Gamifying Fitness or Fitnessifying Games: a Comparative Study Andreas Leibetseder, Mathias Lux Institute of Information Technology Alpen-Adria University, 9020 Klagenfurt Austria firstname.lastname@itec.aau.at access replace tangible city maps, on-demand video streaming gives people the option to watch movies Abstract at home rather than having to visit a theater and action-packed computer games are way more fun than Fitness- or exergames are ubiquitously avail- playing jump-rope outside. These ”advantages” over able, but often lack the main ingredient of the past, however, frequently entail grave detrimental successfully gamified systems: fun. This can e↵ects: social isolation [MSLB06], cognitive laziness be attributed to the typical way of designing [Car08] and physical inactivity [KHKI+ 05] have been such games – highly focusing on specific physi- reported to a↵ect many people’s lives, the latter even cal activities, thus, gamifying fitness. Instead, being deemed the biggest public health problem of the we propose a novel alternate approach to im- 21st century [Bla09]. prove motivation for exergaming, which we Despite current trends, exercising to many individ- call fitnessification: integrating physical ex- uals yet is desirable and oftentimes all it takes to get a ercise into very popular games that have been person started is the right kind of motivation – some- developed keeping fun in mind and frequently thing ironically technology in combination with a spe- are played for long periods of time – so-called cial kind of gamification 2 can provide [WH12]: rather AAA games. In order to evaluate this con- than using specific game design elements to encour- cept, we have conducted a comparative study age sportive activities, as is commonly done in vari- examining voluntary participants’ reactions to ous fitness apps and devices, it is possible to employ testing an ergometer controlled casual game full-blown video games that interact with fitness acces- as well as a modified AAA game. Results in- sories, a popular concept pursued by all major gaming dicate strong tendencies of players preferring consoles as well as arcade machine manufacturers. the newly introduced AAA approach over the By now such exergames are ubiquitously available casual fitness game. and their substantial estimated market share of over 6.6 billion dollars highlights a general public interest in 1 Introduction purchasing them [DGL08]. However, in spite of those facts, the buyers’ initial excitement often quickly fades “Technology brings you great gifts with one hand, and and widely promoted beneficial e↵ects like changing it stabs you in the back with the other.”1 – British sci- peoples’ physical exercise habits are limited to a short entist Charles Percy Snow in his comment to The New period of time, after which the games are shelved3 , York Times addresses a common dilemma in modern making room for new, more exciting games. Several society. Technology regularly is associated with con- studies underline that notion of rapidly declining in- venience because it is able to facilitate and improve terest [BAB+ 12, LML+ 06, MLB13], emphasizing that many aspects of our lives: smartphones with internet ultimately fun plays a vital role in any game – a game 2 For our purposes we define gamification based on Wer- Copyright c by the paper’s authors. Copying permitted for pri- bach/Deterding and colleagues’ definitions [WH12, DDKN11], as vate and academic purposes. the use of game elements and game-design techniques in non- In: F. Hopfgartner, G. Kazai, U. Kruschwitz, and M. Meder (eds.): game contexts to solve problems and engage users. Proceedings of the GamifIR 2016 Workshop, Pisa, Italy, 21-July- 3 Michael Rosenberg for The Conversation, 2014: ”Sorry gamers, Wii 2016, published at http://ceur-ws.org Fit is no substitute for real exercise” (http://tinyurl.com/pzyopx8, 1 Quote first appeared in The New York Times, 15 March 1971. Web. 29 Jun. 2016) lacking fun just turns into a plain task. The root of sufficient instructions to be able to try out the finished this problem becomes apparent when regarding the in- projects for as long as they liked with a nonbindung tuitive but naive design approach for a fitness game: suggestion of five minutes and in immediate succes- gamifying sports strongly advocates a focus on physi- sion were asked to fill out an anonymized questionnaire cal activity – fun games, however, are not really fixated about their experiences, which took place on-site but on anything other than bringing joy and satisfaction unsupervised (see Section 4.1 for more details about to their players. Therefore, a much more interesting the surveys). Each of the studies has been preceded by approach is to shift said focus onto popular games and preliminary trial runs, where students and colleagues enhancing them with fitness elements, i.e. applying have tested the respective prototype and gave valuable what we call fitnessification. feedback, which helped improving it prior to the actual A broad category of such popular games is com- examination. prised of so-called AAA games (pronounced ’triple-A games’), which generally are created by big companies 2.1 Research Goals often investing huge amounts of money in their de- The main purpose of this work is to show that fitness- velopment. Compared to current fitness games they enhanced AAA games are more enjoyable than games do a much better job at entertaining players, as they specifically designed for supporting physical exercise. often bear high replay values, feature the latest ad- Hence, the central evaluation statement S can be sum- vanced technologies and are generally designed much marized as follows: more thoughtfully. Good AAA games maintain their popularity for long periods of time and sometimes even S Fitnessification of games encourages more positive are played for decades4 , while enjoying an active and feedback than gamification of sports. vibrant community of fans. Accordingly, they usually are referred to as the most attractive and anticipated Based on this goal, several research questions can games on the market. be developed, summarizing the purpose of this thesis This work should be regarded as an impulse to sup- as well as related future aspects: port the notion of integrating fitness elements into such fun games and, thereby, increasing the motivation for Q1 How do sportive/unsportive people feel about ex- exergaming. For this purpose we have conducted two ergaming? comparable field studies as a way of proofing our novel Q2 Can exergaming be more fun? concept of fitnessification: an o↵-the-shelf ergometer Q3 Given the AAA game integration, would it be pos- has been utilized to control a purpose-built casual and sible to motivate people to work out more? a modified AAA game prototype – Flappy Cycling Q4 Which conditions must be satisfied for gamers to (FC) versus Quake 3 Arena Fitness Edition (Q3AFE). consider exergaming at home? After trying out these games voluntary participants conveyed their opinions towards sports, gaming, ex- Several hypotheses have been developed based on ergaming and the tested prototypes via identical ques- above questions, listed in Section 4.2. tionnaires for both examinations. Our contributions include: 3 Architecture and Games • Study overview and research goals (Section 2) From a hardware point of view, a fitness device needs • Exergames and their architecture (Section 3) to be suitably connected to a computer and the archi- • Hybrid evaluation results (Section 4) tecture approach taken is depicted in Figure 3. • Related work (Section 5) As mentioned in the Introduction, an ergometer5 • Conclusions and further work (Section 6) has been used as a sports device in both studies, pro- viding two kinds of data – the cycling rate and pulse 2 Study Overview sensor information via two attached handles. To be able to properly capture signals from the device, an The studies have been set up on two di↵erent occasions Arduino Leonardo 6 was used, which is a powerful, pro- in the same public location, the aula of the Alpen- grammable open-source microcontroller board. Addi- Adria University (AAU). Both involved user experi- tionally, a separate pulse sensor7 was employed, as to ence analyses were kept as similar as possible: in a supervised field study, voluntary passers-by were given 5 Model: Ultrasport F-Bike 150/200B (http://tinyurl.com/jlqnbna, Web. 29 Jun. 2016). 4 6 E.g. ”Counter-Strike” (Valve Corporation, 1999: http://tinyurl. Product information page: http://tinyurl.com/qc2xzda, Web. 29 com/ccat4e, Web. 29 Jun. 2016) and ”World of Warcraft” (Blizzard Jun. 2016. 7 Entertainment, 2004: http://tinyurl.com/39vdalk, Web. 29 Jun. Made by World Famous Electronics (http://pulsesensor.com, 2016) Web. 29 Jun. 2016). This prototype represents the conventional way of de- signing an exergame – building it around the fitness activity. 3.2 Quake 3 Arena Fitness Edition (Q3AFE) Quake 3 Arena (Q3A) is the third installment of the popular Quake series10 , which included first-person shooters (FPS) that have been among the first games featuring true real-time 3D game engines, released by id Software11 in 1999. The series achieved tremendous success and even nowadays Q3A still is a popular ti- tle, particularly on so-called LAN parties, where play- ers get together, establish local connections with their Fig. 1: HW Architecture for FC, Q3AFE and both. PCs and hunt each other down in virtual deathmatches with the goal of eliminating as many opponents as pos- enable controlling a game with a joystick while mea- sible, i.e. scoring kills or frags. This AAA game has suring the user’s heart rate through merely applying been chosen for applying fitnessification because back a finger strap (see Section 3.2). Fitness input used for in the 1990s the series’ developers picked up on a newly both games is colored in violet, while FC only elements emerging trend, which they wholeheartedly supported are marked blue and Q3AFE components are shown in Q3A: game modding, i.e. modifying popular games in red. Every signal is translated into keypresses by in terms of visuals, mechanics or gameplay. Q3As en- the Arduino, serving as input commands to the games gine is constructed in a modular manner and parts described in following sections. of it can easily be replaced by customized dynamic- link libraries (DLLs). Thus, standard gameplay could 3.1 Flappy Cycling (FC) comfortably be adjusted to incorporate user-generated Flappy Cycling is inspired by unexpected smash hit fitness information for our study. Q3A basically is casual game Flappy Bird8 and has been implemented a multi-player game but also features a single player from scratch using popular HTML5 game developer’s mode, which has been a requirement for the result- framework Phaser9 . Flappy Bird is essentially a ing (Q3AFE) modification (mod ) in order to be com- smartphone-based endless side-scroller, giving players parable to FC. In this mode, the player’s opponents, control over a bird, which flies through a horizon- so-called bots, are controlled by the computer via an tally scrolling scene and must avoid hitting randomly artificial intelligence (AI) library. appearing green obstacle pipes. Tapping the screen A typical deathmatch begins by randomly putting gives the animal an upward boost, refraining to do the player as well as all bots on distinct starting so lets simulated gravity take its course. Flappy Cy- spots set throughout a virtual level (map), which is cling builds on this idea by boosting the bird’s vertical called spawning. Newly spawned players have a dis- screen position on every fully completed pedal cycle tinct amount of health points (HP) and some starting and instead of dodging randomly appearing static ob- weapons. The goal is to win a tournament by being the jects players must evade enemy birds flying towards first to kill a previously set amount of opponents (frag them in wavy patterns. This can be achieved via limit). Being killed by an enemy triggers a respawn, touching the ergometer’s pulse sensor handles, which again at a random location. Throughout the game, were diverted from the intended use in order to merely weapons and power-ups, boosting a player’s abilities, detecting whether someone is touching either of them, can be acquired, all of which are scattered throughout enabling the controlled bird to horizontally fly for- the map and as well constantly respawn after small wards and backwards. A player’s high-score is cal- timeouts when picked up by players. Study partici- culated by gradually increasing a point multiplier over pants played one such scenario, while being rewarded time, which is reset every time an enemy bird is hit or for performing well on the ergometer. the bird leaves the middle third of the screen on the Compared to FC, a lot more buttons are needed vertical axis. The goal is to rack up as many points to control Q3AFE: walking forwards, backwards or as possible within a five-minute time limit, while the sideways, looking around (panning the camera), jump- amount and speed of enemies constantly increases. ing, ducking, shooting, switching weapons, zooming 8 10 Developed by Dong Nguyen and deemed most addicting mobile Quake series on Steam: http://store.steampowered.com/sub/434, games of 2013 (http://tinyurl.com/z6sls4s, Web. 29 Jun. 2016). Web. 29 Jun. 2016. 9 11 Phaser HTML5 framework (http://phaser.io, Web. 29 Jun. 2016) http://www.idsoftware.com, Web. 29 Jun. 2016 and even more. Therefore, an additional input pos- H1 Gamers would rather play AAA exergames than sibility has been employed: a common gamepad di- casual exergames at home (A✓F C < B✓Q3AF E ). rectly plugged into the PC. Holding such a game con- H2 Active12 people would rather play AAA ex- troller, in turn, renders the ergometer’s pulse handles ergames than casual exergames at home (A✓F C < useless, as touching them requires letting go of the B✓Q3AF E ). gamepad, which would be fatal in a fast-paced FPS H3 Regular gamers who are less active would like Q3A. Since the pedals then would be the sole rather consider working out more playing AAA remaining means of capturing fitness input, a finger- exergames than casual exergames (A✓F C < attachable pulse sensor has been added as well. Al- B✓Q3AF E ). together, the ergometer pedals controlled the speed H4 Active people would rather play exergames at of the player and users were rewarded with in-game home than rarely active people (H4a : A1✓F C < power-ups, when staying in a previously selectable B1✓F C , H4b : A2✓Q3AF E < B2✓Q3AF E ). heart rate zone (recovery, aerobic, anaerobic). H5 Less active people would rather consider work- ing out more by playing exergames than regu- 4 Evaluations larly active persons ( H5a : A1✓F C > B1✓F C , H5b : A2✓Q3AF E > B2✓Q3AF E ). A hybrid approach including a qualitative (Section 4.3) H6 Regularly active occasional gamers would rather and quantitative analysis (Section 4.4) has been chosen play AAA exergames than casual exergames at for evaluating the collected survey data, consisting of home (A✓Q3AF E > B✓F C ). altogether 33 (FC) plus 34 (Q3AFE) records. Before H7 Regularly active people would rather consider outlining discovered results, Section 4.1 describes the working out more playing AAA exergames than essence of the questionnaire and subsequent Section casual exergames (A✓F C < B✓Q3AF E ). 4.2 lists hypotheses formulated prior as well as after H8 Cyclists would rather consider playing AAA carrying out the studies, the latter resulting from con- exergames than casual exergames (A✓F C < ducting a variable correlation analysis. B✓Q3AF E ). H9 Students/gamers/all participants enjoy playing 4.1 Questionnaire AAA exergames more than they do playing ca- sual exergames (H9a : A2✓F C < B2✓Q3AF E , In general, the questionnaire can be grouped into five H9b : A1✓F C < B1✓Q3AF E , H9c : A3✓F C < categories: demographic data (general), user habits B3✓Q3AF E ). and opinions (sports, gaming & their combination) H10 Gamers/all participants have better experience and prototype specific questions. For the quantita- with the ergometer when playing a AAA ex- tive analysis, response statistics, as well as answers ergame instead of a casual exergame (A✓F C < optionally entered into several free text boxes, have B✓Q3AF E ). been examined. A total of 23 relevant variables can H11 Lower ergometer experience influences overall ex- be defined out of the survey questions, a listing of perience negatively (H11a : A1✓F C < B1✓F C , which, however, would go well beyond the scope of H11b : A2✓Q3AF E < B2✓Q3AF E ). this paper. Variables either are of ordinal (5-valued H12 Regular Gamers rate AAA exergames higher than tendency scale) or dichotomous (bi-valued) nature: for casual exergames (A✓F C < B✓Q3AF E ). any quantitative analysis two selected datasets (usu- H13 Less regular active people are less comfort- ally A✓F C and B✓Q3AF E ) either were partitioned by able handling the ergometer than regularly ac- an existing dichotomous variable or via bisecting re- tive persons (H13a : A1✓F C < B1✓F C , H13b : sponses for an ordinal scale variable. Afterwards an- A2✓Q3AF E < B2✓Q3AF E ). other ordinal variable has been used to evaluate ten- dencies in both resulting groups. 4.3 Qualitative Analysis 4.2 Hypotheses Collected data for general type questions for both studies reveals a predominant male participation (> Six hypotheses (H1-H6) have been developed prior to 70%). Therefore, conducting any gender-specific anal- conducting the studies and seven (H7-H13) resulted yses on the accumulated records would not be feasi- from post-study analysis of variable correlation co- ble. The average specified ages range from 25 to 30 efficients: -coefficient [Cra45] for two dichotomous and in particular, almost all participants were below features, Kendall rank correlation with tie correction 35 years old with altogether five exceptions. Close to [Ken38, Ken48] for ordinal data comparison and the half of the volunteers, in equal proportions for both rank-biserial correlation [Cur56] for pairs of both vari- able types. 12 Active refers to practicing sports in all hypotheses. studies, were students and the majority of the remain- Tab. 1: Hypotheses MWW Evaluations. Set partitions: ing people were employed in a technical field. This is A & B, MWW p-value: p, test statistic: z, MWW observed U-value: expectable, since study setups, such as the ones con- u, U-value std. error: u. ducted for this thesis, especially appeal to people inter- H1 |A| |B| p z u reject H0 ested in technology. Despite the AAU’s multicultural u student base13 , most of the test subjects have been H1 16 21 0.038 -1.774 118 168 for ↵ = 0.05 H2 23 23 0.087 -1.360 211.5 264.5 for ↵ = 0.10 from Austria (FC: 76%, Q3A: 70%) and in general, H3 4 10 0.243 -0.698 17 20 no almost all participants were from Europe with two ex- H4a 5 19 0.349 -0.388 43 47.5 no ceptions. H4b 9 16 0.437 -0.159 69.5 72 no When regarding the gathered study-specific statis- H5a 13 11 0.057 1.581 50 71.5 no (U↵=0.10 < u) H5b 17 8 0.089 1.344 55 68 no (U↵=0.10 < u) tics, both studies influenced user experience positively H6 3 8 0.882 -1.186 7.5 12 no as well as negatively, although Q3AFE seems to have H7 11 8 0.187 -0.888 35 44 no been better accepted. The reason both experiments H8 20 10 0.171 -0.952 85 100 no overall received fairly high ratings could be due to their H9a 15 14 0.001 -2.980 42 105 for ↵ = 0.01 H9b 22 27 0.030 -1.880 219.5 297 for ↵ = 0.01 novelty e↵ect: in both cases, only 15% claimed to have (u < U↵=0.01 ) partaken in similar studies and comparable home de- H9c 33 34 0.006 -2.500 385.5 561 for ↵ = 0.01 vices are still rare. Nevertheless, the biggest issues re- H10 22 27 0.003 -3.481 140.5 297 for ↵ = 0.01 ported were control-specific, i.e. difficulties with FC’s H11a 21 12 0.145 -1.058 101 126 no H11b 8 26 0.010 -2.320 54.5 104 for ↵ = 0.01 pulse handles and Q3AFE’s use of a gamepad instead H12 9 16 0.167 -0.967 56.5 72 no of keyboard/mouse, which in future versions should be H13a 17 16 0.051 -1.639 96 136 for ↵ = 0.10 improved as to prevent any biased opinions caused by H13b 23 11 0.179 -0.918 104.5 126.5 no these inconveniences. Furthermore, the most impor- tant future aspect should be testing long-term user en- the cumulative normal function for calculating the p- gagement in playing these games. Outcomes for such value in order to determine if H0 can be rejected in experiments probably depend on how regularly users favor of the alternate hypothesis H1 . However, if the play games, but judging by several user remarks FC cardinality of both sets is below or equal 20, precalcu- might not perform as well as Q3AFE: “Game seems a lated tables listing exact U-values must be consulted to bit random.” (FC), “It is a fun game, but I think it ensure a correct evaluations, which has been applica- will get boring after a while.” (FC). ble in several hypotheses and of particular importance in hypotheses H5a , H5b and H9b , as relying on the 4.4 Quantitative Analysis p-value alone would have yielded di↵erent results in Most of the questions prompted the users to rate an terms of significance. aspect on a 5-valued ordinal scale, which overall did Significant outcomes of H1, H2 and H9 confirm prior not yield normally distributed results, as is mostly qualitative findings, indicating strong tendencies to- the case when employing variables not describing in- wards preferring AAA exergames over casual fitness tervals [Sca13]. Therefore, analogous to previously games when examining students, gamers and active listed correlation coefficients, a non-parametric evalu- people. Moreover, controlling Quake 3 Arena with an ation method has been chosen for all hypothesis eval- ergometer seems to have been received far better than uations: the Mann-Whitney-Wilcoxon (MWW) test using it to operate Flappy Cycling (H10). However, it [Wil47] with significance ↵ = 0.05. The collective re- can not be confirmed that the people’s fitness level nor sults are listed in Table 1, pointing out for which hy- their specific interest in bike riding are indicators for potheses H0 can be rejected on significance levels of them to consider using fitness gaming devices at home. up to ↵ = 0.10. Nevertheless, clear tendencies can be discovered (H5, The MWW test is used to determine if two popu- H7, H8, H11, H12, H13) and thus such assumptions lations (A, B) di↵er in their distributions, i.e. the as- should be of interest for further investigations. sumed null hypothesis H0 : A = B can be rejected in favor of an alternative hypothesis H1 : A 6= B, A < B 5 Related Work or A > B. It accomplishes this by ranking the ele- Gamification for health and fitness is an emerging ments of both sample sets (lower value are assigned trend that is frequently incorporated into smartphone smaller ranks), summing up those ranks for both sets apps or Internet portals. Games for fitness are closely (tied ranks are averaged) and computing the di↵erence related to those applications as they share the same in rankings (U-value). It assumes an approximate nor- goal: motivating people for a healthier lifestyle. This mal distribution of U-values to be able to make use of is accomplished by integrating exercise into games, a 13 cf. AAU report 2015/16 (http://tinyurl.com/jqq2c7h). concept widely adopted in development for home gam- ing consoles but also scientific research. Most popu- custom solutions. Conclusions are of mixed nature lar fitness games are part of series that are still on- – partly positive [MVHV03, LML+ 06, PBBvDN09, going to this day. It is noteworthy that all of those BC10, BVSZ13], indicating an increase in physical ac- games specifically are tailored for exercising with var- tivity, and partly inconclusive, reporting that no sig- ious fitness devices or motion controllers and no at- nificant behavioral change can be identified [YSG08, tempts have been made to integrate exercising into BI10, BAB+ 12]. Moreover, in addition to a question- games that were made without physical activity in able incentive for long-term engagement, the games mind, let alone into AAA games. Such experiments often simply lack fun [MVHV03, MLB13], which can only seem to be undertaken by hobby craftsmen at be addressed using a variety of approaches: psychologi- the moment14 . However, as virtual reality (VR) equip- cal or physiological considerations [MLB13], increasing ment like the Oculus Rift15 goggles are on the verge of the level of perceived accomplishment [MH06], intro- becoming publicly available, steps are made towards ducing social and narrative components [KR09, LT13] increasing player immersion by devising novel game etc. input methods: for example, Virtuix16 develop a vir- Work most closely related to the projects presented tual reality motion platform that lets players wander in this thesis comprises user studies that investigate through the worlds of their favorite 3D games using the e↵ects of home fitness devices connected to ex- their own feet. Similarly, VirZoom17 uses an ergome- ergames. Darren Warburton et al. examined 14 low- ter to achieve said e↵ect, although currently only with active young males partaking in a specifically devised custom made games. Using these devices, of course, exercise program over a 6 week period. They dis- does not necessarily implicate conducting a thorough covered a 30% higher attendance to ergometer train- workout, but definitely increases the amount of phys- ing sessions when video games were involved as op- ical activity involved when playing games. posed to exercising the traditional way [WBH+ 07]. Many academic studies examine the e↵ect of ex- Ahn and colleagues turned a treadmill into an in- ergames on people’s energy expenditure (EE), a mea- teractive running entertainment system, featuring a sure that can be increased by physical activity and built-in computer, user-wearable sensor bracelets and should not be exceeded by dietary energy intake, in an ultrasonic sensor for tracking a player’s position order to prevent gaining weight. Jennifer Sween et [AKP+ 09, PYC+ 12]. They further developed ”Swan al., studying 27 papers on this subject, found a strong Boat”, a competitive team-based multiplayer racing correlation between fitness games and EE [SWS+ 14]. game, in order to evaluate the entertainment sys- Most of the research they investigated discovered that tem. A two-week study with 12 participants revealed introducing such games resulted in test subjects meet- an overall positive reception, beneficial social e↵ects ing the American College of Sports Medicine (ACSM) due to the necessity of team synchronization and in- guidelines for health and fitness [oSM+ 13]. The games creased engagement caused by high competition be- used for the studies included interactive aerobic, cy- tween teams. Yue Gao et al. also utilized a treadmill cling, motion capture, dance simulation as well as iso- as well as a Kinect camera with a casual exergame metric resistance titles of all current major consoles – to confirm that short exercise breaks throughout the Microsoft XBox, Nintendo Wii and Sony Playstation. day yield temporary benefits of cognitive nature, such Although dance simulation games caused the most sig- as improved concentration [GM12]. Göbel and col- nificant EE increases of about three times of a person’s leagues developed three exergames aimed at improv- resting expenditure, not a single study reported any ing the long-term motivation of players: an ergometer inconclusive or negative results. Furthermore, similar controlled set of minigames incorporating vital signals results were discovered in a literature review target- from players as well as two other games tracking user ing the physical e↵ect of exergames in healthy elderly movement via accelerometers and video recognition, [LSLL13]. Such outcomes, on one hand, emphasize the respectively [GHW+ 10]. All captured sensor informa- potential of fitness games, but on the other hand, do tion directly influenced the gameplay, which generally not account for their ability to change people’s atti- had a positive impact on player motivation and ex- tude towards exercising. perience, as they found in various focus group tests. Scientific studies examining fitness games’ e↵ects Finally, Hoda at al. evaluated how user experience in on user behavior, depending on their research ob- ergometer-based exercise depends on the application jectives, either use commercial exergames or create context [HAES13]. They examined 20 participants 14 E.g. ManaEnergyPotion on Instructables, 2009: ”Hooking Up A in five di↵erent cycling settings: no addition, game, Treadmill to PC Game” (http://tinyurl.com/zd2s36f, Web. 29 TV, game plus TV and game plus movie. Results Jun. 2016). showed that integrating workout into a game has been 15 https://www.oculus.com/en-us/rift, Web. 29 Jun. 2016 16 http://www.virtuix.com, Web. 29 Jun. 2016 the most encouraging and motivational environment, 17 http://www.virzoom.com, Web. 29 Jun. 2016 while biking without any entertainment was mostly considered tedious. Arena mod has been better accepted. Most issues re- ported during the experiments were regarded control- 6 Conclusions specific, which could have introduced a bias, hence, should appropriately be addressed in future experi- This paper has introduced fitnessification as a way of ments. Twofold quantitative analyses follow a con- improving upon current exergaming design concepts. firmatory approach as well as investigating data cor- Its major scientific contributions are prototype imple- relations in an explorative manner. Since most of the mentations, user experience studies and their evalua- questionnaire variables are ordinal, a Mann-Whitney- tions. Wilcoxon test is applied for verifying research-relevant hypotheses. Findings include strong tendencies to- 6.1 Research Questions Answered wards preferring AAA exergames over casual fitness games when examining students, gamers and active Q1 How do sportive/unsportive people feel people. Moreover, controlling Quake 3 Arena with about exergaming? an ergometer seems to have been received far better Although only very few participants actually play than using it to operate Flappy Cycling. However, it exergames, responses are overly positive and the can not be confirmed that the people’s fitness level majority of persons view exergames as healthy, nor their specific interest in bike riding are indicators entertaining, motivating and a novelty. Given for them to consider using fitness gaming devices at that many would consider working out with either home. Nevertheless, clear tendencies can be discov- of the studies’ prototypes, it can be presumed that ered and thus such assumptions should be of interest there is still potential in fitness gaming. for further investigations. Q2 Can exergaming be more fun? Judging by the significant results (H1, H2, H9, H10) this seems to be the case. However, biases 6.3 Limitations can not be ruled out, as control issues seemed to Although a population of 67 people is seemingly suf- be disturbing factors in both studies. ficient to conduct a thorough study, it certainly must Q3 Given the AAA game integration, would it be considered that they are equally split between two be possible to motivate people to work out separate studies, which renders the samples less com- more? parable and through partitioning both data sets often It appears that less sportive regular gamers are very few samples remain, occasionally not enough for not very much inclined to do so (H3), however, a meaningful evaluation. Also, the typically short test a tendency can be found when regarding both durations are not necessarily beneficial for getting ac- collected data sets separately (H5), which is defi- curate ratings: as expressed in various comments, of- nitely appropriate for justifying future research. ten more time is needed in order to adapt to the con- Q4 Which conditions must be satisfied for trols. Additionally, a change in user behavior due to gamers to consider exergaming at home? being observed can not be ruled out but it is mitigated The overall positive results convey that the most by not explicitly supervising the act of completing the important element is fun, as is reflected by par- evaluation questionnaire. Lastly, a presumed novelty ticipants’ comments. Furthermore, additionally e↵ect could have been influential and should prospec- to being responsive in terms of controls, games tively be addressed by introducing longer test periods. ought to present an individually adopting chal- lenge to their players. Lastly, volunteers’ feed- 6.4 Future Work back suggests that they must be a↵ordable, which as well underlines the importance of fun, because Several hypotheses suggest that there nevertheless hardly anybody would want to buy expensive de- seems to be some kind of relation between persons’ ex- vices that are boring to operate. ercising habits and their ratings for fitness games. It would be interesting to see if longer test periods could help improving the data collection process in order to 6.2 Summary discover more significant evidence in this regard. It A hybrid approach has been chosen for evaluating sur- further would be beneficial to set up several ergome- vey responses produced by 67 study participants. The ter stations, maybe even letting volunteers compete initial qualitative data inspection suggests that the against each other to increase the experienced social volunteers – predominantly young male students or factor. Finally, long-term engagement should be ex- technicians – mostly were committed to giving hon- amined since this would more accurately describe peo- est and constructive answers. Generally, the studies ples’ preferences and tendencies towards exergaming were very well received, although overall the Quake 3 at home than the presented short-term approach. 7 Acknowledgments and race/ethnicity-united states, 1994-2004. Mor- bidity and Mortality Weekly Report, 54(39):991–994, We would like to thank Oliver Vitouch, rector of the 2005. [KR09] Ekaterina Kurdyukova and Matthias Rehm. Interac- Alpen-Adria University, who came up with the orig- tive fitness game for public places. In Proceedings of inal idea for this work that we could subsequently the International Conference on Advances in Com- puter Enterntainment Technology, pages 437–438. elaborate on. This work was supported by Univer- ACM, 2009. sität Klagenfurt and Lakeside Labs GmbH, Klagen- [LML+ 06] James J Lin, Lena Mamykina, Silvia Lindtner, Gre- gory Delajoux, and Henry B Strub. Fish’n’steps: En- furt, Austria and funding from the European Regional couraging physical activity with an interactive com- Development Fund and the Carinthian Economic Pro- puter game. In UbiComp 2006: Ubiquitous Comput- ing, pages 261–278. Springer, 2006. motion Fund (KWF) under grant KWF 20214 u. [LSLL13] Lisbeth H Larsen, Lone Schou, Henrik Hautop Lund, 3520/26336/38165. and Henning Langberg. The physical e↵ect of ex- ergames in healthy elderly – a systematic review. GAMES FOR HEALTH: Research, Development, References [LT13] and Clinical Applications, 2(4):205–212, 2013. Feng Lu and Kimberly Turner. Improving adolescent [AKP+ 09] Miru Ahn, Sungjun Kwon, Byunglim Park, Kyung- fitness attitudes with a mobile fitness game to com- min Cho, Sungwon Peter Choe, Inseok Hwang, Hyuk- bat obesity in youth. In Games Innovation Confer- jae Jang, Jaesang Park, Yunseok Rhee, and Junehwa ence (IGIC), 2013 IEEE International, pages 148– Song. Running or gaming. In Proceedings of the In- 151. IEEE, 2013. ternational Conference on Advances in Computer [MH06] Soh Masuko and Junichi Hoshino. A fitness game re- Enterntainment Technology, pages 345–348. ACM, flecting heart rate. In Proceedings of the 2006 ACM 2009. SIGCHI international conference on Advances in [BAB+ 12] Tom Baranowski, Dina Abdelsamad, Janice Bara- computer entertainment technology, page 53. ACM, nowski, Teresia Margareta O’Connor, Debbe Thomp- 2006. son, Anthony Barnett, Ester Cerin, and Tzu-An [MLB13] Robin Mellecker, Elizabeth J Lyons, and Tom Bara- Chen. Impact of an active video game on healthy nowski. Disentangling fun and enjoyment in ex- children’s physical activity. Pediatrics, 129(3):e636– ergames using an expanded design, play, experi- e642, 2012. ence framework: A narrative review. GAMES FOR [BC10] Fabio Buttussi and Luca Chittaro. Smarter phones HEALTH: Research, Development, and Clinical Ap- for healthier lifestyles: An adaptive fitness game. plications, 2(3):142–149, 2013. Pervasive Computing, IEEE, 9(4):51–57, 2010. [MSLB06] Miller McPherson, Lynn Smith-Lovin, and [BI10] Elaine Biddiss and Jennifer Irwin. Active video games Matthew E Brashears. Social isolation in amer- to promote physical activity in children and youth: a ica: Changes in core discussion networks over two systematic review. Archives of pediatrics & adoles- decades. American sociological review, 71(3):353– cent medicine, 164(7):664–672, 2010. 375, 2006. [Bla09] Steven N Blair. Physical inactivity: the biggest public [MVHV03] Sari Mokka, Antti Väätänen, Juhani Heinilä, and health problem of the 21st century. British journal Pasi Välkkynen. Fitness computer game with a bodily of sports medicine, 43(1):1–2, 2009. user interface. In Proceedings of the second inter- [BVSZ13] Philipp Brauner, André Calero Valdez, Ulrik national conference on Entertainment computing, Schroeder, and Martina Ziefle. Increase physical fit- pages 1–3. Carnegie Mellon University, 2003. + ness and create health awareness through exergames [oSM 13] American College of Sports Medicine et al. ACSM’s and gamification. In Human Factors in Computing guidelines for exercise testing and prescription. Lip- and Informatics, pages 349–362. Springer, 2013. pincott Williams & Wilkins, 2013. [Car08] Nicholas Carr. Is google making us stupid? Yearbook [PBBvDN09] Marco Pasch, Nadia Bianchi-Berthouze, Betsy van of the National Society for the Study of Education, Dijk, and Anton Nijholt. Movement-based sports 107(2):89–94, 2008. video games: Investigating motivation and gaming [Cra45] Harald Cramér. Mathematical methods of statistics, experience. Entertainment Computing, 1(2):49–61, volume 9. Princeton university press, 1945. 2009. + [Cur56] Edward E Cureton. Rank-biserial correlation. Psy- [PYC 12] Taiwoo Park, Chungkuk Yoo, Sungwon Peter Choe, chometrika, 21(3):287–290, 1956. Byunglim Park, and Junehwa Song. Transforming [DDKN11] Sebastian Deterding, Dan Dixon, Rilla Khaled, and solitary exercises into social exergames. In Proceed- Lennart Nacke. From game design elements to game- ings of the ACM 2012 conference on Computer fulness: defining gamification. In Proceedings of the Supported Cooperative Work, pages 863–866. ACM, 15th international academic MindTrek conference: 2012. Envisioning future media environments, pages 9–15. [Sca13] Likert-Type Scales. Analyzing and interpreting data ACM, 2011. from. Journal of graduate medical education, page [DGL08] A Donner, D Goldstein, and J Loughran. Health e- 541, 2013. + games market report: Status and opportunities. San [SWS 14] Jennifer Sween, Sherrie Flynnt Wallington, Vanessa Francisco, CA: Physic Ventures, 2008. Sheppard, Teletia Taylor, Adana A Llanos, and Lu- [GHW+ 10] Stefan Göbel, Sandro Hardy, Viktor Wendel, Florian cile Lauren Adams-Campbell. The role of exergaming Mehm, and Ralf Steinmetz. Serious games for health: in improving physical activity: A review. Journal of personalized exergames. In Proceedings of the 18th physical activity & health, 11(4):864, 2014. + ACM international conference on Multimedia, pages [WBH 07] Darren ER Warburton, Shannon SD Bredin, 1663–1666. ACM, 2010. Leslie TL Horita, Dominik Zbogar, Jessica M Scott, [GM12] Yue Gao and Regan Mandryk. The acute cognitive Ben TA Esch, and Ryan E Rhodes. The health bene- benefits of casual exergame play. In Proceedings of fits of interactive video game exercise. Applied Phys- the SIGCHI Conference on Human Factors in Com- iology, Nutrition, and Metabolism, 32(4):655–663, puting Systems, pages 1863–1872. ACM, 2012. 2007. [HAES13] Mohamad Hoda, Reem Alattas, and Abdulmotaleb [WH12] Kevin Werbach and Dan Hunter. For the win: El Saddik. Evaluating player experience in cycling How game thinking can revolutionize your business. exergames. In Multimedia (ISM), 2013 IEEE Inter- Wharton Digital Press, 2012. national Symposium on, pages 415–420. IEEE, 2013. [Wil47] Frank Wilcoxon. Probability tables for individ- [Ken38] Maurice G Kendall. A new measure of rank correla- ual comparisons by ranking methods. Biometrics, tion. Biometrika, 30(1/2):81–93, 1938. 3(3):119–122, 1947. [Ken48] Maurice George Kendall. Rank correlation methods. [YSG08] Stephen Yang, Brian Smith, and George Graham. 1948. Healthy video gaming: Oxymoron or possibility? In- [KHKI+ 05] J Kruger, Sandra A Ham, Harold W Kohl III, et al. novate: Journal of Online Education, 4(4):5, 2008. Trends in leisure-time physical inactivity by age, sex,