publisher Blizzard a million dollars a day [1]. So it's fair to say that Diablo Immortal and games like it are here to stay. The game was subject to massive backlash, but it died down within a few days of the PC release.

I want to know how we got here, and when I say that I don't mean pointing and laughing at the horse armour DLC. I'll be doing that too, but what we see in Diablo Immortal is just the newest face of a technology that's much older than video games. So I'm going to review Diablo Immortal, but it's going to be a bit of a journey. This is the first in a series of videos about making money from games, and how monetization affects players and game design.

This video was a lot of fun to make despite the sometimes heavy subject matter, so I hope you enjoy it. Later parts are going to get pretty crazy, but this story has humble beginnings.

Part 1: The Evolution of Gambling

Sports Betting (All the pretty horses)

{== betting is a form of gambling, almost universally lumped in with casinos and lotteries. See [2]. ==}

{++ SHOT: zoom to Churchill Downs, name and state on screen, then talk. ++}

Before video games, there were horses. And with horses came racing and betting. The cash flows directly tied to the race are pretty simple, but horse racing is different from other sports because the people who win the money are two steps removed from the action; the horse does all the work, and the jockey gives all the directions, but the owner gets paid. A jockey who wins the prestigious Kentucky Derby will get paid around $200,000 before taxes and fees, while the owner will earn around 1-and-a-half million.

That makes the sport a perfect fit for betting, another way to profit from a game without being good at it. Horse betting is probably around 400 years old, and it was a feature of some of the first organized horse races in England. Apparently King James liked to race horses [3]. Horse races and gambling are inextricable, to the extent that betting may well be more interesting to attendees than the race itself.

I think most people know more or less how betting works; the bettor believes that a particular horse is going to win. They give their prediction to a bookie along with some amount of money. If their prediction comes true, they get more money than they started with. If it's wrong, they get nothing. The amount you win from a bet is determined by the odds. There are a whole bunch of different ways to express betting odds but they all tell you how much a winning bet will pay back.

In horse betting, odds are usually calculated based on the bets themselves, which is counterintuitive. Betting odds are a way of encoding risk; if a lot of people bet on the same horse, that horse is considered a safe bet. More risk means more reward, so betting on a safe horse means a smaller payout. Odds are calculated as a fraction of the total amount of money that people bet.

This system is called Parimutuel--I'm just gonna say para-mutual--which is French for mutual betting. It was invented in 1867 [4], and it is used for a few reasons. Odds are not an objective measure of anything, so having the bookie set the odds doesn't make sense; they would just be guessing. Parimutuel also protects the bookie from having to pay more money than they have; payouts are always a fraction of the bookie's total income, a piece of the pie, so to speak.

To understand why this is useful, here's an exaggerated example. Let's say the bookie is setting the odds themselves, and they offer 7:1 odds on a horse called Lady Luck. If I bet a dollar and she wins, I get seven dollars. If everyone at the race bets on Lady Luck and she wins, the bookie now has to pay seven times the amount of money they earned, and that could be a problem. Parimutuel betting makes sure that doesn't happen. The parimutuel odds for Lady Luck would be 1:1. {== text: for simplicity this assumes the bookie doesn't take a cut. ==}

In a parimutuel system, every bet changes the odds, so parimutuel odds were a huge pain to calculate in the early days. Horse betting makes a lot of money, so people invested in the technology surrounding it. Special calculating machines called automatic totalisators were introduced in 1913 [4] by a company called Autotote, which now goes by the name Light and Wonder [5]. {== this is not a great way to characterize an acquisition but it's for foreshadowing purposes. ==}

In 1970, the writer Hunter Thompson reported on the Kentucky Derby [6]. What he captured wasn't the race itself, but the ugly spectacle of the thing. He describes the crowd as a drunken, puking, stomping organism whose parts are only distinguished by how expensive their shoes and hats are. Thompson characterizes the face of the Derby as the mask of the Whisky Gentry; the inbred remains of American generational wealth. Horse racing started as, and continues to be an expensive hobby for rich people. Who else could come up with such stupid names for horses? Names like...

Racehorses Trained in the UK Racehorses Trained in the US

Al Capone II
Risk of Thunder
Doctor Syntax
Gods Solution
On the Fringe
Shotgun Kowboy
Tie the Knot
Star Guitar
Beef or Salmon
Boobe Grand
Cottage Rake
Game on Dude
Brody's Cause
Cosmic Bomb
Countless Diana
Declan's Moon
Discreetly Mine
Dream Empress
Eskimo Kisses
Fit to Fight
Forever Together
Dark Ronald
Exotic Dancer
Fallen For You
Gay Donald

By the way, there are thousands of these on Wikipedia and every single one of them has its own page.

Anyway, what Thompson found through his journalistic method of getting really drunk is that the fervor of the Kentucky Derby swallowed him up. Thompson is trying to find this archetypal Whisky Gentry face throughout his whole piece, and he only finds it at the end, staring back at him the morning after the race.

The value of the Kentucky Derby is intimately tied to the spectacle of it. You can only make money on parimutuel betting if a lot of people are doing it: remember, a winning bettor gets a piece of the pie. The size of that pie depends on how much people bet. Bettors are incentivized to whip themselves and others up into a frenzy, to release themselves and each other from decorum and draw in new participants.

Whether this manifests itself in arguing about your predictions, dressing up in fancy clothes, or getting blackout drunk, the way that attendees perform at the Derby fulfills a sort of prophecy embodied in its reputation. {~~ It's like "the most photographed barn in America" from Don Delillo's book White Noise, the thing is transformed by people's perception of it, and establishes a kind of feedback loop ~> The most photographed barn in America got it's reputation because people took pictures of it, and then people took pictures of it because of its reputation. ~~} [7, pp. 12-13].

Oh, by the way, I know I'm writing about this in a sort of clinical and detached way but the festive atmosphere is not a bad thing. There's a mainstream tendency to distrust anything that happens in a crowd because crowds are unruly, they "look bad." I think this comes from TV and especially the news, where real events just become things for you to look at and judge. Same deal with social media.

Anyway, the Kentucky Derby is at once a festival, a gambling event, and a sporting event, and these three aspects feed into each other.

Pachinko (Machine dawn)

Short History of Pachinko

{++ shot: move from Churchill Downs to that Pachinko parlor I found ++}

In the U.S., gambling was illegal for most of the 19th century, although it was obviously still going on. Officially sanctioned horse betting began in the 1920s, when states realized that legalizing and regulating gambling was a good way to make money. Around the time it was legalized in the States, a new form of gambling was taking hold in Japan.

Pachinko evolved from the 1700s French game Bagatelle, which also happens to be the progenitor of pinball. In all three of these games you shoot a steel ball from the bottom of the playfield to the top, then it falls down the playfield and interacts with some pins. The only obvious difference with pachinko, versus bagatelle, is that pachinko uses many small balls while bagatelle and pinball opt for larger ones.

The modern pachinko machine took shape after World War 2, when a playfield layout called the Masamura Gauge was invented [8, p. 164],[9]. It was designed to make the game more exciting; the balls bounce around and make the game's namesake sound as they fall down the field. This new design made pachinko massively popular, even though the machines operate in a legal grey area.

The goal of pachinko is for the balls to fall into one of several cups scattered around the playfield. If a ball reaches the bottom, the player loses it, but if it makes its way into a cup they are rewarded with more balls, usually around ten.

The Machine Zone

{== show the old machine footage ==}

Pachinko machines have all the basic elements of a 21st century gambling machine. The game is quick and has a kind of rhythm to it. The act of playing is tactile; the player shoots each ball with a flick of the wrist and the result is that satisfying pachink-pachink sound. Wins are small but they come often and they trigger exciting lights and bells.

Pachinko is a hundred horse races a minute; the tension and dopamine rush of a single bet is instead divided into hundreds of tiny games, each ball a small wager. What was once a drunken, sweating mass at Churchill Downs is mechanized and individualized in the bright chaos of the machine.

The gambler is pulled close to the action, and the randomness of the machine begins to take on the shape of a game. Their hand fires each ball, but the action is rendered nonconsentual by the inertia and rhythm of the machine. It has a logic all its own and by the time this dawns on the gambler it's too late.

A book called Addiction by Design, written by Natasha Schull {== shool? ==} and published in 2012, opens with the author talking to someone named Mollie. It's 1999 and they're in a Vegas hotel room, Mollie's kid is playing a gambling game on PS1, but she's there to play video poker.

>"When I ask Mollie if she is hoping for a big win, she gives a short laugh and a dismissive wave of her hand. "In the beginning there was excitement about winning," she says, "but the more I gambled, the wiser I got about my chances. Wiser, but also weaker, less able to stop. Today when I win-and I do win, from time to time—I just put it back in the machines. The thing people never understand is that I'm not playing to win." Why, then, does she play? "To keep playing—to stay in that machine zone where nothing else matters."

It's like being in the eye of a storm, is how I'd describe it. Your vision is clear on the machine in front of you but the whole world is spinning around you, and you can't really hear anything. You aren't really there—you're with the machine and that's all you're with.

I think pachinko machines were the world's first encounter with the machine zone. The machines have been declining in popularity for years, but they still take in more money than Macau, Vegas, and Singapore combined.

A lot of people don't really get the appeal of pachinko. As a westerner, it's hard to get over how overwhelmingly Japanese modern pachinko machines are. It doesn't help that pachinko parlors are so loud that most players wear earplugs. But early pachinko machines make sense; they put players in that storm of light and noise without the confusing cultural stuff. Players get their own Kentucky Derby; the sport is the random motion of the balls, the gambling is the cost of the balls and chance for a win, and the spectacle is the light and sound of each win.

Cashflows in Pachinko

As far as gambling machines go, pachinko is not very evil. It's more a form of entertainment than a moneymaker. Everyone that plays will lose money on a long enough timescale, but people aren't expecting to get rich on pachinko. Balls are cheap, payouts are low, and playing is a nice distraction.

And if you do win you can't just get your money out. Gambling is sort of illegal in Japan, so players have to trade their balls in for a prize which can then be sold for money at a nearby shop. It's a long pain-in-the ass game of gambling telephone, so you know that it's really about the love of the game. Conveniently for parlor owners, the whole process obscures the fact that balls sell for a lot less than the player pays for them. A paper I read put the exchange rate at about 62% of what the balls cost [8].

I modeled a really simple pachinko machine based on videos of the 70s-style machines. If the player buys 100 balls at, say, a dollar each, they will need to finish with 163 balls to break even, because they sell for less than a dollar. I can stop my simulation if the computer breaks even. This happens about one in four games, and the player finishes with zero balls the rest of the time.

{== According to my research, 8.25% of balls make it into a cup, and this result was very consistent across machines. My machine rewards ten bonus balls per hit, and my pachinko parlor will buy the balls for 62% of what the player paid for them. ==}

Playing Pachinko

I don't have access to a pachinko machine, but owing to its massive popularity there are a whole bunch of Japanese pachinko video games out there. I played fourteen of them on a whole variety of consoles; these aren't seventies-style machines but instead show off how pachinko evolved into the early 2000s. The first thing you probably notice is that pachinko has joined forces with its American brother, the slot machine.

Pachislots trigger a spin of the slot machine whenever a ball reaches the center cup. Some machines even have stories; they loop over short video clips until the player hits a certain jackpot, after which the story progresses. Our friends at Konami famously made Silent Hill and Metal Gear Solid pachislot machines which united the peoples of the east and west in hating Konami. According to my admittedly limited evidence, the Japanese gaming community found the machines to be condescending [10].

In modern pachinko, balls are fired automatically, and the player decides how hard they're fired using a little wheel. That's pretty much it for mechanics; machines have sweet spots that increase your chances of winning, but whether you win or lose is ultimately up to the machine. Pachislots change the pachinko meta, incentivizing players to always aim for the center cup. Once you get the wheel in the right spot, you can queue up four spins at a time and just watch the machine go.

And that's pachinko. I was able to get most of the machines I played to spin the slots continuously, and once you find the sweet spot there's not much to do. There's no way to make a pachinko game difficult; players get hundreds of balls to test the machine with so even if the sweet spot is very precise the player has plenty of time to find it. Since the speed wheel has an optimal position, it may as well just be a button; the wheel gives players an illusion of control even though the optimal strategy is easy and obvious. The difficulty curve would look something like this. {== show it. ==}

Pachinko looks like a game but you really just watch it. After you find the right position, you don't even need to do that. {== reading a book ==}

If you hit a jackpot on the slots, you can rack up thousands of balls in a few minutes, but pachinko is typically just a slow and steady loss. Manufacturers know that you're going to be spinning the slots hundreds of times per game, so jackpots don't come easily.

One other feature shared by many of the games I played is some sort of an adventure mode, where you take up the noble career of pachinko player in a simulated world. Each of these is very depressing. In real life, there are lots of people that do nothing but play pachinko all day, and it's disturbing to see that lifestyle depicted completely uncritically in a game. Pachinko games are usually licensed, so manufacturers are happy to promote addictive behaviour.

Parlor Pro 64 is a typical example. You start off in your room, usually with the ability to wait, skip a day, or go out and play pachinko. The game runs a clock as you travel to different parlors and play the same three machines over and over. It does a great job putting you in the shoes of a pachinko addict: all your character can do is travel from parlor to parlor and watch time tick by. Their room is small, dark, and practically empty.

The pachinko experience is all about continuation. The machine beats out a rhythm in the pinging of balls and the spinning of slot reels that the addict can't help but gaze at and pay tribute to. The machines are mind-numbing, and the sugary slot machine jingles cause a lurch of dopamine that I found more nauseating than enjoyable. If you learn to like that feeling, then the only goal is to keep the beat of the machine going.

Pachinko adventure mode draws our attention to the consequences of addiction; it depicts a life consumed by the light and noise of this game. All we can do is gamble or sleep. The Pachinko Industry Web Reference, which is run by a machine manufacturer, makes specific references to people leaving their infants in hot cars to go and play pachinko [11].

Oh by the way, Recently-exploded politician Shinzo Abe took significant steps to open up the country's gambling laws [12, 13].

{== show the following on screen: ==}
> When the local pachinko parlor opened in the morning she'd be outside waiting, and she'd still be at it at closing time. "All I could think of was pachinko. I wanted to do it 24 hours a day – the thrill, the feeling of liberation. It was like being on drugs."

Japan Today, Pachinko addiction a growing problem for Japanese women


Mechanical Slots

The first slot machine was developed in New York in the 1880s. It was a strange machine by modern standards that had a poker theme and encoded a deck of cards on its reels. These machines didn't even pay out money, although some establishments would give winners free drinks or smokes. The machines proved very popular (and profitable), and the mechanics of slot machines were simplified and perfected over the next few years.

A classical slot machine uses physical reels with symbols printed on them that spin and stop in a way the manufacturer can't predict. If a certain pattern of symbols shows up along a line called the payline, the player wins some money. There can only be so many symbols, and each has an equal chance of showing up, so the odds of winning are obvious to anybody that plays. Early machines were also very slow compared to modern ones, and were basically one-spin, novelty games.

Pachinko machines haven't fundamentally changed since the 1920s, but slot machines have made strides in lying to gamblers. In the early days, all the mechanics of a slot machine had to be built with actual mechanisms. The technical limitations of the time kept the machines fair; since they already made tons of money, further developments weren't worth the investment. Making small, precise parts is expensive, after all.

Weighted Reels

While these early machines were honest, there was room for improvement. Players were unlikely to use a machine with tons of symbols or tons of reels, because more possible combinations means fewer winning combinations. On the other hand, a machine that has fewer symbols hits the jackpot more often, so they couldn't pay out huge amounts of money either.

One early innovation was making the player's viewing window larger [15, p. 80], so they could see not only the symbols on the payline, but those above and below it. This created the opportunity for near-misses, where the player just misses a jackpot {== show this ==}. It's technically illegal {== text: (in Vegas) ==} to program a machine to deliver near misses on purpose [16] but there are tricks to make it happen and they create a rush in the player: maybe next time it'll be a win.

The biggest change came in 1982, when a patent was filed that unleashed slot machines from the limits of physics {== The first RNG machines appeared in 1978; they were Fortune Coin machines purchased and modified by IGT. [15, p. 84] [17] ==}. It described a machine whose outcome was determined by a random number generator rather than the random action of a brake. That might seem inconsequential, but the inventor realized that, with a computer in the mix, they could now set the odds of each symbol however they wanted, and that's the real innovation.

The patent explains this idea through the metaphor of invisible, virtual stopping points. While the physical machine might have ten symbols, the computer that controls it can map millions of points on the reels. If, for example, the cherry symbol has three hundred virtual stops out of a million total, the odds of it showing up are now .03%.

Just looking at a slot machine, we get an intuitive idea of what the odds are. If there are ten symbols, and one of them is a cherry, then obviously we have a one-in-ten chance of getting a cherry. We would also assume that each reel on the slot machine has that same, one-in-ten chance of delivering a cherry.

With virtual stops, those assumptions are no longer true. The machine looks the same to the gambler, but the appearance of obvious fairness is just an appearance. The physical size of the machine stays the same, but it can promise much larger jackpots. Addiction therapist and consumer advocate Roger Horbay identifies this dissonance between expectations and outcomes as the biggest driver of gambling machine addiction [18].

I've taken to calling this concept deceptive skeumorphism. Skeumorphism is when a new version of some object retains the look of its original, like those electric candles they sell at dollar stores. Mechanical slot machines imply a lot of things about the odds of winning, but weighted reel slots benefit by bringing those same implications to a fundamentally different game.

To relate this back to a more familiar subject, imagine you sit down in front of a screen and see this. What kind of game do you think this is? The UI creates a whole set of expectations about what we're going to control in this game. OK, so unpause it.

{== mario heh heh heh==}

I've made the dissonance obvious on purpose, so you might look to me with hurt and betrayal as tears well in your eyes and whisper "why have you done this to me? This is not Road and Track presents: the Need For Speed (1994) for MS-DOS." But a clever designer can do something more subtle, and bring the UI and the game out of line in ways that are almost invisible. And this is exactly how weighted reel slots use skeumorphism.

As we get further into the impact of these machines on people's lives, consider that in the industry RNG has been called the Really New God {== maybe get a shot of my EEPROM with "THE REAL GOD" above it, and a text note about EEPROM vs. EPROM ==}. On paper, slot technology has always been developed to increase security and predictability for casinos, but they unleashed a monster in the form of RNG. In the distant past, we attributed the confusing motions of the natural world to gods of nature, projections of human order that we eventually grew out of as we began to tame the world.

Computerized slot machines reinject gambling with that old mysticism; they are designed so that the gambler never knows the odds, so that the machine becomes an unintelligible force. Skeumorphism helps create the effect, but RNG is what really directs the game. The random nature of prehistory is replaced with random numbers. Popular games like poker and blackjack have been totally mapped out by math, but the slots are kept hidden. From the mouth of a gaming regulator in Nevada:

>There is only one game in Nevada where the player doesn't know what his odds are...there isn't an establishment that would agree with posting those odds on that slot machine because you are going to take away the mystery, the excitement and entertainment and risk of playing. [15, p. 78]

{-- Although you might see a blackjack table and a slot machine at the same casino, the two are very different things. The God of table gambling is the God of the Old Testament; every play and bet leaves the gambler at his mercy. But his judgments are intelligible, through the language of cards, competition, and rules. The Really New God of the slots is a total mystic, their language is a mystery to the player and the rules of the game are opaque. Somebody at a slot machine gets struck by lightning without advance warning. --}

The Rise and Rise of Slots

Just as weighted reels came to dominate slots, slots came to dominate casinos. The machines were originally perceived as games for the wives of so-called real gamblers at the tables. Addiction by Design credits their ascendancy to a perfect storm: video games were becoming more popular, so video slots looked a little bit familiar; an early 90s recession had many states scrambling to make money; and the slot machine's status as a wife-game spared it from the taint of moral failure that people associated with table games [15]. I would also add that the new technological developments made the machines a lot more appealing to both gamblers and casino owners.

Electronic Slots

Like it's relative pachinko, slots have tended to electrify over time, but they've held onto the concept of reels because it promises a level of fairness that just isn't there. Pachislots are an interesting subset of slot machines, because the player can actually see some of the RNG; the slot machine is triggered by a randomly bouncing ball which the player gets to watch. The randomness becomes explicit, which is a huge difference from the mystical RNG of American slots.

Slots differ from pachinko in another very important way: they're a lot riskier. A pachinko ball costs around three cents, but a credit on a slot machine can cost anywhere from a penny to a hundred bucks [15, p. 5]. And spins usually cost more than one credit.

Video slot machines are also fast, and this makes them fundamentally different from the older style of machine. Classical slot machines consisted of one event, a single spin. It was slow. You had to manually put coins in the machine and pull a lever. If you won a lot of money, an attendant would have to pay you since early machines could only dispense 20 coins or so. This slowed the game down and suggested a convenient end point to the player.

As the machines sped up the experience started to resemble pachinko. Modern slots carry over a kind of momentum between spins, so that they transform into something different: a game of modern slots is made up of many connected betting events rather than disconnected ones. Credit can be loaded using bills or more often magnetic stripe cards at the beginning of play, and the player can spin many times in quick succession using a button or touch screen rather than a cumbersome lever.

The patterns of use and addiction we see in modern slot machines also mimic pachinko. Pachinko makes its money on users who play the games all day every day, and the continuity we get with modern gambling machines invites players to stay in the machine zone for longer and longer. Casinos make the most money on these returning players.

Addiction and Slots

You might get the impression that I have a bone to pick with these machines, and that's because I do. Games can use some of the design techniques I've talked about for the purposes of fun or immersion, but gambling and gaming are very different. For example, we can look at how fast slots are and say that it's just giving players what they want, that speed makes the game more exciting. But that's a way to conveniently ignore the money.

The house always wins in the end, and they win faster the more bets a player can place in a given time. The casino is incentivized toward making betting as fast as possible, and making the machine as satisfying as possible. The player, ensnared by the machine, also sees these as positive outcomes.

That bothers me for a couple of reasons. If we were all perfectly rational decision-making machines, it might be possible to just gamble for fun. I can only speak for myself, but we are usually better at rationalizing our decisions than making rational decisions; I may go into a casino for a bit of fun, but the design of the machines, and of the casino itself, can hijack that and manipulate my behaviour.

The most common piece of advice people give about gambling is to only bring money that you are expecting to lose. So by going into a casino you are accepting that it will manipulate you to such an extent that it's dangerous to bring too much money. When individuals exercise that kind of control, we call them cult leaders rather than business leaders.

Practically all the research I read for this video points to the fact that gambling machines create gambling addicts, also known as whales. As you probably know they are called whales because they are rare, but bring huge amounts of money into the casino.

The most recent paper I could find, from June of 2022 {== written July 6th 2022 ==}, shows that after the COVID-19 pandemic {== it's not over, it will never be over, the west just normalized it, etc. etc. ==} around 84% of the revenue from casino slot machines came from problem gamblers [19, p. 15] {== good pie chart opportunity ==}. Slot machines are by far the worst for this, but across the industry 14% of gamblers account for 44% of gambling revenue. You might argue that the pandemic skewed these results, but many older studies have found that 60% of slot machine revenue comes from problem gamblers [18] [15, p. 16] [2]. And even if the pandemic does skew the results, we live in a world where the pandemic happened, and the 2022 study reflects the so-called new normal.

Whether casinos pay lip-service to the idea of helping addicts or not, the ultimate goal of any business is to make as much profit as possible. And if any addict manages to quit, that is potentially a big loss for the casino.

If academic papers don't convince you, just watch slot machine videos on YouTube. {== long edit of the Slot Bros. video? ==}

Any fun that the player has on a slot machine is only a means to an end, and this makes them fundamentally different from games. That doesn't mean that slot machine design is simplistic though. Slots are a site of profit, and like a factory a lot of money and research goes into making slot machines more efficient and effective. We've seen already how computers became a bigger and bigger part of the games over time, and now I want to get into how video slot machines are actually programmed.

{++ shorten this section ++}
### How Slots Work

Videos and articles produced by casinos and manufacturers will assure you that there's some very complicated math going on inside a slot machine. That's not really true; a lot of research goes into making the machines as compelling as possible, and that involves a delicate balance of payouts, but the basic math of the machines is very simple.

Let's forget about any appearances the machine has; every spin comes down to a win or a loss, so we can basically ignore the process in the middle. That patent from 1982 tells us that the look and feel of the game is completely separate from how it actually functions under the hood.

In analyzing modern slots there are two important pieces of information, the par sheet and payback percentage. P.A.R. stands for probability accounting report [2], and a par sheet tells you how likely every possible outcome on the machine is. I'm going to go through an example in a minute, so don't worry if this seems abstract.

The other important metric is the payback percentage. You'll see these published a lot more often than par sheets, because they sound really good. The payback percentage is the average share of money the machine will give back to players, and it varies anywhere from 80-98% between machines. So over millions and millions of spins, some machines will pay back 98% of the money they take in [20]. Since we're dealing with casinos, we shouldn't take this at face value.

Payback percentages make it look like the margin between winning and losing is thin, and like the player can beat the machine. Let's see where this number comes from.

Here's a par sheet that I put together for my Lucky Slugcat video slot machine.

R = [1000, 30, 12, 0];
P = [0.00001, 0.01, 0.05, 0.93999];

The first column here is the payout, and you can think of it as a multiplier. If I bet a dollar then hit the jackpot, I'll get $1000 out. The second column is the probability that a player will hit each multiplier. For example, there is a 1-in-a-hundred, or 1% chance that we'll multiply our bet by 30.

Notice how none of this has anything to do with the number of reels or symbols on the machine; we can use skeumorphic design to make the odds look better than they really are. Let's give our machine three reels and ten symbols. The probability of a jackpot should be 1%, but I made it a thousand times less likely than that.

Looking at these odds, it seems unlikely that anyone will make money on this machine. There's a 94% chance to lose, after all. Take a second to consider how high the payback percentage could be on a machine where you lose nine out of ten times.

It's 91%. This is how you calculate it {== visuals ==}. Since the multiplier for a losing bet is 0, it completely disappears from the math and we can say our machine pays back 91%. There's a disconnect between the chance of losing and the payback percentage because games aren't played on average; they are played one spin at a time.

We know already that whales are the best customers. I wanted to find a measure that shows how much money problem gamblers are losing. I call it the rate of profit or, more often, the rate of loss. It tells us how much money the player makes, on average, from each spin, and it takes into account the payback percentage and the probability of losing. After a large number of spins, the rate of profit will virtually always be negative, meaning the player is losing money.

{== Text: Math, if you're interested: ==}

{== Slot machines are random in some ways so there is no definite symbolic function for profit on a given spin, p(n); the rate of profit would be the derivative of this function so I calculated it in the simulation using finite differences over the simulated profit data. Profit is roughly linear over many games so even a linear approximation (a slope calculated using the first and last points) works very well for a simulation with 100k+ spins. ==}

{== The profit ratio definition starts with what I call the Unit Bet Profit Ratio, because it assumes $1 bets for each spin but it is trivial to generalize it, and the simulations agree with my math in all cases, as close as you would expect with a random variable. UBPR is the payback percentage minus loss percentage; -3% on Lucky Slugcat. ==}

{== Something I did not account for is the partial-loss, where the player wins but wins less than the cost of a bet. This would be an interesting line of inquiry for anyone who wants to develop slot machine math further. The MATLAB code is available on my website. ==}

The rate of profit exposes several truths about slot machines. First, anybody who plays these machines a lot is losing money. Researchers say that video slot machines can do about 1200 spins per hour, so even one weekend in Vegas is very likely to lose a slot player money. Somebody who plays for an hour only has a 37% chance of breaking even, and somebody who plays for 24 hours has a 16% chance. It's worse than a coin toss no matter what.

Another interesting point is that the rate of profit gets slower as the payback percentage increases. This means that players still lose money, but over a longer period of time. In the short term the gambler is winning more often, so they might profit for a little while, but their bankroll is still slowly going down.

For long-term business, a payback percentage of 98 is better for the casino than 80. A higher payback percentage keeps the whales on the machines and strengthens their connection to the machine zone, even as they lose money long-term.

Electronic Slot Design

Creating long-term addictions to video slots is helped by a slow and steady rate of loss punctuated with many small jackpots. It's impossible to talk about slots without mentioning the Skinner box, so let's get it out of the way. B.F. Skinner was a psychologist who was interested in manipulating or conditioning behaviour. He created something called an operant conditioning chamber, or Skinner Box, which conditions a response to some stimulus. For example, if a green light comes on and the subject hits a lever, they will get some food.

What Skinner discovered is that using simple systems of reward and punishment is a very effective way to manipulate how animals and people make choices. More crucially, he found that the best way of getting somebody to pull the lever over and over is to give them a reward after a random number of pulls. The free market discovered this principle decades before Skinner in the slot machine. While gambling addicts are technically free, in the sense that their hand is putting money into the machine, the reality is that slot machines are incredibly effective conditioning systems, to the extent that they undermine the actual desires of the addict--namely, the desire to stop gambling.

While behaviour can be conditioned in both people and animals, only people have the ability to reflect on their conditioning. Addiction by Design describes a sort of nihilistic euphoria that machine gambling can create; the machine zone is something that envelopes the player and the only goal is for the feeling to never stop. {== clip about getting mad when you win ==}

Machine gamblers are attracted not by the risk of gambling, but the familiarity of the mechanism itself; the light and sound demands nothing but the gambler's full attention. {== narcoleptic clip? ==} Player-centered design is the current trend, and it's less of a buzzword than it sounds. Cashless gambling, where cards are used rather than cash, let the player get started more quickly and they help to abstract away the money they're losing. The pace of machines is now set by players, and the switch from cumbersome lever-pulls to buttons and touch screens also opens up gambling to wider demographics. So at least some AAA developers care about accessibility.

Research and theory on gambling machines shows pretty conclusively that they create addicts, but classifying people that way implicitly shifts the blame onto them rather than the conditions that create them. Our new habit of calling them "whales" fully dehumanizes them.

That's not to say that learning from individual gambling addicts is a bad thing, and there has been some progress toward actually understanding the addiction. Gambling addiction was recently re-classified as an addiction, rather than an impulse control disorder [21], which is a fundamentally different thing. Bad impulse control is when I eat a whole bag of chips, addiction is when I center my entire life around eating chips. Addiction by Design is framed entirely through the experiences of real gambling addicts, and it's a landmark book because Schull learns about her subjects by actually talking to them, which is sadly a pretty uncommon approach.

I've been focusing mainly on the design of the machines because that's what I'm good at, but if you want to know what these machines do to people Schull's book is indispensable. It is packed with stories of vulnerable people whose lives are consumed by gambling machines, and Schull uses these stories to frame her discussion of slot machine and casino design. If you find this topic interesting you should check it out.

King Neptune: A Guided Tour

Knowing what we know about slots, it's time to play a machine. Casinos aren't allowed to operate where I live; video gambling machines are allowed in some bars but I couldn't find one. Thankfully, there's an emulator available online for a series of slot machines produced by the Australian company Aristocrat Leisure. To reproduce the slot experience I turned my speakers up to embarrassingly loud levels and pulled a big bright TV from my pile of obsolete tech. The emulator has some issues so I apologize for the awful sound in this section. It was built for Windows 2000.

I'll focus on the flow of a single machine, King Neptune. There are lots of variations in machines but also a few basic design patterns that show up over and over. This machine helpfully displays our bet in dollars, but many don't. Abstracting the player's money as credit is a way to distract us from how much we're actually betting. This machine, King Neptune, has twenty-five pay lines and a whole bunch of convoluted, poorly explained rules.

(slots-02.mp4, 09:05)
>When the Each Way Bonus symbol occurs, the following rules apply: Highest win paid left to right and highest win paid right to left except tridents which pay any.

That's not a rule, it's just a random statement. The inscrutable logic of the machine is meant to overwhelm and numb the player, we're just supposed to bet max and enjoy the show. Overwhelming the brain is actually a pretty common technique to put people in a hypnotic trance [22], and the machine is purposely obnoxious and confusing so that you'll slip into a sort of go-with-the-flow state of mind.

In terms of pacing, the machine moves at its own speed once you press the button, but between spins it is silent. The machines are designed this way so that the player can feed themselves dopamine in their own perfectly addictive rhythm. The machine suggests a pace, set by the spinning of the reels, but in the end the player has to unleash the spectacle of the machine, and this is the real meaning of player-centered design. The machine gives a feeling of control. Combined with the trance effect, players tend to press the button according to an internal rhythm that they aren't directly aware of.

It creates the satisfaction of killing a boss perfectly in time with the music, but this effect is hijacked and used to manipulate players.

It's hard to put into words what a slot machine feels like to play, but it snared me almost immediately. After putting a ton of damning research about slot machines into this script, my immediate response to the machine was that I won more often in the first few spins of a session.

So it took like, two minutes to trick me. Slot machine outcomes are absolutely random by law and by design. There's an apparent innocence to the machines; they look like stupid, chintzy little toys whose sugary music couldn't fool anyone. There's also an apparent order and set of rules that gives us an excuse to project strategies and patterns where there really aren't any. They are designed to make you think you can beat them. {== this is a fucking big difference between gambling and grinding; the grind acknowledges itself as such. ==}

That might be part of why they've proven so addictive, but all of the research on gambling machines points to a deep level of immersion in players. It's not immersion into Neptune's Kingdom or anything like that, but a connection with the machine that drowns out everything else. Mollie, interviewed in Addiction By Design, says that her addiction to slots is about being trapped; she describes gambling machines as "a series of entrapments that you can't appreciate from inside them" [4, p. 24]. So while the spectacle and the logic of the machine contribute to the machine zone effect, the ensemble has to be taken as a whole.

Honestly, I'm not sure how to make sense of electronic gambling machines. If you have the constitution for it, track down that emulator and play a machine until you run out of credit. Even on a fake machine, running out of money makes you feel like you've done something very wrong. The playing experience is strangely comforting, the machine builds a sense of familiarity in the span of seconds. Maybe they make players feel like they're good at playing, providing the satisfaction of mastery with none of the practice. The interviews in Schull's book talk about this in more detail, but alongside the comfort of playing there's also a clear feeling that something really fucking bad is going on.

The Aristocrat machines you can play on that emulator are from the early 2000s, and they actually have a vestigial mechanic that I think is interesting. After a winning spin, you can gamble those winnings in a guessing game that can double your win, or lose it. Over time, machines have tended to eliminate anything that acts as a break in play; they still have special features but they are triggered by certain symbols rather than being directly controlled. The Aristocrat wagering minigame is unsatisfying because it breaks the flow.

Modern machines, at least the best ones, are centered entirely on the reels, and several different games can play out within that same, familiar space. Here's a later Aristocrat slot machine, Buffalo Cash. They kept the same awful jingles from the old machines, but all of its bonus games are triggered by the reels and take place on the reels. This guy is betting $100 at a time and spends the entire video trying not to throw up.

Buffalo Cash also has a feature that is showing up on more and more machines, a bonus that triggers rarely at a random time. This counter goes up whenever a buffalo appears, and at some random time it triggers a bonus feature. It's another layer of Skinner Box design; slot players often have a superstition that a particular machine is "ready" to give up a jackpot, and features like the buffalo counter give that idea a substantial form.

Designing machines like this always keeps the player feeling like a jackpot is just around the corner. For the addict is has a different appeal, but all addicts start out as players.

Rhythm and pacing are more important than the content of the games themselves. It could be video poker, slots, or keno. It's all a means to an end, a way to get into the machine zone. Players, and addicts especially, tend toward playing as quickly as possible. I think there are two reasons for this: of course familiarity lets people operate the machines more quickly, but more importantly the natural pace of the machine is set to be as fast as possible. On King Neptune for example, if you bet at the first possible opportunity the reels never stop spinning.

As we know, every bet loses money on average, so more bets means more profit for the casino. Playing fast also perfects that machine zone experience; the slot machine opens up into an endless chain of events with no breaks in between.

The reason slots unsettle me is because they represent an entire course of development, a whole history geared toward turning people into lever-pulling rats. Somebody discovered that wagering money makes pretty much anything compelling, and they ran with this idea to create the most efficient possible gambling device.

Pokemon Cards (First gen ocean pack opening s+ gem dx)

While the adults were playing slots, their kids were collecting Pokemon cards. Prestige in collectible card games comes from buying tons of card packs, and they helped to normalize the idea of artificial scarcity, which video games later took to its logical conclusion. In Pokemon, cards are rare so that not everybody can have them, and the limits on how many exist are totally arbitrary. The scarcity is artificial.

This isn't a new thing: collecting is an enduring hobby and collectible cards are also very old. But Pokemon, which was developed by Wizards of the Coast and debuted in 1996, was a phenomenon, and its influence has a direct line to today.

Pokemon cards, especially at the peak of their cultural relevance, worked a bit like table gambling. Which cards you pull are basically random, but there is a high status associated with the winners. Having a rare card makes you royalty in third grade.

What sets card games like Pokemon, Magic, or Yu-Gi-Oh apart from table gambling is that the cards have specific functions, and collecting them is also buying into a sort of community, because having cards lets you play against other people.

I was into Yu-Gi-Oh but I didn't know how to play or anything, I just liked the anime and the art on the cards was cool. If you're an insulated white trash five year old boy in the cultural void of 2004, you don't just see something like Red Eyes Black Metal Dragon and go on with your day. Nothing can compete with a concept that cool. Like, damn, I can't wait for Shrek 2 to come out.

Yu-Gi-Oh rocks. Anyway, card collecting isn't exploitative the same way gambling is. These cards are used in a game of skill, and I imagine a lot of kids' desire for more cards was checked by their parents, so they aren't continuously betting and losing money by opening them. These games create communities with hierarchies that are determined by a roll of the dice, but they gate access to an actual game, so the content is much more substantial than something like a slot machine.

The cashflows associated with card games are consistent, small purchases that don't guarantee value. I think gambling is a fair word for that, but the way kids interact with card packs is nothing like a slot machine; they might get one or two packs at any given time and opening card packs very much is about the outcome. You wanna to catch 'em all.

But the line between a so-called healthy gambler and a problem gambler is thin. Card packs have become grist for the content mill, and I think a representative example is this video by Leonhart {== play the same distorted-sound video of him screaming every time you say his name? Maybe tell him to chill the first time har har ==} on YouTube. The setting, his performance, and the editing all create an aesthetic you might see from a modern pachislot machine; it's bright, loud, and fast. When Leonhart opens packs he cuts the stack of cards so that a low value card is shown first, and he does this to build up the tension. It's the same reason the reels on a video slot machine still spin; the rhythm of tension and release gets the viewer invested, and keeps people coming back. Pure tension-and-release, without any thematic content is like a story arc in a book or TV show stripped down to its ugly metal skeleton. All of the dopamine with none of the thinking.

Something I discovered while working on this video is that you can get into the machine zone just watching other people gamble; slot machines can really suck you in even in videos. People who open card packs for their underage audiences are creating a whole new generation that think the machine zone is a normal, healthy place to be. Leonhart even plays a slot machine jackpot sound when he pulls a good card.

I don't want to veer into any 'think of the children' bullshit, but normalizing gambling behaviour is very different from normalizing something like video game violence, because real violence will never be normal whereas gambling has been worming its way into the mainstream. I'm not picking on Leonhart because I think individual YouTubers are destroying young minds or whatever, but because his project, and collectible card games themselves, are an early form of socially-acceptable gambling that extends beyond the racetrack or casino. In 2022, gambling is everywhere, all the time, but we'll get to that later.

Collectible card games, like so many other fun things, have been marketized; the value of Pokemon cards has nothing to do with how useful they are in the game or even how cool they look. The value of Pokemon cards is their rarity, and it's weirdly normal even for younger people to talk about them in terms of quality grades and auction values.

Which kind of brings us full circle; spending $50,000 on Pokemon cards, like owning a race horse, is a hobby for rich losers. There's a change here, though. The Kentucky Derby was a balance of spectacle, gambling, and sporting. It might not be the prettiest thing in the world, but it's a genuine social activity. With card packs, we see that the sporting part, the game itself, isn't really necessary. If you're just opening packs the game isn't in the loop anymore. Also, the other elements, the gamble and the spectacle, are now entirely individual; even if you mine your own addiction for content you're still opening card packs alone, screaming and sweating for an imagined audience {== leonhart, mask him into a black void? ==}. Opening card packs is a horse race with one horse in an empty stadium.

For people that play these games, the card packs exist in tension with the game. It's like having to buy expensive gear to play hockey, except there's a 50% chance the pads are papier-mache, or you get 100 cheap sticks and no helmet. If you like playing the game, it's not a good thing that a player with money can steamroll a player without money. The skill of a card game is in building a deck to counter an opponent, and executing a strategy in spite of the randomness of the shuffled cards. As a competitor or spectator, I want to see the full range of possible cards, because that makes the game more fun.

For those who collect and sell the cards, the game doesn't even exist. The logic of gambling itself has replaced the game; it's the tension and release of revealing the cards, the tension of the slot reels spinning and coming to rest, that constitutes a game.

The fun part of collectible card games is in a contradictory relationship with the way the cards are distributed, and this is by design. In the beginning, the creators knew that players would have to waste some money to get cards they wanted. However, as time has gone on, a kind of pathological nostalgia developed largely around Pokemon that eschews the game for the masturbatory act of just opening the cards. To put it another way, the money-value of the cards has overcome their play-value, and opening the cards has become a fetish totally disconnected from the game of Pokemon.

A Note About Table Games & The Story so Far

So that covers a few gambling topics. Table games have more in common with sports than slot machines, so I skipped over them. Poker is all about making the best of random cards, and a game like blackjack at least has an element of strategy, even if perfectly played blackjack is basically a dice roll. Table games are also slow, they stretch out the tension-release cycle of gambling over many minutes. From the perspective of the casino they're vestigial instruments.

Machines make money, and they get people hooked. Bill Friedman, the casino designer par excellence and author of the definitive book Designing Casinos to Dominate the Competition, centers his entire design philosophy on machines [15, p. 40].
Images on screen are from [15, p. 42], text quotes are from the same chapter.

The whole purpose of a casino's interior design is to draw people to machines, and make them feel safe while they gamble there.

{== VO session 5 was corrupted from here to the end. Thanks Adobe.==}
What we see with most forms of gambling is a tendency to dehumanize gamblers. I presented in-person horse betting to introduce gambling as a social activity. Gambling at its best is a social activity, a few dollars passed between friends or staked on a game of poker. Games played between people. But machine gambling, and individual gambling more broadly, use betting and randomness as tools that make players little more than rats.

A skinner box is a systematic way of manipulating a creature's desires. A whale is technically free to stop pressing the button, free to stand up and never enter a casino again, but the machine has hijacked their desires; all they want is to play slots and to stop playing slots. Slot machines strengthen that animal, dopamine-maximizing urge that we so often find ourselves in conflict with. The urge can't be overcome just by thinking it over, it's too strong for that.

That's why people who are struggling with addiction seek help in groups of other people, that's why the Alcoholics Anonymous program has you submit to a higher power. That's why slot machines must be a private game; if slots were a social activity people would help one another, and most players could never reach the point of addiction. The machines prey on people who are too weak, whose lives are too empty to resist them. And I haven't even talked about online gambling.

We're going to check back in on the world of gambling in a bit, but first we have to trace the evolution of video games, from arcades to the invention of the internet. The next episode is gonna be fun, for real.


N. Long, "Blizzard earned $49m from Diablo Immortal's first month, with 10m downloads to date,", Jul. 04, 2022. (accessed Jul. 09, 2022).
Lucas, Anthony F., and Katherine A. Spilde. "Examining Expanded Differences in Slot Machine Pars: An Analysis of Revenue Impacts and Player Sensitivity to 'Price.'" International Journal of Hospitality Management 86 (April 2020): 102450.
"A History Of Horse Gambling," Jul. 31, 2019. (accessed Jun. 05, 2022).
"Parimutuel Betting - History | Technology Trends." (accessed Jun. 20, 2022).
"Light & Wonder," Wikipedia. Jul. 05, 2022. Accessed: Jul. 06, 2022. [Online]. Available:
H. S. Thompson, "The Kentucky Derby is Decadent and Depraved," Scanlan's Monthly, vol. 1, no. 4, Jun. 1970.
D. DeLillo, White noise. New York: Penguin Books, 1986.
G. Rockwell and A. Keiji, "Pachinko: A game studies perspective | Kinephanos," Kinephanos journal of media studies and popular culture, vol. 5, pp. 161–174, Dec. 2015.
"Pachinko History." (accessed Jul. 21, 2022).
"Efforts for the Prevention of Pachinko Addiction | Pachinko Industry Web Reference." (accessed Jul. 23, 2022).
K. Sasaki, "Japan's Integrated Resort Promotion Law," International Masters of Gaming Law, p. 26, May 2017. Available:
"Pachinko and Slot Machine Addiction in Japan,", Aug. 02, 2018. (accessed Jul. 23, 2022).
"Pachinko addiction a growing problem for Japanese women," Japan Today. (accessed Jul. 23, 2022).
N. D. Schüll, Addiction by design: machine gambling in Las Vegas. Princeton, NJ: Princeton University Press, 2012.
"The Near Miss Slot Programming." (accessed Jul. 06, 2022).
"History Of Slot Machines - Slot Evolution: 1891-2022." (accessed Jul. 06, 2022).
R. Collier, "Do slot machines play mind games with gamblers?," CMAJ, vol. 179, no. 1, pp. 23–24, Jul. 2008, doi: 10.1503/cmaj.080870.
H. Wardle, A. Kolesnikov, I. Fielder, and N. Critchlow, "Is the economic model of gambling dependent on problem gambling?," p. 22.
AcgAdmin, "Slot Machine Payback Percentages - Can They Help You Win?," American Casino Guide Book, May 07, 2021. (accessed Jun. 19, 2022).
L. Rennert, C. Denis, K. Peer, K. G. Lynch, J. Gelernter, and H. R. Kranzler, "DSM-5 Gambling Disorder: Prevalence and Characteristics in a Substance Use Disorder Sample," Exp Clin Psychopharmacol, vol. 22, no. 1, pp. 50–56, Feb. 2014, doi: 10.1037/a0034518. Available:
"Mental Confusion Inductions." (accessed Jul. 06, 2022).