PLAYFUL METADATA Between Performance Careers and Affect Modulation Pablo Abend / Max Kanderske ABSTRACT In the field of specialized hardware for digital gaming, an increasing num- ber of products not only promise ever-increasing precision, but also pro- vide self-tracking functions intended to quantify the player’s gaming ac- tivities and actions. We position these developments at the intersection between the Quantified Self movement and the tradition of playful self- measurement. Building on practice theory, we raise the following ques- tions concerning the datafication of gaming practices and the use of what we call playful metadata: What do players and game developers do with data that is generated within, and in relation to, games? How does the emergence of playful metadata modify interactions, both between play- ers and between the players and the game? By analyzing exemplary quantifying practices found in the contexts of speedrunning, competitive gaming and game streaming, we identify three central motives for quan- tified gaming: 1) the appropriation of games’ spaces and goals by players who define their own parameters of success by quantifying their game- play; 2) the production and communication of individual performance ca- reers aimed at modulating the player’s affects towards their own perfor- mance; 3) the production of data for competitive comparability and/or cooperative sharing of knowledge. Keywords: self-tracking, quantified self, quantified play, practice theory Pablo Abend / Max Kanderske Playful Metadata 1 . THE COMPUTERIZATION OF GAME PRACTICE “The Naos QG is a next generation gaming mouse that measures the user's biometric information and movement data. This allows the Naos QG to provide valuable, interesting and fun insights that creates a richer user experience.” (Mionix 2020) As the manufacturer’s description suggests, the Naos QG mouse is a gadget for generating data about one's digital gaming activities. The input device, which comes in the shape of a conventional ergonomically de- signed computer mouse, measures in-game actions per minute based on click frequency and the distance travelled across the mousepad. Further- more, it provides information about stress levels during gameplay via heart rate monitoring and skin conductivity measurements. Fig. 1: Naos QG settings menu. (Source: Mionix, author’s screenshot) We posit that the Naos QG and its features are representative of a wider range of recently developed products that monitor and quantify digital gaming: At the hardware level, built-in sensors directly or indirectly meas- ure physical reactions to the game or record eye movement.1 At the soft- ware level, we see a proliferation of applications and platforms that col- lect, calculate and visualize data about the user’s performance in the 1 See interfaces like the SteelSeries Sentry or the Tobii EyeX. 88 Spiel |Formen Special Issue: Ludomater ial it ies game. While there are major differences in terms of functionality, all these technologies datafy the act of playing by translating the players’ actions and physical reactions during the game into numerical and statistical form. In this article, we will examine different practices and interfaces of quantified gaming and take a look at different forms of interaction that emerge in the context of datafied games. Following a media-as-practice (Couldry 2004) approach to games, we do not limit our analysis to a par- ticular title or genre, but instead ask what players do with games and the data generated in and around them. In doing so, we hypothesize that dis- tinct forms of interaction can be observed in the context of datafied games. In addition to the practices themselves, the necessary interfaces, the hard- and software as well as the wider network and platform infra- structures come to the fore. For as Schatzki points out, practices are by no means to be considered in isolation, but always in relation to the various “material arrangements” with which they maintain a reciprocal relation- ship. Neither is isolated; rather, practices and materialities form what Schatzki terms “bundles”: “By ‘material arrangements’ I mean linked bodies, organisms, arti- facts, and things of nature. […] The idea that practices and arrange- ments form bundles implies that practices and arrangements inter- relate. Practices and arrangements form bundles in that (1) practices affect, alter, use, and are directed toward or are inseparable from arrangements; while (2) arrangements channel, prefigure, and facil- itate practices.” (Schatzki 2016, 32) This occurs first within a field of practice (Schatzki 2006) in which individ- ual media use organizes itself into larger contexts. The analytical challenge here is the scaling: Which forms of use manage to imprint themselves on the larger context and for what reasons? According to Swidler (2006), there are “anchoring practices” that play a key role in the reproduction of larger systems of discourses and practices. Applied to quantified gaming, the question is consequently what significance this form of gaming prac- tice has for digital gaming culture as a whole. To approach this question, it will not only be necessary to answer how individual actions and opera- tions scale into practices, but also how these practices stabilize, how they 89 ... Pablo Abend / Max Kanderske Playful Metadata are enabled by material arrangements, and how they are involved in bringing forth further material arrangements. In order to accomplish this through a praxeological investigation, we will analyze three bundles with regard to their quantifying anchoring practices: speedrunning and its prac- tices of timing and sequencing, competitive gaming and its practices of logging and stat tracking and game streaming and its practices of visuali- zation. Commercially available self-monitoring tools can be expected to change both the meaning of play and the games themselves. Conse- quently, Ben Egliston employs the concept of “quantified play” (Egliston 2020) to shed light on how quantification transforms gameplay and what effects it has on users. From a phenomenological point of view, he asks what new ways of playing are created by self-monitoring and where tra- ditional ways of playing are displaced or even prevented. (Egliston 2020, 3) In a similar fashion and also in relation to digital gaming, James Ash de- scribes how technologies “recalibrate” the perception of the here and now through quantification. (Ash 2012) He illustrates this with the example of the fighting game Street Fighter IV, and observes that particularly skilled players break the game down into individual frames, measuring the time between animations in individual frames. In this way, the frame rate be- comes a new way of dividing and measuring time, which historically should not and could not actually be perceived by the recipients. (Ash 2012, 193) In this sense, quantified play introduces an additional layer of datafi- cation between interface and body that renders previously inaccessible information about the player’s actions visible. We propose to conceptual- ize the data that emerge from and feed back into ludic environments and situations in that fashion as playful metadata, with the prefix ‘meta’ de- noting that it is additional data about the player’s actions, both within and outside of the game, that are generated. In ascribing a certain kind of play- fulness to these data, we build on Deborah Lupton’s (2018) notion of “lively data” produced by self-tracking technologies: 90 Spiel |Formen Special Issue: Ludomater ial it ies “The digital data that are generated by self-tracking may be con- ceptualised as ‘lively’ in various ways. First, these data are generated from life itself, in terms of documenting humans’ bodies and selves. Second, as participants in the digital data economy they are labile and fluid, open to constant repurposing by a range of actors and agencies, often in ways in which the original generators of these data have little or no knowledge. Third, these data are lively by vir- tue of the advent of algorithmic authority and predictive analytics that use digital self-tracked data to make inferences and decisions about individuals and social groups. These data, therefore, have po- tential effects on the conduct of life and life opportunities. Fourth, by virtue of their growing value as commodities or research sources, the personal data that are derived from self-tracking practices have significant implications for livelihoods (those using these data in the data mining, insurance and data science industries, for instance).” (Lupton 2015, 563) Likewise, the data we are concerned with in this article can be considered as ‘playful’ in three distinct ways. First of all, they emerge from within games. As they document the player’s (re)actions, they open them up to practices of evaluation and spectatorship. Second, they inform playful practices of altering the rules of the game which can be carried out by both the developers and the players. Third, they relate to practices of ludic bi- ography, that is, to the writing of individual performance careers that un- derpin real or perceived life opportunities connected to playing digital games. Rendering the imperceptible perceivable via playful metadata often follows an economic impetus and can thus be theorized as part of the on- going dissolution between the domains of work and leisure. As diagnosed by Rhee: “[...] work no longer happens just at work; it also happens when- ever we engage our devices, when we look up restaurants online, stream a movie, send an email or play a video game.” (Rhee 2018, 46) Specifically addressing the sphere of play and games, Abend et al. employ the inter- dependent concepts of “laborious play” and “playful work” in this context. (Abend et al. 2019; 2021) Seen in light of professional streaming and the ever-growing esports scene, the industry’s promise of increasing individual player performance through quantified gaming seems to suggest the possibility of a seamless professionalization of one's own gameplay. Thus, our thesis is that the 91 ... Pablo Abend / Max Kanderske Playful Metadata technologies used in quantified gaming serve as mediators between indi- vidual performance careers and a broader culture of the professionaliza- tion of gaming. In this sense, the quantification of individual performances is an important factor that contributes to the professionalization of a prac- tice formerly understood as a leisure activity. (Guttmann 1978) 2. THE TAUTOLOGY OF QUANTIFIED PLAY With regards to digital games, talk of quantified play is akin to a tautology. Digital games have always been quantifiers of human action. In order to function, they process the input of players by quantifying movements, thereby rendering them machine-readable. The machine then generates an interpretable output, which in turn serves as the basis for the next player input. This output often takes the form of unnecessary obstacles that players have to overcome in order to win the game. (Suits 2002, 55) The attraction of a game is that this process cannot be fully anticipated. Players find themselves within a situation of an artificially created contin- gency that nevertheless “generates interpretable outcomes.” (Malaby 2007, 96) This “interpretable output,” which Malaby considers central to the game definition, simultaneously acts as an indicator for success or fail- ure and enables comparability between players. Thus, in order to render visible success and failure, victory and defeat, the input must be made measurable through the game. While the need for creating interpretable outputs also exists for analog board and card games, as well as for sports competitions, the practice of quantification is usually triggered by certain key events (e.g., goals in soc- cer). It is therefore possible to perform game actions that do not entail any immediate quantifiable output. In the case of digital games, however, any participation in the game means subjecting one’s body – or at least the body parts acting on the interface – to a system of measurement and evaluation. The player operates within a feedback loop in which the ma- chine continuously processes the inputs and generates corresponding outputs. 92 Spiel |Formen Special Issue: Ludomater ial it ies From this technology-centered perspective, the players of digital games have always been quantified. Playing digital games is thus always a datafied practice. However, this datafication does not necessarily result in a human-readable output of numerical values. Whether numbers are shown to the player and what meaning is ascribed to them in the context of the game strongly depends on the respective genre – from arcade titles, whose high score values signify success or failure without having an im- mediate effect on the player’s actions, to simulation games whose game- play centers around interpreting and manipulating numerical values dis- played across a multitude of tables and charts. Between the game’s invisible underside and its visible surface, the out- put can take on a range of different forms. For example, success can also unfold spatially or narratively: a new area becomes unlocked, or the game’s story progresses. Whatever shape the output may take, the appeal of playing lies in overcoming the initial, artificially created contingency. This requires a) un- covering the operational logic of the game, that is the relation between input and output and b) adapting one's own play to the routines of the machine: One plays and is played. By directing one's input towards achieving desirable game states, playing becomes a permanent “accom- modation to the machine” (Pias 2000, 232). However, playing “in the form of adaptive action in the designed game space” (Hawranke 2018, 45) is not the only way to deal with digital games. Just as the rules of an analog game can be negotiated and changed during play, this also happens when playing on the computer. Such forms of ap- propriation in and through play can be called “transformative” (Salen/Zimmerman 2003) or “transgressive” (Aarseth 2007): “Transformative play is a special kind of play that occurs when the free movement of play alters the more rigid structure in which it takes shape. The play actually transforms the rigid structure in some way. Not all play is transformative, but all forms of play contain the potential for transformation.” (Salen/Zimmerman 2003, 311) 93 ... Pablo Abend / Max Kanderske Playful Metadata Aarseth in particular sees transgressive play as the conflict between the “ideal type” of player assumed by the developers and the individual play- ers who bring their own ideas and purposes into the game – as a “symbolic gesture of rebellion against the tyranny of the game” (Aarseth 2007, 132). 3 . PRACTICES AND METRICS While digital games have always been quantifying machines for human action, quantifying hardware and software ensure that additional game in- formation that normally remains invisible and imperceptible to players is collected, sorted, and presented in discrete numerical and dominantly vis- ual form. In terms of game actions, quantifying tools consequently enable the storage of fleeting interactions that can become action-guiding as pre- dictions of future events. In the following sections we will introduce ex- emplary bundles of quantifying practices around which larger systems of discourse and practice have formed. We will specifically focus on the an- choring practices of sequencing, logging/accounting, and making visible, as well as their relationship to specific forms of play located between the poles of transgression and professionalization we have identified above. SPEEDRUNNING One bundle of playful practices particularly relevant to the subject matter of this article is speedrunning. In speedrunning, the players’ goals substi- tute the criteria of success imposed by the original design. Speedrunning can thus be described as an appropriation of game space, in which even narrative-driven games are re-interpreted as sprint competitions (Knorr 2009, 223). During a speedrun, playing is no longer a matter of advancing the story, but rather of exploiting all possible means to traverse the game (ibid.) as quickly as possible and set a new record time: 94 Spiel |Formen Special Issue: Ludomater ial it ies “Speedrunning is not about breaking down the general rules of the game, rather these are tested for their interpretative and configura- tional flexibility. [...] The original goal of the game is overwritten by the self-defined goal. The actual run is documented on video and shared within the community. On relevant Internet platforms, these videos serve as proof of the runner's masterful performance.” (Hawranke 2018, 46, author’s own translation) T I M I N G AN D SE QUE NCI NG Timing and (re-)sequencing here emerge as the anchoring practices around which other strands of the bundle, such as streaming, maintaining leaderboards and performing speedrun historiography and forensics, co- alesce. Performing these anchoring practices requires the software equiv- alent of a stopwatch: applications like LiveSplit allow the users to deter- mine split times for discrete game sections (splits) that constantly relate the ongoing speedrun to previous attempts and/or online leaderboards. To partition the game into sections, players first pick clearly identifiable measuring points like cutscenes or boss fights. This practice is usually part of a preparatory phase which can also involve mapping out the game and the fastest routes to its completion. Once a checkpoint is reached during the actual run, the player can stop the split times manually by pressing the corresponding key. The software then calculates the time lag or lead over the comparison run and outputs it on the screen. For the sake of better accountability, runners who play the same game usually choose the same partitioning, enacting a form of canonization that spreads from the fastest runners to the rest of the community. This practice is not only promoted by sharing runs via live streams or videos, but is now firmly embedded in the split software’s functionality of downloading record holder’s partitions and split times. Not every run involves testing the rules for interpretative and configura- tive flexibility: especially in games that have been ‘ran’ for a long time, in- terpretative closure occurs, as individual runs approach the pre-stabilized ideal of the supposedly perfect, i.e. shortest possible run. Accordingly, new runners have to adopt the routes and techniques already worked out by the community in order to be able to participate in the competition at 95 ... Pablo Abend / Max Kanderske Playful Metadata all. They are still transgressive2 when compared to the gameplay originally envisioned by the developers, but in terms of game style and interface configuration they are bound to the established conventions of the speedrun community. This homogenization of game actions primarily rests on what James Ash refers to as “spatialization of time” (2015, 67): the partitioning of the total distance one needs to cover to successfully com- plete the game into individual sequences, which are subsequently as- signed numerical values. Returning to Aarseth's hopeful prospect of a revolution led by transgres- sive players, it may seem as if speedrunners have indeed broken the tyr- anny of their game’s original metrics of success. But – as one might po- lemically add – that achievement comes at the cost of having installed a new and possibly even stricter ruler: the temporal regime produced by quantified gameplay. Fig. 2: Screenshot of a successful world record attempt by speedrunner Lozoots in the game OCARINA OF TIME (Nintendo, 1998). 2 Since a vivid community of speedrunners dramatically increases a game’s longev- ity by attracting new players long after the initial release, some developers – espe- cially in popular speedrunning genres runs such as jump & run – have started im- plementing speedrun modes into their games from the get-go. The interpretative flexibility of the rules thus gives birth to a new practice which in turn becomes a set of rules to be reincorporated into the software. 96 Spiel |Formen Special Issue: Ludomater ial it ies The display of split times generated by the split software (Fig. 2, top left) is part of the basic inventory of speedrun video aesthetics. It is not a mere visual recall to similar representations employed in televised racing, but a constitutive element of the practice, as the display of time is what allows the gameplay to be immediately perceived as a race at all: The running timer signifies a race against the clock; the split times are colored red or green depending on the gap or lead, signifying a race against an absent competitor. In addition, runners usually show the gameplay in their videos or streams (Fig. 2, right), as well as their face, the input devices held in their hands, or an abstracted representation of the input commands (Fig. 2, bottom left). On the one hand, this configuration serves to substantiate the measured times, as viewers can verify for themselves whether reach- ing a checkpoint within the game actually corresponds with the time of measurement.3 On the other hand, the visual arrangement allows for the communication of one's body and game knowledge to the community by revealing the inputs necessary to execute the virtuoso game action. While the splits, understood as an abstraction of these inputs, act as the central metric for competitive comparison, the gameplay visuals facilitate the co- operative advancement of routes and techniques, as they provide expla- nations akin to a live-video tutorial. Speedrunning’s visual documentation far exceeds the singular value of traditional high scores, which serves to position one's own game performance within a field of (possibly absent) competitors but plays no role in knowledge transfer beyond pure proof of feasibility. Understood as a form of transgressive play, speedrunning exhibits a disparity between the information displayed by the (unmodified) game and the information required to compete for the self-set goal. To employ Aarseth's terminology: The speedrunner is not the kind of player tacitly assumed by the developers, (Aarseth 2007, 132) therefore most games’ interfaces are not designed to meet the needs of speedrunners. This defi- ciency is best illustrated by the total time spent on the current play-through, 3 It thus enables practices of “speedrun forensics,” which can identify cheating at- tempts by pointing out the fragmentation of the video material (Jobst 2020). 97 ... Pablo Abend / Max Kanderske Playful Metadata a metric that is rarely used – and seldom displayed – during normal game- play, but which acts as the pivotal playful metadata underpinning the whole practice of speedrunning. Consequently, players took it unto them- selves to time and quantify the progress of ongoing and recorded runs by developing their own software tools and the adherent interfacing prac- tices. It turns out that here – in line with the bundling of practices and ma- terial arrangements described by Schatzki (2016, 33) – practices, goals and media emerge at the same time. Quantified play is not merely a tool to overcome the artificial contingency of play. Rather, quantification and the ensuing playful metadata ensure that practices which originally were not covered by the game’s output can now stabilize. By logging ephemeral gameplay actions and generating meta-information about the actual practices of play, quantification creates the conditions for the emergence of specialized communities of practice (Lave/Wenger 1991; Wenger 1998) that cooperatively build and maintain assets of knowledge. Within the speedrunning community, the anchoring practices of sequencing and tim- ing allow members to flexibly shift back and forth in an alternating mode between cooperation and competition. (Hawranke 2018, 46)4 COMPET IT IVE GAMING We consider competitive gaming to be another bundle that is constituted by and constitutes specific quantifying practices. While speedrunning’s transgressive anchoring practices of timing and sequencing radically alter the nature of games, turning them into competitive races in the process, the quantifying practices found in competitive gaming are more closely aligned with the respective games’ already competitive structure and of- tentimes rely on built-in functionalities provided by the developers. 4 Drawing on Huizinga, Schemer-Reinhard (2020, 103) likewise describes the rela- tion between players who share the same game (or its components) while acting as opponents within the scope of the game as being connected in a “spirit of en- mity and community.” The production of cooperation and consensus by dividing a game into sections and sharing those sections within the community adds another layer to this dynamic. 98 Spiel |Formen Special Issue: Ludomater ial it ies Though there is significant overlap, both bundles form separate arrange- ments of practices and materialities as they differ in terms of their in- tended effect and purpose pursued. To illustrate this point, we will shed light on the anchoring practices of logging and stat tracking. LOGGI N G Especially in games that require fast reactions and complex input se- quences, such as fighting games, competitive gamers and streamers often display additional information via a variety of interface layers. For exam- ple, STREET FIGHTER V’s players can tap into the game’s input log, a real- time record of all player commands that is available in training and replay modes.5 Fig. 3: Training mode in STREET FIGHTER V (Capcom 2016). The input log represents player commands as symbols arranged in a se- quence diagram, (Fig. 3) providing a visual link between the player’s phys- ical movements, the actions performed by the game character, and spe- cialized knowledge about the game. The inputs, which usually are not ren- dered within the game image and have to be inferred from the gameplay actions, are thereby operationalized: Their visualization allows the players 5 Many fighting games provide the corresponding function themselves; external software solutions include the applications Gamepad Viewer and OBS Display Fightstick motions. 99 ... Pablo Abend / Max Kanderske Playful Metadata to relate them to the notational system used by the developers and the community to communicate certain techniques, such as ‘special moves’ and chains of commands that are deemed most effective. Accordingly, they play a central role both in checking one's own movement sequences for the sake of error analysis and in conveying input schemes to inexperi- enced players. The purpose here is to log physical actions and reactions that occur so quickly that they can be traced back to a form of embodied knowledge that operates in parts below the threshold of consciousness. (Ash 2012) S T A T AN D M A T CH T R ACK I N G In other competitive games, especially within the shooter and MOBA gen- res,6 a more sophisticated form of logging can be found. Here, both the developer studios themselves – in the form of monetized add-on services – and third-party platforms offer the functionality of statistically pro- cessing data generated by the players’ actions (Fig. 4), tracking various metrics throughout individual matches or lifetime careers. Egliston sees this quantification of gaming practices as a form of “sur- veillance capitalism,” (Zuboff 2019; Zuboff 2015) a way of exercising power and control based on the aggregation and circulation of data col- lected through surveillance technologies. He differentiates between three forms of surveillance practices enabled by statistics portals: “self-surveil- lance,” meaning the control of one's own performance parameters for the purpose of self-optimization, “lateral surveillance,” (Andrejevic 2004) meaning the mutual control and disciplining of competing players among each other, and “machine surveillance,” the analysis of data material sup- ported by machine learning algorithms that generate an ideal concept of good game actions, on the basis of which concrete suggestions for im- provement are made to the players who pay for this service. (Egliston 2020a, 9-13) In this process, the data are also used to generate an ideal concept of “good” gameplay actions. Even the data of players who do not make use of statistical services themselves eventually become the basis of the statistical evaluation, since 6 Short for “Multiplayer Online Battle Arena.” 100 Spiel |Formen Special Issue: Ludomater ial it ies statements about the efficiency of concrete game actions (such as the se- lection of items or abilities) can only be made if the largest possible basis of comparison of games is available. Accordingly, one could speak of a permanent “cooperation without consensus,” (Star/Griesemer 1989) in the context of which a community of practice that is heterogeneous with respect to its own playful ambitions jointly creates a database of played games. In concrete terms, this means that even players who show no in- terest in the practices of self-monitoring and external monitoring (or who are not even aware of their existence) can participate in the project of quantifying or optimizing game actions. Fig. 4: An excerpt of the guides section of the statistics portal Dotabuff (author’s own screenshot). The goal of the survey is the automatic formulation of game instructions or guides (Fig. 4), which are supposed to relieve the players of pivotal game decisions. The form of these guides, as well as the aesthetics of the se- quence of item and ability symbols attached to them, can be traced back to the early MOBA prototypes, which were still modifications of the game WARCRAFT 3 (Blizzard 2002). Based on “build orders” 7 that originated within the strategy game genre, players communicated their game knowledge in the form of so-called “skill and item builds,” sequences of game decisions formulated in the style of illustrated guides akin to cook- ing recipes, which were shared and discussed in community forums. Ac- 7 This is the optimal order in which one's foundation should be built in the context of a certain strategy. 101 ... Pablo Abend / Max Kanderske Playful Metadata cordingly, the statistical platforms under consideration are to be under- stood as material arrangements that support (or at least promise to sup- port) pre-existing practices of gameplay quantification carried out with the explicit purpose of generating and sharing game knowledge – only this time in a fully automated fashion. When players try to improve their performance via quantification and logging, they are confronted with two fundamental problems: First, data sets are usually incomplete, distributed across many players and plat- forms, and may already be outdated by the time of analysis. These diffi- culties, summarized by Pink et al. under the term “broken data,” (Pink et al. 2018) occur especially in games whose ideal gameplay8 is in a state of constant flux due to frequent updates. While updating a game’s ruleset is a strategy purposefully employed by the developer studio to keep players interested over long periods of time, it also undermines the community’s efforts to “figure out the game,” as both the data that has already been collected and the optimization strategies derived from it become unusa- ble in regular intervals. Data evaluation and the appropriate (and timely!) adjustment of input thus become a substantial part of maintaining one’s relative “skill” within the ever-changing landscapes of continuously up- dated games. Second, isolating the parameters that are relevant for (self-)optimiza- tion from the amount of collected data is no trivial task. In the context of complex games – and especially for inexperienced players – it is not im- mediately obvious which recorded parameters correspond to “good” game actions, that is, those that lead to victory. Hardware and software manufacturers take advantage of this circumstance by promoting the sim- plistic formula “the more data, the better”, while remaining intentionally vague about the actual relationship between data and skill. This approach, which addresses the potential customers’ desire for improvement by promising a utility value that could – but it is no way guaranteed to – 8 Often referred to as “metagame,” which denotes strategic decisions or certain styles of play that are temporarily considered optimal. 102 Spiel |Formen Special Issue: Ludomater ial it ies emerge from the captured data, is characteristic for the commodity aes- thetics (Haug 1971) that hardware manufacturers and platform operators employ in the field of quantified play. Let us now return to the Naos GQ mouse for a moment. While its ca- pability of measuring the distance covered by and with the mouse initially seemed pointless, we can now see that it exhibits the same logic outlined above. The manufacturer’s ability to tap into the player’s fundamental de- sire for self-optimization and advancement within the gaming competi- tion hinges on hinting at a relationship – one that may or may not exist – between actual skillful playing practices and the supposedly useful met- rics provided by their product. L IVE ST REAMING In the context of our investigation, live streaming could be considered a ‘meta bundle’, as speedrunning, competitive gaming and a plethora of other gaming-related activities share the same material structures, com- mon live streaming practices, and overlapping communities. Neverthe- less, it is possible to differentiate between these bundles by acknowledg- ing the intent behind their quantifying practices, as we will show in the following section. H E A R T R A T E V I S UA LI ZA T I ON In the context of live streaming, the practice of heart rate measurement and visualization focuses on the numerical abstraction of physical exer- tion.9 When combined with a Bluetooth heart rate monitor, the PULSOID application allows the heart rate to be displayed in real time during game- play. Even though this is reminiscent of monitoring vital functions with fit- ness wristbands and watches, it does not involve evaluating data for train- ing purposes. While sharing stats online is part of many practices of self- 9 Data obtained by measuring heart rate and skin conductance is also increasingly used as an argument within a discourse of nobilitation: Here, an equation of sports and esports is to be achieved via the common factor of physical exertion. (Krell 2019; Wolmarans 2016) 103 ... Pablo Abend / Max Kanderske Playful Metadata quantification, here the feature is exclusively directed at an audience. Ac- cordingly, the app is primarily intended to appeal to streamers and uses the advertising slogan “Add your live heart rate to your broadcast. Be closer to your viewers!” (Pulsoid 2020) The measurement of vital signs is correspondingly linked to the promise of taking the parasocial relationship between streamer and stream viewers to a level of physical proximity. The FAQ pages also state: “Our approach gives the best accuracy, wide customization and simple interface for users. At the same time, Heart Rate Widget is a great way to make your broadcast more interesting and interactive, you can use it to increase viewer engagement or make the stream more realistic.” (Pulsoid 2020a) Heart Rate Widget’s purpose is not to monitor one’s health while gaming, but to increase the engagement of potential spectators. The audience should be able to read within the data how the players are affected during gameplay (Egliston 2020) It is therefore a matter of rendering the player’s affective involvement visible to increase the entertainment value of the stream. Depending on the context, different patterns of effect and evalu- ation can be identified. In an esports environment, a lower heart rate is valorized, as it seemingly shows that players can keep calm in stressful situations.10 In contrast, a higher heart rate shows the wearer’s tension and involvement, thereby communicating which game situations are per- ceived as crucial by the participating players. Here, the heart rate, which is usually tracked in a chart, becomes playful metadata for structuring the viewing experience by accentuating individual ‘plays’ or situations. Horror game streams exhibit a different dynamic, in which the heart rate monitor renders the player’s fear tangible and attests to the visceral effect of the game. In the context of live streaming, the notion of entertainment value cannot be separated from the competitive dynamics inscribed into the streaming platforms themselves, as the streamers reveal data about their 10 Following this logic, the rofessional league for battle royale game H1Z1 made their players wear heart rate monitors. (Cameron 2018) 104 Spiel |Formen Special Issue: Ludomater ial it ies own body in the hope of gaining an advantage in the competition for the streaming audience’s attention. E Y E T RA CK I N G In the case of eye-tracking interfaces, two possible use cases are adver- tised by manufacturing companies: the recording of eye movements for demonstration, analysis and training purposes, and the use of the eye- tracking hardware as a supposedly efficient input interface that can be op- erated intuitively and at a high speed. (Amazon 2020; Amazon 2020a) Fig. 5: Use of eye tracking during game review of a COUNTER STRIKE match. Following this pattern, the use of the technology in the context of esports commentaries can be interpreted as a way of simultaneously offering credibility to the players’ skill, which is rendered visible by the device, and to the measuring apparatus itself; the latter being usually provided by a manufacturer of gaming hardware who also acts as the event’s sponsor. However, insights gained from the eye tracking data rarely go beyond what the transmitted game image already conveys to the audience11: The measured player’s focal point (see Fig. 5, light blue area at the bottom of the screen) usually jumps to the enemy characters during moments of 11 This finding seems trivial when one considers that the speed and precision of eye movements are reflected in the game actions that immediately follow them, meaning the movement and aiming processes. 105 ... Pablo Abend / Max Kanderske Playful Metadata confrontation, and otherwise moves back and forth between the interface elements relevant for gameplay, those being the counters for ammunition and health points. Eye tracking thus advertises an ideal of technically me- diated visualizations of embodied knowledge as well as the hard- and software products brought to bear for this purpose. Crucially, it cannot de- liver on the promise of visualizing concrete decision-making patterns and thus fails to improve the audience's understanding of the game.12 Fundamentally, the practices of visualizing and optimizing movement patterns can be seen in the tradition of scientific management's move- ment studies: For example, the eye movements depicted as ghostly traces are reminiscent of Frank Bunker Gilbreth's long-exposure film recordings for the analysis of work processes. (Hoof 2015) In the context of compet- itive gaming commentary, however, the practices of making bodily states and bodies of knowledge visible do not follow the telos of sequence opti- mization usually found in movement studies. Instead, they are employed in the service of an economically motivated affect modulation aimed at gaining and maintaining viewership numbers. 4 . QUANTIFIED GAMEPLAY BETWEEN SELF-MEASURE- MENT AND AFFECT MODULATION It seems obvious to relate the quantification of gameplay to the overarch- ing practices of a data-based lifestyle. The purposes also seem to be sim- ilar at first sight. Especially the sensors involved are comparable to those used in the Quantified Self movement (motto: “Self-Knowledge Through Numbers”) and in the field of so-called personal informatics, (Lupton 2016; Abend/Fuchs 2016). Self-measurement activities with the help of digital sensors and mobile technologies such as smart watches can be considered modern, i.e. digital techniques of the self. (Foucault 1993, 26) These techniques of the self have a history that can be told along the 12 It is fair to say, however, that the shooter genre offers little room for surprising eye movements due to the focal point (the crosshairs in the center of the screen) being firmly inscribed in the game image. The situation is different in the strategy game, where a larger space, which is doubled once again by the mini-map, must be cap- tured with the gaze. 106 Spiel |Formen Special Issue: Ludomater ial it ies changing ways we take care of ourselves and the media practices we em- ploy to that end. In this context, self-observation through quantifying technologies is not to be seen so much in the tradition of (technologically supported) observations of consciousness and the mind but rather of medical practices that monitor vital functions and bodily responses. What most forms of self-measurement have in common is that this monitoring of vital functions is supposed to lead to an optimization of everyday rou- tines in the sense of a healthier life. Such somatization of everyday practices, where introspection refers not to work on the inner mental life but to self-engineering aimed at the body, can also be observed in quantified play. Consequently, the add-ons and peripherals used to computerize the game are primarily presented and marketed as performance-enhancing. In addition, monitoring is sup- posed to offer an unspecified enrichment of the gaming experience, which presumably appeals to the ideals of total control and efficiency com- monly associated with the accumulation of data. At the same time, it promises a component of generating entertaining insights about one’s own game – insights whose appeal might be grounded precisely in the fact that the game itself does not provide this kind of information. Another commonality shared between the practices of quantified play and Quantified Self is the transformation from a “technology of the self” to a “technology of the social,” (Lemke 2011) from self-measurement as an individual action to the sharing of acquired data with others (a func- tionality supported by the majority of commercially available tracking and tracing technologies). While terms such as self-tracking and personal in- formatics attribute self-monitoring to the sphere of private media use, the insights gained do not remain tied to the individual: Data is shared locally (with other members of the QA scene or with friends on social media) or circulates (semi-)publicly on digital platforms, some of which are provided by the technology providers. Quantifying gaming also initially seemed a practice taking place exclu- sively between the user, the game, and the quantifying interface. How- ever, since increasing one's own performance is also about creating com- parability with other players, it is not surprising that practices of self- 107 ... Pablo Abend / Max Kanderske Playful Metadata measurement can be found in well-networked communities, especially in the field of competitive gaming. On streaming platforms such as Twitch, but also on statistics pages like Dotabuff, individual self-observation be- comes a social technique of the body and can thus be understood as a form of dressage of the body. (Mauss 1974, 208) Technologies of quantified play, such as the Naos QG mouse men- tioned at the beginning of the text, represent a trend in digital gaming cul- ture to monitor one's own performance on a small scale and to optimize it in order to increase efficiency. The manufacturers of quantifying hard- ware propagate that this is a way of reflecting on one's own gaming and thus also improving it. (Egliston 2020, 2) As a rule, this is done by means of visualizations that are displayed during gameplay or that can be ac- cessed afterwards. This creates a second feedback loop to the game that adds further parameters to its output, allowing one's own playing to be adjusted to the displayed values. Depending on the genre and type of quantification, this adaptation can be done in quasi real-time or in a sub- sequent reflection phase. Ash speaks of an exteriorization of gameplay through proprietary tracking platforms. (Ash 2015, 109) According to him, the quantification of gameplay provides contextualization within an ini- tially individual performance career. To exaggerate, one could say that by providing the tools to describe such careers, the corresponding measure- ment, documentation, and comparison technologies and practices make their existence possible in the first place. The decisive factor here are au- tomatic documentation mechanisms that draw statistical connections be- tween matches that exist separate from each other on a gameplay level. The selection and visual representation of the displayed data decisively influence how individual performance careers - and by extension one's own relationship to the games played - are perceived. The manufacturing companies exploit this connection in various ways. In the simplest case, absolute numerical values, which necessarily increase over time (e.g., the total number of games won), are placed prominently on the player's pro- file, while other – potentially demotivating – relative values (such as the percentage of games won) remain “hidden” in submenus. Here we can speak of targeted affect modulation on the part of the developers and 108 Spiel |Formen Special Issue: Ludomater ial it ies platform operators: The data is used as material to generate positive af- fective states, highlighting one's own skill development in particular, in or- der to encourage the continuation of one's gaming career. At the same time, negative affective states associated with personal mistakes and losses are cushioned by a narrative of long-term improvement against which failures take the shape of temporary set-backs.13 5 . CONCLUSION Quantification makes it possible to connect individual performance ca- reers to larger digital economic contexts: The measurement data of quan- tified gameplay does not remain in the feedback loop between the game and the player but is displayed and adapted for (affective) economic pur- poses of players, manufacturers, and platform operators. Fig. 6: Statistics banner in APEX LEGENDS showing the leading player. (Respawn Entertainment 2019) In this regard, the collected playful metadata contribute in various ways to the formation and development of the material arrangements from which they emerge and in which they are embedded. For example, they can form the basis for adjustments to game balance or – visualized as a 13 Ben Egliston describes these mechanisms with the conceptual pair of proximity and distance. (Egliston 2020a, 10) 109 ... Pablo Abend / Max Kanderske Playful Metadata hybrid of in-game scoreboard and player profile – reinforce competition among players (see Figure 6). By offering a trajectory for the professionalization of play, playful metadata undermine established notions of a strict separation of play and labor, (Huizinga 1956) contributing to the increasing diffusion of both spheres that is expressed in hybrid concepts such as “playbor” (Küklich 2005) and “laborious play” (Abend et al. 2016). Accordingly, the profes- sionalization of play can be related to the gamification of work processes since both are underpinned by infrastructures and practices of measure- ment, quantification, and calculation. The permeability between private play-as-leisure and professional play-as-income that is inscribed into both streaming and professional play ensures that players become part of potentially exploitative structures of data aggregation from the get-go. However, as the analysis of speedrunning practices has shown, playful metadata can also become an instrument for transgressive or transform- ative play, as provides metrics and goals not envisioned by the original developers. By investigating the anchoring practices of sequencing, logging/calculat- ing and visualizing, we have shown that the player’s appropriation, devel- opment and refinement of gameplay actions and goals is mutually de- pendent on the (re-)formation of material arrangements. It is characteris- tic that the playful metadata collected by the players is simultaneously used for cooperative knowledge transfer (e.g., in speedrunning or in the fighting game community), but also for competitive comparison. Playful metadata enables communities of practice to jointly undertake the pro- ject of approaching their ideal of good gameplay, while it allows the play- ers to compete more effectively with each other in individual games or races. 110 Spiel |Formen Special Issue: Ludomater ial it ies REFERENCES Amazon (n.a.) (2020): Product Description of the Tobii Eye Tracker. Online: https://www.amazon.com/Tobii-Eye-Tracker-Game- changing-Peripheral/dp/B01MAWPMXQ [Accessed 10.03.2020]. Amazon (n.a.) (2020a): Product description of the SteelSeries Sentry Eye Tracker. Online: https://www.amazon.de/SteelSeries-Sentry-Gam- ing-Eye-Tracker/dp/B00SB8YKVI [Accessed 10.03.2020]. Aarseth, Espen (2007): I Fought the Law: Transgressive Play and The Im- plied Player. In: Proceedings of the 2007 DiGRA International Confer- ence: Situated Play, pp. 130–133. Abend, Pablo/Fuchs, Mathias (Eds.) (2016): Quantified Selves and Statis- tical Bodies. In: Digital Culture & Society, vol. 2, no. 1. Abend, Pablo/Fizek, Sonja/Fuchs, Mathias/Wenz, Karin (Eds.) (2019): La- borious Play. In: Digital Culture & Society, vol. 5, no. 2. Abend, Pablo/Fuchs, Mathias/Wenz, Karin (Eds.) (2021): Playful Work. In: Digital Culture & Society, vol. 7, no. 1. Andrejevic, Mark (2004): The Work of Watching One Another: Lateral Sur- veillance, Risk, and Governance. In: Surveillance & Society, vol. 2, no. 4. DOI: 10.24908/ss.v2i4.3359. Ash, James (2012): Technology, Technicity, and Emerging Practices of Temporal Sensitivity in Videogames. In: Environment and Planning A, vol. 44, no. 1, pp. 187–203. DOI: 10.1068/a44171. Ash, James (2015): The Interface Envelope: Gaming, Technology, Power. New York, NY: Bloomsbury 2015. Cameron, Zac (2018): H1Z1 Pro League Heart Rate Monitors Reveal Who Can Keep Their Cool. Online: https://www.twingalax- ies.com/feed_details.php/1180/h1z1-pro-league-heart-rate-moni- tors-reveal-who-can-keep-their-cool [Accessed 16.05.2022]. Couldry, Nick (2004): Theorising Media as Practice. In: Social semiotics, vol. 14, no. 2, pp. 115–132. 111 ... Pablo Abend / Max Kanderske Playful Metadata Egliston, Ben (2020): Quantified Play: Self-Tracking in Videogames. In: Games and Culture, vol. 15, no. 6, pp. 707-729. DOI: 10.1177/1555412019845983. Egliston, Ben (2020): Surveillance Technicity: Affect, Retention and Vide- ogame Analytics. In: Media, Culture & Society, vol. 42, no. 6, pp. 915- 931. DOI: https://doi.org/10.1177/0163443719880139. Foucault, Michel (1993): Technologien des Selbst. Frankfurt a.M: S. Fischer. Guttmann, Allen (1978): From Ritual to Record. The Nature of Modern Sports. New York: Columbia University Pres. Haug, Fritz (1971): Kritik der Warenästhetik. Frankfurt a.M.: Suhrkamp. Hawranke, Thomas (2018): Modding – Künstlerische Forschung in Compu- terspielen. Weimar: Bauhaus-Universität Weimar. Hoof, Florian (2015): Engel der Effizienz. Eine Mediengeschichte der Unter- nehmensberatung. Konstanz: Konstanz University Press/Fink Verlag. Huizinga, Johan (1956): Homo Ludens: Vom Ursprung der Kultur im Spiel. Hamburg: Rowohlt. Jobst, Karl (2019): The Worst Fake Speedrun on Youtube. Online: https://www.youtube.com/watch?v=AFrQ1_2bbsI [Accessed 10.03.2020]. Knorr, Alexander (2009): Die kulturelle Aneignung des Spielraums. Vom virtuosen Spielen zum Modifizieren und zurück. In: Bopp, Matthias/Wiemer, Serjoscha (Eds.): Shooter. Eine multdisziplinäre Einführung. Münster: Lit pp. 217–246. DOI: http://dx.doi.org/10.25969/mediarep/1875. Krell, Jason (2019): Is esports a sport? Researchers undecided. Online: https://globalsportmatters.com/culture/2019/09/18/is-esports- a-sport-researchers-undecided/ [Accessed 16.05.2022]. Kücklich, Julian (2005): Precarious Playbour. Modders and the Digital Games Industry. In: Fibreculture Journal, vol. 5, no. 1. Online: http://www.journal.fibreculture.org/issue5/kucklich_print.html [Accessed 11.03.2020.]. 112 Spiel |Formen Special Issue: Ludomater ial it ies Lave, Jean/Wenger, Etienne (1991): Situated Learning: Legitimate Periph- eral Participation. Cambridge: Cambridge University Press. Lemke, Thomas (2011): Beyond Foucault: From Biopolitics to the Govern- ment of Life. In: Bröckling, Ulrich/Krasmann, Susanne/Lemke, Thomas (Eds.): Governmentality. Current Issues and Future Challenges. New York: Routledge. Lupton, Deborah (2018): Lively data, social fitness and biovalue: the inter- sections of health and fitness self-tracking and social media. In: Bur- gess, Jean/Marwick, Alice/Poell, Thomas (Eds.): The sage handbook of social media. Thousand Oaks, CA: SAGE Publications, pp. 562-578. https://dx.doi.org/10.4135/9781473984066.n32. Lupton, Deborah (2016): The Quantified Self. A Sociology of Self-Tracking. Cambridge, UK/Malden, MA: John Wiley & Sons. Online: http://gbv.eblib.com/patron/FullRecord.aspx?p=4678321 {Ac- cessed 06.02.2020]. Malaby, Thomas M. (2007): Beyond Play: A New Approach to Games. In: Games and Culture, vol. 2, no. 2, pp. 95–113. Mauss, Marcel (1974): Die Techniken des Körpers. In: id.: Soziologie und Anthropologie, vol. 2. Munich: Carl Hanser, pp. 197–220. Mionix (2020): Description of the NaosGQ mouse. Online: https://mi- onix.io/collections/naos/products/naos-qg [Accessed 10.03.2020]. Pias, Claus (2000): Computer Spiel Welten. Weimar: Bauhaus-Universität Weimar. Pink, Sarah/Ruckenstein, Minna/Willim, Robert/Duque, Melisa (2018): Broken Data: Conceptualising Data in an Emerging World. In: Big Data & Society, vol. 5, no. 1. DOI: 10.1177/2053951717753228. Pulsoid (2020): Pulsoid Homepage. Online: https://pulsoid.net/ [Ac- cessed 10.03.2020.] Pulsoid (2020a): Pulsoid FAQ. Online: https://pulsoid.net/ [Accessed 10.03.2020.] 113 ... Pablo Abend / Max Kanderske Playful Metadata Rhee, Jennifer (2018): The Robotic Imaginary – The Human & the Price of Dehumanized Labor. Minneapolis MN: University of Minnesota Press. Salen, Katie/Zimmerman, Eric (2003): Rules of Play: Game Design Funda- mentals. Cambridge MA: MIT Press. Schatzki, Theodore (2006): Introduction: Practice Theory. In: Schatzki, Theodore R./Knorr-Cetina, Karin/von Savigny, Eike (Eds.): The prac- tice turn in contemporary theory. London: Routledge , pp. 10–23. Schatzki, Theodore R (2016): Practice Theory as Flat Ontology. In: Spaar- garen, Gert/Weenink, Don/Lamers, Machiel (Eds.): Practice theory and research. Exploring the dynamics of social life. London, New York: Routledge Taylor & Francis Group, p. 28-42. Star, Susan Leigh/Griesemer, James R. (1989): Institutional ecology,trans- lations’ and boundary objects: Amateurs and professionals in Berke- ley's Museum of Vertebrate Zoology, 1907-39. In: Social studies of science, vol. 19, no. 3, pp. 387-420. Suits, Bernard (2002): The grasshopper. Games, life and utopia, Peterbor- ough: Broadview Press. Swidler, Ann (2006): What Anchors Cultural Practices. In: Schatzki, Theo- dore R./Knorr-Cetina, Karin/von Savigny, Eike (Eds.): The Practice Turn in Contemporary Theory. London: Routledge, pp. 83–101. Tobii (n.a.) (2022): Tobii homepage. Online: https://www.tobii.com/ [Ac- cessed 16.05.2022]. Wenger, Etienne (1998): Communities of practice: Learning as a social sys- tem. In: Systems thinker, vol. 9., no. 5, pp. 2–3. Wolmarans, Kyle (2016): A German scientific study shows eSports profes- sionals are athletes. Online: https://www.criticalhit.net/gam- ing/german-scientific-study-shows-esports-professionals-ath- letes/ [Accessed 16.05.2022]. Zuboff, Shoshana (2015): Big Other: Surveillance Capitalism and the Pro- spects of an Information Civilization. In: Journal of Information Tech- nology, vol. 30, no. 1, pp. 75–89. 114 Spiel |Formen Special Issue: Ludomater ial it ies Zuboff, Shoshana (2019): The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. London: Profile Books. GAMES APEX LEGENDS (RESPAWN ENTERTAINMENT, 2019) COUNTER-STRIKE: GLOBAL OFFENSIVE (Valve/Hidden Path Entertainment, 2012) DOTA 2 (Valve, 2013) THE LEGEND OF ZELDA: OCARINA OF TIME (Nintendo, 1998) STREET FIGHTER V (Capcom, 2016) WARCRAFT III (Blizzard 2002) ABOUT THE AUTHORS Pablo Abend is Professor for Design Theory at Burg Giebichenstein Uni- versity of Art and Design Halle, Germany. Before that, he worked as a sci- entific coordinator of the interdisciplinary research school “Locating Me- dia” at University of Siegen. He received his PhD in 2012 for his theses on the history and present of cartographic media and practices. His research interests include participatory design, qualitative methods, geographic media, gaming cultures, and Science and Technology Studies. Max Kanderske is a PhD candidate at the chair of Science, Technology & Media Studies at the University of Siegen and a research associate in pro- ject A03 – “Navigation in online/offline spaces” of the Collaborative Re- search Centre “Media of Cooperation”. His research interests include me- dia geography, media history, navigational media, STS, and Game Studies. Contact: max.kanderske@gmail.com 115 ...