Aesthetic Details and Player Behavior : a study on how the colors of checkers in backgammon affect players’ choices of action

(1)

Aesthetic Details and

Player Behavior

A study on how the colors of checkers in

backgammon affect players’ choices of action

Andreas Bergqvist 4/30/2013

Bachelor Thesis, 15hp

Game Design and Graphics, third year Supervisors: Hedda Gunneng, Adam Mayes Examiner: Masaki Hayashi

(2)

i

Abstract:

Based upon earlier studies in the fields of psychology and game design by Frank and Gilovich (1988) and Yee (2007), this paper examines the influence of color on players of backgammon. The paper examines the probability of a player hitting an opponent’s piece when possible in backgammon depending on the color of the checkers. The data sample was compared using a t-test but showed no significant correlation. Based upon the results it was concluded that the Proteus Effect is not as effectively achieved through units in board games compared to avatars in virtual worlds.

Keywords: Backgammon, Color, Behavior, Choice of Action, Player, the Proteus Effect, Board Games, Self-perception, Self-representation

(3)

ii

Acknowledgements:

I wish to acknowledge the help provided by various people to this project.

I am grateful for the patient supervising and guidance given by Adam Mayes as well as Hedda Gunneng and Jakob Berglund.

I would also like to express my gratitude to Amanda Larruy-Bergqvist who supported me through the project. My gratitude extends to Alexander Eriksson and Johan Sjöström for their help.

(5)

1

1 Introduction:

In 1988, Mark G. Frank and Thomas Gilovich published the paper The Dark Side of Self- and Social Perception in which they presented the hypothesis that the colors of uniforms in sports would affect how players act during games. This might not be surprising since how color is often associated with feelings and emotions; red is the color of love while blue is often referred to have a valence of sadness. Today, 25 years later, many contemporary competitive video games allow their players to buy

aesthetic choices with real money for their avatars. Recent studies have shown evidence for the effect of the visuals and colors of avatars on players’ behavior in non-competitive virtual worlds. (Peña 2006; Yee 2007)

This paper will examine how the colors of the checkers in backgammon affect the player’s play style. This will be studied through the probability of players hitting as well as the amount of opportunities they are given to hit.

If this effect is evident within competition-based board games, it can be assumed to exist in other competitive games as well. The choice of board game was backgammon. This was due to backgammon being symmetrical and the color of the checkers not having effects on the game’s mechanics. Backgammon includes risk and reward which encourages different kinds of play styles. A match is played between two players and that it is a renowned game which is played at a professional level.

The study’s result does not show any significance of the effect of checkers’ colors on players’ actions in backgammon.

(6)

2

2 Basic Rules and Knowledge of Backgammon:

As the rules for backgammon are extensive, only the most necessary terms will be explained. If you as a reader want to study them more in depth, there is a link, after the sources, to a webpage which has the rules for the game

The board game backgammon is approximately 5000 years old. (Bray 2009:2) It consists of letting two players race with 15 checkers per player against each other on a 24 tiles long board. The players use the rolled numbers of the dice in order to decide how to move their checkers. When all the checkers have reached the last stretch of the board, the goal changes to removing the checkers from the board. (Bray 2009:9) Backgammon is usually played several games in a row until one of the players reaches a predetermined score.

Pieces:

Point:

The board you play backgammon upon is divided into 24 sharp triangles. These are referred to as points.

The Board and the Bar

Backgammon is played at a foldable board. The elevated edge in the middle of the board is called the bar. It separates the so called home boards from the outer boards. The home boards and outer boards are a quarter of the whole board each. Each of them includes six points.

Checkers:

The checkers are the pieces you control in the game. Each player has 15 of them. The checkers are often white, black or red, but they are not required to have those colors. Each point is a valid position for the checkers. A checker can never be standing on more than one point at any time. Checkers are only allowed to be placed on free points. A free point has no more than one enemy checker upon it.

Rules:

Dice:

At each player’s turn two dice are rolled. The players will use the numbers of the roll as the movement of the checkers across the board, each point on a die represents a point on the board. Each die will only affect one checker at each turn. Both dice can be used at the same checker thus letting it first go the distance of one of the dice and then of the other one. Thus on a normal roll of the dice you are allowed to either move one or two checkers. The exception is when both dice show the same side, the player is allowed to do four moves instead of two.

Blot:

When a single checker is left alone on a point it is considered a blot.

Hitting:

When one of the players moves one or more of their checkers to a point where the other player has a blot, the blot is hit. The hit checker is moved to the bar and the hit player needs to re-enter the hit checker before any other move can be made. Because of hitting players usually try not to leave blots.

Owning a Point:

If a player has two or more checkers at the same point, that player owns the point. Owned points cannot be hit and are therefore an important defensive placement of the checkers. When a second checker is placed upon a point it is referred to as making a point.

Bearing off:

When all 15 of a player’s checkers have arrived in the home board, the player tries to bear them off. The first player to bear of all their 15 checkers wins the game. If one of the players’ checkers are hit during the bearing off, that checker must be moved back into the home board before the bearing off can continue.

(7)

3

Doubling:

Backgammon has an additional set of optional rules which allow and encourage gambling called doubling. It consists of giving the players the options to double the stakes or forcing the other player to forfeit the game.

Tactics:

Prime:

A prime is when one of the players owns several points in a row. This limits the opponent’s possibilities to move forward since one player can’t move onto a point owned by another player.

Blitzing:

Blitzing is an offensive tactic that tries to hit an opponent as many times as possible in your half of the board.

Anchors:

An anchor is when a player owns a point in the opponent’s home board. Leaving an anchor is a defensive measure to be sure to get a hit checker back into play, as well as counter-defensive in how they block the movement for the opponent.

Building:

Players are building when they focus on optimizing the placement of the checkers on their half of the board. Building usually, but does not need to, consist of a prime.

Racing:

Racing is when players focus on trying to move their checkers as fast as possible into their own home board.

Pips and Pip-count

By doing a pip-count you measure how it goes for one player compared to the other player. In its simplicity you count how many pips, dots on the die, are needed to move all your checkers of the board and compare it towards the opponents’ amount of pips. A rule of thumb is to double when you are 10% ahead in pips.

Backgammon Notation:

The notation explained here and used in this paper is based upon the version of backgammon notation Chris Bray uses and explains in his book Backgammon for Dummies (Bray 2009:22-24). Backgammon notation was created to let games be recorded. This paper will use notation of the moves to make it clear for the reader when moves are described. Slight differences can be viewed in different versions of backgammon notation.

Each point is numbered. The first point (1) is the innermost point on your home board and the last point (24) is furthest from your home and innermost in the opponent’s home. The notation has two different variants. Either it switches the order of the numbers so the current player always will go lower while moving the pieces, or the numbering of the points is static which results in one player going with the increasing numbers. The latter will be used in this paper.

A move is recorded as XY a1/a2 b1/b2. X and Y are the values the dice have shown. The highest value is always written first. The notation a1/a2 and b1/b2 refers to the movement of the checkers. The a1 refers to the starting position while a2 refers to the position moved to. The same is the case for b1/b2 but for the second moved checker.

If both dice’s values are used to move the same checker, it is written as a1/a2/a3. A2 refers to the first move and a3 to the second.

(8)

4 When a double is rolled, the player is allowed four moves instead of two. This is noted by adding a parenthesis with the number of checkers that do the move. This will be written as a1/a2 (z) in which z is the number of checkers that moved.

A hit is noted by adding a * after the move, for example a1/a2* or b1/b2*/b3.

When a checker has been hit it is placed on the bar. The move to take it from the bar is noted as bar/a2. If it can’t enter the board with the dice rolled it is noted as bar/fan.

When a checker is borne off in the end of the game it is noted as a1/off. Sometimes there are no possible moves to do. This is noted with no play.

If players play with the gambling rule a double is noted as double. The opponent’s answer is noted as either take or pass depending on what they choose to do.

At some points it will be impossible to see the top side of the die and the player cannot use the die to move. In those scenarios the die will be written as a zero (0) due to it is impossible to know what the die showed. This is not included in the standard of the notation.

Figure 2: Backgammon example – Red has rolled a 4 and a 3. White has left blots on point 2 and point 4 in red’s home. Red can hit white both by moving checkers from point 6 and point 8 but due to that both those moves require a 4 red can only hit one of them.

(9)

5

3 Background and Previous Research:

Through the latest century researchers have done an extensive set of studies examining the effect colors have on people. As of the growth of the game industry in the latest years, more studies of the effects on the player have been conducted.

3.1 Color theory:

A color is the perceived effect of light bouncing of a surface. A color space is a range of colors built up by a set of parameters. There are a lot of different versions of color spaces, each with its own use and dependent on different parameters.

In the early 20th century Albert Munsell redefined the way people considered color spaces in his book A Color Notation. It was revolutionary for its time; Gordon McKay describes it in its introduction as a “decided step in advance over any previous proposition” (Munsell 1919:5) and portrays how it goes away from the previous attempts to define color from the wavelength. (Munsell 1919:5-6) It became the base for the color model Hue, Saturation, Value (HSV) which Alvy Ray Smith developed during his work on SuperPaint (Richard Shoup, Xerox PARC, 1973) in 1974. (Smith 1997) HSV can today be found in most of the contemporary digital paint systems. Munsell’s definition of color space consists of the parameters hue, value and chroma. (Munsell 1919:18-19)

3.2 Psychology:

3.2.1 Priming and Behavioral Assimilation:

Priming is a large field of study within psychology. Thus only a small selection of studies will be mentioned. The selection will cover different areas within priming. Kelley (1950) showed that describing a stimulus person as rather warm or rather cold gave big differences in how people perceived the stimulus person. People given the warm description were more likely to talk with the stimuli and rated the stimulus person’s attribute higher.

Bargh & Pietromonaco (1982) tested if priming could be done without the subject being aware of it. They had words of different valence flashing at a monitor at speeds which made them unreadable by the participants. The study examined how often words were detected by the subjects to make sure that the subjects were not aware of it. On another group of participants similar words flashing at

unreadable speeds were presented, the participants were then assigned the task to describe their impression of a stimulus person. The results show that the valence of the primed words affected the impressions of the stimuli.

Both Automaticity of Social Behavior: Direct Effects of Trait Construct and Stereotype Activation on Action (Bargh et al. 1996) and The Relation Between Perception and Behavior, or How to Win a Game of Trivial Pursuit (Dijksterhuis et al. 1998) give evidence that priming has effects on a person’s behavior, an effect which is referred to as behavioral assimilation. Bargh and his colleagues (1996) describe three different studies of behavioral assimilation. The first experiment examined how long it would take a subject primed with rude, neutral or polite words to interrupt a discussion between the experimenter and a confederate acting another subject. The study showed that a subject who was primed with a rude attitude was more eager to interrupt the experimenter than a subject primed with a polite attitude. The second study Bargh and his colleagues did examined if priming a subject with words connected to an elderly stereotype would affect the speed at which they walked down a corridor from the experiment. The study gave obvious results that the subject’s speed was affected by the priming. The third study examines the level of hostility in a Caucasian subject after being primed with a picture of either an African American or Caucasian face when encountering a computer error that requires them to do a tedious task again. While I am not endorsing the study due to its level of

institutionalized racism, it gives an example of behavioral assimilation based on an existing bias in the subject. The study let two assistants blind to the rest of the study rate the level of hostility observable in the subjects’ expression upon encountering the error.

Dijksterhuis and van Knippenberg (1998) examine how priming of roles with valences connected to different levels of intelligence and knowledge affected the results on a multiple choice question sheet.

(10)

6 The authors perform four different experiments on behavioral assimilation. The experiments were similar and examined different parts of the theory. All the experiments gave evidence that by priming a trait it was possible to increase that trait in the participant and the time spent at being primed was important for the level of effect. They conclude that by priming stereotypes and traits it is possible to increase and decrease those traits in a subject. This however does not mean that a subject knows more by being primed with the trait knowledgeable, but it is possible that it could be explained either by an increased confidence in their own knowledge, by a higher amount of effort spent or by utilizing a higher level of thinking.

3.2.2 Colors and Valence:

An extended study of how colors affect people covers the results of 89 earlier studies on color and affect. (Adams et al. 1973) In the paper A Cross-Cultural Study of the Affective Meanings of Color, Adams et al. covered participants from 20 countries and examined if there existed a cross-cultural relation between valence and basic color terms. In the study 23 groups of 35 to 40 participants were given the task to rate the basic color terms on twelve seven point scales. The scales were “equivalent in affective meaning” (Adams et al. 1973:136). According to the authors the result indicated a cross-cultural relation between color and affective meaning. The study showed that independent of culture, people will in general associate with colors in the same way. They conclude that red is strong and active, black is bad, strong and passive while white is good, weak and active Adams et al. limits colors into groups by seven of the basic color terms. This causes the paper to partially be based more around the naming of colors and general assumptions of a color based on the basic color terms. Even so the authors indicate a relation between the colors of the basic color terms and association.

A study of how people perceive attributes through color indicate that brightness of colors causes overall more positive assumptions in humans compared to less bright colors. (Meier et al. 2004) Meier et al. examined if people preferred words with positive valence in white over black and words with negative valence in black over white. Their results were positive and they concluded the following from their result.

Our results support the existence of an automatic association between brightness and affect. The association suggests that in making affect judgments, people attend to the physical features of an object in such a way that irrelevant physical metaphors (e.g. related to color) bias their response. (Meier et al 2004:86)

The brightness Meier et al. use is a measure of the amount of light reflected in regards to an extreme black or white. Thus they describe lightness instead of brightness. They also limit the research to the extremes of value, thus only discussing and studying black and white. Even if it regards the extremes of lightness, the study suggests that a relationship exists between high lightness and positive valence, and low lightness and negative valence.

3.2.3 Self- perception in Sports:

An early study of how color affects behavior in sports is a study which concludes with how professional football and hockey teams wearing black clothes appear more malevolent and become more penalized than teams not wearing black. (Frank et al. 1988) The paper The Dark Side of Self- and Social Perception consists of four different studies. The first tests if people who don’t know much about sports find dark sports uniforms to have more negative valence than light uniforms. (Frank et al. 1988:75) The second part of the study examines the amount of penalties in the records of hockey and American football. (Frank et al. 1988:76) The third study examines how referees judged games in which players wore black or white uniforms. (Frank et al. 1988:79) The games were staged so that no differences except the colors of the uniforms would be present. The fourth study examined how subjects sought after possibilities to be aggressive by letting them rank a group of activities which they would most want to do. (Frank et al. 1988:81) The fourth study also included a Thematic

Apperception Test (TAT). The TAT consisted of subjects describing a picture on a card and being graded by the level of graphical violence they used in their description. The paper gives evidence that individuals in black uniforms seek more aggressive challenges than those in white, as well as dark colors having a more negative valence than light colors and that referees judge players in dark

(11)

7 uniforms harsher. Aggression was defined as the desire to do harm to others. Frank et al. also state that the observed increase in aggression is presumed to be connected to the uniform and the surrounding factors of the competitive sport. (Frank et al. 1988:83-84)

A later study reexamines the topic of the Frank and Gilovich study further and studies the latest 25 years of hockey. (Webster et al. 2012) The study examines the records of 25 years of hockey and the penalties for teams playing in dark uniforms as well as teams playing in light uniforms. In the study the authors reach the conclusion that there is a correlation between black uniforms and penalties but the paper concludes that the reason for the penalties cannot be defined from results of the study. Either the penalties depends on referees judging players in dark uniforms harsher or on individuals in dark uniforms playing with more malevolence.

A study that was done during the 2004 Olympic Games indicated that in boxing, tae kwon do, Greco-Roman wrestling and freestyle wrestling, the competitors in red win more than those in blue clothes. (Hill et al. 2005) Hill et al. studied how often players of the same level won when having a red uniform compared to when wearing a blue. The results showed that a red uniform gives a better chance of winning when the players being at a similar level. The authors also state that the

improvement achieved by the color of the uniform is very small. It will not affect the outcome if the opponent is better than the subject in the red uniform. The paper does not come to a conclusion why this is. They portray two possibilities. The first being an effect of self-perception, that the player in red will do better due to wearing red or that the player in blue will do worse due to wearing blue. The second alternative is that the player is being affected by the opponent’s uniform. This is portrayed in similarity to red being a color showing of testosterone in animals when animals compete over for example courtship. Hill et al. concludes in that colors of sporting attires may affect how humans play sports. The paper states that it needs to be studied more in depth and that the regulations controlling sporting attire could suffer implications.

3.2.4 The Proteus Effect in Virtual Worlds:

There have been studies of how visual cues of avatars affect the behavior of players in virtual worlds. The effect is in those studies is referred to as the Proteus Effect. (Yee 2007) Two of the studies focus primarily on how colors of avatars’ clothes affect behavior in groups. (Peña et al. 2006; Peña et al. 2009) In Avatar Color and Social Identity Effects, Peña et al. (2006) examine how participant’s behavior was affected when assigned with avatars dressed in black and white with either a group or individual identity. The participants were assigned to discipline an alleged criminal within a game system. Before and after the discipline, the group was asked to discuss the case. The study’s result shows that there was a difference between how the participants with an avatar in white and in black acted. According to the study the avatars in white caused players to care more about what the group thought and spend a longer time discussing the issue than the players with avatars in black. (Peña et al. 2006:14) The players with avatars in black thought the crime was less severe and felt more likely to have done it themselves than the players with avatars in white robes. (Peña et al. 2006:13)

The Priming Effects of Avatars in Virtual Settings (Peña et al. 2009) also discusses the effects the clothes of an avatar have on its user. The study is based upon two different experiments. The first experiment continues from their previous study (Peña et al. 2006) to examine differences in behavior when playing with an avatar in black or white. The study is very similar to their previous study but does not consider the group or individual identity and instead focuses more on the color of the avatar’s clothing. The results fit with the author’s previous study. (Peña et al. 2006)

The second experiment studies the effect of avatars with attires that has a symbolic meaning; doctor coats and Ku Klux Klan garbs. In the experiment the participants were either given an avatar dressed in either a doctor coat or a Ku Klux Klan garb, or did the experiment without being showed an avatar. The participants were then showed two TAT cards and were then given the task to describe them. Two judges uninformed of the purpose of the study and of the identity of the participants were assigned to grade the personality from the descriptions. The descriptions were graded on affiliation, aggression, achievement and power. The results showed an existence of differences in the valence of the avatar in

(12)

8 the TAT descriptions. The conclusion of the study states that the participants were unaware of that their behavior was affected by the avatars.

There has also been a set of studies of how the appearance of an avatar affects behavior in the virtual world (Yee 2007) (Yee et al. 2007) and also the effects on behavior after stopping playing in a virtual world (Yee et al. 2009).

In The Proteus Effect: Behavioral Modification via Transformations of Digital Self-Representation (Yee 2007) Yee does his first assessment of The Proteus Effect. The paper discusses four pilot studies and two different dissertation studies that were done on The Proteus Effect. The first pilot study places the participant in a room wearing a head-mounted display. (Yee 2007:33) On the display a 3d room with a mirror and a confederate, an avatar belonging to a research assistant, is presented. In the mirror, the participant can see their avatar. Different participants are given different avatars that are either attractive or unattractive. When the test starts a research assistant asks the participant to verify what they see in the mirror. The study examined how close to the confederate the player moved and how much they told about themselves to the confederate. The study shows in its result that the participants given an attractive avatar moved closer to the confederate than the one given an unattractive avatar. (Yee 2007:36) The attractive avatar also made the player speak more about themselves than the unattractive avatar. The attractiveness of the avatars was assessed through earlier testing in an online survey. The attractiveness was based upon the average of the 14 answers that was given to the survey. The second pilot study of The Proteus Effect: Behavioral Modification via Transformations of Digital Self-Representation (Yee 2007) examines the effects of the length of the avatar on the player. The test consists of an ultimatum game in the same virtual setting as the previous experiment. Two players bet over a pool of money. They take turns in deciding on how to split the money. The other player can then accept or reject the money. If he accepts both get the decided split, if he rejects none of them gets any money. In the study, the participant was either given a tall or a short avatar. The splits and how the participant accepted or rejected the splits were recorded. The results showed that the players assigned with a tall avatar asked for splits in their favor and did not accept deals which were not in their favor as often as those with shorter avatars.

The third pilot study examined how the progression for players in World of Warcraft (Blizzard Entertainment, 2004) was affected by the attractiveness and length of the avatars they play. The study researched the levels, which is the “main advancement metric” (Yee 2007:43) in World of Warcraft, of players’ avatars against the length and attractiveness of the fantasy race of that avatar. The races were dwarfs, gnomes, humans, night elfs, orcs, taurens, trolls and undeads. In the study participants were first assigned to order a total of 64 characters of the eight races that exists in World of Warcraft by how attractive they appeared. The mean attractiveness and height of the avatars was then compared to the average level of the avatars of those races. The result showed that the avatars whose race appeared more attractive and were taller had a higher average level than the avatars of shorter and less attractive races. These results conclude that The Proteus Effect is present in real virtual worlds and not only in test environments as well as the fact that The Proteus Effect keeps affecting the players over longer spans of time.

The fourth pilot study of The Proteus Effect: Behavioral Modification via Transformations of Digital Self-Representation (Yee 2007) iterates upon the second study. As in the second pilot study, the fourth pilot study examines the effects of the height of an avatar on its player through an ultimatum game. The main difference was that after doing the ultimatum game inside a virtual setting it was done again face to face with the assistant who played the confederate. The results showed that inside of the virtual setting participants with tall avatars were more likely to give splits in their own favor. The second interesting data that was found out was that players which had been assigned a tall avatar did in the face to face ultimatum game bolder splits than the players of shorter avatars. This concludes that the avatars affect a player “conception of social identity” (Yee 2007:49) even after exiting the virtual setting.

(13)

9 The first dissertation study of The Proteus Effect: Behavioral Modification via Transformations of Digital Self-Representation (Yee 2007) is an iteration of the first pilot study. The main difference between this study and the pilot study is that it tries to find evidence for the Proteus Effect not being a version of behavioral assimilation. Yee explains behavioral assimilation as priming a person’s

behavior. (Yee 2007:56-57) Overall the study was done as the pilot study. To be able to detect if The Proteus Effect is not behavioral assimilation half of the participants were presented with a virtual setting with a mirror, as before, while the other half were presented with a virtual setting in which the mirror was exchanged with a monitor playing a recording of how a previous participant moved their head. After that followed the same procedure as it was in the first pilot study. After the sequence within the virtual reality, the participants were asked to help with a study of whether dating sites can conclude on good partners by only seeing the picture of them. This followed with the participant filling out personal information. The participants were then given nine photographs of individuals. The assistant asked the participant to pick the two individuals which seemed most interesting and the one that was most likely to be attracted to the participant. Through this method, beyond the same results that were recorded in the pilot study, the photos of the participants and the chosen photos were compared. The participants’ height was also measured as well as asked for in the personal information acquired. The results of the study gave evidence for a continued effect on the subject. Participants, who had an attractive avatar and had seen it, chose more attractive persons in the second part of the study. In the setting with the monitor the result was not equally evident. The virtual setting with the mirror caused a notable difference in the length of the participant compared to the setting with the monitor.

A post-hoc comparison with Fisher’s LSD [Least Significant Difference] test showed that in the mirror condition, participants in the unattractive condition were significantly more likely to increase their reported height (M = 1.17, SE = .33) than participants in the attractive condition (M = .17, SE = .39), p = .05. (Yee 2007:67)

The results of the interpersonal distance showed that it only caused a notable effect when the mirror was within the virtual space. Through these results it concludes that “The Proteus Effect” is not a result of priming.

The second dissertation study goes further in depth than the second pilot study. The experiment iterates on the previous study on a larger scale, as it examines if the effects of an avatar’s length affect a player when bargaining inside of a virtual setting. As the results are similar to the pilot study and do not bring forth anything that will affect this study, the second dissertation will not be looked into any deeper.

3.3 Game Design:

3.3.1 The Player and the desire of Emotion:

A challenge is a non-trivial task that requires the player to provide an amount of effort to be

accomplished. (Adams 2010:10) Many games give the player a choice in how to complete challenges as well as giving the player a specific set of challenges. (Adams 2010:256) There are a lot of different challenges that in their basic state require the player to do different actions to complete them. (Adams 2010:261) This gives a player a lot of options of how to play a game. But a player of games most often wants to have fun. Fun is in most cases the desired and aimed for emotion when playing a game. (Bartle 2004:129) Players will act differently and aspire to complete different goals in games as each player has his or her own desires in the game and different ways of achieving their desired emotion. (Adams 2010:253; Bartle 2004:130; Fullerton 2008:312) Even though players aspire for different goals, most games have a final winning condition and in the end this condition is where many of different ways of playing a game reunite. Different researchers and professors define these groups of players and aspired ways of playing differently. There are two main ways of defining these trends; either by the commitment and intent of the actions the players seek to do while the action itself becomes of inferior value or by the players’ relation to the virtual world and the other players. For example, Fullerton (2008:92) describes a large set of roles players might fall into while the MDA

(14)

10 framework (Hunicke et al. 2004) has a different approach. Instead of categorizing the players, they categorize the games through the way they inspire different emotions.

Bartle (2004:130), Yee (2002) and Lazzaro (2004:3-5) divide these group similarly to each other. All of them describe an archetype that finds the best kind of fun in achievement. Being an achiever is about being on the top of what the game lets you do. It should not be easy as a certain level of talent or dedication is needed. All of them also describe an archetype that fills his or her needs in socializing, being a part of a group and interacting with it. They value friendship and relationships. Yee also gives the facet he calls Leadership. It is similar to the socializers in its need of others around him or her but does not to be part of the group. It is about being above it, governing it and leading it and thus is a sort of social achiever. Bartle presents the type Explorers. An explorer seeks to experience the

geographical content as well as the physics of the virtual world. They can also search to explore stories and lore. Yee and Lazzaro call their similar version for Immersion or Easy Fun. According to them the importance is not in the exploring but in the immersion without having to prove what you can do. They put importance into storytelling and role playing. Both Bartle and Yee present the players that live on others’ misfortune. The killers seek to afflict grief in others; either by direct aggression or through indirect manipulation. Lazzaro also presents the Altered States key. This group includes players who play games for their potential as therapy. The game is a method of meditation for relaxation and serenity.

Randall (1992) explains his perceived social archetypes in the game Habitat (Lucasfilm Games, Quantum Link, 1986) as social dimensions. In Habitat, Randall describes five kinds of dimensions. The Passives are players who only want the easy and cheap entertainment. The Actives are the group that played more than they maybe should as they lose track of time. They prioritize the game but do not contribute to more than their enjoyment. The Motivators are the players that take the game a step further and contribute in a way that improves others enjoyment of the game. “They throw parties, start institutions, open businesses […]” (Randall 1992). Randall also explains that the Motivator’s also partake in contests. The Caretakers are the ““mature” Motivators” (Randall 1992). They assist new players and help the community in becoming a better place. The Geek Gods are the players that control the game. They are the moderators and operators that control, build and balance the game. Based upon the fact that different players get their fun from different actions in games and those players play games with different intentions and habits, it is possible to state that players have their relatively unique play style. People play games in different ways. These play styles make it possible to generalize players into groups of behaviors.

3.3.2 The Player and their Game Pieces:

Bartle (2004:154) describes immersion from the point of the player and the avatar at the point which the boundary between them has been removed. According to him the levels of immersion are player, avatar, character and persona. The player is outside of the game as it is being played. The avatar is the tool the player uses to play the game. The avatar can become the body of the character when the player lets it assume a role. It has gotten depth and a personality. The player has still distance between the character and itself. At the point where the distance is removed, the character has become the persona. The persona is the player itself within the game.

Both Yee’s research (2007) and Peña and his colleagues in their studies (2009) show through their results that the Proteus Effect can affect players’ behaviors without the players being immersed at the level which causes the avatar to become a character. (Bartle 2004:154) Bartle’s levels of immersion are not needed for an avatar to affect us. Thus something in the avatar causes an effect in us as players. An avatar refers to the game piece that takes the shape of a person, although it does not need to be human, and the player act on within the game through. (Adams 2010:128) In some games the player controls a party of avatars.

There is a lack of scientific and philosophic research and discussion on the role of the player piece in the relation to the player in board games. There have been extensive studies of the relation to the

(15)

11 avatar in computer games but on similar player representations in games without avatars. The player piece in board games comes in many different variations.

In the new edition of Arkham Horror (Chaosium, 1987; Fantasy Flight Games, 2005) the player’s piece is a portrait of a person while in the old edition it was a plastic abstraction of a human. In both versions of Arkham Horror, the players are given a sheet which holds a portrait and information of the person the piece represents. In Arkham Horror the player piece acts as an avatar, being the tool which the player plays the game through, although it almost tries to force the player to make it into a

character by giving the player the amount of information and lore to role-play through the avatar. In Monopoly (Charles Darrow, 1933; Parker Brothers 1935) each player has one unique piece which is represented by an inanimate object. The pieces only act as markers which show your position on the board. They are player tokens. They are not avatars due to them being inanimate objects and not a unit due to the mechanics; they are only used to mark a position and they are not a resource used to achieve a goal.

In Carcassonne (Klaus-Jürgen Wrede, Hans im Glück, 2000) the player controls a small set of pieces in the shape of stylized humans. Although they resemble humans they are not a player’s avatar. They are units in the sense of them being a resource used to gain points.

In backgammon (Unknown, earliest found pieces from around 3000 BC (Schädler et al. 2009)) the player controls a set of round abstract counters with a color that represents each sides’ units. They are units. The only similarity to the avatar is that both are representations of the player.

(16)

12

4 Hypothesis:

Earlier studies have shown that color has inherited meaning and the effect that behavioral assimilation and priming of simple roles and traits can have on a subject (Adams et al. 1973; Bargh et al. 1996; Dijksterhuis et al. 1998). Other studies show that the colors of uniforms affect self-perception of players of sports and that through the Proteus Effect the colors of avatars affect players’ behavior. (Frank et al. 1988; Webster et al. 2012; Peña et al. 2006; Yee 2007) The color red is connected to traits like strong, active, recklessness, salient, power, anger and dominance. (Adams et al. 1973; Hill et al. 2005) The color black has the inherit meaning of evil, negative, aggressive, strong and passive. (Adams et al. 1973; Frank et al. 1988; Meier et al. 2004; Peña et al. 2006; Webster et al. 2012) White is connected with good, weak, calm and active. (Adams et al. 1973; Frank et al. 1988; Meier et al. 2004; Peña et al. 2006; Webster et al. 2012) This study will assume there are no cultural differences in the way we perceive color. (Adams et al. 1973)

Through these valences of color and the earlier research on effects on behavior, I hypothesized that: 4.1 H1:

The color of the player’s game pieces in backgammon should either through behavioral assimilation, self-representation or self-perception cause a limited effect on players’ choices and play style in board games.

4.2 H1.1:

Red checkers will cause the player to play more daring and with higher risk than white and black checkers. Red will be more aggressive than white. The results will show red has a high probability to hit and has given a high amount of chances to the other player to hit.

4.3 H1.2:

Black checkers will cause the player to play more aggressive than with white checkers but not as stubbornly as red checkers. The passive aspect of black will not override the aggressive and evil valence. The results will show black has a high probability to hit and has given a low amount of chances to the other player to hit.

4.4 H1.3:

White checkers will cause the player to play friendlier and less aggressive than red and black checkers. The results will show white has a low probability to hit and has given a low amount of chances to the other player to hit.

(17)

13

5 Method:

To be able to research if there exist any effects of color in board games on players the existing studies regarding the effect on behavior of color, self-perception and self-representation will be studied. Prior to the study, two things needed to be decided; which game to examine the effects of color on behavior and how to gather data for the study.

5.1 Reasons for choosing Backgammon:

Backgammon is a renowned competitive board game for two players. The choice of backgammon was made due to several reasons.

1. Backgammon is a symmetrical game, and the only difference between the different sides is the color of the checkers. If comparing backgammon to chess, chess bases who is the first player on the color of the pieces. In backgammon the player who will start and what color they use is randomized. The color does not change between each match in a set of games.

2. Backgammon includes a level of risk versus reward. The game has a variable of chance affecting each turn through the dice and promotes gambling through the doubling cube as well as different styles of play. Players can go for a reckless race, an offensive blitz, a counter-defensive anchor or a counter-defensive prime

3. Backgammon is always played between two players. Each player that partakes in a game adds more outside-factors which are affecting the results and the collected data. A game that only needs two players is better at limiting the factors and variables that can affect the data than a game which needs more players.

4. Since backgammon is well-known and competitions in it are held, it would be easy to find recorded games to base the study on.

5.2 Video material:

There were three ways to gather material for the study; tests on participants, transcribed games and video recordings of games.

The initial plan was to do a test on participants, but due to several problems it was not possible to achieve. First, it proved to be nearly impossible to find enough participants who were willing to spend an hour on playing a board game. Second, few of the participants which could be found had a decent knowledge of the game. They also had very different styles of play from the start. This made the participants and the data unreliable; the data would give reproduction errors due to its low precision. The second option was gathering the data from already transcribed games of backgammon. Many notated matches could be found on the internet, but a vast majority of them lacked information on the colors of the checkers. This caused the transcribed games to be unusable in the study.

The last alternative was to transcribe games which had been recorded on video. A large quantity of recorded games from championships and tournaments exist on YouTube. The recordings show both the colors of the checkers and as the players were playing at a professional level, the play style and lack of knowledge of the game is not an issue.

5.3 Procedure:

All material that was used needed to fulfill a set of requirements. The recording was examined on the following points.

 The video needed to include the whole game. This includes that all the players’ moves can be perceived. Reflections in the dice were not considered a problem since the move of the checker makes the die roll perceivable. Recordings that were considered unusable included either movement in the camera, switching camera angles during the move or was badly cut. Recordings missing parts of the game could be used if an annotation of the game existed. Also, videos which were unwatchable due to low frame-rate or resolution were not used.

(18)

14

 Based upon the amount of existing studies on colors and the norm in colors of checkers in backgammon, the colors of the checkers that were used in the recordings needed to be red, black or white. Checkers with a medium level value, a high chroma and a red hue are

considered red. Checkers with a low value, a low chroma and of any hue are considered black. Checkers with a high value, a low chroma and of any hue are considered white.

 The players in the recordings needed to play at a professional level. This was because of two reasons. Games at a professional level are primarily played by players who follow the types of Bartle’s Achievers (2004:130) and Randall’s Motivators (1992). They are assumed to be playing to win. Since being of the Achiever type and playing for the same goal, they should all have similar play styles. The second reason for using games played at a professional level is due to the regulations and strict controls of professional games. A game played at a

professional level holds through this a higher level of reliability than a game of unknown origin played by unknown players.

If a recording fulfilled the requirements it was categorized into one of three groups dependent on the colors of the checkers in the matchup. The three groups of matchups were therefore red versus black, black versus white and red versus white.

The recordings were transcribed using backgammon notation. A note was added on each hit that was done on an opponent’s checker. At each turn, the maximum additional amounts of hits that could be done from the current position of the checkers and the rolled dice were also recorded.

In the study a sample of 84 matches was studied; 40 between red and black checkers, 24 between red and white and 20 between black and white.

(19)

15

6 Results:

The results were compared both between the base color terms of the checkers as well as between the groups of matchups on the probability of a player hitting when possible and the amount of

opportunities to hit. The results were compiled both per turn as well as to per game. This was due to the fact that the different compiled versions of the results gave different degrees of significance on a student’s t-test. The average of hits and possibilities to hit per game disregards the length of the games while the average per turn disregards that the opportunities and the hits took place in different games. All numbers are rounded down to three decimals in the tables.

A student’s t-test (t-test) was used to determine if any of the results showed any significant evidence. The test compares two sets of data and examines the likelihood of them being part of the same set of data. The t-test shows significance if the value of the t-test is below a critical probability, denoted as p. A t-test was run with the colors of the checkers as the between-subject factor and the dependent variable being the probability of a player hitting. All games in which a player could not hit will not be used. This is due the mathematical problem which arises when dividing by zero. It showed no

significance in a red versus black matchup. It showed no significance in the red versus white match-ups. It showed no significance in the black versus white match-match-ups.

A t-test was run with the colors of the checkers as the between-subject factor and opportunities to hit as the dependent variable. It showed no significance in a red versus black matchup. It showed no significance in the red versus white match-ups. It showed no significance in the black versus white match-ups.

These can be seen in table 1-3 on page 16. The tables include the average hits per game and per turn, the average amount of opportunities to hit per game and per turn, the probability to hit when possible and the standard deviation and the values of the t-test of each color in the match-ups.

(20)

16 Hits Oppertunities Probability of Hit

Red Black Red Black Red Black

Games: 40

/Game 2,775 3,175 5 5,55 0,578 0,562

SD (σ) 2,281 2,772 3,623 4,088 0,281 0,287

T-Test 0,483 0,526 0,799 p:0,05

Turns: Red: 694 Black: 697

/Turn 0,160 0,182 0,288 0,319

SD (σ) 0,386 0,412 0,519 0,524

T-Test 0,298 0,278 p:0,05

Table 1.

Results in Red and Black Match-ups in Backgammon

Hits Oppertunities Probability

Red White Red White Red White

Games: 24

/Game 4,083 3,458 6 5,375 0,668 0,605

SD (σ) 4,568 4,128 5,838 4,968 0,366 0,297

T-Test 0,621 0,691 0,539 p:0,05

Turns: Red: 382 Black: 385

/Turn 0,257 0,216 0,377 0,335

SD (σ) 0,483 0,436 0,565 0,558

T-Test 0,218 0,302 p:0,05

Table 2.

Results in Red and White Match-ups in Backgammon

Hits Oppertunities Probability White Black White Black White Black

Games: 20

/Game 2,5 2,05 4,7 4,25 0,532 0,482

SD (σ) 2,417 2,282 3,643 3,768 0,320 0,301

T-Test 0,549 0,703 0,850 p:0,05

Turns: White: 314 Black: 314

/Turn 0,159 0,131 0,299 0,271

SD (σ) 0,384 0,347 0,530 0,486

T-Test 0,326 0,480 p:0,05

Table 3.

Results in Black and White match-ups in Backgammon

(21)

17 The data was also compiled by the basic color terms and can be seen in table 4. The data was

processed with a single factor ANOVA, due to having three sample groups. ANOVA is similar to the t-test but allows more than 2 groups of data samples. All 14 different ANOVAs that were done can be found in the appendix.

/Game Black Red White Hits 2,80 3,27 3,02 SD (σ) 2,65 3,35 3,45 Oppertunities to hit 5,12 5,38 5,07 SD (σ) 4,00 4,56 4,38 Probability of hitting 0,53 0,61 0,56 SD (σ) 0,29 0,31 0,31 Being hit 2,68 3,28 3,16 SD (σ) 2,31 3,32 3,81 Oppertunities to being hit 4,90 5,48 5,20 SD (σ) 3,60 4,40 5,03 Probability of being hit 0,55 0,58 0,58 SD (σ) 0,29 0,29 0,35 /Turn Black Red White

Hits 0,19 0,17 0,19 SD (σ) 0,43 0,39 0,41 Oppertunities to hit 0,32 0,30 0,32 SD (σ) 0,54 0,51 0,55 Being hit 0,20 0,19 0,16 SD (σ) 0,43 0,42 0,39 Oppertunities to being hit 0,33 0,32 0,29 SD (σ) 0,53 0,54 0,52 Table 4. Results of Data Compiled by Color

Not any of the ANOVAs showed any significance. The closest was the ANOVA that examined the relevance of color when being hit. It gave a probability of being significant at 0,075 and the needed probability for being significant being 0,05. This can be seen in table 6.

(22)

18 SUMMARY

Groups Count Sum Average Variance

White 696 139 0,19971 0,18596

Red 1082 210 0,19409 0,17691

Black 1008 161 0,15972 0,14825

ANOVA

Source of Variation SS df MS F P-value F crit Between Groups 0,874 2 0,43677 2,58751 0,075 2,999 Within Groups 469,8 2783 0,1688

Total 470,6 2785

Table 6. Being Hit Dependent on Color per turn - ANOVA Single Factor

(23)

19

7 Discussion:

The results give no evidence that the assumption of the hypothesis being significant. Although no significant evidence for causality between color and an effect on human behavior was evident, a small shift towards a suggestion of an effect can be noticed in all sets of data, varying between different levels. Due to the very small amount of data and the small ranges of the values, not much can be said at this point. For it to be possible to make sure that causality exists in the relation between color and human behavior in board games, a larger sample of studies will need to be collected.

The hypothesis had the assumption that the differences between the colors would primarily affect recklessness and aggressiveness. A high level of risk and recklessness would give the other player more opportunities to hit than a safer and more secure style of play. Aggression was theorized to cause a player to have a higher probability of hitting.

After several matches between red and black the most evident difference would be in the amount of given opportunities to hit the other player’s pieces. As can be seen in the results, red gave more opportunities for black to hit than black gave to red. Red did play overall with a more reckless and a higher risk play style than the player with the black checkers. But due to lack of evidence to assert significance, this could be an accidental result of the data. A larger sample of data for the study would be needed in order to get conclusive evidence.

Even if the results of the study on the red and white match-ups show no significance in a t-test, there are differences in the data. Hypothesis 1.1 and 1.3 assumed that red checkers would give more opportunities to hit than white checkers as well as having a higher probability to hit. The higher probability to hit appears to fit with the collected sample of data as red have a higher hit-ratio than white. But, white gives red more opportunities to hit than red gives to white. As with the red and black match-ups, a larger sample of data for the study would be needed to be able to get conclusive

evidence.

The results of the match-ups between black and white checkers show the least significance of the three match-ups. The small differences that were noticed did not fit with the hypothesis. White had a higher amount of opportunities to hit while the hypothesis did not assume there to be any difference. White also had a higher probability to hit, which also goes against the hypothesis.

When viewing the ANOVAs of each color independent of the match-ups, an interesting detail can be observed. Black pieces are being hit less then red and white checkers. (Table 6; Table 10) Black gives fewer opportunities to the opponent to hit the black checkers. (Table 8; Table 12) Even if not being at the same level of evident, red appears to hit the most and have the highest probability to hit and white appears to hit the least and have the lowest probability to hit. (Table 5; Table 9; Table 13)

7.1 The interpretations of the effects of color

Due to the fact that red got more opportunities to hit white than white got to hit red, it is plausible to assume that a part of the hypothesis is incorrect. The most likely aspect to be incorrect which would cause this outcome is either if an effect on behavior presented itself in another way then was assumed or that the valence of the color is not correctly connected with the way it affects players. When deducing from earlier research and deciding on the hypothesis a choice of how different traits of the colors’ valence would be visible in a game of backgammon. The hypothesis assumed that due to the decreased aggression and the calm nature of white, the player would be more considerate and taking less risks. But Adams (et al. 1973) describes white as an active color. If the trait active is applied to a player, it is possible that white would race more and take more chances since the applied trait could cause an eagerness to act. These contradictions were thought about during the decision of the hypothesis. The choice was based upon which trait seemed the most evident in the earlier research. That black is considered passive is further evidence for. The ANOVAs (Table 6; Table 8; Table 10; Table 12) show that black is letting it self being hit less than the other checkers, thus suggesting the concept that the passive trait of black lets the player with the black checkers stay safer.

(24)

20 If this is the case, it is possible that the study examines the results and the transcribed games at the wrong premises. Instead of only examining one aspect that is assumed to be the outcome of the hypothesized effect, several aspects which would be affected could be studied. For example, players do pip-counts to know both when the timing for a double would be good and when it is wise to take the double. If a player would play with a more reckless style it can be assumed that they would be more likely to double when being less ahead than a player who play more controlled and securely. Instead of measuring how many opportunities to hit each player gives to the other, the amounts of blots after each move could be measured. This would be a more direct method to examine the recklessness of moves than to examine the opportunities to hit.

7.2 The Player-Checker Relation

In a similar way as the uniform of a hockey or American football player can affect the player’s behavior (Frank et al. 1988) and the effect behavioral assimilation has on a subject (Dijksterhuis et al. 1998), the Proteus Effect (Yee 2007) causes the player to be affected by the sole game piece in a game they are not even immersed into. But how does a checker in backgammon adhere to this? It appears to be no scientifically discussed definition of the players’ relation to the pieces of a board game, while there are many definitions that exist which describe the relationship between the player and the avatar. The checkers in backgammon are not avatars. The player pieces in Arkham Horror come close to fill the description of avatars even though the game forces it on the player. The player pieces in Monopoly are just tokens shaped as items and used to mark your position. In Carcassonne the player controls a set of humanoid tokens. These tokens are just units used in a larger scale. Compared to these player tokens, the checkers in backgammon does not fulfill the requirements inherent in avatars. The pieces in backgammon are not a direct factor of self-perception, the adhesive priming that causes behavioral and perceptive influences nor are they the self-representations of the player. Due to the fact that the results of this study do not show a significant relation between the colors of checkers and player behavior, it can be concluded that the distance between the player and a checker or unit is further than the distance between a player and an avatar.

7.3 The Other Side of Achievers:

The material used in the study was limited by the player type the participants was. This was to limit and eliminating differences in the players to get more streamlined data which would not be affected by strong variance in play styles. Because of this the study would be limited to one player type. The chosen player type was a combination of Bartle’s Achievers (2004:130) and Randall’s Motivators (1992). They would play the games to win and thus aim to do the move which caused the best outcome. They would also be invested into the game and partake in championships. Before the study, this seemed like a good idea since the actions would be more directed and thus easier to detect differences. After the study was completed, a realization of the questionability of this assumption arose. Professional players have decided in their mindset that they are going to do what they can to win. Through this, they have set their minds on doing optimal moves and actions. If this is the case, professional players would have made themselves immune to any or all of the effects that were hypothesized to exist. The results of Frank and Gilovich’s study (1988) and the study of Webster and his colleague’s study (2012) say otherwise. Both of their studies show an effect on professionals by the simple colors of their uniforms. This suggests that even professionals can be affected by self-perception. But there is an important difference between hockey players and backgammon players. Backgammon is a very strategic and logical game while hockey, even if it includes an element of strategy, is mostly about team coordination and athletic contribution. Being part of a team causes less individuality and thus lets people do things they would not do otherwise. Also the athletic effort and the physical force in hockey causes an increased release of adrenaline which can cause instinctive behavior. It is possible that these two factors cause the differences between a professional

backgammon player and a professional hockey player. The fundamental differences in the design of both games could be possible to prime the players to allow different levels of further effects of stimuli. 7.4 Outside factors

In the end of the game 2012 LA Open between Justin Nunez and Bill Riles, the former of them throws the dice cup as he realizes that he loses. The more games that players play against each other, the more

(25)

21 emotions seem to arise. This was not considered during the study and all the games from each

recording were therefore used. The effect on the data of the uprising emotions should be possible to decrease by limiting the study to only use the first or the first few games of each recording.

(26)

22

8 Conclusion:

Studies have shown that colors have an inherited meaning and valence. (Adams et al. 1973; Meier et al. 2004) Frank and Gilovich (1988) succeeded with finding a cause and effect of colors on subjects’ behavior in professional sports and Webster and his colleagues (2012) verified Frank and Gilovich study. Later studies have shown similar effects of color and player-representations affecting players’ behaviors in virtual worlds. (Peña et al. 2006; Yee 2007) Through this it was assumed that the colors of checkers in board games would affect their players. This paper studied how the colors of checkers in backgammon affect players’ actions. The study examined matches between professional players and noted how hitting and the opportunities to hit were affected by color. My results do not show that there is a significant relation between the checkers’ colors and player behavior in backgammon. Even if the results did not show any significance, the results show a small difference between the different sets of colors. The existence of a Proteus Effect or equivalent effect cannot be concluded or dismissed through the collected data sample.

8.1 Further Studies:

As of this study, there appears to be no significance of the color of the self-representations in board games. But due to the fact that it is assumed to only be a small difference in the result, a more

extensive sample of data will be needed if the results are supposed to show any significance. A deeper understanding of the effects of color on checkers can be achieved through further studies.

Due to the possibility that professional players’ behaviors are set to win, there is a probability that professionals would not be affected by the valence of colors. If this group of players would be immune to the Proteus Effect, further studies examining if it affects other types of players would be needed to give evidence of the significance of the hypothesis.

Many contemporary competition-based computer games follow a free to play model with micro-transactions. Because of this they often have purchasable aesthetic choices. The reasoning behind them usually consists of the idea that it is alright to buy virtual goods that do not affect the mechanics of the game. But if the Proteus Effect (Yee 2007) can affect players in these games, the choice of aesthetics would be a choice of the behavioral effects. Even if my study did not find any evidence of an effect within board games, it is possible that it can be achieved with in other competitive games. A further study into the effects of the avatars in these games would give a deeper insight into this.

If any evidence of effects on behavior through units and board game tokens would be found in further studies, it would be wise to examine the cause of the effect. The three different previously described causes affect people very differently; The Proteus Effect is described in its regard to the avatar-player relation, priming affects through thought or unconsciously and self-perception causes changes in behavior and attitude based on the current state of the self-image. Neither of them fits perfectly with units or player tokens. But if a further evidence for an existing effect on player behavior would be found, a reason for it should be examined.

(27)

23

Sources:

Adams, Ernest 2010. Fundamentals of Game Design Second Edition Berkeley, California, United States: New Riders Publishing

Adams, Francis M., Osgood, Charles E. 1973. A Cross-Cultural Study of the Affective Meanings of Color I: Journal of Cross-Cultural Psychology Volume 4 Issue 2 pp. 135-156

Bargh John A., Pietromonaco, Paula 1982. Automatic information processing and social perception: The influence of trait information presented outside of conscious awareness on impression formation I: Journal of Personality and Social Psychology, Volume 43, pp. 437-449

Bargh, John A., Chen, Mark, Burrows, Lara 1996. The Automaticity of Social Behaviour: Direct effects of trait concept and stereotype activation on action. I: Journal of Personality and Social Psychology, Volume 71, Number 2, pp. 230-244

Bartle, Richard A. 2004. Designing Virtual Worlds Berkeley, California, United States: New Riders Publishing

Berlin, Brent, Kay, Paul 1969. Basic Color Terms Their Universality and Evolution Berkeley, California, United States: University of California Press

Bray, Chris 2009. Backgammon for Dummies Chichester, United Kingdom: John Wiley & Sons, Ltd Collins, Allan M., Lotfus, Elizabeth F. 1975. A Spreadin-Activation Theory of Semantic Processing I: Psychological Review Volume 82, Number 6. pp. 407-428

Dijksterhuis, Ap, van Knippenberg, Ad 1998. The relation between perception and behavior, or how to win a game of Trivial Pursuit I: Journal of Personality and Social Psychology, Volume 74, Number 4, pp. 865-877

Fairchild, Mark D. Color Appearance Models: CIECAM02 and Beyond. I: Slides from a tutorial at the IS&T/SID 12th Color Imaging Conference. 9 November 2004. URL:

http://www.cis.rit.edu/fairchild/PDFs/AppearanceLec.pdf Retrieved 7 May 2013

Farmer, F. Randall 1992. Social Dimensions of Habitat’s Citizenry I: Loeffler, Carl E., Anderson, Tim 1994. The Virtual Reality Casebook New York City, New York, United States: Van Nostrand

Reinhold

Frank, Mark G., Gilovich, Thomas 1988. The Dark Side of Self- and Social Perception: Black Uniforms and Aggression in Professional Sports I: Journal of Personality and Social Psychology, Volume 54, pp.74-85

Fullerton, Tracy 2008. Game Design Workshop Boca Raton, Florida, United States: CRC Press Hill, Russel A., Barton, Robert A. 2005. Red enhances human performance in contests, I: Nature, Volume 435, pp. 293

Hunicke, Robin, LeBlanc, Marc, Zubek, Robert 2004. MDA: A Formal Approach To Game Design and Game Research I: In Proceedings of the Challenges in Games AI Workshop, Nineteenth National Conference of Artificial Intelligence

Keiser, Ross E., Prather, Ellen N. 1990. What Is the TAT? A Review of Ten Years of Research I: Journal of Personality Assessment, Volume 55 Issue 3-4, pp. 800-803

Kelley, Harold H. 1950. The warm-cold variable in first impressions of persons I: Journal of Personality, Volume 18, pp. 431-439

(28)

24 Lazzaro, Nicole 2004. Why We Play Games: Four Keys to More Emotion Without Story Oakland, California, United States: XEODesign Inc.

Meier, B. P., Robinson, M. D., Clore, G. L. 2004. Why good guys wear white: Automatic inferences about stimulus valence based on brightness. I: Psychological Science, Volume 15, pp. 82-87.

Munsell, Albert H. 1919. A Color Notation Fifth Edition New York City, New York, United States: Munsell Color Company

Peña, Jorge, Hancock, Jeffert T., Merola, Nicholas A. 2006, June. Avatar color and social identity effects on attitudes and group dynamics in online video games I: Paper presented at the 56th annual meeting of the International Communication Association, Dresden, Germany.

Peña, Jorge, Hancock, Jeffert T., Merola, Nicholas A. 2009. The Priming Effects of Avatars in Virtual Settings

Robertie, Bill, 2003. Backgammon for Serious Players Las Vegas, Nevada, United States: Cardoza Publishing

Rogers, Scott, 2010. Level Up! Chichester, United Kingdom: John Wiley & Sons

Schädler, Dunn-Vaturi, Ulrich, Anne-Elizabeth 2009. Board Games in pre-Islamic Persia URL:

http://www.iranicaonline.org/articles/board-games-in-pre-islamic-persia Retrieved 7 May 2013 Smith, Alvy R. 1997. Digital Paint Systems URL: http://www.alvyray.com/awards/default.htm

Retrieved 7 May 2013

Yee, Nick 2002. Facets: 5 Motivational Factors for Why People Play MMORPG’s I: URL:

http://www.nickyee.com/facets/home.html Retrieved 7 May 2013

Yee, Nick 2007. The Proteus Effect: Behavioral Modification via Transformations of Digital Self-Representation

Webster, Gregory D., Urland, Geoffrey R., Correll, Joshua 2012. Can Uniform Color Color

Aggression? Quasi-Experimental Evidence From Professional Ice Hockey I: Social Psychology and Personality Science, Volume 3 Issue 3 pp. 274-281

Wegner, Daniel M., Schneider, David J., Carter, Samuel R., White, Teri L. 1987. Paradoxical effects of thought suppression I: Journal of Personality and Social Psychology, Volume 58, Number 1, pp. 409-418

The Rules of Backgammon: