Influence of Non-Task Related Feedback on Enjoyment, Competence and Challenge in a Competitive Game

Influence of Non-Task Related Feedback on Enjoyment,

Competence and Challenge in a Competitive Game

SUBMITTED IN PARTIAL FULLFILLMENT FOR THE DEGREE OF MASTER OF SCIENCE

Hendrik Engelbrecht

11388803

M

I

S

G

S

F

S

U

A

June 23, 2017

1st _{Supervisor 2}nd _Supervisor

Dr. Frank Nack Daniel Buzzo ISLA, UvA ISLA,Uva

Influence of Non-Task Related Feedback on Enjoyment, Competence and

Challenge in a Competitive Game

Hendrik Engelbrecht

University of Amsterdam

Amsterdam, The Netherlands

engelbrecht92@gmail.com

ABSTRACT

Engaging design principles for single player experiences have been refined for the past couple of decades. With competitive online multiplayer games occupying a large part of the market today, design for the interaction between a vast amount of diverse players needs to be considered more thoroughly. Skill differences in online matchmaking pose a threat to new player retention. For unexperienced players losses can result in substantial demotivation and therefor need to be buffered. The current study tries to utilize task irrelevant feedback as a possible buffer for negative affect resulting from losing a competitive game. Although findings do not differ significantly for losers and winners in the experiment, a positive effect of task irrelevant feedback on enjoyment and competence was observed. Several design implications resulting from these findings are discussed.

Keywords

Games, competition, feedback, player retention

1 INTRODUCTION

Gaming itself has long surpassed its rival mediums. With revenues totaling around 99.6 billion in 2016 [1] alone, videogames have become a tremendous juggernaut of economic growth and is one of the most used forms of entertainment.

Except for co-located multiplayer experiences, videogames used to be a form of solitary entertainment for the most part. Since the turn of the century internet broadband has become more widely available and affordable leading to the rise of interconnected gaming experiences. From the popularity of massively multiplayer online role-playing games in the early 2000’s [2] to the mass market of online games today [3], gaming has become a social phenomenon whose main draw is no longer the isolated solitary experience, but the interpersonally shared experience. The inclusion of social gaming aspects, such as leaderboards, shared resources or other cooperative and competitive features, extends to almost all games today. Mobile games are often integrated in a larger ecosystem that promotes social features and cooperative or

competitive gameplay. Candy Crush by King [4], for example, is a game that can only be played alone, though it incorporates many social features. Connecting the game to his social media account, the player can compare his score on each level with friends, can share his scores and even help out friends by gifting lives and power ups. On home systems the same trend occurs and massively multiplayer games are consistently among the big sellers. In 2016 all five top grossing games on mobile as well as consoles were predominantly multiplayer oriented titles [5].

To exemplify the current landscape, we would like to take the example of League of Legends (LoL), a multiplayer online battle arena game (MOBA) developed by Riot

Games [6]. Since the launch of LoL at the end of 2009 the

player base has grown to over 100 million concurrent players per month in 2016 [7]. LoL is, as of right now, the most played game in the world, creating over 1.6 billion dollars in revenue in 2015 alone [8]. LoL has one of its biggest draws in the emphasize on competitive multiplayer. In fact, LoL can not be played as a single player game. Players compete in five on five teams, which emphasizes competition as well as cooperation. As of the writing of this paper, LoL is also the biggest electronic sport in the world with 43 million unique viewers tuning in for the latest world championship in 2016. With a total of 6.7 million dollars awarded as prize money to players, LoL rivals even traditional sporting events [9].

The success of LoL across demographics, countries, cultures and media exemplifies an important development in the gaming space. Humans, and players for that matter, are social animals and they have an inherent need to compete [10]. The competitive online gaming market will only expand in the coming years and research in the domain of player motivation and engagement for these integrated social experiences has become an important focal point for designers.

Aim of the Study

With multiplayer games being a driving force in the market, early player retention is one of the most important aspects in design. Purely multiplayer focused experiences need to

adjust for the imbalance amongst players with differing levels of experience. A new player will be easily demotivated when facing off against experienced players. Without careful balancing and matchmaking between players, multiplayer focused games can fall prey to rapidly shrinking player bases. This will have a ripple effect for new adopters. If retention of new players is not achieved, the player base will shrink eventually and the imbalance of highly skilled versus low skilled players will deter new players even further. The current study aims to investigate the design of feedback mechanisms as a way of buffering losses. The aim is to investigate whether non-task related feedback can buffer the negative effects of losing by retaining player enjoyment, competence and the perceived challenge. Design recommendations for effective feedback will be derived based on the findings.

2 LITERATURE REVIEW

2.1 Why do Humans Like to Compete?

Humans are inherently social mammals and this does extend to the way people play [11]. The innate drive to play serves a learning function, to explain and explore the social world [12, 13]. Furthermore, play serves an important function in establishing status. As can be seen in primates, establishing hierarchies is done by so called ‘playfighting’, a rough form of physical play, in order to retain or gain status [14]. In the same manner, humans are hardwired to compete to create social hierarchies [15].

Play itself can be described as “the voluntary activity to

overcome unnecessary obstacles” [16, p.55]. As Zilman

[17] describes it, overcoming these obstacles leads to an excitation transfer which elicits high arousal. In the face of a positive outcome this high arousal leads to enjoyment of the activity. Transferring this to games, the obstacle in a single player game is represented by the computer controlled environment. The obstacles can be the environment itself or enemies controlled by artificial intelligence (AI).

Games utilize this by strategically building the experience towards greater and greater challenges. The player learns the mechanics, trains on them and uses them with more competence in each new trial, while the game itself adjusts the challenge accordingly. The challenge to overcome the threat leads to higher enjoyment of the successive outcome than e.g. having the ability to explore an environment in a game [18]. Direct competition in a game, be it versus a human or AI controlled enemy, provides a direct threat that leads to high excitation transfer due to the nature of the challenge and makes competitive games so rewarding for players. Additionally, competition has numerous positive psychological effects on the player. Players seek out social competition to elevate their mood [19] and to improve their self-esteem [20].

Despite all the benefits of competitive games and their popularity due to human’s innate drive, the danger for an unenjoyable experience lies in an unavoidable part of competition and that is losing.

2.2 Effects of Losing on Player Motivation

Intrinsic motivation is the most powerful determinant of motivation when it comes to challenging tasks [21]. Games often try to serve players with adjusted gaming experiences to offer multiple opportunities for overcoming challenges. This takes place in the form of adjusted difficulties, which may even adjust dynamically based on performance [22], and skill based matching during online competitive play [23, 24, 25]. However, losing can not always be avoided and this can have a tremendous effect on intrinsic motivation. As shown by Reeve, Olson and Cole [26] losing does not only diminish intrinsic motivation more than winning, but also has a negative effect on competitive performance. This leads to a loss of perceived competence which again affects enjoyment [27]. Successive losses will result in very little motivation to continue playing and therefore need to be buffered for successful player retention in multiplayer games.

The concept of flow, as first described by Czikszentmihalyi

[28], further adds another important element linked to

enjoyment and that is challenge. In order to sustain motivation, by inducing a flow state, the player needs to be optimally challenged [28]. Reaching a failure state and retrying needs to be carefully balanced to achieve the perceived challenge as rewarding for the player.

The solution to the dilemma of losing has received a lot of attention throughout any competitive field. But the solution might not lie in how to avoid a loss, but how to motivate the player to keep going despite it.

2.3 Using Feedback to Retain Motivation

Feedback is an important aspect of any competitive form of play. Providing information about performance and progress is essential to foster competence and intrinsic motivation in players [29].

Offline competitions often take place in a social situation as a spectator event. The resulting direct social feedback is obfuscated in online games due to the lack of an audience. With verbal, as well as non-verbal, feedback from opponents and audience often being non-existent, it leaves the autonomy on feedback to the game itself. This creates a novel situation in which the developer has a lot of freedom on what to communicate to the player and how to do so. Intrinsic motivation and competence have been shown to increase when competitors receive positive feedback [30, 31]. Success feedback is effective in retaining intrinsic

Motivation for competition [32]. This provides a problem for losing, since no success feedback is available, i.e. no success has been achieved, but at the same time can be part of the solution.

3 RESEARCH QUESTION

Utilizing what was described previously we can take the example of Overwatch by Blizzard Entertainment [33] as the basis for this study. Overwatch obfuscates any direct post game comparisons between players (see appendix Figure 7). A loser will not be confronted with direct comparisons which, as elaborated above, would lead to a loss in intrinsic motivation. Instead, fact-based success feedback based on other metrics is shown. Metrics like ‘Damage Blocked’, ‘Offensive Assists’ or ‘Healing

Provided’ are important for success, but only indirectly

connected to winning the game. This means that while a player on the losing team can not get positive feedback on the match outcome, he still can be highlighted for performance on a subset of indirectly related metrics. Obfuscation of facts and careful selection of specific metrics provide a motivational basis even for a losing player.

For the current study I aim to take this a step further by utilizing task and skill unrelated metrics for feedback. Doing this I hypothesize that the positive nature of the comparative feedback will act as a buffer for game

enjoyment, competence and perceived challenge in the case

of a loss and strengthen the same factors in case of a win. This leads us to the following research question.

Research Question

Does the provision of positive or negative non-task related comparative feedback have an impact on enjoyment,

competence and challenge for winners and losers in a

competitive game?

4 METHOD 4.1 Design

To investigate the possibility of using non-task related feedback to buffer the negative impact of losses as well as strengthen the positive impact of wins in a competitive videogame, participants were instructed to play a game of pong against an AI opponent. Pong was chosen as a game to alleviate differences in skill based on previous experience and further provides non-task related metrics that can be measured during the game. Furthermore, programming a balanced AI, to obtain roughly equal numbers of winners and losers, was achievable due to the simplicity of the mechanics. This made pong the most suitable game for this study with regards to the aim of the investigation.

Based on the literature review above enjoyment, perceived

challenge and competence were assessed in questionnaires

to see whether task irrelevant feedback can be used as an effective way of motivating players. These measures will be compared to a baseline condition where no additional feedback is provided (aside from the score).

Additionally of importance is the general competitive orientation. Trait competitiveness has been thoroughly studied and is linked to seeking out more challenging jobs [34], job performance in competitive climates [35] and increased performance in sport competitions [36]. The same holds true for the research conducted in the domain of video games. People higher in trait competitiveness enjoy competitive games more and perform better [37]. With much evidence on the influence of trait competitiveness in competitive situations, it was decided to account for it for the current study.

Furthermore, gender was accounted for during analysis. Research in evolutionary psychology and biology show clear preferences of men seeking out competition more than women, being more motivated by challenging situations and enjoying competition more [38, 39, 40]. This has further been proven in the investigation of competitiveness with regards to videogames. Hartmann and Klimmt [41] found significant evidence that males tend to score higher than females on all measurements related to trait competitiveness. Based on the expansive and conclusive research on gender differences it was decided to account for gender during the study.

4.2 Experimental Setup

The study was conducted using a 3x2 between subjects design in order to cover all feedback conditions (neutral, positive and negative) and to collect data for winners as well as losers. The factor feedback was varied as neutral, positive and negative. The variable outcome was not varied by the researcher. Participants were assigned to one of the three conditions and naturally split into either the winning or losing (outcome) condition based on the result of the game played during the experiment.

The questionnaire part, recording demographics as well as the DV’s and possible confounds, of the experiment was hosted on Qualtrics, an online survey platform. The manipulation, in form of a game of pong, was hosted externally.

During the game the participant controlled the left paddle on the screen with his keyboard. The game consisted of a game of best of nine with three rounds of trial being played at the beginning of the game in order to get used to the controls and overall feel. After the AI, or the player, scored a total of five points the game was over.

Three different metrics were measured during the game: number of wall hits after last paddle contact, distance travelled with the paddle and accuracy of hitting the ball relative to the center of the paddle. These metrics were chosen as they are not related to the game outcome. Furthermore, testing these metrics with people unrelated to the study showed that the metrics were simple to understand as feedback measures. All three metrics were continuously measured for both the human and the AI player.

Three different versions were deployed for the three different conditions. In the first condition participants would be prompted with no additional feedback and solely the score would be displayed (see appendix Figure 3). In the second condition, participants were given positive non-task related (ntr) feedback on their performance. Out of the three metrics mentioned above, the game would chose to display the metric on which the participant scored higher than the AI every time at the end of a round (see appendix Figure 4). In the third condition, the game would display negative ntr-feedback by choosing a metric that the AI performed better on at the end of every round (see appendix Figure 5).

Depending on the outcome of the best of nine game, the players would naturally fall into one of the six conditions:

1. Win x neutral feedback 2. Win x positive ntr-feedback 3. Win x negative ntr-feedback 4. Lose x neutral feedback 5. Lose x positive ntr-feedback 6. Lose x negative ntr-feedback

In case the game was not able to find a positive / negative comparison metric in the respective condition, the participants were shown an altered version of the accuracy measures that fit the condition. Due to carefully programmed AI behavior this did not occur more than about once in 30 trials. Accuracy is very difficult to assess visually alone by the participants, thereby warranting that the manipulation was perceived as actual measures even if altered.

4.3 Participants

The sample in this study consisted of 185 participants. All participants were recruited using the Amazon Mechanical

Turk. They were compensated with 0.40 €. Due to the large

number of conditions of the experimental setup, and in order to not solely rely on a student population, this method of data aggregation was chosen. The sample consisted of US citizens.

4.4 Procedure

Participants were first asked to fill out a short questionnaire inquiring about general demographics and previous videogame experience as possible confounds. Next, they were redirected to a website where the game was hosted. After the game was finished, participants continued with the second part of the questionnaire.

To measure trait competitiveness items from the competitiveness subscale of the WOFO [42] were used, because of their high alpha coefficient and strong construct validity. Furthermore, items proved to be reliable across genders, which is of importance for this study, as elaborated earlier.

The dependent variables enjoyment (6 items), competence (4 items) and challenge (4 items) were measured using items from the Game Experience Questionnaire [43]. Due to the higher risk of dishonest and inattentive participants on Amazon Mechanical Turk [44] two attention checks were included in the questionnaire. Another attention check was included after the end of the game (see appendix Figure 6), prompting participants to remember a specific animal, which they were asked to recall later in the questionnaire. Together with a timer, tracking time spent playing the game, this ensured that only participants who read all questions and played the game until the end without losing intentionally would be included in the analysis. Several reverse coded items were used in addition to assure reliable measures.

5 RESULTS

Out of the initial sample of 184 participants 13 were excluded due to either failing an attention check or completing the game in less than 40 seconds. Initial tests revealed that participants not losing intentionally would need at least 40 seconds to complete the game successfully. The following analysis was done using data collected from the remaining 171 participants. With 97 male and 74 female participants there is a roughly equal gender distribution in the sample. The age ranges from 18 to 76 with a mean age of 34.87 years (SD = 11.35).

QQ-plots were created to test for non-normality of the data. Visual assessment indicated no strong variation from normality. Levene’s test for equal variances [45] lets us assume equal variances across groups for all three dependent measures, F(5, 165) = .676, p = .642 (enjoyment) , F(5, 165) = 3.139, p = .1 (competence) and F(5, 165) = .581, p = .714 (challenge).

To check for possible confounds correlation tables were calculated. Competitive personality shows significant positive correlations with enjoyment and competence (r = .185, p = .015 and r = .229, p = .003). Looking at previous

Table 1. ANOVA for DV challenge

Table 2. ANCOVA for DV competence

game experience, competence (r= .235, p = 0.002) and enjoyment (r = .155, p = .042) were also positively

correlated. Despite evidence in literature, no correlations were obtained between gender and the dependent measure. Due to the significant relationships of two of the DV’s with

competitive personality and previous game experience, both

were included as possible covariates for further analysis. A two-way ANOVA was run to test if the independent variables (outcome and feedback) had a significant effect on

challenge. The dependent variable showed no significant

main effects for both independent variables. Furthermore, no interaction effect was observed (Table 1).

Two two-way ANCOVAs were run, including competitive

personality and previous game experience as covariates, to

assess the effect of the manipulations on enjoyment and

competence. For both dependent variables, the IV feedback

indicates a significant main effect. No main effects were found for the effect of outcome on both variables enjoyment and competence. Furthermore no interaction effects were observed for enjoyment or competence (Table 2 & 3).

6 DISCUSSION

The current study has not found evidence for a significant relationship between the independent variable outcome and the three dependent measures challenge, competence and

enjoyment. Despite accounting for possible confounds for

two of the three DVs, no significance was achieved. Using pong as a game for the experimental manipulation did help to lower the skill barrier enough to achieve a good distribution of winners and losers between participants with varying experience. Yet, the non-significance of the results indicate a possible problem with this approach. General challenge perceptions might have been affected by the simplicity of the game mechanics. This might not have offered enough variation in strategy and applicability of skills to be seen as challenging and a reflection of competence, which in turn did not affect enjoyment of the game strongly enough to be significant.

The low means for challenge measures throughout conditions further support this possibility. For the participants losing, for example, the means behave similarly to how it was expected, with participants in the positive ntr-feedback condition feeling less challenged than in the negative ntr-feedback condition. Similarly, the means, although not significant, between outcome conditions do show a slight increase for the winners on both enjoyment and competence, indicating a weak manipulation (see appendix Table 4).

To conclude, it is advised to increase personal involvement for the competition scenario, by e.g. adding extrinsic rewards, and choose a slightly more challenging game to strengthen the manipulation. Extrapolating this to the

design of multiplayer games, we expect the heavily integrated reward structures of most games, in the form of social and progress rewards (e.g. leaderboards and item unlocks), to be strong enough to strengthen the observed effects and create a significant difference between losers and winners.

Surprisingly, gender did not have a significant effect on any measures taken. What has caused the non-significance of results for the outcome conditions could have at the same time diminished the effect of gender influence for the sample. The weak manipulation might have led to a less competitive environment where the differences in gender were not pronounced enough to be detected.

Significant results were found for the relationship between

feedback and the two DVs competence and enjoyment.

For participants in the lose condition, enjoyment was strengthened by positive ntr-feedback and weakened by negative ntr-feedback as compared to the neutral feedback condition (see appendix Figure 1). This is in line with the research question and provides evidence for the value of non-task related feedback. For participants winning, the positive ntr-feedback also strengthened enjoyment, but surprisingly the negative ntr-feedback did so as well compared to the baseline in the neutral feedback condition (see appendix Figure 2). These results give us evidence to believe that feedback seems to behave differently for losers and winners. A possible answer could lie in the interpretation of the feedback provided. While participants who lost have perceived the negative ntr-feedback as another source of negative affect, the winners might actually interpret the results as further confirmation. Due to the salience of the overall score in their favor, the negative ntr-feedback might be interpreted as irrelevant. To put it simply, the negative ntr-feedback emphasizes the irrelevance of their opponents’ actions (and the metrics shown) which leads to more enjoyment of the game. This is further underscored by looking at competence means.

Competence behaves in the same way as enjoyment does,

where negative ntr-feedback acts as a buffer for participants losing the game and negative ntr-feedback amplifies feelings of incompetence. Surprisingly, negative ntr-feedback strengthened competence for the winners, compared to the neutral condition, in the same way as it did for enjoyment. These results emphasize an important point for further investigation. Interpretation and type of feedback are different for winners and losers.

In conclusion, these significant findings provide a compelling case for the use of ntr-feedback for player retention. This study presents a very isolated case for a mechanically simple game. Extending this research to other games there are a lot more possibilities to apply these findings. The broader the strategic possibilities, the more metrics can be gathered during the game. This means that

designers have a lot more freedom to utilize semi-outcome related measures to motivate players and to retain new players by making them feel competent. Even in the absence of semi-related measures, feedback using skill unrelated metrics can be used as an effective tool for player retention by buffering loss of intrinsic motivation through increased enjoyment and competence.

Important to note however is that design for player engagement can not be standardized for winners and losers. A novel, though ethically questionable approach, could be to highlight how badly the winners opponent is doing to increase enjoyment and competence. The effect of ridiculing the winners opponent seems effective in this study, but possibly has too many negative side-effects for actual use in competitive games, as this might foster anti-social behavior amongst players.

Important for future investigation is further also the feedback effect over the long term on motivation. This study only assessed the effect in one best of nine game, but it is hard to speculate whether the same effect could hold true in many consecutive games. As has been discussed in the beginning of this paper, the more consecutive games are lost, the more the motivation to continue diminishes. The positive ntr-feedback buffer, although seemingly effective in the short term, might lose its effectiveness over time. The salience of the irrelevance of the feedback measures might become stronger and possibly even lead to a ridiculing effect that enhances negative affect.

Another possibly fruitful endeavor for future research lies in the obfuscation of feedback for relevant outcomes. While the present study simply adds other irrelevant metrics on top of the existing score, the trend in modern games has been going towards hiding game relevant metrics. Using a more complex game, with feedback on metrics related to performance, might be of interest to investigate how obfuscation of performance related metrics influences players.

The current study provides compelling evidence for the use of ntr-feedback as a means of player retention in competitive videogames. While significant effects were obtained only for some measures, this study can serve as a good basis for future research.

REFERENCES

1. Anon, (2017). [online] Available at:

https://newzoo.com/insights/articles/global-games- market-reaches-99-6-billion-2016-mobile-generating-37/# [Accessed 23 Jun. 2017].

2. Olson, R. (2017). MMORPG Popularity, 1998-2013. [online] Dr. Randal S. Olson. Available at:

http://www.randalolson.com/2014/11/12/mmorpg-popularity-1998-2013/ [Accessed 21 Jun. 2017].

3. Superdataresearch.com. (2017). SuperData Research | Games data and market research » The MMO & MOBA Games Market Report, 2016. [online] Available at:

https://www.superdataresearch.com/market-data/mmo-market/ [Accessed 21 Jun. 2017].

4. King.com. (2017). King.com. [online] Available at: https://king.com/ [Accessed 21 Jun. 2017].

5. Anon, (2017). [online] Available at:

http://scrolltoday.com/top-grossing-games-2016/ [Accessed 23 Jun. 2017].

6. Riot Games. (2017). Riot Games. [online] Available at: https://www.riotgames.com/ [Accessed 21 Jun. 2017]. 7. Anon, (2017). [online] Available at:

http://www.riftherald.com/2016/9/13/12865314/monthl y-lol-players-2016-active-worldwide [Accessed 23 Jun. 2017].

8. Dotesports.com. (2017). Report: League of Legends made $1.6 billion in revenue last year. [online] Available at: https://dotesports.com/

league-of-legends/league-of-legends-2015-revenue-2839 [Accessed 21 Jun. 2017].

9. Lolesports.com. (2017). LoL Esports. [online] Available at: http://www.lolesports.com/ en_US/articles/2016-league-legends-world-championship-numbers [Accessed 23 Jun. 2017]. 10. Christiansen, F.B., & Loeschcke, V. (1990). Evolution

and competition. In K. Wohrmann & S.K. Jain (Eds.). Population Biology , 367-394..

11. Young, S. N. (2008). The neurobiology of human social behaviour: an important but neglected topic. Journal of Psychiatry & Neuroscience : JPN, 33(5), 391–392.

12. Ruckenstein, M. (1991). Homo ludens: a study of the play element in culture. Leisure and Ethics, 237. 13. LaFreniere, P. (2011). Evolutionary Functions of Social Play: Life Histories, Sex Differences, and Emotion Regulation. American Journal of Play, 3(4), 464-488.

14. Paquette, D. (1994), Fighting and playfighting in captive adolescent chimpanzees. Aggr. Behav.,, 20, 49–65.

15. Zink, C. F., Tong, Y., Chen, Q., Bassett, D. S., Stein, J. L., & Meyer-Lindenberg, A. (2008). Know your place: neural processing of social hierarchy in

humans. Neuron, 58(2), 273-283.

16. Suits, B. (2014). The Grasshopper-: Games, Life and Utopia. p.55. Broadview Press.

17. Zillmann, D. (1996). Sequential dependencies in emotional experience and be- havior. Emotion: Interdisciplinary perspectives, 243-272

18. Malone, T. W. (1981). Toward a theory of intrinsically motivating instruction. Cognitive Science, 4, 333-369. 19. Colwell, J., Grady, C. & Rhaiti, S. (1995). Computer

Games, Self-Esteem and Gratification of Needs in Adolescents, Journal of Community and Applied Social Psychology, 5 (3), 195-206.

20. Zillmann, D. (1988). Mood management: Using entertainment to full advantage. Communication, social cognition, and affect, 31, 147-171.

21. Dickey, M. D. (2007). Game design and learning: A conjectural analysis of how massively multiple online role-playing games (MMORPGs) foster intrinsic motivation. Educational Technology Research and Development, 55(3), 253-273.

22. Hunicke, R. (2005, June). The case for dynamic difficulty adjustment in games. In Proceedings of the 2005 ACM SIGCHI International Conference on Advances in computer entertainment technology (pp. 429-433). ACM.

23. Delalleau, O., Contal, E., Thibodeau-Laufer, E., Ferrari, R. C., Bengio, Y., & Zhang, F. (2012). Beyond skill rating: Advanced matchmaking in ghost recon online. IEEE Transactions on Computational Intelligence and AI in Games, 4(3), 167-177. 24. Myślak, M., & Deja, D. (2014, November).

Developing game-structure sensitive matchmaking system for massive-multiplayer online games. In International Conference on Social Informatics (pp. 200-208). Springer International Publishing.

25. Véron, M., Marin, O., & Monnet, S. (2014, March). Matchmaking in multi-player on-line games: studying user traces to improve the user experience.

In Proceedings of Network and Operating System Support on Digital Audio and Video Workshop (p. 7). 26. Reeve, J., Olson, B. C., & Cole, S. G. (1985).

Motivation and performance: Two consequences of winning and losing in competition. Motivation and Emotion, 9(3), 291-298.

27. Vallerand, R. J., Gauvin, L. I., & Halliwell, W. R. (1986). Effects of zero-sum competition on children's intrinsic motivation and perceived competence. The Journal of Social Psychology, 126(4), 465-472. 28. Czikszentmihalyi, M. (1990). Flow: The psychology of

optimal experience. Praha: Lidové Noviny. 29. Hagger, M. S., Koch, S., & Chatzisarantis, N. L.

(2015). The effect of causality orientations and positive competence-enhancing feedback on intrinsic

motivation: A test of additive and interactive effects. Personality and Individual Differences, 72, 107-111.

30. Burgers, C., Eden, A., van Engelenburg, M. D., & Buningh, S. (2015). How feedback boosts motivation and play in a brain-training game. Computers in Human Behavior, 48, 94-103.

31. Senko, C., & Harackiewicz, J. M. (2005). Regulation of Achievement Goals: The Role of Competence Feedback. Journal of Educational Psychology, 97(3), 320.

32. Harackiewicz, J. M. (1979). The effects of reward contingency and performance feedback on intrinsic motivation. Journal of personality and social psychology, 37(8), 1352.

33. blizzard.com. (2017). Blizzard Entertainment. [online] Available at: http://eu.blizzard.com/ [Accessed 23 Jun. 2017].

34. Houston, J. M., Harris, P. B., Howansky, K., & Houston, S. M. (2015). Winning at work: Trait competitiveness, personality types, and occupational interests. Personality and Individual Differences, 76, 49-51.

35. Fletcher, T. D., Major, D. A., & Davis, D. D. (2008). The interactive relationship of competitive climate and trait competitiveness with workplace attitudes, stress, and performance. Journal of Organizational Behavior, 29(7), 899-922.

36. Hellandsig, E. (1998). Motivational predictors of high performance and discontinuation in different types of sports among talented teenage athletes. International Journal of Sport Psychology, 29, 27–44.

37. Song, H., Kim, J., Tenzek, K. E., & Lee, K. M. (2013). The effects of competition and competitiveness upon intrinsic motivation in exergames. Computers in Human Behavior, 29(4), 1702-1708.

38. Vugt, M. V., Cremer, D. D., & Janssen, D. P. (2007). Gender differences in cooperation and competition: The male-warrior hypothesis. Psychological science, 18(1), 19-23.

39. Kivlighan, K. T., Granger, D. A., & Booth, A. (2005). Gender differences in testosterone and cortisol response to competition. Psychoneuroendocrinology, 30(1), 58-71.

40. Taylor, S. E. (2011). Tend and befriend theory. Handbook of Theories of Social Psychology: Collection: Volumes 1 & 2, 32.

41. Hartmann, T. and Klimmt, C. (2006), Gender and Computer Games: Exploring Females’ Dislikes. Journal of Computer-Mediated Communication, 11: 910–931.

42. Helmreich, R. L., & Spence, J. T. (1978). The Work and Family Orientation Questionnaire: An objective instrument to assess components of achievement motivation and attitudes toward family and career.

American Psycholog. Ass., Journal Suppl. Abstract Service.

43. IJsselsteijn, W., Poels, K., & de Kort, Y. A. (2008).

The Game Experience Questionnaire: Development of a self-report measure to assess player experiences of digital games. TU Eindhoven, Eindhoven, The Netherlands.

44. Rouse, S. V. (2015). A reliability analysis of

Mechanical Turk data. Computers in Human Behavior, 43, 304-307.

45. Brown, M. B. and Forsythe, Robust Tests for the Equality of Variances. A. B. (1974), Journal of the American Statistical Association, 69, pp. 364-367

Appendix

Table 4. Overview of means between outcome conditions

Figure 2. Comparison of means between win conditions and dependent variables

Figure 4. Screenshot of positive ntr-feedback condition

Figure 6. Screenshot of manipulation check at the end of the game