Fairness in Games, Metagames, and the “Real” World

February 4, 2012 at 17:16 (Game Design, Law, Philosophy, Politics)

Leave it to a stimulating conference to bring me back from thesis-induced blog-cation. Two weeks ago, the Peter Wall Institute hosted an exploratory workshop on fairness, and I was most fortunate to attend as a graduate student commentator for a panel on fairness and economic advantage. In the course of reading over the papers for the panel – an eclectic assortment of pieces ranging from insolvency law to the Eurozone crisis to some of the work I’ve helped carry out looking at attitudes towards distributive justice issues in cognitive enhancement – I perceived several recurring themes that I drew out and elaborated on in my commentary. One in particular, it seemed to me, emerged frequently throughout the rest of the conference as well: thinking of fairness in terms of games and rules, especially with an eye to zero-sum vs. positive-sum games, and often with similarities to the Prisoner’s Dilemma or other game-theoretic scenarios. In particular, some of the discussion later in the conference that happened to deal with fairness in sporting events set me thinking about how this approach to conceptualizing fairness might be quite concretely useful when redeployed in the realm of rather more “serious business.”

Let’s begin with an example – a clear, if not exactly commonly occurring, example of an unfair game. Imagine a marathon about to begin. The various competitors are poised and ready at the starting line … but one of them is not on foot. He’s sitting in a Formula One car. It should be clear enough that this is a pretty terrible marathon that none of the on-foot participants will be altogether keen on going through with. But though the example is whack-you-over-the-head obvious in its unfairness, things get a little more interesting when we do some proper philosophy and try to clarify just what about the nature of the situation is constitutive of its unfairness.

I think we can say at least two things on this topic that will turn out to have sufficient generality. The first crucial feature is roughly this: that the participants are, in effect, no longer playing the same game. The racers on foot are playing “how fast can you run to the finish line?” The racecar driver is playing “how consistently can you floor the accelerator?” (Put down the pitchforks, NASCAR fans; I know there’s more to it. Simplifying here.) So, in a fundamental way, the possibility of it making much sense to evaluate the competitors’ performance in achieving the goal of the game has broken down; what’s going on in each case is just so different that what it is to succeed for each competitor is no longer recognizably similar. I tend to feel that this is a rather powerful and useful way of thinking about fairness in games, especially in sport, because it helps boil some of the problematic normative components away, leaving room for disagreement only at the conceptual or empirical level. Should Oscar Pistorius, “the fastest man on no legs,” be allowed to compete in a given event alongside able-bodied runners? If blood doping with EPO were perfectly safe, would there be a leftover reason to bar cyclists from using it if they so please? Instead of appealing to likely ill-suited moral language or high-minded pablum about the “integrity of the sport,” one could make these types of discussions much more productive by asking “what does the game look like from the perspective of each of the players, and how interchangeable is that perspective?”

Of course, it is not as though we must reject every instance in which the game differs for different players; tennis for Patrick Rafter (nimble master of the serve-and-volley) was never the same as tennis for Bjorn Borg (effectively a human cannon from the baseline).  But it is the difference between what tennis is for me (in this example, having to consciously think through a whole array of rules, techniques, and strategies that are not second nature to me – to say nothing of the vast differences in effort required relative to physical ability) versus what it is for Borg that would make a singles match between us completely unfair.

The second, and arguably more important, feature that we can identify as underlying (or at least bound up with) the unfairness of games concerns what I’ll refer to as the metagame: the evolving set of strategies and skills that define optimal play and influence the dynamics of balance within a game. (See also metagame analysis.) In the case of our Formula One marathoner, it should be quite clear that the metagame for this marathon has shifted to a place that is uninteresting or worse. Namely, the metagame has re-centered around what equipment you bring to the track, which probably in most cases boils down to how much money you’re willing to throw at the game, and that’s not very fun or interesting, to put it mildly.

Moreover, this kind of shift in the metagame directly subverts the reason we were playing the game in the first place. While I am not exactly an avid sports fan, I’m still happy to wax poetic about why sport is such a big deal: it’s a stage and a medium for the cultivation of various kinds of excellence that matter to lots of human beings. To be perfectly Aristotelian about it, these are aretaic pursuits whose value (to the rest of us not playing) is largely aspirational, and hey, that’s pretty cool! So of course we have reason to worry when the metagame of a sporting event shifts in a way that centralizes anything but the diligent development and skillful deployment of natural ability, both strategic and physical. If tennis games are ever won sufficiently largely on the basis of racquet selection, the whole point would be gone – unless, of course, we re-centered the metagame around the more interesting and desired elements by providing everybody with this same wonder-racquet. At that point, tennis would have become something of a different game, but it would be different in the same way for everyone. (In this way, I disagree with commentators like Sandel, who maintain that, e.g., genetic enhancement or doping necessarily corrode the cultivation of excellence in sport. I would demur that they merely threaten it – by way of re-centering the metagame and hence unbalancing the game – but only if unevenly distributed. Create a “steroid league” for baseball that uses some impossibly hypothetical safe and cheap steroid, and nobody has any leftover grounds to object.)

One interesting dynamic I would highlight here is that the relationship between the fairness of a game and the quality of its metagame is rather tight. (I might not even disagree if someone argued that these concepts were two sides of the same coin.) We just went over an example of how a shoddy, unbalanced metagame (in which tennis match outcomes are largely decided by racquet choice) was decentralized and thereby improved by standardizing a particular element – by making the game more similar for all players. Still, not all metagame problems stem from dissimilarity amongst different players’ experiences, and accordingly not all metagame imbalances are rectified by standardizing them. Take the case of competitive Pokemon back in the original Game Boy generation: for a while its metagame centered strongly on dealing with the ability of some characters to boost their evasiveness, which frequently resulted in protracted arms-race-like stalemate battles in which neither player could land an attack on the other and each was largely concerned with neutralizing the other’s evasiveness advantage. This made for a fair game (once you deduced the state of the metagame and built your team around evasion-boosting), but a pretty shallow and annoying one. The quality of the whole experience was concretely improved by banning use of the “Double Team” ability responsible for the evasiveness arms race.

You might think of it as analogous to how a game of chess can center around neutralizing the threat of the opposing player’s queen and most effectively deploying one’s own, only to a much greater extent, one at which chess stops being about “checkmate” and becomes all about “garde.” (Or, as my roommate once put it while playing a familiar console fighting game, “This isn’t Super Smash Brothers anymore. This is Super-Play-As-Meta-Knight-Or-Lose-To-Him-Brothers.”) That all being said, the most surefire way to shoot the metagame straight to hell is to offload too many of the factors that dictate the game balance dynamics into the realm of things already determined before the game begins. If some competitors show up already having lost in virtue of their failure to stack the deck in their favor – to create, in effect, an easier version of the game for themselves – then the whole point of having everybody play by ostensibly the same set of rules has been thwarted. Especially when you consider that winning at the game of “stacking the deck in your favor” often has very little to do with the kinds of excellence the actual game is meant to cultivate and much more to do with leveraging unrelated – often unearned – advantages.

But what, you might ask, does any of this have to do with fairness in “real life?” Actually, if you ask me, this kind of analysis is every bit as useful in the design of real-world institutions and regulations as it is in the crafting of fun and balanced games. After all, these things share important features. Without running afoul of Wittgensteinian concerns about the ineffability of a family-resemblance cluster concept like “game,” we can still safely say that much of what we affix that lexical label to can be described with terms like “rule-bound,” “goal-directed,” “collaborative or competitive or both,” and “metagame-generating.” And all of these terms are equally applicable to vast swaths within the spheres of human activity, especially those involving markets and those involving law. Indeed, we can think of markets in particular as uncannily game-like: where sports function as aretaic engines of human excellence, markets (ideally) function as utilitarian engines of efficiency and wealth. They are rule-bound, with some of the rules emanating from brute game-theoretic facts as apprehended by boundedly rational agents, and others being imposed top-down via state regulation. They are goal-directed, in the sense that individual agents go into them with specific objectives that don’t necessarily have anything to do with the overall justificatory basis for the market’s existence as such. They are sometimes collaborative and almost by-definition competitive. And they quite clearly have metagames – devilishly intricate ones, for that matter, the kind that very clever people are paid staggering sums of money to deduce and exploit.

Extending the analogy even further by way of bringing up a frequently-revisited theme from the Peter Wall conference, the pathology of unbalanced games and badly re-centered metagames is as baneful to the constructive operation of markets as it is to that of sports: one need look no further than the recent financial meltdown to witness what were intended to be great turbines of prosperity and opportunity running off the rails and exploding in our collective face, to the benefit of only a privileged few. The dynamics are all familiar. The root of the problem in each case, for instance, is not so much the presence of “bad seeds” – participants in a given game will always seek to optimize their play, and invariably some succeed in figuring out the best strategy for winning even at the expense of other players and of the whole point of the game. This seems to me exactly what was going on with many of the practices, e.g. of predatory lending, toxic asset shenanigans, brazen executive compensation, et cetera, that are rightly faulted for the economic crisis. We allowed a system to develop in which this was, for a given competitor, what optimal play looked like; while this of course is not meant to be some flat-footed moral absolution of corporate greed or Gordon Gekko values, it speaks to the fact that we ought to approach our financial systems like game designers – with a careful eye on the metagame – because there’s no limit on what people will do when they’re channeling Charlie Sheen and fixated on “duh, winning!”

The other crucial component of that analogy is the role of background conditions of fairness in creating a game that plays similarly for all competitors, since, as we’ve seen, a metagame that has been carefully crafted to operate in a positive-sum manner can easily be shifted to any number of more easily achieved zero-sum equilibria if the conditions of the game become unfair. Too often in today’s financial system, we see ample opportunities for shysters to take advantage of suckers. What Panglossian free-marketeer fundies so often miss about this kind of situation is that the conditions under which markets live up to their “rising-tide-lifts-all-boats” existential justification are not axiomatically guaranteed simply by virtue of their being markets. Those positive-sum effects are the emergent properties of very specific metagame dynamics, and as we have learned, those dynamics are fragile and especially vulnerable to unpredictable re-centralization, which in turn is so often a product of unbalanced dynamics of play between participating actors. To put it in the plainest English possible: a game that one player (or class of players, in this analogy entities like hedge funds and investment banks) can reliably win largely in virtue of what they bring to the table is extraordinarily unlikely to be a game worth playing from the point of view of society. It is not going to benefit everyone, and therefore the justificatory pretext for its existence has evaporated: it has become a culturally legitimated stage for nakedly predatory, dog-eat-dog dominance contests. Few things could be more counterproductive to the large-scale societal goals that most people can agree to value – fairness, justice, the common good. Maybe some folks are not so enamored with those high ideals and would prefer a more “law-of-the-jungle” setup – but at least, on this view, we can call a spade a spade and characterize that as fundamentally defecting on the basic prisoner’s dilemma of human social existence.

One big counter-argument against taking steps to rectify unbalanced conditions of background fairness in economic “games” invokes the notion of desert. It runs something like, “look, everybody has got enough basic rational self-interest that they ought to be expected to figure out how to succeed in whatever financial venture they freely choose to enter into, and it is fitting that anybody who fails to play the game at a winning level should have to suffer the consequences.” This actually breaks down into two claims, one empirical, the other normative. I won’t waste any time carving up the empirical one – the old chestnut of homo economicus, and its quixotic kissing cousin characterized by a refusal to acknowledge the variation in people’s opportunity and ability to actually optimize their financial strategies – as these have been dead horses since Daniel Kahneman. The normative claim is more interesting: why not accept the distribution of winners and losers created by the operation of a given market as morally appropriate? Again, the analogy with more familiar types of games comes in handy: we see a kind of fittingness to the outcome of a sporting event only when, and only to the extent that, the game was balanced. Moreover, this fittingness can only legitimately underwrite the distribution of benefits for winners if the metagame was actually producing some worthwhile emergent effects: nobody would feel particularly inclined to greatly praise or reward the winner of a Scrabble tournament in which all (or some) of the players used a smartphone app that computed the optimal move in every case from the current state of the game board. Q.E.D. – when the whole purpose of a game-like activity is subverted by the exercise of an optimal strategy, the whole conceptual substructure that underwrites the validity of these notions of “deserving” to win or lose is totally undone. Put another way, if blaming and humiliating the individuals who were roped into subprime loans actually did a damn thing in the world to facilitate a smoother-running system on the whole, then we’d have a defensible rationale for doing so. As it stands? Not so much.

Now, I have jawed on quite a bit about economics in this post – somewhat at my peril, being admittedly a naïf in this area – but where I am really interested in applying these principles in pursuit of better systems design is actually law. (This couldn’t possibly be at all related to the slew of law school applications I recently sent out, no siree!) Games are so often defined by their rules, and the law is in the unique position of being able to impose rules on just about any kind of human activity, especially the game-like kind. One could imagine this approach being fruitfully applied in many legal areas. In criminal law, how do we set up rules and structure consequences in a way that effectively and fairly modulates human behavior for the better? In tax law, how do we craft a metagame where the optimal strategies are actually conducive to the public goods that tax codes are meant to incentivize? In constitutional law, how do we most prudently demarcate the boundaries of the democratic legislative process? In general, how do we cobble together institutional systems in which the metagame dictates that any optimal strategy favors positive-sum collective benefit over free-riding or zero-sum competing? All of these questions are systems design questions: their simplified analogues are tackled routinely on a smaller and lower-stakes scale by the people who plan out the mechanics of a Magic: the Gathering expansion or decide on balancing tweaks for a World of Warcraft patch. My suggestion is that we can fruitfully import principles from the creation of good games in order to make such questions more tractable.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: