Sailing the Endless River of Games: The case for Historical Design Patterns

Jesper Juul: Sailing the Endless River of Games: The case for Historical Design Patterns. 1st International Joint Conference of DiGRA and FDG 2016. Dundee. (Presentation version.)


This paper makes two arguments: First, that video game history fundamentally proceeds by changing in ways we had not predicted, as new developments (casual games, independent games, free to play games) force us to consider facets of video games that we had previously ignored. We therefore continually have to reconsider what our object of study is; to reconsider the ontology of what we are studying. Second, how can we then grasp, or model, video game history? This paper therefore proposes that design patterns, originally defined as an atemporal list of patterns, can be reconceptualized as a historically evolving ontology of game design. This approach has the potential to allow us to explore the multifaceted history of video games in a single framework. This approach has the potential to allow us to explore the multifaceted history of video games in a single online system.


Design patterns, game history, ontology


What are we talking about, when we talk about video game history? I will argue that the major events in video game history concern things that had previously been taken for granted: MMOs like World of Warcraftmoved the role of the player community to the forefront; casual games reconfigured the audience; mobile games reconfigured distribution and business models; independent games set up a new relation between developer and audience.

In this way, video game history is pushing game studies towards interdisciplinarity, and thereby towards redefining what is relevant to studying games. In more technical terms, this interdisciplinarity forces us to reconceive the ontology of the object of study (Barry, Born, and Weszkalnys 2008). This is not a relativist position, but an observation that ontology can be “a way to structure the knowledge that emerges from specific perceptual, cognitive, and operational relationships with a world” (Gualeni 2015), which means that when video games change in ways we hadn’t anticipated, we have to reconsider what it is we are looking at in the first place. Video game history continually forces us to reconsider what it is we are studying, when we study video games.

Given this shifting ground, how can we then talk about video game history? One possible answer to this conundrum is to consider video game history from the point of view of Design patterns (Björk and Holopainen 2004). Design patterns are a tool for describing game elements shared between games, such as the common paper rock scissors pattern, where three types of units each dominate only one other unit type. Design patterns have proven useful for many types of game design analysis (Smith et al. 2011) (Hullett and Whitehead 2010) as well as for procedural content generation (Dahlskog, Togelius, and Björk 2015). Design patterns are the most prominent of a group of similar, though not identical, concepts. These also include mechanics (Hunicke, LeBlanc, and Zubek 2004) (Sicart 2008), and ludemes (Parlett 2015) (Bojin 2010)[1]. The core assumption shared by these concepts is that games contain identifiable units that can be shared between multiple games, and which can be combined in different ways. This has mostly been discussed in a synchronic fashion, i.e. as an analysis of the current state of game design. In this paper I would like to extend the theory of design patterns (on which I focus) to be able to account for historical change in video games.

Note: I am not addressing the general question of how to write video game history, but working from the specific question of how to capture the aspects of game design that different games share, and how to use this to understand how video games change over time. To examine this, it is instructive to consider some possible ways to structure such an endeavor, based on recent work in video game studies.

5 solutions to the problem

1) A straightforward solution could be to examine history from existing game definitions, such as (Juul 2003) and to simply see video game history as the continued expression of a few underlying principles. For example, we could look at how a particular component of a game definition, such as player effort, is expressed over time. Alternatively, we could examine how recent games challenge such a definition by virtue of toning down the role of player effort (see Dys4ia (Anthropy 2012) or Proteus (Key and Kanaga 2013)). Such an approach is in some ways useful, especially for examining how the concept of game may change over time, but it falls short a) in being unfit for examining changes that do not directly work to challenge our existing notions of what games are, and even then, b) only being able to discuss games that challenge the assumptions that we were already aware of.

2) A second, and well-known, model is to assume that history is really the expression of a particular binary opposition[2]. We could think of video game history as being the expression of the conflict between freedom and constraint, between rules and fiction (Juul 2005), or between abstraction and representation. This faces a problem similar to game definitions, in that we will only be capturing aspects of video game history that we identified from the outset.

Figure 1: A typology of video game topography (Aarseth, Smedstad, and Sunnanå 2003)

3) A more detailed approach is to create a general typology of video games (Aarseth, Smedstad, and Sunnanå 2003), as shown in Figure 1, but this is similarly unable to describe video game history for a simple reason: general typologies assume a particular perspective and level of detail for the description of video games, but the truth is rather that historical developments in video game design by themselves create a need for new distinctions. For example, it was only when games became three-dimensional that distinctions between types of three-dimensional camera movements became relevant. This makes static typologies unable to grasp video game change. This shows that an examination of video game history cannot work from a static list, but must work with continually changing field that continues to spawn new internal (and external) distinctions.

Figure 2: The segmentation of gameplay sub-hierarchy of the Game Ontology Project (Zagal et al. 2005)

4) Moving into more open-ended conceptions of video game history, The Game Ontology Project (Zagal et al. 2005) provides one possible solution to this problem, where the ontology of game design is seen as having some top level categories (interface, rules, goals, entities, entity manipulation), yet the ontology is meant to be developed organically, as new games are analyzed or created. Furthermore, individual ontology entries are hierarchically organized as either being subtypes or as compound (Zagal, Fernández-Vara, and Mateas 2008, 177), as shown in Figure 2. The Game Ontology project acknowledges that history is about introducing new elements, but it still makes an atemporal ontology to describe it, meaning that the project easily ends up with large static list of descriptors, which will eventually become unmanageable.

5) As stated, my focus here is on design patterns. Game design patterns were originally developed by Staffan Björk / Jussi Holopainen as a description of game elements shared between games. (Björk and Holopainen 2004). I am using design patterns in a broad sense, focusing not so much on the typical formulation of patterns as being designs with particular consequences, but more on design patterns as the identification of game elements. Design patterns share a problem of scope with the Game Ontology Project: what is our criteria for identifying game elements, and when can we claim that our list is exhaustive?

The Pay once & Play problem and the coastline paradox

Figure 3: Pay once & Play Apple App Store category

The challenge of open-ended lists for the description of games is easily illustrated through the example of the Pay Once & Play category (Figure 3), first introduced in the Apple App Store in early 2015. Though video games were for a long time, at least from 1985 to 2015, mostly sold in boxes for upfront payment, this business model was not actually named as “Pay once & Play” until after the emergence of the free-to-play or freemium business model. This demonstrates that although the Ontology Project or Design Patterns aim to identify the components that games consist of, it turns out to be impossible to be exhaustive.

The second, and related, problem is the coastline paradox. Using design patterns, Canossa, Bjork, Nelson provide an interesting in-depth analyses of the recent reboot of the X-COM franchise (Canossa, Björk, and Nelson 2014). As it turns out, even the comparison of two games in the same genre (and of the same name) yields the introduction of dozens of new design patterns. In fractal theory, the coastline paradox refers to the fact that a coastline will be longer the more detailed our measurement is, as a detailed measurement will capture more features of the coastline than a high-level measurement will. As we can see, this is also the case for design patterns: the closer we look, the more design patterns we will find.

These two examples illustrate that we need to reconsider what the Game Ontology Project and the Design Patterns are accomplishing: they are in practice not about the identification of the foundational components of games; they are rather driven by the need for differentiating between games. As the Pay once & Play example shows, it is difficult to identify patterns that all games share (and hence difficult to be exhaustive), but easier to identify patterns not shared.

Design patterns and genre

With that, let us consider how genre theory – often applied to literary or film history – can explain how patterns can change historically. Recent work on design patterns and genre (Lessard 2014) proposes that a game genre can be understood as a collection of patterns, and as allowing for a number of patterns to be replaced without breaking the genre. Borrowing from computer science, Lessard gives the example of how the adventure game is an architecture (which corresponds to a genre), with the input type being a design (which corresponds to a sub-genre), and a particular puzzle being an implementation (which is title-specific). This is illustrated in Figure 1.












Adventure games


Sub-Genre specific


Parser input vs. point-and-click input


Title specific


The cat-hair


puzzle in

Gabriel Knight


Table 1: Lessard's model of genres and design patterns

The advantage of this approach to genre and patterns is that we describe design patterns not just as a list with some internal relations, but as a hierarchy. This gives structure to design patterns and place them in relation to genre. However, there is a problem in Lessard’s account of genres and design patterns: if we concede the possibility that new (named) genres may appear, it becomes clear that we cannot know ahead of time where the distinction between genre and subgenre will be located.

Hence Lessard’s model would have a problem accounting for change because we cannot decide for all eternity whether something is a genre, sub-genre or implementation, as this is subject to historical change. Any implementation can become the basis of a genre. For example, match-3 video games were originally considered a subset – an implementation - of matching tile games, but can by now be seen as a separate genre in their own right. And before that, matching tile games were probably seen as a subgenre of puzzle games.

It follows that genre / pattern / history descriptions have to be scale-free (Barabasi and Albert 1999), never assuming that any particular combination of patterns forms a genre, or that any particular hierarchy is staple or has a top level, because we cannot know ahead of time which pattern combinations will become stable, and then later elaborated with sub-genres. There is no baseline for determining what is the top- or bottom of a pattern hierarchy.

Let us therefore examine how this applies to a particular subset of video game history.

Design patterns and video game history

Family tree 19v2b

Figure 4: A visualization of the history of matching tile games (large version)

In previous work, I examined the history of a particular small game area – that of matching tile games (similar to Tetris or Bejeweled) (Juul 2007), using an informal concept of innovation to describe change. Figure 4 shows how design influences and innovations could be traced over a period of 20 years. The analysis of this area of video game history was used as a starting point for revealing both larger historical trends such as the tension between games with time pressure and games without, the status of such “smaller” games in video game culture, and the way developers in interviews exhibited an “anxiety of influence” in denying inspiration from earlier games. It also showed that while it could be tempting to discuss video game history in biological terms[3], video game history is not like a branching tree of life, but rather a history of splitting and merging of minor game types, and of ideas that are forgotten, and sometimes reused or reinvented decades later.

What are design patterns?

I am proposing that we can expand on this history using design patterns. Yet, what are design patterns? Björk and Holopainen define game design patterns as “semiformal interdependent descriptions of commonly recurring parts of the design of a game that concerns gameplay” (2004, 34). This doesn’t quite provide us with guidance: are they objective facts about a video game, or are they subjective observations? The relation of game design patterns to software engineering patterns suggests that game design patterns have an objective quality, but the fact that we have no fixed heuristic for identifying patterns suggests that patterns are subjective. But there is a third option: design patterns may be similar to genres (though on a more granular level), being used by both designers, players, and reviewers to describe games. Genres are commonly assumed to have both a “design” component and a discursive component, i.e. for something to be a genre, it has to be commonly discussed under that name. As literary theorist Todorov states:

[W]e can attest to the historical existence of the genre "tragedy" in France during the seventeenth century thanks to the discourse on tragedy (which begins with the existence of the word tragedy itself); but that is not to say that the tragedies themselves do not have common features and that it would not be possible to give an other than historical description of them. (Todorov 1976, 162)

In fact, regular language already operates with notions of design patterns. A collection of card games will often include a glossary which introduces a number of high-level patterns that can be then be used to describe individual games as consisting of these patterns. Video game reviews will refer to existing design pattern-like concepts (“capture the flag”, “turn-based”). The analysis of games using design patterns is then a combination of design patterns used in common parlance and among game developers, and analytically created design patterns created for the purpose of comparing games.

Yet, what aspects of video games are captured by game design patterns? The previously quoted definition asserts that design patterns focus on gameplay, but some of the canonical design patterns described by Björk and Holopainen actually concern player psychology (identification) or narrative (narrative structures). This suggests that actual identification of design patterns is less based on the specific exploration of gameplay or game rules, and more based on the identification of salient features of particular games. In order to examine video game history, we can therefore use design patterns in a broader sense than is common, expanding them include, for example, game rules, visual and auditory design, and fiction (Juul 2005, chap. 4). Furthermore, in order to be able to discuss individual games, we should be able to make at least provisional description of design pattern based on a single sample[4]. Still, wouldn’t this make the identification of design patterns even more unmanageable?

Retaining a focus: Perspectivalism and Adequatism

One solution for managing the amount of design patterns needed is to be inspired by work in ontological-building in computer science. Arp, Smith and Spear argue for the principle of perspectivalism, meaning that:

Ontology developers should not seek to represent all portions and features of reality in a single ontology, but should seek, rather a modular approach, in which each module is maintained as far as possible by experts in the corresponding scientific discipline,” (Arp, Smith, and Spear 2015, 45)

In effect, this means that we are not required to add every possible design pattern imaginable, but rather be clear about what our perspective is, and only add patterns accordingly. A visual designer may identify different patterns than a computer scientist or a social scientist will. This is not a problem – we can simply let several parallel sets of design pattern annotations (modules) co-exist. In addition, the authors argue for adequatism, that “entities in any given domain should be taken seriously on their own terms” (2015, 46) rather than attempted to be reduced to any underlying constituents, such as would come from a game definition. This means that we are not obliged to identify any bedrock of fixed ontology, but we can rather take specific patterns seriously on their own level, without trying to reduce them to any underlying “atoms” or components.

In practical terms, the main perspective for examining video game history must be to introduce new patterns only so far as they are needed to compare and distinguish between games. Design patterns must therefore be introduced on an as-needed basis, and it is possible to write several different design pattern histories, each of which will have its own emphasis. Yet, it would still be possible to gather these in a shared system for modeling video game history.

Historical design patterns: a case study


Figure 5: Jewel Quest (iWin 2004) introduced background tiles that must be cleared


Figure 6: 7 Wonders (Hot Lava Games 2006) introduced triangular "cornerstones" to move through a level

Take the case of matching tile games, again. Following the popularity of Bejeweled (PopCap Games 2001), the match-3 game, of swapping tiles to make matches of three or more similar tiles, developed quickly over a short period of time. Typically, new games emulated earlier games, but added particular design changes. Where the original Bejeweled set the per-level goal of clearing a specific quota of tiles, Figure 5 shows Jewel Quest (iWin 2004), which introduced the level goal of clearing all the background tiles in order to complete a level. Seen as design patterns, we could name the Bejeweled goal counter, while the Jewel Quest goal could be match background. Similarly, Figure 6 shows 7 Wonders (Hot Lava Games 2006), which introduced cornerstones (cornerstones design pattern) that the player needed to move through the playfield in order to complete a level.

Candy crush score points Candy crush clear jelly Candy crush playfield jelly

Figure 7: Candy Crush (King 2012) point goal and clear jelly goal

Candy crush ingredients Candy crush ingredients playfield

Figure 8: Candy Crush collect ingredients [a cherry] goal

While matching tile games for a period of time dominated the downloadable casual game charts, they were later supplanted by time management and hidden object games. However, around 2012, matching tile games became prominent on the mobile game charts. Candy Crush (King 2012) is a particularly interesting example, where each level has its own success criteria, such as counter goals and background-clearing goals (Figure 7) and the “cornerstone” goal of moving ingredients through the playfield (Figure 8).

Success criteria

Figure 9: The historical evolution of matching tile game goal design patterns

We can therefore illustrate the evolution of success criteria in matching tile games as in Figure 9, where the design pattern of the counter goal is supplanted by the match background goal of Jewel Quest and the cornerstones goal of 7 Wonders. Candy Crush then uses a multiple criteria design pattern that allows for different success criteria per level, but incorporating success criteria from earlier games (counters, cornerstones, match background).

Matching tile games as an evolving ontology

Figure 10: A history of matching tile games (top) with historically developing ontology (bottom) (large version)

Based on these principles, Figure 10 shows a study of how to represent the history of matching tile games as the development of a genre-specific ontology whose components change over historical time. The top part of the illustration shows an updated history of matching tile games, and the bottom part shows which design patterns where introduced at particular times, and are in use at particular times.

The study was developed by following the history tree shown in Figure 4 and graphically representing the appearance of individual new design patterns historically. In addition, categorically related design patterns (such as success criteria, discussed above) are marked as related using arrows. This has the interesting effect that the historically developing list of design patterns (bottom part of Figure 10) by itself becomes a history of game design, and a window onto which design ideas were in use at a given time.

Here, I have kept a focus on matching tiles through the lens of gameplay, but we can easily see how this could be augmented with additional perspectives, such as business models (pay once & play, advertising, and free to play) and visual style. The point is that once we accept that video game history will always challenge our preconceived notions of what we are studying, it becomes perfectly possible to include new facets of the games we are examining.

Modeling all of history

By redefining design patterns as historical, and by introducing them only as needed to make distinctions, design patterns can thus be used to chart video game history in a new way.

I propose 5 principles for using design patterns to model video game history:

  1. Design patterns are derived from particular perspectives. My focus here has been on the rules; game mechanics, of matching tile games, but nothing prevents us from expanding design patterns to include fiction, visual style, audio, controls, business models, or anything else. Indeed, video game history is in itself about changing perspectives: casual games introduced a focus on player and time investment (Juul 2009); independent games involve a new focus on developers (Juul 2014) (Ruffino 2013).
  2. Design patterns are adequate (only identify the patterns needed for comparing games, and without trying to describe an underlying general ontology).
  3. Design patterns are an evolving ontology, historical in that certain game ideas only appear at a particular point in time. At a given point in time, and in particular genres, game types and distribution channels, particular sets of design patterns are available, some of these being associated with the history of games, and some seen as contemporary.
  4. Design patterns are grouped in the sense that particular patterns fall into a particular category, such as success criteria. Counter, match background and cornerstones are paradigmatic in that they can replace each other. The category is not obvious before multiple patterns exist within the category: as long as there only was the counter success criteria, there was no obvious need for a success criteria category. We only need to introduce categories to the extent that they are needed to gather paradigmatic design patterns that can replace each other.
  5. Design patterns are hierarchical in the sense that particular design patterns are elaborations of earlier patterns, in the way that multiple counters is an elaboration of counter. (This is described as modulates in the original definition of design patterns.).

In turn this makes the use of design patterns to annotate history possible because there is no requirement of creating a general ontology, or type of description, that is supposed to capture all aspects of game history. As an added bonus, the design patterns will themselves become a history of game design.

As a practical matter, we can imagine this as a web-based timeline of video game releases, where different researchers can annotate their own video game histories depending on their perspective. A particular game may then be annotated both as part of an overall video game history, as a particular genre history, as part of a casual game history, as an independent game history, as a history of the discourses around video games, or indeed as part of any other imaginable history.

Video game history continually asks us to reconsider what it is we are studying, when we study video games. What I have proposed here is to take this challenge seriously, and to examine video game history as an object of study that continues to change. For this, design patterns can be a tool for understanding video game history, collecting video game history in a shared system, while being open to the multiple perspectives, and acknowledging that video game history is by itself about evolving new categories, distinctions, and about forcing us to acknowledge facets of games that we had previously overlooked.


Thanks to José Zagal, Jonathan Lessard, Staffan Björk, Nanna Debois Buhl, and the DiGRA reviewers for constructive comments and criticism.


Aarseth, Espen, Solveig Marie Smedstad, and Lise Sunnanå. 2003. “A Multidimensional Typology of Games.” In Level Up Conference Proceedings. Utrecht.

Anthropy, Anna. 2012. dys4ia. (Flash).

Arp, Robert, Barry Smith, and Andrew D. Spear. 2015. Building Ontologies with Basic Formal Ontology. Cambridge, MA: MIT Press.

Barabasi, Albert-Laszlo, and Reka Albert. 1999. “Emergence of Scaling in Random Networks.” Science 286 (5439): 509–12. doi:10.1126/science.286.5439.509.

Barry, Andrew, Georgina Born, and Gisa Weszkalnys. 2008. “Logics of Interdisciplinarity.” Economy and Society 37 (1): 20–49. doi:10.1080/03085140701760841.

Björk, Staffan, and Jussi Holopainen. 2004. Patterns in Game Design. 1sted. Rockland, MA: Charles River Media.

Bojin, Nis. 2010. “Ludemes and the Linguistic Turn.” In Proceedings of the International Academic Conference on the Future of Game Design and Technology, 25–32. Futureplay ’10. New York, NY, USA: ACM. doi:10.1145/1920778.1920783.

Canossa, Alessandro, Staffan Björk, and Mark Nelson. 2014. “X-COM: UFO Defense vs. XCOM: Enemy Unknown.” Proceedings of the 9th International Conference on the Foundations of Digital Games.

Cardellicchio, Cosimo. 2014. “Evolution for Games.” Board Game Studies Journal, no. 8: 1–2.

Dahlskog, Steve, J. Togelius, and S. Björk. 2015. “Patterns, Dungeons and Generators.” In Proceedings of the 10th Conference on the Foundations of Digital Games.

Gualeni, Stefano. 2015. “Enlarge Your Mesoscopy: A Philosophical Reflection on the Human Scale and Projectual Ontologies.” euSLSA conference, Furjana, Malta.

Hot Lava Games. 2006. 7 Wonders of the Ancient World. (Windows).

Hullett, Kenneth, and Jim Whitehead. 2010. “Design Patterns in FPS Levels.” In Proceedings of the Fifth International Conference on the Foundations of Digital Games, 78–85. FDG ’10. New York, NY, USA: ACM. doi:10.1145/1822348.1822359.

Hunicke, R., M. LeBlanc, and R. Zubek. 2004. “MDA: A Formal Approach to Game Design and Game Research.” Proceedings of the AAAI Workshop on Challenges in Game AI, 04–04.

iWin. 2004. Jewel Quest. Gamehouse (Windows).

Juul, Jesper. 2003. “The Game, the Player, the World: Looking for a Heart of Gameness.” In Level Up: Digital Games Research Conference Proceedings, 30–45. Utrecht: Utrecht University.

———. 2005. Half-Real: Video Games between Real Rules and Fictional Worlds. Cambridge, MA: MIT Press.

———. 2007. “Swap Adjacent Gems to Make Sets of Three: A History of Matching Tile Games.” Artifact 1 (4): 205–16.

———. 2009. A Casual Revolution: Reinventing Video Games and Their Players. Cambridge, MA: MIT Press.

———. 2014. “High-Tech Low-Tech Authenticity: The Creation of Independent Style at the Independent Games Festival.” In Proceedings of the Foundations of Digital Games Conference.

Key, Ed, and David Kanaga. 2013. Proteus. (Windows).

King. 2012. Candy Crush. King (iOS).

Kirkpatrick, Graeme. 2015. The Formation of Gaming Culture. Basingstoke, UK: Palgrave Pivot.

Lessard, Jonathan. 2014. “Game Genres and High-Level Design Pattern Formations.” In Proceedings of the 2014 Foundations of Digital Games Conference. Florida.

Magnus, Olsson Carl, Björk Staffan, and Dahlskog Steve. 2014. “The Conceptual Relationship Model: Understanding Patterns and Mechanics in Game Design.” In Proceedings of DiGRA 2014.

Parlett, David. 2015. “What’s a Ludeme? And Who Really Invented It?”

PopCap Games. 2001. Bejeweled. (Windows).

Ruffino, Paolo. 2013. “Narratives of Independent Production in Video Game  Culture.” Loading - The Journal of the Canadian Games Studies Association 7 (11).

Sicart, Miguel. 2008. “Defining Game Mechanics.” Game Studies 8 (2).

Smith, Gillian, Ryan Anderson, Brian Kopleck, Zach Lindblad, Lauren Scott, Adam Wardell, Jim Whitehead, and Michael Mateas. 2011. “Situating Quests: Design Patterns for Quest and Level Design in Role-Playing Games.” In Interactive Storytelling, edited by Mei Si, David Thue, Elisabeth André, James C. Lester, Joshua Tanenbaum, and Veronica Zammitto, 326–29. Lecture Notes in Computer Science 7069. Springer Berlin Heidelberg.

Todorov, Tzvetan. 1976. “The Origin of Genres.” Translated by Richard M. Berrong. New Literary History 8 (1): 159–70. doi:10.2307/468619.

Zagal, José P., Clara Fernández-Vara, and Michael Mateas. 2008. “Rounds, Levels, and Waves The Early Evolution of Gameplay Segmentation.” Games and Culture 3 (2): 175–198.

Zagal, José P., Michael Mateas, Clara Fernández-Vara, Brian Hochhalter, and Nolan Lichti. 2005. “Towards an Ontological Language for Game Analysis.” In Proceedings of DiGRA 2005 Conference: Changing Views – Worlds in Play. Vancouver, Canada.


[1] For a discussion on the relation between these concepts, see (Magnus, Staffan, and Steve 2014).

[2] I leave it to the reader to find examples in other fields.

[3] See, for example, (Cardellicchio 2014).

[4] The original definition of design patterns emphasizes that a pattern is ”shared between games”. We can compare this to Parlett’s phonology-inspired discussion of ludemes as being “non-contrastive”, meaning that similar aspects of game design will be grouped under the same ludeme. Parlett argues that the concept of suit hierarchy in card games is a ludeme, while a particular suit hierarchy is not, as “The thinking involved would not be altered.” (2015).