ABSTRACT
Starting out with the idea that within the field of (entertain-ment, but most notably serious) gaming each game brings with it an intense and personal emotional experience, this thesis explores the possibilities for doing so. Instead of re-mediating other media, this thesis is aimed to show that the unique potential of games should be capitalized upon. After establishing what has already been accomplished and which challenges we are presented with, this thesis forms a recom-mendation for the use of player modelling and abstract nar-ratives structures in addition to biofeedback based content to create an intense emotional experience. By taking into account all aspects of emotion (both physical and mental), and subsequently valence in addition to arousal, through the use of narrative in its abstract, almost formulaic, form, and the addition of a workable playermodel in the shape of a player type and limited understanding of the social cultural background, a first tentative step towards affective gaming can be made.
Keywords
Affective gaming, Biofeedback, Immersion, Player modeling
1.
INTRODUCTION
Literature and films have long since been recognized for their ability to emotionally affect us, eventhough they make us sympathize rather than experience the emotion first hand. Games differ slightly, they put us in the middle of the action, no longer are we spectators, we are agents within the story1. Often the most memorable games are those that moved us emotionally in some way. Ideally, within the field of (entertainment-, but most notably serious-) gaming, each game has such affect.
Games currently remediate other media such as film and lit-erature to create emotionally compelling stories, rather than capitalizing on their own unique potential2. This potential can be used to advance the multidisciplinary knowledge on emotion, due to its interactive nature, as will become clear through the course of this thesis.
1.1
Research Questions
This research started out within the context of a project aimed at establishing an intense and personal emotional ex-perience. The initial approach of this project involved using arousal measurements to predict the most arousing audio-visual content, with the purpose of maximally arousing a person in the shortest time possible. Where this thesis was first focused on finding the ideal way of presenting such con-tent, it became apparent after some research that a more general approach was needed. As such, the focus on arousal has been relinquished for the time being, giving way to a more general question of how to create an intense and per-sonal emotional experience within a game setting.
The first step into doing so is knowing how to measure,
in-1
though in general there is still a distinction between our-selves and the gameworld; we are conscious of the fact that we are playing a game.
2which is logical, from a design perspective; as they need to
be intelligible and have some degree of recognisability
terpret, and affect emotion. In particular, it is necessary to frame the research question in the context of pertaining design challenges to this end. While the origin and back-ground of these challenges will be explored in the upcoming Related Work section, the research questions that address these challenges are formulated here:
A. How can be made sure the measured physiology is in-deed a representation of an emotional state, and not a conscious manipulation (of the player’s own physiol-ogy)?
B. Which parameters should be altered?
C. How to make sure arousal results from a reaction to the game?
D. How to circumvent arbitrary reactions by: • minimizing the Kuleshov effect,
• making use of social/cultural background, • and focusing on aspects of emotion other than
those that can be measured in real time? E. How to make the game compelling, interesting and
per-sonal,
F. but not too compelling, leading to desensitization? G. How to decrease the (emotional) distance between the
player and the game?
1.2
Thesis Outline
With the above mentioned questions in mind, this thesis is structured as follows. In section 2 related endeavours and literature are reviewed, to establish what is already known, successfully applied, or a presents a challenge. At the end of this section, it will become clear where the above men-tioned challenges (A to G) stem from. With these challenges established, section 3, a literature study, will provide tools which contribute to overcoming one or more of the found challenges. As to be expected, some of the solutions impli-cate one another, and some even give rise to new challenges, which will be detailed in section 4.
2.
RELATED WORK
In this section different perspectives on emotion will be dis-cussed and examples of related endeavours, in order to es-tablish what the challenges and accomplishments are when it comes to measuring, interpreting and affecting emotion. First the idea of emotional modalities and the arousal va-lence model will be discussed, after which two examples; a scientific research, and an entertainment game will be re-viewed. At the end of this section (in section 2.3) a list of challenges, partially synthesised from the preceding exam-ples, will be discussed. After which, in section 2.4, a few tentative solutions will be detailed.
2.1
Perspectives on Emotion
Even though different disciplines have a relatively good un-derstanding of emotion, there is one problem; in emotion there is a distinction between four modalities, the links be-tween which still remain somewhat of a mystery.[5]
Figure 1: A specific combination of arousal (vertical axis) and valence (horizontal axis) represents a spe-cific emotional state. For instance, high arousal and a positive valence indicate an exited state.
Especially this last century there have been many researches into emotion, using different perspectives and approaches, not all humane (e.g. the little Albert experiment [14]). Over the past decades however, emotion research within the af-fective sciences have led to some relevant key findings [5], such as a distinction in types of emotional modalities; (1)be-havioural/ expressive, (2)somatic/ physiological, (3)cogni-tive/ interpretive, and (4)experiential/ subjective.
There seems to be a consensus in both scientific research and entertainment, with regards to the first step into finding these unknown links, namely focusing on that which can be measured real-time: (2) physiology.
There are multiple models on classifying emotion, one work-able model of classification, which appears to be the basis of this consensus, is that of arousal, modality 2, versus va-lence (arguably modality 3 and 4), also known as the Russel model. According to this theory most, if not all emotions can be classified as a certain combination of valence and arousal values (see Figure 1)[1].
Videogames are good at eliciting emotions, some games more effectively than others. One emotion that they excel in, ac-cording to Richard Rouse (amongst others), is that of ten-sion and fear3. In the following subsection two examples are given of video games that use physiology (such as arousal measured through heartrate) as an input.
2.2
Examples
Games that employ biofeedback have been around for some time; there was a game in 1983, for instance, that already
3
unfortunately, he doesn’t give an exclusive reason, we can however reasonably assume that this success is partially due to the already visceral nature of the genre in other media
Figure 2: demonstration of the Atari Mindlink sys-tem in 1984, an example of a biofeedback loop; the player interacts with the system through brow movement that is picked up by the headband and translated into game actions
used biofeedback instead of a joystick to control certain as-pects within the game. Most, if not all, of these games are examples of ‘simple’ biofeedback loops; in which player in-put is partially, or even completely based on biofeedback. In these cases the user is only consciously interacting with the game through his/her physiology(Figure 2). The mea-surement is then not representative of modality 2, therefore eradicating the ability to measure modality 2.
But there have also been scientific endeavours, one research by Dekker and Champion in 2007, for instance, used biofeed-back to dynamically adapt the gameplay of an existing game. One of their findings was that audio cues were the most effective in depicting biometric information. Specifically the sound of the avatar’s heartbeat made their participants more anxious. The visual changes they made to the game were less successful; changing the colours to black and white, made players calmer, whereas a red filter had no significant effect, and white only confused them. One visual tactic that did elicit strong reactions was the shaking of the screen, this was due to unpopularity [3]. Choosing the right parameters to change might not be as straightforward as we might have
thought.
Combinations of certain parameters, such as sound- and/or visual content may add up to meanings other than the ones expected. In the next section this will be discussed further, along with other related challenges.
2.3
Challenges
Before getting further into the combining of content, we need to remember the arousal/valence model, as the above ex-amples show, the current trend is to focus on modality 2, and mainly arousal 4 When focusing on arousal, it makes sense to try to increase it5, as Dekker and Champion did. However, when one succeeds in steadily increasing arousal, either desensitization or termination of engagement (of the user) will occur, due to the (rapid increase of ) intensity[7]. Additionally, quite similar to both desensitization and ter-mination of engagement, a lack of control/agency can lead to distantiation.
One of the problems in using arousal as an indicator is that the measured arousal might not result from a reaction to the game itself at all. As stated in the previous subsection, reactions related to the game might be arbitrary. This can be explained by a film theory, known as the kuleshov ef-fect. This theory dictates that two seperate images create new meaning when combined (see Figure 3). When the suc-cession of content is solely based on measured arousal (from now on referred to as biofeedback), there will be no intended logic perceivable to the ‘user’. As humans have a tendency to apply causal logic (narrative) to all they perceive, a player will also do this with regards to the biofeedback based con-tent presented. This will interfere with the measurement; content that first might have been arousal increasing may be interpreted differently based on the combination of/with previous content. However, it is more likely that the user will lose interest before long, due to the same lack of logic or narrative structure. The interpretation can also differ between different players, this can be due to differences in social or cultural background.
To recapitulate, there are seven challenges (see list below) that need to be taken into account (I also like to refer the reader to the questions marked A to G stated in the intro-duction), in addition to the main question, which can now be best described as ‘how to incorporate valence and modal-ity 1, 3 and 4 in addition to arousal and/or modalmodal-ity 2, in order to create a truly affective game?
A. From the mindlink example we have learned that the biofeedback loop needs to be made unconscious. B. The research by Dekker and Champion show that there
is no real grasp of which parameters should be altered. C. It needs to be made sure that arousal results from a
reaction to the game
4The mindlink example is not directly based on arousal,
but is chosen to rather show the age of this idea of using biofeedback.
5
low arousal is associated with emotional states such as boredom, which might compromise an experiment as de-creases interest and thus the engagement of a test subject.
Figure 3: eventhough the left image remains the
same, each horizontal pairing has a different mean-ing. The top pair shows a hungry man, the middle, a grieving man, and the lower pair shows a man being attracted to a woman.
D. Arbitrary reactions due to the following need to be avoided:
• the kuleshov effect
• differences in social/cultural background • sole focus on aspects of emotion that can be
mea-sured in real time
E. In order to ensure participation, the game should be compelling, interesting and personal
F. but not too compelling, leading to desensitization or loss of interest [7]
G. (Emotional) distance between the player and the game should be decreased without endangering the former points (through the use of agency)
The term ‘affective game’ is in itself already a solution, which will be explained further in the next subsection, along with some other solutions.
2.4
Tentative Solutions
The first challenge [A] is resolved by making this interaction unconscious, the game becomes what Gillaede et al. call an ‘affective game’ [4]. They distinguish between three different types; assist me, challenge me, and emote me. Based on the biofeedback received, each of these types will act differently6. The latter one is of interest to this thesis. Not surprisingly, Gilleade et al. remain vague in their description of how the third type works, hence this thesis.
As it is an affective game, this resolves challenge [G], given it provides enough agency; another emotion that games elicit better than other media is that of pride and accomplish-ment[9, ]. Another way of looking at this statement is by looking at determination theory. According to self-determination theory, there are three key motivations for participation: competence, autonomy and relatedness. Com-petence and autonomy are both closely related to pride, and to another point Rouse tries to make: games are extremely immersive and effective because the player determines the actions (to some extent) of the protagonist in a game. “ A game that leaves the player in control of the camera feels more immersive than one where the camera control is con-stantly taken away” [9, 24].
The other challenges (B to F) need some more delibera-tion, in the following section solutions will be discussed that counter one or more of these challenges, in addition to lead-ing to answerlead-ing the main research question.
3.
LITERATURE STUDY
In the following subsections ways of solving the remainder of the challenges are presented through discussing how to incorporate valence, and/or modality 1, 3 and 4 (in addition to modality 2). One of the most important ingredients for a successful affective game is a playermodel, which is the first to be discussed in the following subsections.
6the first two change the difficulty level, (the first) based on
frustration levels, and (the second on) the amount boredom.
Figure 4: The brain hex player types and associated physiology. The relative position of each player type is linked to the localisation of predominant brain activity of each specific type
3.1
Player Types & Models
Apart from informing on the preferences of a player, knowl-edge of player type can help quantify modality 1. To pick up on the idea of player motivation, from the former section, it is important to note that not every player has the same drives and motivation.
There are some different approaches to defining player types, most of them rely on, or are similar to, Bartle’s four cate-gories; killer, achiever, explorer and socialiser, and add to these.
Bart Stewart has made a comparison of some of the models, as is shown in Table 1, the four Bartle archetypes are clearly represented [13].
Additionally, Stewart conceptualized some design recom-mendations (Table 2)[13].
Apart from design, these player types show different types of behaviour in different contexts, with different meaning. If gameplay behaviour is to become a predictor of the emo-tional state of the player, each form of behaviour needs to be interpreted within the context of the player type and sit-uation.
In this respect the Brainhex model gives some insight, they distinguish between seven types of players; seeker (focused on the world), survivor (intense experience), mastermind (strategy/ solution), conqueror (challenge), socializer (in-teraction), daredevil (risk/ thrill) and achiever (completion/ achievements) (Figure 4) [8].
Survivors and daredevils fall under the first category on the unified model; artisan/killer/experientialist. Conquerors and achievers under the second, seekers and masterminds under the third, and socializers under the fourth.
Keirsey Bartle Callois Lazzaro GNS+ MDA+ Handy Gallup Covey Motivation Problem-Solving Overall Goal Artisan
(Tactical) Killer
(Manipulator) Ilinx Serious fun Experientialist Kinetics Power Impacting Power Power
(Manipulative sensation) Performance Do Guardian
(Logistical) Achiever Agˆon
Hard fun
(’Fiero’) Gamist Mechanics Role Striving Security Security
(Competitive accumulation) Persistence Have Rational
(Strategic) Explorer Mimesis Easy fun Simulationist Dynamics Task Thinking Wisdom
Knowledge
(Logistical rule-discovery) Perception Know Idealist
(Diplomatic) Socializer Alea People fun Narrativist Aesthetics People Relating Guidance Identity
(Emotional relationships) Persuasion Become
Table 1: Comparison of Game Models into the unified model, by Bart Stewart
UNIFIED Play Style Associated Gameplay Features
Artisan/Killer/Experientialist action, vertigo, tool-use, vehicle use, horror, gambling, speedruns, exploits
Guardian/Achiever/Gamist competition, collections, manufacturing, high scores, levels, clear objectives, guild membership, min-maxing Rational/Explorer/Simulationist puzzles, creative building, world-lore, systems analysis, theorizing, surprise
Idealist/Socializer/Narrativist chatting, roleplaying, storytelling, cooperation, decorating, pets, social events
Table 2: Design recommendations
When the general motivations and adhering playstyles are known, ingame behaviour (1) can be explained. This be-haviour can even be used as a proxy for measuring valence, the extent to which this valence is related to the game itself however, is, amongst other things, dependent on immersion.
3.2
Immersion
Immersion increases the first-hand experience rather than that of being a passive spectator, which relates to the chal-lenge presented by question C.
Even though immersion is not synonymous with close to real experience, it is an important aspect. Immersion is not only established through application of multiple different modalities (e.g. vision, audition, and tactition), but also through means of aspects of narrative and agency (more on these in following subsections).
While immersion is closely tied to the senses, suspension of disbelief and identification, while being related to the con-scious experience are almost identical to it.
3.3
Suspension of Disbelief & Identification
In addition to facilitating a first-hand experience, suspension of disbelief and identification (with the ingame self) create an emotional engagement (modality 3 and 4) and give the player a goal/purpose, thus help resolve challenge E. By repeating types of content, anticipation can be created, thus heightening arousal. This anticipation is dependent on belief, in order to create this, the least that needs to be achieved is a suspension of disbelief, a term more often used in the context of film. Suspension of disbelief, entails the administering of ‘a human interest and a semblance of truth’, and takes as long as the unravelling of the ‘story’ [2]. According to Murray Smith, this human interest is often achieved through characters, creating empathy and/or sym-pathy through exposure (to those characters and their ac-tions, and sometimes even personal thoughts)[12].
In interactive environments, this identification goes one step further; the player controls (some of) the behaviour of his/her avatar (character), and becomes more attached the more he/she plays with this character (exposure).
In certain game genres, a cinematographic method is used; point of view shots, also known as ‘first person’ (in game context). Thus the player (or the audience in case of a film) is effectively put into someone else’s shoes (again reinforcing empathy/sympathy).
The semblance of truth is needed to make it intelligible, as well as rationally acceptable as reality (temporarily). As seen with various games, photo realism is not always needed to induce both human interest and a semblance of truth. A similar concept to identification and suspension of disbe-lief is that of emotional contagion, which leads to some new insights with regards to the social substance of an affective game.
3.4
Emotional Contagion
Emotion is contagious, if others are in a heightened state of arousal, the player will get more aroused as well (not neces-sarily to the same extent, with the same valence). Emotional reaction isn’t only involved in first hand experiences, seeing something (emotionally compelling) happen to someone else will also entice emotion, which can be partially explained through the theory of Murray Smith discussed earlier. Emotions, such as fear, rely on social factors, such as con-tagion. It comes to no surprise that Dekker and Champion mention multiplayer possibilities in their future work and recommendations ; “where the game player 1 is given infor-mation of biofeedback of player 2 and uses this knowledge in an attempt to trigger the phobias of player 2.”[3, 558] Other than this interesting suggestion, emotional contagion offers a semi controllable variable, without needing real-time valence assessment, to induce certain emotional states, e.g. through the use of (real) actors.
Most of emotional contagion can be embedded in a narrative, while suspension of disbelief and identification are largely dependent on it, which will be further discussed in the next subsection.
3.5
Narrative
Besides enabling suspension of disbelief and identification, and therefore resolving issues D and E, narrative also has
the capability of regulating arousal levels (resolving issue F), through creating anticipation, suspense or using surprise for instance. Well executed narrative helps make ‘the trance deeper and deeper without the emotions becoming hotter and hotter’[7]. When immersion is increased, so can the emotions.
There is a limited amount of ‘master narratives’, and in addi-tion there are also limited narrative events. It is not unthink-able that storytelling can be done by computer algorithms, there are already numerous attempts; from video games that change their storyline based on player choices and actions, to completely generated narrative (using a database), to user generated (constructivist) narrative.
Different narratives appeal to different emotions and reflec-tions of the self, but most importantly to different types of people (challenge E). For instance: some people like stories of conflict, involving overcoming an enemy, or the odds (re-member table2).
Narrative is in itself not interactive, which (re)constitutes challenge G, which brings us to the notion of agency.
3.6
Agency
Earlier it was stated that the playertype helps quantify mea-sured behaviour, however this can only be done, when the player has agency. In an interactive setting, ideally, a player has control over his/her movement and interaction with the environment. Inherently, the player has the ability to react to those parameters changed on the basis of biofeedback. In some cases, when it involves a negative valence for instance, the player might use this agency to avoid those parameters that heighten his/her arousal. To reduce this effect, or even eradicate it, the illusion of control must be maintained while maximising exposure. This can be done through adjusting the mechanics; when the player avoids certain content; put other, maybe even more arousing, content on his/her path. As has probably already become apparent, the above subsec-tions are closely interrelated, which will be briefly clarified in the following subsection.
3.7
Interrelations
Playertype and narrative are related to each other as they explain one another; the playertype determines the type of narrative, and the narrative defines the player as an actor within the game. Additionally, the playertype determines what kind of agency should be given to the player, and together they constitute player expression. Narrative fur-thermore enables identification (which is further increased through agency) and, similarly, emotional contagion. Not only are these elements closely interrelated, but they also implicate one another, and bring new challenges with them, Which will be discussed in the following discussion section)
4.
FURTHER IMPLICATIONS
Here, existing solutions that were highlighted in the related work will be synthesized with new connections that were drawn in the literature study. Specifically, in order to yield
Figure 5: The elemental tetrad model by Jesse
Schell: ‘Mechanics are the procedures and rules.,. The story refers to the sequence of events that
un-fold. Aesthetics describe how the e game looks,
sounds, smells, tastes and feels. Lastly, the technol-ogy constitutes any materials and interactions that make the game possible.’ [10]
generalizable insight, we proceed as follows. In addition to discussing the player in terms of a playermodel with special attention to emotions, different aspects of an affective game will subsequently be discussed following Jesse Schell’s model (Figure 5), for more details I would like to refer the reader to one of his works; ‘The Art of Game Design: A book of lenses’ [10]
4.1
The Player
As each player is an individual with personal preferences and background, a playermodel is needed to anticipate some specifics in creating a personal intense experience. However, before going into this player model, I want to recapitulate how to address all modalities of emotion.
4.1.1
Emotion
As stated in the related works section, emotions can be di-vided into four modalities (see Introduction), to reiterate: behavioural / expressive, somatic / physiological, cognitive / interpretive, and experiential / subjective.
The first two modalities can be ‘measured’ and used as feedback, the first through analysing the interactions of the player with the game, and additionally expressions and be-haviour outside of the game7, the second through the appli-cation of technology, such as EMG, BCI or EKG for instance
8.
The latter two cannot be measured in real-time (yet?), but
7
Though this could involve violation of privacy and/or might otherwise not be preferable if applied in a personal environ-ment such as a player’s own home
8there is a trade-off; more accurate technologies are often
based on literature, they can be affected with some speci-ficity, by making use of a playermodel, consisting of a play-ertype and an indication of social/cultural context.
4.1.2
Playermodel
The extent of arousal (as reaction to certain stimuli) of play-ers will differ based on social and cultural background, thus a player model is needed. This player model can be par-tially constructed through the integration of social network-ing sites, such as Facebook, in addition to usnetwork-ing a playertype based estimate (such as the Brainhex model).
Other player modeling can be done by comparing measure-ments of multiple players and subtracting common predic-tors based on clustering.
Another aspect that needs to be considered is ethics. As this type of games seems to be still in its infancy, it might not be advisable to submit non-adult players (and those belonging to psychological risk categories) to such games. Apart from differences in reaction, Interpretation of the game is depen-dent on multiple factors including for instance: emotional state, cognitive load, social/cultural context (e.g. biased in some way), physical state (e.g. tired or full of energy), and effectiveness of sensory information. As such a choice must be made in which factors can be actively monitored, and which to ignore. The factors that could feasibly be included are those aforementioned.
4.2
The Game
To make things more clearly structured, the following sec-tions describe the game elements, as proposed by Schell starting with mechanics, followed by story, aesthetics and technology, respectively.
4.2.1
Mechanics
Mechanics are crucial in maintaining the illusion of control, when implemented correctly, they can reinforce the suspen-sion of disbelief.
As players differ from each other, the mechanics need to be adjustable to suit their motivations and goals. Not so much their workings, as their meaning (much like swapping and endorsing abilities in rpgs). This can be staged in character choosing, or assuming a credible role within a bigger whole as ‘oneself’.
Some of the playertypes mentioned depend on other play-ers, and additionally, other findings suggest that emotional response is largely non-independent of other humans (and positively anthropomorphised animals).
Based on these findings we may reasonably assume that a multiplayer environment, or hiring extremely gifted CGI personnel, is most preferred when it comes to creating an emotion inciting game. As this ensures the most controll over all variables.
When incorporating other human players, however, there is a choice in either letting them all directly participate, or using them to set the scene. As Dekker and Champion suggest, other human players can be used to work ‘against’
one player [3]. The drawback of such a system is that quality of experience cannot be guaranteed, nor controlled. In both cases, it is important to simulate real emotional affect, to ensure the process of emotional contagion, identi-fication and suspension of disbelief.
When direct participation is chosen the environment can be either player vs. player or player vs. environment9. In both cases there is a problem with the personal nature of emotion and what triggers each specific player; if there are multiple players in the same environment, how could they be offered personalized content?
There are several options, amongst others: keeping each player in a separate environment while still enabling so-cial interaction. They can even be virtually present in each other’s environment, however what they encounter is tai-lored to their specific emotional makeup.
Not all players are socializers, having to socialize could affect the effectiveness of the game, and therefore if a multiplayer gameplay is chosen over single player, interplayer communi-cation should be optional.
Whichever choice in mechanics is made, this choice needs to be supported by the story.
4.2.2
Story
Story is an important aspect, not only in regulating arousal, but also in constituting suspension of disbelief and inher-ently the engagement of (in this case) the player.
As mentioned earlier, there are multiple types of players, each with their own motivations, which all fit with different preferred stories (e.g. a socializer wouldn’t particularly like pacman).
As there is a limited set of narrative formulae, these can be used to serve as the basis for generating personalized content and inherently serve as a ways of giving meaning to mechanics.
One general narrative structure that seems especially suited for sustaining almost any form of narrative is that of the labyrinth or rhizome. According to Schell, for somewhat similar reasons, every game designer should consider the heroes journey structure [10]. There are, however, count-less more narrative structures, all with their strengths and weaknesses, which should be carefully weighed.
Mechanics and a story need to be conveyed in some way, and that is where the aesthetics come in.
4.2.3
Aesthetics
Even though there is no evidence that suggests high fidelity is needed, if the environment should not be made multi-player, the human characters within the environment need to be as credible as possible. This involves adding enough
9this doesn’t necessarily mean competitive, a puzzle also
detail to show facial expressions and enabling eye contact, supported by a credible AI.
An additional factor that needs to be taken into account is theatrics; creating anticipation and participation. As emo-tion is contagious, the mood can be already set by other players and humans alike, within the direct context of the game.
As a general rule, the aesthetics should initially correspond to the measured state of the player. When that state is not the one desired, the aesthetics can be adjusted towards a setting corresponding to the desired state. Dekker and Champion, for instance, used an audio feedback of the heart rate, with success [3].
Aesthetics are dependent on the way they are delivered; the technology, which will be detailed in the final section of the game model.
4.3
Technology
Creating an immersive experience relies on suitable tech-nology. In theory, the more modalities are used, the closer the experience resembles reality. Although reality is not the direct goal, sensory realism can further help induce immer-sion, suspesion of disbelief, identification and as such help constitute a first-hand emotional experience.
5.
CONCLUSION & FURTHER RESEARCH
Starting out with the idea that within the field of (entertain-ment, but most notably serious) gaming each game brings with it an intense and personal emotional experience, this thesis explored the possibilities for doing so. Instead of re-mediating other media, this thesis is aimed to show that the unique potential of games should be capitalized upon. Foremost, because this potential can be used to advance the multidisciplinary knowledge on emotion, due to its ability to act upon all modalities of emotion through its interactive nature.
To do so, focusing on the question of how to incorporate all of the emotional modalities in game design, we have first looked at what is known about and accomplished in measur-ing, interpreting and affecting emotions in section 2. From this section it became clear that there were a few challenges; A. The biofeedback loop needs to be made unconscious B. There is no real grasp of which parameters should be
altered
C. It needs to be made sure that arousal results from a reaction to the game
D. Arbitrary reactions due to the following need to be avoided:
• the Kuleshov effect,
• differences in social/cultural background, • sole focus on those aspects of emotion that can
be measured realtime
E. The game should be compelling, interesting and per-sonal,
F. but not too compelling, leading to desensitization or loss of interest
G. (Emotional) distance between the player and the game should be decreased without endangering the former points
With the above mentioned challenges in mind, through means of a literature study, it can be established that there are mul-tiple tools to overcoming these challenges.
By taking into account all of the modalities, and subse-quently valence in addition to arousal, through the use of narrative in its abstract, almost formulaic, form, and the addition of a workable playermodel in the shape of a player type and limited understanding of the social cultural back-ground, a first tentative step towards affective gaming can be made.
To summarize more explicitly, challenge [A] was solved through creating an affective game, in which the biofeedback loop is not made central to the interaction between the player and the game. Challenge [B], needs further research, what how-ever can be established, dependent on the scope of the target audience, different mechanics and narrative structures need to be used based on the player’s preferences (which can be monitored real-time, in addition to a classification prior to playing the game). Based on the technology used and the physical context, challenge [C] can be more or less overcome, ideally physical movement is incorporated in the game. The use of abstract narrative structures, can help overcome chal-lenge [D to G]. And lastly, the application of agency can es-pecially help decrease the distance between player and game (challenge G).
The next step that needs to be taken is to create a minimal viable prototype, consisting of some, if not all elements, in order to establish baselines (especially with regards to the implications of the social/cultural background and the po-tential correlations with the player types), through means of approaches such as neural networking. Certain theoreti-cal assumptions, such as the preferences to certain narrative structures of certain player types need to be validated. What is, furthermore, yet to be researched, is which other parameters, apart from those discussed in the former sec-tions, can be effectively altered to generate the desired out-come.
6.
REFERENCES
[1] Ahn, J., Gobron, S., Silvestre, Q., and
Thalmann, D. Asymmetrical facial expressions based on an advanced interpretation of two-dimensional russell’s emotional model. Proceedings of ENGAGE (2010).
[2] Coleridge, S. T. Biographiae litteraria of
biographical sketches of my literary life and opinions. Rest Fenner, 23 Patermaster Row, 1817.
[3] Dekker, A., and Champion, E. Please biofeed the zombies: enhancing the gameplay and display of a horror game using biofeedback. Proc. of DiGRA (2007), 550–558.
[4] Gilleade, K., Dix, A., and Allanson, J. Affective videogames and modes of affective gaming: assist me, challenge me, emote me.
[5] Hudlicka, E. Affective computing for game design. [6] Lang, B. Oculus shares 5 key ingredients for presence
in virtual reality. http://www.roadtovr.com/
oculus-shares-5-key-ingredients-for-presence-in-virtual-reality/, September 2014. (Visited on 07/06/2015).
[7] Murray, J. H. Hamlet on the holodeck: The future of narrative in cyberspace. MIT Press Ltd, 1998.
[8] Nacke, L. E., Bateman, C., and Mandryk, R. L. Brainhex: Preliminary results from a neurobiological gamer typology survey. Springer.
[9] Rouse, R., and Perron, B. Match made in hell: The inevitable success of the horror genre in video games. Horror Video Games: Essays on the Fusion of Fear and Play (2009), 15–25.
[10] Schell, J. The Art of Game Design: A book of lenses. CRC Press, 2014.
[11] Schell, J. Gamasutra: Jesse schell’s blog - making great vr: Six lessons learned from i expect you to die. http://www.gamasutra.com/blogs/JesseSchell/ 20150626/247113/Making_Great_VR_Six_Lessons_ Learned_From_I_Expect_You_To_Die.php, June 2015. (Visited on 07/06/2015).
[12] Smith, M. Engaging characters: Fiction, emotion, and the cinema.
[13] Stewart, B. Gamasutra - personality and play styles: A unified model, 2015.
[14] Watson, J. B., and Rayner, R. Conditioned emotional reactions. Journal of experimental Psychology 3, 1 (1920), 1.
