Last week, Valve announced and made available for purchase the Dota 2 Fall 2016 Battle Pass. Rejoice! Alongside more familiar compendium-ish fare (unique items, quests, etc.), the Battle Pass marks the return of Rylai’s Battle Blessing. The minigame is a lovely bit of skeuomorphism; the digital roulette wheel whirls round and round, awarding virtual goods ranging from the almost-worthless to the slightly-less-worthless. And the more money you spend on the Battle Pass, the more chances you have to win. Or lose.
Hardly anyone thinks these games aren’t rigged.
Anyway, it goes without saying that the Battle Blessing AKA Wheel of Fortune is not actually a physics simulation; it generates an animation based on an outcome chosen by a random number chosen at the moment the user “pulls” wheel’s lever. Likewise, Lina’s Battle Blaze from The International 6’s compendium is not really simulating a two-dimensional ball plunking down through an array of pegs. Hardly anyone, I hope, thinks these games aren’t rigged. But the ways in which the simulation manipulates the outcomes of either game are interesting, in part because the psychological tricks that underlay the design are drawn from the (marginally more sordid?) world of machine gambling.
In her 2012 book Addiction by Design: Machine Gambling in Las Vegas, Natasha Schüll examines the history of gambling devices, from the mechanical slot machines of the late nineteenth century to the finely-tuned video machines of the present. This arc of innovation has produced some of the most addictive and lucrative machines ever built; in terms of revenue, today’s machines are several hundred times ‘better’ than those produced even 50 years ago. (Factum: An experienced gambler in the early 1980s could play around 120 hands an hour; these days, it’s more like 1,500. And at, say, five bucks a hand, well … ).
Now, whether or not you think that’s “evil” or “genius” probably depends on if you own a casino. But ethics aside, Schüll’s “addiction by design” is the result of constant iteration in the design of gambling machines. It’s a great, multifaceted story, but the most significant innovation came in 1978, when the digital microprocessor was introduced. Prior to its debut, slot machines had largely been mechanical or electro-mechanical, and were therefore subject to all kinds of physical interventions – tilts, shakes, etc. – from players seeking to influence the outcome of a spin. They were also limited in the number of slots that could fit on a reel; standard machines had only 22 slots per reel because that’s the most a machine could take without becoming too cumbersome for its own good.
Microprocessor-based video gambling put a stop to both of these “problems.” The move to digital gambling also enabled what’s known as “virtual reel mapping,” a technique in which the microprocessor maps multiple outcomes to the same digital “slot.” This enabled both the manipulation of the results of any given pull, as well as the ability to boost jackpots to previously inconceivable levels. Consider: a standard mechanical slot machine from the 1960s had three reels with 22 slots each, meaning that the chance of hitting the jackpot was one in 10,648. Thus, $10,658 is the most a single machine could offer on a one dollar pull while still guaranteeing a net profit for the house. But with virtual reel mapping, which is unencumbered by those tiresome laws of physics, the number of possible outcomes can effectively be limitless. As Schüll explains,
“On a machine whose virtual reels have 64 stops each with only one stop mapped to a jackpot symbol, the chances of hitting that symbol on all three reels to win the jackpot are 1 in 643, or 262,144. The machine could therefore offer a jackpot of up to $262,144 without losing money. On a machine with 512 virtual stops, the odds of a jackpot would be as rare as 1 in 137 million — giving the house a safe enough edge to offer $20 or $30 million prizes and still ensure long-term profit.”
Despite the longer odds, players have consistently been willing to spend more money and more time on machines with larger jackpots, in large part because virtual reel mapping can distort the apparent odds of winning on a given machine. Consider: visually speaking, the chances of winning an Arcana on Rylai’s Wheel of Fortune look to be about 1 in 60 (i.e. there are 20 different “slots,” and the Arcana takes up ⅓ of one of them). But, thanks to virtual reel mapping, the chance of actually getting an Arcana could be (and probably is) a lot higher; all Valve would need to do is map a single win onto the Arcana slot, and map all other possible outcomes onto less valuable slots. There’s a gap, in other words, between the player’s perception of probability based on the visual information available to them and the actual probabilities determined in unseen computational processes. In short, virtual reel mapping can make a win seem a lot more likely than it is.
Virtual reel mapping can make a win seem a lot more likely than it is.
What’s even more sinister than virtual reel mapping’s distortion of apparent probabilities, though, is its ability to manipulate players’ perceptions of losses in very seductive ways. Thanks to a technique known as “clustering,” designers are able map a disproportionate number of losing outcomes to slots directly next to winning ones. Thus, insofar as player perception goes, it’s far more likely to almost win than it is to lose outright, and the sense of having almost won correlates strongly with playing again. So, if it feels like you get “spin again” on Rylai’s Battle Blessing a lot more than you should, or that worthless 100 point payout on either side of the 24,000 point jackpot in Lina’s Battle Blaze, well … that’s not coincidence. It’s virtual reel mapping at work.
Finally, the move to digital gambling allowed designers to present the price of a given machine in “credits” (or “plays” any other number of terms) rather than the real monetary value. By abstracting the actual cost of gambling, they found that players were willing to spend much larger amounts (it’s a lot easier, it turns out, to bet to 20 credits worth $5 each than $100). Imagine a Rylai’s Battle Blessing that shows the real-world value of prizes and calculates how much your spin is costing you; all of the sudden, it sounds a lot less enjoyable.
In one form or another, each of these techniques – virtual reel mapping, clustering, the abstraction of costs, etc. – is present in both Rylai’s Battle Blessing and Lina’s Battle Blaze. And while this isn’t meant to be taken as a 1:1 comparison – both minigames pay out a prize, however modest, every round, and the prizes are virtual goods rather than a common currency – it’s still worth thinking about the design decisions that underlie the games, the history of these design choices, and the psychological processes that have made them so effective in the world of machine gambling.
The boundary between minigame and gambling gets rather porous indeed.
For now at least, no one can seriously claim that these chance-based minigames are anything but harmless. Still, Valve has a strong incentive to ensure that future compendiums and battle passes generate more and more revenue to maintaining the perception of Dota 2’s endless growth. Whatsoever tricks Valve might someday call upon to better dredge the wallets of Dota 2 fans is, of course, an open question. But, if the Rylai’s Battle Spin and Lina’s Battle Blaze are any indication, they, like many other game companies, will likely look to the world of machine gambling for inspiration. Imagine being able to purchase individual drops and spins, win ever more elaborate (and improbable) prizes, or increasing the number of chances to play to the point where a systematic program of intermittent reinforcement is feasible. At some point, and insofar as you have to pay to play, the boundary between a benign minigame and bonafide gambling gets rather porous indeed.
But in the meantime, if you’re a Dota 2 player and you feel like risking some money, I hear Team Secret is looking for recruits. Kappa.