Analysis: 'Socially Awkward' - The Return of Skill

[Continuing his eccentric series on social games, game designer Patrick Dugan predicts that skill games, though seemingly "banished to the desert of freeware and table-top conventions", could make a triumphant comeback. Here's "Socially Awkward VIII".] A friend of mine used to have a mohawk. He used to date daughters of Columbian drug lords, and now lives with his girlfriend in a cramped, one bedroom apartment. By day he runs teams in a virtual office, producing everything from slot-machines to Facebook games. The mohawk is gone. But one day he became a living legend in World of Tanks, a very hardcore, freemium game based on a 3D downloadable client. The majority of this game's revenue comes from speed-ups and 120 percent damage ammunition. They are apparently doing quite well off a high activity, steady niche of a few hundred thousand daily players -- which is beside the point. The point is he needed no special ammo to face down six enemy tanks when his entire team was destroyed and single-handedly win the match. He was able to do this because he has sociopathically honed his brain over thousands of hours of competitive gaming across genres, over hundreds of hours spent studying military history for fun, because of situational awareness to twist his tank's geometry so no opponent could score anywhere near a perpendicular hit in the time intervals it took him to aim and fire. And so he picked them off, with 2 percent health remaining, singing Panzerlied into his VOIP headset with manic baritome:

He would be forever known to the 15 other people he out-classed that day, out-classed in every way, their minds marked indelibly, as if by branding iron, that this individual is an Ubermensch. As he would say: soooo eepic! Nintendo's $1 Billion Dollar Gorilla In 1981-1982 Nintendo made over $180,000,000 selling Donkey Kong cabinets. Adjusted for hyperbolic, late-civilization inflation, we're talking about something on the order of half a billion 2010 dollars. That was Nintendo's gross revenue selling these machines for four grand a pop. Ostensibly the proprietors of ice cream parlors, bars, pizza shops, family restaurants, dedicated arcades and souvenir shoppes had to collect more than 16,000 quarters to make a profit on their investment, so the total gross of the game is probably a lot more, maybe a billion 2010 dollars. We look at top grossing social games and marvel at a half billion in gross revenues. We speculate that, since its in Flash, the operating margins must be huge, more than some clinky physical machine. Zynga's S-1 reveals that, perhaps the margins aren't SO awesome but still pretty good. At least this time around we're saving trees by caking our margins with the cost of virtual goods rather than physical ones. A key difference is that nobody will ever make a documentary about an epic struggle between two people trying to master FarmVille. Note I didn't use the term "compete"; I understand that instead of frustrated house husbands competing against megalomaniacal hot sauce mavens, it's frustrate house wives co-operating with megalomaniac friends from high school who just seemed to drift away but still need horse shoes. The distinction is that in FarmVille, you win for showing up; it's like games took a page from the U.S. public school system. Here's an "A" for effort. I've been modeling games like FarmVille for about a year, and when I say "like FarmVille", ho ho ho, let me tell you, I mean that quite literally. At least, that's how it started; I've since taken baby steps toward more complex and interesting games, but still games where the model calculates a deterministic progression over time and the only variable is how often people come back to click stuff. Yay. A Game Of Skill The word "skill" has not been mentioned once by any speakers in any social game-related conference material I've heard in the past two years. Anecdotal, sure, so let's see what comes up on a Google Trends query of "skill games". There was an uptrend through early 2009 as momentum built for "skill-games", that is games that monetize from people betting on their own performance, the subject of my second attempt at entrepreneurship. After 2009 though, whew, the term seems to have been pinned down to the mat. Before there was Google Trends, "skill" was a word you could count on two hands and ten toes in any enthusiast publication, middle-school cafeteria conversation, or even here on Gamasutra. It had a certain sanctity to it, now banished to the desert of freeware and table-top conventions. But what if skill were about to make a triumphant comeback? It's easy to forget that once people made games to challenge you from the frontal lobe down to the lizard brain, ludology recapitulating phylogeny, and those games made bank just as much as any latter day, watered down, behaviorist lever-contraption. Bejeweled Blitz is still, a year after launch, the #1 Facebook game in terms of the ratio of daily to monthly - shall we say - players, and its skill is just a pattern recognition exercise that lets you roll dice as many times as possible with a less impactful adjunct skill of recognizing and prioritizing multiple patterns. Not to allude that richer skill-based Facebook games would clean the competitions' big black clocks, though you could do worse than to Pig Up yourself. Clearly the way forward here is multiplatform and mobile. I got as high as I could go in the established milieu and then looked for new opportunities, an excuse to get paid to research a new way of balancing a free-to-play economy around deeper, more skillful gameplay. It is indeed possible to model with enough precision how variations in a game, including both consumable and permanent paid-items, will affect the probability distribution of outcomes, and then you can use post-hoc data to refine the constants you use for different skill levels and gauge how quickly people's brains adapt along the logarithmic scale from initiation toward mastery. The cliche'd phrase, "easy to learn, hard to master" conforms to that kind of logarithmic advance, it's just a question of what your "r" is in e^(rx). Let's Be Discrete Modeling real-time games is more complex than games that wait patiently for the next server message; you have multiple discrete events occurring simultaneously and making a mathematical impact on the underlying game state, whether that be measured in points, resources, levels, whatever. I started with a fairly simple, arcade-style game, and intend to try more complex tactical combat games, logistical games, and racing games in the future. I'm wondering when I'll break the limits of what can be done in a spreadsheet and have to start coding abstract simulations from scratch. Once you've got your idealized simulation, you can whip out some handy, off-the-shelf spreadsheet functions and get a probability distribution of player performance. There is an unspoken axiom that the unit of activity being simulated will be repeated a large number of times (to allow mean-reversion to make meaningful any probability distribution derived) and there are explicit axioms of performance constants for each skill benchmark you're using, which you have to verify with post-hoc data. Most scrappy indie developers with a few Android devices lying around don't have the resources to even design their games with an essential server-side component, much less the matrix-esque data infrastructure of large social game developers; this is why I think the good 'ol fashioned (i.e. 2006), Nicole Lazarro-style playtesting lab is going to make a comeback. You just need to run 1200 people or so through a live test of the game, set an egg timer, and measure precisely how often they do each crucial thing -- I'm getting 1200 from 4 skill benchmarks multiplied by 300 people required to get a statistically robust sample. You can roll that data up in the 4-6 weeks of closed Beta for the cost of bulk-rate Coca-Cola, just get the QA guys to do it. Naturally they'll try to recruit interns from the local high school to both run the tests and submit to them, but fuck it, you're practically running Vertica. Trust Your Brain Most people don't bother with the quantitative stuff so much and chalk it up to intuition, this is because human intuition is more or less the above process of honing in on something through a vast, probabilistic sea. There's no reason why one can't augment the other, and actually playtesting the game yourself should be the guiding reference through the whole process. A key insight that came to me the other day while waiting for a Peruvian-style sandwich to be served to me with fries in the middle along with an assortment of mayonnaise variations, is that there should be an inverse proportionality between the Kurtosis of the probability distribution of your skill-gameplay's outcomes and that of LTV. Instead of wasting space trying to verbally describe what the above picture does almost instantly, I'll just say in plain english that for a casual game, you want most people up on the plate, scoring either the average, a bit below average, or a bit above average, so you can sell them on breaking free of the plate. For a more competitive game, you want a thin slice of your population batting average, and a steep plain of highly skilled contenders duking it out, always facing someone better then them around the corner. Either way, you surely want the distribution of your player's life-time-value to you having a leptokurtic distribution; imagine the y-axis is the portion of your audience and the x-axis is a log scale of how much money a player funnels into your accounts receivable, either through paying directly or sending signals that motivate others to come back and pay. Because the middle part is so skinny, the tails of your probability distribution are fattened like turkeys waiting to finally be appreciated on an upcoming holiday. So you've got a big chunk of people not contributing anything, so what, who gives a doodle, it's a free-to-play game, that's part of the deal. In the middle you've got a high probability of people contributing a few bucks, not much, but hey, it's something. Then on the right you've got this other fat tail of people paying $10, $100, $1000, and they aren't .01% of the audience like in these skill-less social games we've been seeing, they're more like .1% Leapin' langosta! Doesn't seem like an impressive number, but booking ten times the monthly pull per active player as your average click-timer game and twice as much as your average casino is what we're talking about here. There's good money in making good games! Hope For Us All I just made the business and math case to sedate all the 800-pound gorillas in the audience, let's talk about something worth being passionate about. It is possible to have your cake and eat it too, no longer does indie have to mean naively exploitable with self-isolation being the only defense. I mean, why even have a cake if you can't eat it? You're really going to let the gate-keepers at XBLA/PSN/WW/FB/AAS/GAS/Steam neglect your distribution and eat 30 percent of the scraps? You can make the game you want, a deep game, a game with power, and you can go out and connect with hundreds of thousands of people who love your game specifically, and you can rake it in the way indie hits do without needing to be such a rare hit. You just have to have something good enough that some niche can love it, use the free-ness to allow like-minded folk to promulgate it amongst themselves, and use statistical analysis to balance your up-sell just right without needing big boy infrastructure. We're talking about another Cambrian explosion of innovation, akin to the manic experimentation of the early-to-mid 80s: free-to-play, on mobile, full-screen experience, interact with it wherever you go and on the web, strike up some multiplayer with your friends or suitably matched strangers, no DRM malware getting in the way, no ads in the corner of your eye. Just good fucking gameplay made great by all the new advantages available to game developers. It doesn't matter what demographic you belong to, we all have brains that adapt patterns and learn, that hone in on things, that get conditioned into habit by dopamine release schedules. One thing that I've tickled at being able to model but probably never could, is burn-out. How do people "consume" content, how do they grind through novel variations in information, how does their perception of novelty warp and shift over time? I mean, I thought WoW would peak years ago and I called tops in this manufactured stock market rally for a couple years as well. Don't underestimate human stupidity; it's the greatest power source since solar. On the other hand, especially for those in power, it's all too easy to be cynical and underestimate human intelligence, our capacity to overcome, to prevail, to make radical jumps in what we're willing to put up with and in what we can understand. Dictatorships seem stable until they don't, and I believe the probability distribution of how people capitulate with the current crop of popular games will also go leptokurtic over time. Even if it's just a few percent of the audience, there's going to be a lot of fat tails to chase, a lot of particular niches to be cultivated, a LOT. How many 100k DAU games could be supported? There are going to be several billion people in the total global audience within the next few years, so I would say at least one or two thousand. One or two thousand unique games that don't step on anyone's toes, that do their own thing, that provide a very nice profit to their creators. It will take skill to make these games, it will invite skill to play them, and maybe somewhere along the line, people can actually augment their consciousness while being entertained. Crazier things have happened, a game about a plumber timing jumps to confront an 800-pound gorilla once made a billion dollars and decades later gave meaning to a middle-school teacher's life. If there were no mountains to climb, we'd have to invent them. [You can also read the rest of Dugan's series about Metal Gear Solid 2 predicting Facebook, the industry's Wall Street envy, the joy of vector meme, peripheral visions and dreams, metrics, F*book, and mobile social local games]