Abstract
The New York Attorney General filed a lawsuit against DraftKings and FanDuel in 2015 accusing them of operating illegal gambling platforms. Using actor–network theory, we show how DraftKings, FanDuel, the New York Attorney General, critics, and legislators were preoccupied with how much agency players possessed. They also understood agency as emerging from a sociomaterial arrangement of human and nonhuman entities and saw how agency could be enhanced or limited by introducing new rules, technologies, and constraints. We compare this perspective on agency with theories typically used in sociology of sport and we consider how sociologists can intervene in sport agency.
Introduction
Questions of agency are foundational to sociological inquiry in sport. Agency is associated with the existence of free will and self-determination; it is the cause of social change; and agency presupposes peoples’ ability to make a difference in the world. It stands as a metaphysical hinge for debates around subjectivity, structuralism, and human action. Agency is the justification for critical inquiry into sport and physical culture because critique is based on the potential for change, reconciliation, and betterment (Sage, 1998). Indeed, agency is central to most sociological theory—including C. Wright Mills’s (1959) sociological imagination, Anthony Giddens’s (1979, 1984) theory of structuration, and Bourdieu’s (1977) habitus.
Despite the centrality of agency to social theory, commentators have noted that many researchers do not adequately define agency (Ahearn, 2001; Emirbayer & Mische, 1994). Even when agency is defined, there has been substantial debate over the theoretical foundations of agency (Ahearn, 2001; Sewell, 1992). Sociologists of sport, for example, tend to identify agency where actors are resisting social structures, norms, or ideologies (Cooky, 2009; Laurendeau & Sharara, 2008). Dworkin and Messner (1999) distinguished between reproductive agency and resistant agency, where resistant agency involves transforming unequal or constraining structures and reproductive agency involves using a structure to benefit oneself while reaffirming power relations. However, Emirbayer and Mische (1994) argued definitions of agency based on structure lead to impoverished conceptions of agency and failed to distinguish its affirmative characteristics and Abu-Lughod (1990) cautioned against the “romance of resistance” generally. Sociologists have also been critical of social scientists’ tendency to define agency as synonymous with free will, noting human efforts and intentions are shaped by historical norms, practices, and discourses (Ahearn, 2001).
In this article, we use actor–network theory (ANT) to examine agency as expressed in the legal debate over daily fantasy sports (DFS). In 2015, the New York Attorney General filed a lawsuit against DraftKings and FanDuel and the House of Representatives Committee on Energy and Commerce held a subcommittee hearing on DFS. DraftKings, FanDuel, the New York Attorney General, the subcommittee, experts, and critics debated the nature of fantasy players’ agency as evidenced in skills, control over the outcome, and unbalanced abilities. We treat this lawsuit as a critical case to examine how agency is contested in the sport industry (Flyvberg, 2011). Although ANT is often used to make theoretical arguments about the existence of nonhuman agency (Esmonde, 2018; Weedon, 2015), we argue ANT scholarship can be seen as methodological propositions for studying agency generally (Sayes, 2014). Following Latour (2005), we use ANT as akin to empirical metaphysics, allowing scholars to examine how the basic components of the world such as truth, reality, or—in this case—agency emerge from the daily practices of scientists, lawyers, and software engineers. Our case study shows how agencies were configured in DFS, which will help scholars reconcile theoretical debates over the nature of agency. We also create theoretical and methodological insights serving those hoping to harness agency for institution building, social movements, and political causes.
Daily Fantasy Sports (DFS)
DFS evolved from season-long fantasy sport leagues, which have existed since the 1960s (Hruby, 2013). Season-long fantasy was traditionally played by groups of friends, who would draft teams at the start of the season and play through to the championship. In season-long fantasy, participants earn points according to the performance of athletes on their team, who are playing in real events, but may be distributed across many different real teams (Edelman, 2012). The goal is to accumulate more points than one’s peers by drafting a team of athletes that perform better than others. The fantasy player can win even if the real teams his athletes are on lose, because his or her success is based on a collection of individual players’ performance statistics.
The internet made fantasy sports easier to manage and fostered larger leagues. In the United States, the internet also became a tool for offshore poker and sports betting sites, leading Congress to hold hearings concerned with internet gambling in 1997. In 2006, Congress passed the Unlawful Internet Gambling Enforcement Act (UIGEA), which targeted gambling companies and payment processers. Presumably due to lobbying efforts from the sports industry and in particular the Major League Baseball Players Association, the UIGEA included an exemption for “fantasy sports.” The UIGEA does not modify existing State or Federal law, so it does not permit fantasy sport in situations where it would otherwise be illegal, such as if it were used for gambling purposes; nevertheless, some started to talk about the “fantasy sport carve-out” as though it were a loophole to prohibitions on sport gambling (Holden, 2018). In 2010, FanDuel launched a series of competitions under the moniker of “daily fantasy sports.” DraftKings followed shortly after and emerged as their main competitor (Edelman, 2016).
Daily fantasy games differ from season-long formats because players pick lineups for games that will be played during the week or even the same day. Players typically pay fees to enter their lineups into competitions with predetermined prizes. For example, “Head-to-Head” matches pit one player against another, with the winner earning a prize equal to the entry fees minus a portion kept by the company. In “50/50” games, entries finishing in the top-half of the field win cash prizes. “Guaranteed Prize Pool” contests pay an entire predetermined prize pool and typically have thousands of entries (Ehrman, 2015). Players pick a lineup of athletes with artificial salaries subject to the constraints of a salary cap. Like traditional fantasy, they score points when their athletes perform well on the field.
The creators of these games initially explained they were designed to blend traditional fantasy with sports gambling, based on assumed protection under the UIGEA carve-out (Edelman, 2016). Jason Robins, the CEO of DraftKings wrote, Our concept is a mashup between poker and fantasy sports. Basically, you pick a team, deposit your wager, and if your team wins, you get the pot. Fantasy sports has a carve-out from the 2006 gambling regulation because it’s considered a game of skill. This concept where you can basically “bet” your team will win is new and different from traditional leagues that last the entire season. (quoted in Bird, 2015, p. 1)
Although Robins and other industry leaders later distanced themselves from gambling (Holden, McLeod, & Edelman, 2020), other commentators and stakeholders saw DFS as a thinly veiled attempt to skirt gambling laws. In October 2015, the New York Attorney General’s office launched an inquiry into DraftKings and FanDuel. Soon after, his office filed a lawsuit accusing the companies of illegal sports betting and fraudulent advertising. Officials from other states followed, with letters from Illinois, Hawaii, Mississippi, Nevada, Tennessee, and Texas declaring daily fantasy illegal under their states’ laws. The New York lawsuit, however, became a pivotal trial for determining the legality of daily fantasy sport.
Although gambling laws differ by state, lawmakers typically define gambling as consisting of three elements: a prize, some degree of chance, and a bet or payment. Contests such as daily fantasy, which have an entry fee and a prize, are illegal if the degree of chance is more than that permitted by state law. States use different tests of chance (Holden & Brandon-Lai, 2017). The dominant factor test considers a game lawful if a player’s skill is at least 51% responsible for a game’s outcome; the material element test, on the contrary, only tolerates chance that is less than a “material element” in the game’s outcome, and some states disallow any chance whatsoever. Thus, in New York, and across the United States, the argument around DFS turned to whether players won due to the degree of skill they exercised. Moreover, in states like New York, which uses the material element test, it became necessary to establish exactly how much of the game was determined by skill. Thus, the courts, regulators, and companies debated the nature of players’ abilities and how much players’ abilities affected the outcome of daily fantasy competition (Holden et al., 2020).
Using ANT to interpret the legal argument around daily fantasy sport, we suggest that players’ agency became a matter-of-concern to companies, regulators, and the public. From a sociological perspective, these legal debates about daily fantasy hinged on (a) how much control the individual participant had in making choices or taking actions that might lead to certain outcomes (winning or losing money, for example) and (b) how the governance structures protected or should protect those with limited agency. Whereas previous studies on DFS and gambling examined sports gambling-related social risk (Frey, 1991; Smith, 1990), motivations (Curry & Jiobu, 1995), identity formations and politics (Davis & Duncan, 2006), market activity (Lopez-Gonzalez et al., 2017; Lopez-Gonzalez & Griffiths, 2018), and negative social outcomes (Martin & Nelson, 2014), this case provides an important opportunity to extend analyses of the interactions between state and individual actors. It provides a unique case to rethink and reconceptualize broader social, political, and juridical definitions of human agency within sporting contexts. Thus, we use the New York lawsuit to examine how various actors enhanced, contested, and regulated a new type of agency.
Literature Review
Agency in Sociology
While theories of agency are diverse and hotly debated, most sociologists agree “theories of practice” are the best starting points, especially Giddens’s (1979, 1984) theory of structuration and Bourdieu’s (1977) theory of practice (Ahearn, 2001; Emirbayer & Mische, 1994; Sewell, 1992). Practice theories explain how human agency is influenced by social structures and may even reproduce social structures. Practice theories describe agency as habitual, repetitive, and taken-for-granted (Emirbayer & Mische, 1994). And they appeal to sociologists’ skepticism regarding the existence of free will while acknowledging the minute ways people contribute to the world around them.
Giddens (1979, 1984) saw structures as “dual” in the sense that structures shape people’s practices while also being the outcome of human action. Giddens argued human agency and structure presuppose one another. Knowledgeable actors derive meaning and purpose from social structures so that, in acting, they put into practice necessarily structured knowledge. As a consequence, their actions reinforce social structures. Agency is not an expression of free will, but instead an act made possible by structure and constrained by it, although people’s actions also change social structures.
A similarly recursive relationship between structure and agency can be found in Bourdieu’s (1977) habitus. Habitus is a scheme of thought and action people acquire while being socialized into a class location: “As an acquired system of generative schemes objectively adjusted to the particular conditions in which it is constituted, the habitus engenders all the thoughts, all the perceptions, and all the actions consistent with those conditions and no others” (Bourdieu, 1977, p. 95). Habitus generates an infinite but bounded selection of actions and perceptions, each imbued with culturally constructed meanings derived from socialization into a given class position. Or in Korp’s (2008) words, “habitus expresses the idea that people make choices, but their choices are always constrained and influenced by the resources and social identity of the actor” (p. 19).
Whether drawing on Giddens and Bourdieu or other theorists, sociologists studying sport often echo their conclusions when discussing agency. Sociologist of sport show how participants, coaches, and spectators are agents who both resist and reinforce ideologies, discourses, and institutions (Cooky, 2009; Dworkin & Messner, 1999; Esmonde, 2019; Laurendeau & Sharara, 2008; Singer & May, 2010). Such observations are warranted because sociologists often focus on how people interpret and act within broader social forces that structure sport. However, critics have argued that the interweaving of agency and structure fails to distinguish agency as an analytical category in its own right. For example, Sewell (1992) criticized Bourdieu’s habitus for retaining the “agent-proof quality that the concept of the duality of structure is supposed to overcome” (p. 4). Emirbayer and Mische (1994) critiqued the “flat and impoverished conception” of agency that remains “so tightly bound to structure that one loses sight of the different ways in which agency actually shapes social action” (Emirbayer & Mische, 1994, p. 963). Therefore, alternative approaches, not based on the determinacy of social structures, may help sociologists comprehend agency and harness it for change, resistance, and other political goals.
Agency in ANT
Although ANT and the writings of its main protagonists (Bruno Latour, Michel Callon, and John Law) cover many important, controversial, and well-debated topics, we focus on the key concept of an actor–network and explain its implications for studying agency. In so doing, we follow a number of scholars contributing to this and other sport sociology journals in applying Latour and colleagues’ actor–network approach to the study of sport and physical culture (Darnell et al., 2018; Esmonde, 2018; Esmonde & Jette, 2018; Kerr, 2018; Kerr & Obel, 2018; McLeod et al., 2018; Weedon, 2015). The actor in actor–network can be thought of as a termination point: an actor borrows action from a host of entities that innervate it. Because an actor is equivalent to its innervations, both in terms of what it is capable of doing and in terms of its being, any given actor can only be described by also tracing the entities that associate with it, as though these entities formed a network with the actor at the center (Latour, 2005). The actor–network concept allows us to inquire into agency by tracing the network of entities that make agency possible. Entities are usually also actors in their own right, with their own traceable networks.
The actor–network concept introduces four changes to how sociologists think of agency. First, whereas social scientists often use agency to mean causal agency, ANT scholars use agency to study a broader, more inclusive range of capacities. Latour (2005) described agency as anything “making some difference to a state of affairs” (Latour, 2005, p. 52). According to Sayes (2014), this is a “minimal” conception of agency because “it catches every entity that makes or promotes a difference in another entity or in a network” (p. 141). For example, Latour (2004a) wrote, “there might exist many metaphysical shades between full causality and sheer in-existence: things might authorize, allow, afford, encourage, permit, suggest, influence, block, render possible, forbid, and so on” (p. 226). ANT includes all these verbs and more under the term agency.
Second, because ANT uses a minimalist conception of agency, it also accepts nonhumans have agency. Examples include larvae that anchor to devices (Callon, 1986), road bumps that slow motorists (Latour, 1999), microbes that kill sheep (Latour, 1988), and cars that only start after fastening one’s seatbelt (Latour, 1989). Some have critiqued ANT for suggesting inanimate objects have human characteristics such as intention or free will. However, scholars like Latour (2005) and Callon (1986) simply argue that nonhumans frequently make a difference in human projects. Thus, rather than anthropomorphizing objects, it is more accurate to see ANT as deanthropomorphizing agency by separating it from uniquely human capacities for abstract thought, reflection, and free will. A network conception of agency also tempers the assumption that human agency is exceptional by showing that humans depend on gadgets, alphabets, calculators, barriers, and more to do most things considered especially human.
Third, ANT scholars see agency as configured by sociomaterial networks of human and nonhuman actors. Whereas some social theories assume people have more agency when removed from language, norms, and objects, ANT scholars assume agency is borrowed from, or distributed through, one’s network meaning agency can be enhanced or at least changed by rearranging networks (Latour, 2005, p. 46). This perspective shares some similarities with Giddens’s theory of structuration in that it sees other entities as enabling or fostering agency. However, ANT scholars disagree with using structures as explanatory variables in the first place, so they are unlikely to describe agency in terms of its reproductive or resistant qualities (Latour, 2005). Another implication of a network definition is seeing agency as collective. While some sociologists argue for collective notions of agency (Sewell, 1992), ANT scholars were the first to argue the collective is composed of nonhumans as well as people.
Fourth, if agency is distributed across a sociomaterial network of human and nonhuman actors, then agency should change as networks change. Thus, whereas some theorists and philosophers describe agency as essential, universal, and timeless, ANT scholars agree with sociologists who assume agency can change over time, between actors, and across situations (Emirbayer & Mische, 1994). It is better to speak of agencies than agency. Each agency is configured within a given sociomaterial network, making infinite possible configurations, each affording a unique action or ability. For example, the daily fantasy sport market can be seen as a sociomaterial arrangement that configures a special type of agency.
Agency in Markets
Callon (1998) applied ANT to markets, arguing markets are unique arrangements of human participants and nonhuman actors that create a special type of collective agency called “calculative agency.” Calculative agency allows people to make economic decisions such as: “Should I buy this jersey?” “Is this exercise worth my time?” “Which material should I use to manufacture this stadium?” and “How valuable is this athlete?” Thus, and in contrast to the debates that have defined economic anthropology since Malinowski’s (1921) and Polanyi’s (1957) substantive explanations of exchange, Callon (1998) argued people are neither inherently economic nor inherently noneconomic, neither rational nor irrational, neither embedded in social relations nor free, but—when arranged with nonhuman actors in a certain way—people form an agency having the ability and motivation to act as economic theory predicts they should act.
To define calculative agency, Callon and Muniesa (2005) used a broad definition that accounted for qualitative judgment as well as numeric calculations. They wrote, “Calculation starts by establishing distinctions between things or states of the world, and by imagining and estimating courses of action associated with those things or with those states as well as their consequences” (Callon & Muniesa, 2005, p. 1231). Calculation depends on three material steps (Callon & Muniesa, 2005). First, entities to be calculated must be detached from their existing associations and arranged in a single space. Second, the entities must be associated with one another so that they can be transformed and manipulated together. Finally, a new entity must be produced (a sum, a list, a choice, or an evaluation) that corresponds with the original entities. These steps are neither purely human nor artificial; instead, they are an achievement made from the careful arrangement of humans and things.
Markets, therefore, are sociotechnical arrangements, which create certain types of agency and, in doing, delimit a space of confrontation where participants compromise over the exchange of goods and then leave (Çalışkan & Callon, 2010; Callon & Muniesa, 2005). The social and technical elements that make up market arrangements are called market devices. Common market devices include property rights, bidding mechanisms, algorithms, displays, prices, spatial arrangements, representations of data, and measurement tools. As there are many market devices, and because they can be arranged differently, markets and market agency can be formatted in endless ways. This raises important questions about what types of agency stakeholders try to create and defend when building markets such as in DFS. It also allows researchers to study agency as an empirical phenomenon that emerges as the result of a network configuration. In this article, we examine how actors created, defended, and contested a specific configuration of agency in DFS.
Method
ANT has been described as a methodology rather than a theory, made of a set of sensibilities and a repository of terms for studying the world (Baiocchi et al., 2013; Law & Singleton, 2013; Sayes, 2014). One of the most important sensibilities is to focus on matters-of-concern rather than matters-of-fact (Latour, 2004b). Matters-of-concern are things whose reality, nature, or existence is uncertain and controversial, leading many other actors to be implicated in the controversy and its remediation. We selected DFS as a matter-of-concern because the industry created substantial controversy, especially in 2015 when the New York Attorney General filed a lawsuit against FanDuel and DraftKings and a subcommittee of the House of Representatives held a hearing on DFS. Much of the U.S. sport industry, media, and consumers were implicated in this controversy. Most importantly, fantasy players’ agency was the key matter-of-concern, as seen in the skill versus chance debate. Therefore, DFS were a matter-of-concern, analogous to a critical case selection (Flyvberg, 2011), which we used to examine how a new type of agency was contested.
The DFS controversy also set clear boundaries for collecting empirical material, which can be a challenge in ANT studies (Latour, 2005). The problem with Latour’s definition of an actor–network is that the network may be incredibly large if subsequent actors’ networks are also implicated by association, and so on, down a chain of actors. The question, then, is how to study an issue if issues are studied by examining networks and if those networks are vast. Latour (2005) recommends “follow the actors,” which means a new entity becomes implicated in a network if an actor points to them as relevant. However, the new actor may also identify other actors, leading to an ever-growing network. For example, whereas we have so far focused on the immediate network innervating a daily fantasy sport player (the market and its devices), he or she is also likely to point to helpful blog posts, analytics classes, and professors in developing their ability to calculate. Does ANT have any rules for limiting the boundaries of a network or a study?
Latour (2005) acknowledged actors’ work, in principle, goes on forever, but he did not provide any strategies for how to demarcate a study (although others have offered solutions, Sayes, 2017). The best suggestion comes from analyzing Latour’s (1988) own approach in his main document-based study, The Pasteurization of France. In this study of Louis Pasteur, Latour (1988) decided to analyze periodical journals where Pasteur and his critics published their work. On the surface, it may look as though Latour neglected many of the actors associated with Pasteur. However, Latour saw periodical journals as the terrain scientists use to fight, form alliances, and convince each other (Latour, 1987). He treated the texts as places where reality was being “enacted” rather than represented (Mol, 2002). While the periodicals were not the whole picture of pasteurization, they were a relatively complete yet circumscribed picture of the scientific debate over pasteurization.
We follow Latour’s (1988) lead by examining the legal proceedings around daily fantasy sport. As legal proceedings require submission of all relevant arguments and supporting material, they serve as a relatively complete, yet delineated, account of the issue. Like the periodical journals Latour (1988) studied, they are also a place where reality is being enacted and debated rather just being represented. Therefore, we collected 150 documents from the New York Attorney General’s case via the New York court’s website against FanDuel and DraftKings in 2015 and 14 documents from the House of Representatives Committee on Energy and Commerce’s hearing on “Daily Fantasy Sports: Issues and Perspectives” in 2016. The examined documents included affidavits, motions, and accompanying reports and studies from the Ney York Attorney General, DraftKings, FanDuel, the Fantasy Sports Trade Association, and expert witnesses. As a consequence of this design, our study lacks a full understanding of how players’ agency is actually configured in daily fantasy, but it is a relatively complete account of the legal debate over players’ agency.
To examine legal documents, we used methodological tenets also outlined by Latour (1988). First, we began by identifying the actors. Actors are anything that makes a difference. In Latour’s analysis, the most important actor in the periodical journals was Louis Pasteur, who discovered the principles of vaccination, fermentation, and pasteurization. However, Latour also identified nonhuman actors, including microbes, lactic acid ferments, and laboratories. In daily fantasy sport, actors that had effects included FanDuel, DraftKings, the Attorney General of New York, scientists, fantasy players, software engineers, social workers but also computers, algorithms, and team lineups.
Latour’s (1988) second methodological tenet is that actors define each other. The third, related, tenet is that actors define others to enroll them in translations. A translation occurs when one actor redefines another’s interests to align them with one’s own. For example, Pasteur successfully redefined the goals of the hygienists in the late 1800s so they would support his efforts to discover microbes. These tenets (interdefinition and translation) are methodological tenets because they imply the reality or the identity of a thing cannot be supposed prior to the analysis. In this case study, we focus on those actors whose interdefinitions and translations revolved around daily fantasy players’ agency.
Findings
Defining Players as Agents
The New York Attorney General accused DraftKings and FanDuel of hosting games that violated New York gambling laws. Illegal gambling in New York contains three elements: (a) the player “stakes or risks something of value”; (b) the player “receive[s] something of value in the event of a certain outcome”; and (c) the player wins or loses depending on “the outcome of a contest of chance or a future contingent event not under [the player’s] control or influence” (N.Y. Pen. Code §225.00 et seq.). As daily fantasy players clearly stake something of value in an entry fee and receive something of value in a prize, the lawsuit focused on whether or not players won or lost based on a contest of chance or a future contingent event not under players’ control or influence.
DraftKings and FanDuel started by defining players in ways that emphasized their agency. For example, Gregory Karamitis of DraftKings said, “sports fans have grown increasingly sophisticated and analytical about the sports they love.” FanDuel stated, “The overwhelming majority of people who participate in DFS are dedicated sports fans who recognize the commitment to knowledge-gathering and making judgments about player capability” and they threatened that if the New York courts stopped daily fantasy, it “would deprive hundreds of thousands of subscribing New Yorkers of the opportunity to pit their skills against the skills of others.”
DraftKings and FanDuel also sought to speak for their players by explaining why and how they played. According to FanDuel’s statement, [P]articipants in competitions for fantasy sports feel like competitors. The evidence indicates that they regularly work very hard to select their teams. They believe that their efforts make a difference. They’re excited about that, they find fun in this separate fantasy context of establishing their team and seeing how it does.
Therefore, DraftKings and FanDuel defined players as agents who found it exciting, fun, thrilling, and challenging to see their work and skill tested in daily fantasy competitions.
Eric Schneiderman, the New York Attorney General, countered this definition of players by arguing many players were there to gamble and that DraftKings and FanDuel misled some players about the extent of their agency. He said, “the compan[ies] deliberately seek to attract persons searching for online gambling” and called on experts on problem gambling who argued “for addicts, fantasy sites can lead to a ruinous path.” Moreover, DFS commercials “further deceived consumers about the chances that an average user will win money.” In other words, whereas DraftKings and FanDuel emphasized the agency of users, the Attorney General deemphasized their agency by arguing players were trying to gamble rather than play a skill game, some were problem gamblers with little control, and those who thought they had control were misled.
Defining Players as Controlling Agents
As New York prohibits games where a player wins or loses based on “a future contingent event not under [the player’s] control or influence,” DraftKings and FanDuel also had to establish their players were agents with substantial control over the event. In contrast to the New York Attorney General, who argued daily fantasy competitions were decided by a real-world sport event outside of the players’ control, DraftKings and FanDuel redefined the relevant event as a competition in the fantasy world. Their lawyers stated, The outcomes of daily fantasy sports contests do not depend on any one real-world team’s winning . . . Instead, daily fantasy sports competitions are decided based on the relative strength of the composite statistics generated by the lineups selected by the fantasy contestants—which bear no relation to any real-world sports team. In short, the “future contingent event” upon which the outcome of daily fantasy sports competitions depend is the competition between two or more fantasy teams.
Thus, “[i]n the fantasy game, you are betting on what happens in the fantasy world.” DraftKings and FanDuel argued, moreover, that fantasy players clearly controlled this event. DraftKings stated, “Because it is the unique combination of athletes on a fantasy team that dictates success, each fantasy sports contestant is able to exert substantial control and influence over the results of the contest by selecting his [sic] fantasy lineup.” Using particularly evocative language, the companies’ lawyers said, “They [the players] are actually the movers, the marionettiers of that outcome,” because “the relevant contest is the contest over who is superior at picking rosters.” Thus, DraftKings and FanDuel tried to establish their players as controlling agents by delineating a fantasy space and event, separating it from real sporting events, and explaining how players had control and influence over the fantasy event.
Defining Players as Skilled Agents
DraftKings and FanDuel also had to establish their players were not winning or losing a contest of chance. Their main strategy was to define fantasy players as skilled agents whose skill had a substantial impact on the outcome of the competition. Representatives described how players used “knowledge,” “skill,” “strategy,” “evidence-based analytics,” “research,” “analysis,” “creative and strategic thinking,” and “hard work.” DraftKings submitted the affidavit of a fantasy player who said, Typically I spend about 70 to 80 hours per week preparing, watching games, researching, and playing DFS contests. During the fall season—from about September to the end of the year—my hours of play and preparation increase to about 90 hours a week because of the high volume of games.
His work included watching tape, watching every game, studying batted ball speed, tracking injures, watching weather forecasts, consulting air density reports, modeling umpire influence and game psychology, and analyzing player performance data. DraftKings also used research, including an article by Miller and Singer (2015) who identified two primary ways skilled users succeed over unskilled users. Skilled users employ lineups that create covariance to produce extreme DFS outcomes necessary to win large-field tournaments. Skilled users also exploit salary cap pricing inefficiencies using sophisticated models to optimize their lineups.
The New York Attorney General accepted daily fantasy players used skill. But, given New York uses the material element test (a higher standard than other states), he argued that chance remained a material factor in the outcome. Thus, DraftKings and FanDuel were tasked with proving players were not only skilled, but that they won because of their skill and not because of chance.
DraftKings and FanDuel used expert testimonies, mathematics, academic research, and data analytics to show how much skill was involved in daily fantasy. DraftKings employed Gaming Laboratories International to compare the performance of actual lineups of DraftKings players against randomly generated computer simulations and against “skilled” computer simulations. Actual lineups won 82% to 96% of the time and outperformed skilled computer simulations. Professor Zvi Gilula analyzed user performance data from DraftKings and found high-performers won more than 80% of head-to-head contests. Professor Anette Hosoi found statistical correlations in player performance indicating that, over time, successful fantasy players tend to remain successful, moderately successful players tend to remain moderately successful, and unsuccessful players tend to remain unsuccessful. FanDuel submitted one analysis comparing daily fantasy to the PGA Tour. They concluded, “For all statistics FanDuel is very clearly on the same level as golf and scores marginally lower only in the matchups category.”
Thus, FanDuel and DraftKings employed a series of mathematical tests and analyses, experts, and published opinions to demonstrate the skill of players. Moreover, they compared skilled players’ lineups with those created by unskilled players and computer programs to establish exactly how influential skill was to winning. Professor Gilula’s conclusion serves as a summary of DraftKings’s and FanDuel’s argument: “[O]ne MUST conclude that the fantasy games offered by DraftKings have an inherent and vast character of skill where chance is overwhelmingly immaterial in the probability of winning in such games” (emphasis in original).
A Game That Enhances Agency
DraftKings and FanDuel clearly championed the free will, intent, and abilities of their human players. But they also went to great efforts to show how their games supported and even enhanced players’ agency. In other words, agency did not belong to players only, but to the configuration of players, games, and other technologies.
First, DraftKings and FanDuel explained how their games separated the fantasy contest from the real-world contest. “The scoring of fantasy sports contests is not based on the scoring of any real-world game. And the relative performances of individual athletes on particular fantasy rosters are detached from the outcome of the underlying sporting events.” The detachment of the game from reality created a circumscribed fantasy space where players could exercise their control. Moreover, DraftKings and FanDuel assigned fictional salaries to athletes and contestants picked teams of athletes subject to a salary cap. Therefore, contestants needed to balance an athlete’s expected value against his salary while accounting for opportunity costs of other athletes not selected. While fantasy salaries are correlated with players’ expected performance, they are not perfect predictions of performance as is the case with a betting line, so there are substantial opportunities for contestants to find underpriced athletes.
Second, DraftKings and FanDuel explained how DFS created better conditions for players to demonstrate their skill compared with season-long fantasy sports. DraftKings argued, “the rules of daily fantasy sports are designed to reduce the already immaterial role that chance plays in season-long fantasy sports.” Such rules included (a) shorter duration of contests, which minimized effects of player injury and unexpected weather patterns; (b) salary caps that required contestants to strategically evaluate athletes’ value; (c) the ability of multiple contestants to select the same athletes; (d) no player draft, which makes every player available for every competition and reduces the chance associated with determining draft position.
Critics of daily fantasy sport also saw nonhumans as playing a substantial role in player agency. Professor Kurt Eggert said, “it would be impossible to capture, remember and analyze the enormous amount of data involved without the aid of computer assistance.” According to Eggert, computer assistance had important implications for player agency: It is important to note that the algorithms do not merely aid professionals in choosing their entries, but instead directly determine what entries to submit. A professional might tinker with the results spit out by the algorithms at the margins, but the algorithms are making real determinations about what players select for lineups.
In other words, he believed the game was played between algorithms, with minimal contributing from human players.
The Problem of Uneven Agency
It was clear to observers, even the New York Attorney General, that certain players won far more than others. One study of daily fantasy baseball found that just 1.3% of contestants won 91% of the prizes. DraftKings and FanDuel used these statistics as evidence players were highly skilled agents. But during the House subcommittee hearing on fantasy sports, regulators and critics began to see unbalanced competition as a problem. Professor Kurt Eggert identified a difference between professionals and recreational players and said, “Daily Fantasy Sports have been unfair to recreational gamblers.” Professionals were armed with computer algorithms, which gave them a substantial advantage over recreational players because algorithms were better able to manage and create portfolios of lineups from data and were better at acquiring data in the first place.
Computer programs not only provided advantages in head-to-head matches, they were used to gain advantages in large-field games with huge prizes. Professional fantasy players discovered the best strategy for beating a large field was to submit a large number of entries filled with players with highly variant performance, meaning they were more likely to accumulate extreme scores (low or high). Programs not only created these lineups, they also helped submit many lineups quickly. Professionals used scripting tools, which made it easy to submit multiple lineups to many competitions at once. In gambling parlance, professionals are called “sharks” who take advantage of “fish” or “whales,” players who regularly lose large amounts of money. The authors of one cited study noted the whole industry depended on 5% of players who contributed 36% of the entry fees and lose on average a third of their money. Another problem was DraftKings and FanDuel employees being able to play on other daily fantasy sites. Employees, who had access to game data, may have had an unfair advantage because they could pick athletes less often selected, increasing their chance of winning.
Critics of DFS companies used House hearings to identify consumer protections that were necessary for the daily fantasy industry. Expert witness Dr. Ryan Rodenberg gave the following list: (i) advertising restrictions; (ii) prohibiting fantasy play by minors via age verification procedures; (iii) making available compulsive gambling and addiction resources (and implementing data-driven procedures to identify problematic behaviors); (iv) geo-location tracking; (v) availability of self-exclusion lists; (vi) restricting or eliminating the use of algorithms, scripts, and bots; (vii) limiting the number entries in certain daily fantasy contests; (viii) including special protections for “head-to-head” daily fantasy contests; (ix) requiring the adoption of “know your customer” guidelines; and (x) mandating explicit accounting procedures to ensure that fantasy player funds are in segregated accounts and not commingled in general operating funds.
Among these protections were some designed to regulate the agency of professional and recreational players, so that certain people would be excluded (minors and problem gamblers) and the agency of professional players could be reduced (restrictions on algorithms, limited entries, and protections for head-to-head matches).
Peter Schoenke, Chairman of the Fantasy Sports Trade Association, stated his support for common sense regulation. Among the measures used to reduce some players calculative agency, DraftKings and FanDuel banned third-party scripting tools, created employee policies stopping play on other platforms, and banned the use or sharing of insider information. Fantasy sports companies also excluded certain people using geolocation technology and self-exclusion options. Collectively, these measures were designed to regulate players’ agency.
In sum, DraftKings, FanDuel, the New York Attorney General, and the House Subcommittee put daily fantasy agency on trial, but rather than debate whether or not agency existed, they debated whether players had enough agency. Consistent with ANT, they described agency as a sociomaterial network of players, algorithms, policies, rules, platforms, teams, salary caps, scripting tools, and geolocation technologies. As a consequence, the companies and regulators believed agency could be enhanced (with algorithms and scripting tools) or restricted (with policies, rules, and self-exclusion options). Indeed, after debating the extent of players’ agency, critics focused on how to balance the agency of professional and recreational players, which suggest they saw agency as an unfinished, actor–network, open to revision.
Discussion
Our aim was to use DFS as an exercise in empirical philosophy to examine how a new type of agency was contested. There are three key findings. First, although actors were fighting over players’ agency, they were not asking whether or not agency existed, but rather what type of agency players had and how much of it they possessed. In response to New York’s gambling laws, which prohibited games of chance outside players’ control, DraftKings and FanDuel described competitors’ agency as skilled and controlling. They also used data, analytics, mathematics, and comparisons to show that players’ skill was the dominant factor in determining winners and losers, whereas chance was immaterial. High-performing lineups were tested (similar to Latour’s, 1987, idea of trials of strength) to see whether they could outperform random, unskilled, and even skilled lineups made by computers and other contestants. It was decided that fantasy players exhibited more agency than computers and about the same agency as a PGA tour athlete.
Second, agency emerged from a specific sociomaterial network. Players only exhibited agency while playing games with artificial salaries and salary caps. They could improve their agency with algorithms and scripts. Thus, a sociomaterial network of players, companies, games, algorithms, and penal codes came together to create fantasy agency. This finding supports theories of agency as collective, historically specific, and changing (Emirbayer & Mische, 1994; Sewell, 1992). Like Giddens’s (1979, 1984) theory of structuration and Bourdieu’s (1977) habitus, we have shown how agency is enabled by certain objects and arrangements, but unlike these theorists, we have avoided using social structures as the counterpoint to agency. As predicted in critical theory, player agency was contested by the New York General Attorney and others, but it emerged at the termination point of a network of effects, rather than in opposition to a structure.
Although DraftKings, FanDuel, the New York Attorney General, and other critics recognized the importance of sociomaterial networks to players’ agency, they disagreed whether nonhumans’ agency was legitimate. The New York Attorney General and others sought to prohibit daily fantasy lineups picked by algorithms, which suggests they thought algorithms were illegitimate actors, supporting Weedon’s (2015) assertion that agency is commonly attributed to humans rather than nonhumans. On the contrary, DraftKings and FanDuel were willing to accept the differences made by algorithms as important components of players’ agency; however, this is probably because the companies needed to prove their games were skilled. Indeed, DraftKings and FanDuel are unlikely to take posthumanist thinking to its logical conclusion by crowning the best-performing algorithm as the winner rather than the best performing human. And yet, despite the tendency to attribute agency to humans, the use of employee policies, geolocation technology, and self-exclusion options show that companies and regulators saw human agency as vulnerable and frail and open to improvement.
Third, outside of the issue of legality, regulators’ other main concern was unbalanced calculative agency of professionals versus recreational players. Critics sought regulations that would reduce the abilities of professionals, exclude minors and people with gambling problems, and would match people with even skills together in head-to-head competitions. According to Çalışkan and Callon (2010), after a market is made, its sociotechnical arrangement may confer different calculative agencies to different market participants, resulting in some participants having enhanced agency compared with others. Unbalanced calculative agency was evidenced in daily fantasy, with “sharks” having greater calculative abilities than “fish” and “whales.” This finding demonstrates how market participants have agency, but some have greater agency than others.
The main theoretical contribution of this study is the picture we have drawn of agency. Actors fought over the nature of players’ agency but focused on what type of agency they had and how much of it they showed when playing. Daily fantasy agency was composed of a sociomaterial network of players, games, algorithms, and rules and, as a result, it could be enhanced or reduced. We recommend sociologists build on this study by asking how much and what type of agency is evidenced in sport and physical culture. They can do this by studying the sociomaterial networks of actors that make agency possible. Examples of this work include Weedon’s (2015) study of camaraderie in Tough Mudder events, Darnell et al.’s (2018) study of sport for development in Kingston Jamaica, and Esmonde’s (2018) study of heart rate technology in physical activity class. Darnell and colleagues (2018), for example, found sport for development organizations in Kingston needed to establish and maintain connections with international stakeholders to host effective development programs. In other words, the sport organizations’ capacity to act was an outcome of the connections they were able to make, meaning their agency was distributed among other human and nonhuman actors.
However, scholars have yet to study how much agency is evidenced in sport and physical culture. By rephrasing agency as a question of “how much?” scholars are led away from structure and toward the things that enhance or inhibit agency. Indeed, if agencies are sociomaterial arrangements, it stands to reason that agencies can be improved or changed by modifying their components and arrangements thereof. DraftKings and FanDuel created platforms and games that enhanced players’ skill and control. Certain players used algorithms and scripts to improve their calculative agency. Regulators and the Fantasy Sports Trade Association used policies, geolocation technology, and self-exclusion options to limit players’ agency. Seeing agency as composed of sociomaterial arrangements of humans and nonhumans has methodological advantages because scholars can examine specific configurations of agency using empirical methods rather than relying on philosophical or theoretical assumptions. This fits with initiatives such as physical cultural studies, which seek radically contextualized analyses of physical culture and sport (Millington & Wilson, 2017).
Callon (2010) took this reasoning further arguing, as agencies can be formatted in multiple ways, “the only valid question is not the existence or not of [agencies] . . . but their design and constitution” and the “modalities of their intervention” (p. 232). If agencies can be changed, enhanced, or limited, as we and Callon suggest, then it raises important questions about sociologists’ rights and responsibilities to intervene. For instance, should sociologists intervene in the daily fantasy sport social-material arrangement to enhance players’ calculative agency? We believe there are many interventions that should be supported by sociologists of sport in addition to what is currently being practiced by the daily fantasy industry. For example, there is clearly an imbalance between the calculative agency of “sharks” and “fish” or “whales.” One option would be to give amateur fantasy players the option of playing head-to-head competitions against players with similar experience or winnings, so they are pitted against competitors with similar calculative agency. Another, common intervention is educating players on how the games really work (i.e., explaining that 1.3% of contestants win 91% of the prizes); in this case, teachers play an important, yet sometimes underappreciated role in students’ actor–networks. A more complicated question is whether there are any interventions that sociologists are particularly suited for? Topics like risk, problem gambling, and sociality are central to sociology and daily fantasy sport. Sociologists may be able to create applications that help players recognize their own problem gambling and have easy access to self-exclusion options. They should also seriously evaluate DraftKings’s and FanDuel’s claims that players use their platforms to socialize and supplement their fandom (Farquhar & Meeds, 2007). Different motivations for playing will determine what types of assemblage adjustments are wanted and needed. In short, ANT raises the possibility of studying agency as we have done here, but it also suggests intervening in its construction.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
