The Speedrunning museum of accidents

Special Issue, August 2018

PDF Version


Rainforest Scully-Blaker
University of California, Irvine


This paper discusses speedrunning, the practice of completing a game as quickly as possible without cheats or cheat devices and how it relates to Paul Virilio’s Museum of Accidents. Speedruns are shown to be an example of what the author calls ‘curatorial play’, or play that is intentionally preserved and organized with a view to being shared. Speedrunning in particular is shown to re-curate a game by exhibiting glitches (accidents) that allow the game to be viewed in a different light by player and spectator alike. This notion of player as curator is then used as a lens to examine various developer responses to speedruns that the author has observed.

Keywords: Speedrun, Curation, Glitch, Preservation, Game studies


Though there are exceptions, it can generally be said that progressing quickly in games is viewed positively. Whether we cite obvious examples such as racing games or Sega’s rhetoric of Blast Processing and Sonic the Hedgehog or not, the fact remains that players who can ‘beat’ a game faster are generally held as more proficient than slower players. Indeed, many games incentivize speedy playthroughs with achievements and unlockable rewards and more still have a built-in timer to help keep track of one’s playtime as a potential substitute for the ‘High Score’ boards in arcade games. But even these games that demand speed and efficiency from players make assumptions about just how fast one can go.

In an effort to complete games as quickly as possible without the use of cheats or cheat devices, speedrunners are continually striving to find new ways to shave precious seconds off of world-record times. To do so, members of the community often develop a more comprehensive knowledge of the worlds found in games than the very people who built these virtual spaces. Games that are designed to guarantee particular experiences are shown to be fallible as apparent paths and boundaries are proven to be mere suggestions of one way to reach the end of a story. Playing games in this way has accrued a large following and the speedrunning community, which began as a small group of players, has come to occupy a significant place within game culture.

And while elsewhere[1] I have discussed speedrunning as both practice and community and how such gameplay relates to game rules, in this paper I am now interested in suggesting that speedrunning is an example of what Paul Virilio calls a museum of accidents, in this case a re-curation of play experiences through their explicit rules. Once explained, this assertion will then be used to account for the various ways that developers, as the original curators in this context, react to their games being played in this way.


Review of literature

This work germinated out of another paper of mine entitled “A Practiced Practice: Speedrunning Through Space With De Certeau and Virilio” (Scully-Blaker, 2014). In it, I frame speedrunning within the works of Michel De Certeau and Paul Virilio and propose a taxonomy of rules in games that aids in encapsulating the speedrun practice – implicit versus explicit rules. I concluded that while a speedrun can be said to trespass against what I call a game’s implicit rules (rules that players generally assume exist, either out of a sense of immersion in a game world or from a sense of video game convention), they do not in any way conflict with a game’s explicit rules (the rules as they actually apply, i.e the code).

Both here and in my earlier paper, I deploy the concept of space in the way that Michel de Certeau discusses it in The Practice of Everyday Life. For De Certeau, space is “a practiced place”. Whereas place refers to “an instantaneous configuration of positions”, space “exists when one takes into consideration vectors of direction, velocities, and time variables”. Put simply, “the street geometrically defined by urban planning is transformed into a space by walkers” (de Certeau, 1980/1984: 117). In a virtual context, if I borrow the phrasing of De Certeau’s street – walker relation, one may say that ‘the game environment coded by programmers is transformed into a gamespace by players’.

As for how these spaces are designed and navigated, De Certeau discusses the concepts of strategy and tactic. He calls a strategy “the calculus of force-relationships which becomes possible when a subject of will and power (a proprietor, an enterprise, a city, a scientific institution) can be isolated from an environment”. It is, in other words, the planning-out of a distinct space by those who have dominion over it – “a victory of space over time” (De Certeau, xix). This begins to sound familiar if we recall Squire and Jenkins’ assertion that “Game worlds are totally constructed environments” (Squire and Jenkins, 2002). Indeed, Game spaces are intelligently designed, strategic spaces that make operational assumptions about how players could or indeed should interact with them, but these assumptions are not always correct.

A tactic, on the other hand, is “a calculus which cannot count on a ‘proper’ (a spatial or institutional localization). Unlike a strategy, “because it does not have a place, a tactic depends on time – it is always on the watch for opportunities that must be seized”. Many activities are tactical in nature for De Certeau, “clever tricks” – in essence, an individual’s intelligent ways of making-do in an existing system much larger than themselves (De Certeau, 1980/1984: xix). This, too, is relatable to virtual worlds. As Katie Salen writes, “creators of emergent systems […] can never fully anticipate how the rules will play out […] the force of play is so powerful that it can change the rule structure itself (Salen, 2002).

In few gameplay practices, I would argue, can one find a more apt example of this than speedrunning. Indeed, runners are able to exploit holes in a game’s programming (strategy) to create drastically shorter paths (tactics) through a game. This is achieved not only through knowledge of the game space, but also through an insistence on moving through the virtual world as quickly as possible. To better understand this drive for acceleration, let us now turn to the concept of speed itself.

My understanding of speed stems from the work of Paul Virilio. “’To conquer is to advance,’” Virilio quotes from Frederick II in The Aesthetics of Disappearance (Virilio, 1980: 97). Those ‘advancements’ that happen more rapidly can be said to dominate in their respective fields, whether one speaks of military campaigns or of technological innovation. For Virilio, this stems from the same drive for far-off lands felt before the world was completely mapped. But now that most spaces on Earth have been explored and mastered, “a desire for the discovery of speed than any far-off elsewhere” becomes the dimension of movement whose boundaries humans have not met (1980: 111).

An essay of his, “The Museum of Accidents” is of note within his body of work in that the piece makes reference to virtual speed, whether it is through simulation technologies or through the mediation of video footage from violent disasters like the Challenger explosion or the Hindenburg disaster. In the essay, Virilio derides the contemporary museum of science and technology for being what he perceives as “a lyrical illusion of progress whose purpose is to continuously mask all that is negative in the name of science” when in fact, he continues, it is folly to suggest that technology “would progress by dissimulation, or censorship of its own errors and false calculations…” (Virilio, 1986: 82). Censoring the potential for and the occurrence of accidents is not only detrimental to researchers, however. Virilio suggests that the call for transparency “should apply not only to professionals, those responsible for the programs and other decision-making executives, but also to the amateurs and naive spectators of recent technological achievements.” (1986: 83)

While the piece is written with reference to the public representation of major pillars of modern society like ‘science’ and ‘technology’ and his ‘accidents’ are large-scale disasters like the Challenger or the Hindenberg explosions, I believe that there is room to read speedrunning into this concept. Indeed, when Virilio writes that “To innovate the vessel was already to innovate the shipwreck, to invent the steam engine, the locomotive, was again to invent the derailment, the rail catastrophe”, I would argue that to innovate the video game was already to innovate the desire for virtual speed, the glitch, and indeed the speedrun (1986: 81).

Newman takes up the player’s drive to glitch in Best Before (2012). He notes that some practices central to the speedrunning museum of accidents such as “the exploitation of glitches […] have become written into the standard lexicon of gameplay” while even specialized language to describe game-specific exploits is “coded” for “more sophisticated players” and “specific types of play” (Newman, 2012: 150). The community hubs for records and videos are specifically cited as examples of preserving the unique gameplay experiences that speedrunning produces. Though I agree with Newman in this regard, the book is of particular use as a jumping-off point for my own argument. For indeed, if play communities can be said to preserve their unique brands of play, then the gap between speedruns and curation is narrowed considerably since preservation is arguably a side effect of curation, even if not all acts of preservation are inherently ‘curated’.


Speedrunning as museum, glitch as accident

When attempting to read speedrunning into Virilio’s writing on the museum of accidents, the greatest obstacle that I perceive is a matter of scope and scale. For example, to mention the infinite sword glitch[2] and the Challenger Explosion in the same sentence risks aggrandizing one and trivializing the other without any context. However, I argue that the text speaks to exhibiting ‘the accident’ more broadly than the context of tragic, real-life disasters. Despite one taking place in virtual space and the other taking place in natural (‘real’) space, the infinite sword glitch and the Challenger disaster are both essentially ‘accidents’.

The only major discrepancy between the museum of accidents in Virilio’s context as opposed to my own is that, within Virilio’s frame of reference, the project emerges out of a sense of moral obligation. Whereas Virilio writes that “pontifical infallibility does not exist when it comes to major catastrophies (sic)” and that such situations “require the utmost vigilence (sic) against routine…”, it is difficult to suggest that one ought to have similar feelings about discovering and exhibiting glitches (Virilio, 1986: 83). Speedrunners have a variety of reasons for playing through games as quickly as possible, but the proceeding analysis does not wish to suggest any ethical imperative behind what speedrunners do. For the moment, I argue that the speedrunning museum of accidents represents a more pointed discussion of what some[3] have already gestured to in game scholarship, namely speedrunning’s place as a community-based form of gameplay preservation and media archiving. With that said, let us proceed with an analysis of the speedrunning museum of accidents, beginning with the concept of ‘accident’ itself.

“To innovate the vessel was already to innovate the shipwreck”, Virilio states in the opening to his essay. He continues, “the beginning of wisdom would be, above all, an awareness of the symmetry between substance and accident, instead of constantly dissimulating them.” (1986: 81). In other words, every object, by virtue of existing, threatens itself and those around it with the risk of the accident, and the video game is no exception. As with any emerging medium, there have been plenty of claims made and studies done based on whether games are harmful to players. However, to avoid treading off our present train of thought and into the realm of moral panic, this paper is only concerned with the threats that a game poses to itself – glitches, both slight and major. If video games are the substance in Virilio’s equation, then I put forth that, so far as the speedrunning museum of accidents is concerned, glitches are the accident. The linkages between the text and speedruns do not end at substance and accident, however.

Just as Virilio argues that his new museology would overturn the “philosophical and scientific positivism” that obscures “all that is negative in the name of science”, so too does speedrunning overturn notions of conventional, immersive play (1986: 82). Still, while Virilio is actively interested in changing how museology is conducted through the development of a “platform for what never exposes itself, yet nevertheless exposes us incessantly to major risks”, speedrunners are, of course, less directly interested in impacting game design. This is not to say that speedrunning has never altered game design, however. As I will discuss below, developers have responded to speedruns in many ways, sometimes making design decisions explicitly based on the community. Both Virilio’s museum and the speedrunning museum of accidents bring about change through a subversion of the norm, both with reference to how science (game design) is done and with reference to how it is curated (played).

Speedrunning as a project is effectively a museum of accidents that consists of curating the fastest playthroughs of games (substance) as well as all the glitches and exploits (accidents) that make these playthroughs possible. By reading speedrunning in this way, it becomes clear as to why one of the few ways that the practice has been discussed by other game scholars is with reference to its potential as one way of presenting an archival game history. Indeed, when Virilio writes that “what is needed is a new scenography where only what is exploding or decomposing is exhibited” we can see that speedrunning falls in line with the text once again (84, italics my own). The history of science and technology is not properly told by the charred remains of the Challenger[4], nor is the history of video games entirely summarized in romanticized stories of the Video Game Crash of 1983, complete with the infamous tomb for copies of Atari’s E.T. (Atari, 1982).[5] These are relics of an old museology, historically relevant, perhaps, but doing less to put the accident on full display and more to help emblematize ‘pivotal’ moments in their respective histories.

Instead, the speedrunning museum of accidents exhibits glitches as they occur in real time. Whether I am referring to a livestream of run attempts, a YouTube video of a completed run or even a glitch tutorial, speedrunners revel in understanding and showcasing how games are broken. This, in turn, creates unique ‘histories’ for particular games as glitches are discovered and applied and world record times are traded back and forth. Collected as a cohesive project, speedrun videos and forums become histories of applied theorycrafting in this sense. It is how these histories are shared that shapes the speedrunning museum of accidents.

Throughout his essay, Virilio uses the phrase “exhibiting the accident” to refer to the main role of the museum (Virilio, 1986: 84). And given his characterization of ‘exhibiting the accident’, that it consists of “exposing what is improbable, what is unusual and yet inevitable”, the speedrunning community’s use of Twitch, YouTube, and forums to share information immediately comes to mind. And when Virilio writes, “the ‘museum of accidents’ already exists, I’ve seen it: it is the TV”, it is simple enough to make the leap to speedruns being broadcast on these websites. In so doing, however, there arises a key difference between natural and virtual accidents that must be addressed.

“’To expose or to be exposed, that is the question’”, Virilio writes, “to be or not to be conscious, scientifically speaking, of risks, of what befalls without consent: accident, the hidden face of all natural or man-made substance” (1986: 83). Ultimately, the answer offered by “The Museum of Accidents” is that, courtesy of television, it is impossible for the advancements science and technology to be curated without their corresponding failures and immoral for any attempts of the sort to be made. As such, we must move museology away from an ideological focus on progress by exhibiting the accident. For games, however, this is not exactly the case.

Players are deeply fascinated with glitches. Indeed, most older videogames could never be curated without mention of these programming oversights. From as early as 1995, the users of websites like GameFAQs have written and shared in-depth walkthroughs of how to complete various games. And, as Newman has noted, “Game Guides often list [glitches] and the conditions under which they can be called into being with as much rigour and detail” as they expand on describing item locations, boss strategies, and so on (Newman, 2008: 114). However, contemporary developers have the ability to ‘patch’ these mistakes, effectively removing them from the game. The erasure of the accident is much more possible in virtual spaces than it is in the natural world.

And yet, within the speedrunning community, the drive to exhibit the accident continues to exist despite the fact that games, unlike live, televised broadcasts, can be edited or ‘repaired’. By either keeping track of speedrun records from various versions of a game’s software, or by deliberately running an obsolete build of a particular game, the gameplay contained by the speedrunning museum of accidents is one of the few instances in game culture where one can still find certain glitches and exploits that have otherwise been erased from existence. This would be reason enough to argue for the importance of speedrunning for game scholars, however the linkage between speedrunning and the museum of accidents can be pushed further.

Speedrunning is a broadcasting of an alternative to conventional, immersive play. The practice is also heavily curated by a tight-knit network of expert communities, making the likening of speedrunning to a ‘museum’ seem natural. And now, with the matters of scope and scale addressed and certain textual parallels established, it has been made clear that speedrunning is a project, a museum of accidents, a re-curating of a game through play according to its explicit rules. What is meant by the term curatorial play and how speedrunning fits this definition must now be discussed.


Exhibiting the accident through curatorial play

I am by no means the first to point out that certain players take joy in exhibiting the accident, although others may not use the same terminology. This assertion lies very much within the tradition laid out by other game scholarship[6], including James Newman’s characterization of glitches as “another attempt to seek alternative gaming pleasures” as well as “further evidence of the way videogames are played with” (Newman, 2005: 63). Within non-competitive gaming environments, glitches are often “collected and curated with the same care and precision […] as the fragments of knowledge pertaining to strategy or the locations of items in the gameworld”, and speedrunners are, if anything, more inclined to share glitches with each other than most players, especially if they have the potential to shorten the necessary time to complete a game (2005: 114). Still, this is not enough to suggest that play itself is potentially curatorial. The pursuit of glitches is a respected part of what ‘knowing’ a game entails, but the pursuit of this knowledge – what I have seen referred to as ‘procedural trivia’ elsewhere[7] – does not inherently come with the overarching organizational vision that one traditionally associates with curation. Indeed, as I will suggest, to re-curate a game through playing it requires intentionality and certain specific contextual elements. But first, a note on curation itself.

What it means to curate, or occupy the role of curator is a subject that museologists have debated at length and presented numerous stances on, many of which extend beyond the basic premise of displaying artifacts in a space. For example, the Raqs Media Collective uses a form of poetic acrobatics to qualify curation with terms from A to Z. Under C is collision – “we come face to face with the ‘curatorial’ whenever we witness within ourselves or around us the collision of artistic forms”. These collisions include but are not limited to “unforeseen accidents” and “the accumulation of readings against the grain of intention” (Raqs Media Collective, 18). Under J is jailbreaking, in the sense of “the liberation of a device from the straightjacket of its prescribed mode of operation” – “one can coax a work out of its accustomed frame, provoke a situation into yielding results other than what its authors, actors and agents intended” (21-22). The curatorial practice, then, is one that has the potential to defy authorial intention and become a reframing, particularly with reference to mediums that make use of the moving image like film or games.

Whether an artist can curate their own work or whether a curator can organize an exhibition in such a way as to become a new author of a work or series of works are both questions raised in Cultures of the Curatorial (von Bismarck et al., 2012). This blurring of roles acts as a helpful backdrop to our discussion in that it allows for an artist to curate how individuals ‘collide’ with their work while simultaneously suggesting that a curator can directly colour how a work is encountered. An example of the artist turned curator can be found in Cihat Arinç’s notion of “the curatorial mode of filmmaking” (Arinç, 2010: 183). He argues that some filmmakers[8] “assign priority to the set design, or the narrative organization of cinematic space, through visible, audible and intelligible objects” (2010: 184). By carefully staging what is shown to a viewer frame by frame, these filmmakers curate the cinematic experience to evoke particular emotional responses.

Although, in games, it is not simple to dictate how a player behaves, the overall ‘narrative organization of the virtual space’ is still a matter of concern for game developers. Even ‘open-world’ games that tout unprecedented player freedom still have to create boundaries in which those choices may be made. These virtual boundaries constitute a form of curation, a velvet rope guiding exactly how players might experience the virtual worlds that are on display. However, there are entire ranges of player experiences that can occur within even one specific game world. Some of these, speedrunning included, are a jailbreaking – a step further along in the artist-curator authorship debate whereby the visitors to these curated virtual spaces offer alternate interpretations of the games and ultimately re-curate them through their play.

It is important to note that I am not suggesting that all styles of play involve some form of curation, nor indeed would I suggest all instances of curatorial play involve re-curating a particular play experience. A casual playthrough of a game that follows the nodes of the plot in an appropriate order is akin to experiencing any other curated experience in an intended way. Such ‘immersive play’ experiences only have the potential to become curatorial if the player shares the account of their play experience, perhaps online in the form of a YouTube video or Twitch livestream. The act of sharing one’s gameplay to the digital commons suggests intent on the part of the player to offer their interpretation of how to navigate the game – a Let’s Play. By splitting gameplay into segments and (potentially) offering commentary as the game unfolds, players who take part in this practice are simultaneously preserving and exhibiting – in other words ‘curating’ – a play experience, but one that does not enter into conflict with the original artist’s intentions.

It is only once the act of play opposes authorial intention that I would suggest a ‘re-curation’ might occur. When speedrunners collide against a game’s implicit rules, thereby revealing the scaffolding of the explicit rules behind them, they suggest an alternate way that the game can be experienced. Like immersive play that goes unshared, this sort of play is not inherently a form of curation, it is simply deviant play. However, when coupled with the collaborative project of the speedrunning museum of accidents, whereby these alternate pathways through a game are not only shared but carefully regulated and classified, speedrun play can truly be said to re-curate the range of play experiences afforded by a game.

To reiterate, then, curatorial play is that play which is intentionally preserved and organized with a view to being shared. If such play also operates according to a separate set of standards that stands opposed to the traditional mode of interaction with a game, then it can be said to re-curate a gameplay experience. While there are certainly other examples of players or play communities that potentially re-curate games in particular ways (machinima and modding come to mind), few have drawn as much of a range of responses from the games’ original curators as speedrunning.


The role(s) of the developer as the original curator

I have observed that developer responses to speedrunning can generally be placed into three categories, which I call (in order from most to least common) prevention, compromise, and endorsement. The first and most common developer response to speedrunning is one of prevention. This refers to any instance of developers attempting to fix glitches that are used in speedruns. The clearest example of this can be found in contemporary console and PC games, particularly those released by large, ‘triple-A’ companies. When new games first come out and runners begin to find game-breaking glitches, it is not uncommon for an update to the software (a patch) to be released that fixes these exploits, forcing runners to make route changes to adjust to this new version of the game.

To expand upon this in relation to speedruns, I will be using the example of Nintendo’s The Legend of Zelda: The Wind Waker (Nintendo, 2002) and its HD remake (2013) for the Wii U. For although the original Wind Waker was released before the onset of online play (and with it the downloadable update), the relatively recent overhaul of the title was put in direct conflict with the speedrunning community’s style of play when it was discovered that a major glitch had not made the jump to the HD version.

The Wind Waker alters the Zelda formula by trading Hyrule’s fields for an imposing ocean. Instead of travelling by foot or by horseback, Link sails from island to island. From a speedrunning perspective, these boating interludes, which could take up to 5 minutes, did not seem to be the fastest way to navigate the game world. Indeed, when the game was first released, even more conventional players complained that the sailing sections felt too long. Soon a glitch was discovered that drastically shortened the trips from one plot point to the next, even cutting out entire segments of the game altogether. Known as ‘super swimming’, the glitch caused Link to gain infinite speed while treading water in the Great Sea. If enough speed was built up and players could control their avatar properly, then Link would be sent rocketing across the sea in a matter of seconds.

Whether Nintendo was initially aware of this glitch or not, when the announcement of Wind Waker HD came, it was clear that they were aware of broader complaints from their playerbase that sailing was potentially tedious (Good, 2013). Along with a graphical update to the game, the HD port offered new items including the Swift Sail, which allowed players to navigate the game world much more efficiently, although still not nearly as quickly as super swimming. At this point in the narrative, details become somewhat speculative but at least this much is clear: prior to Wind Waker HD’s release, Nintendo toured the game at multiple Best Buy locations, at least one of which was visited by members of the speedrunning community who were curious about whether the major glitches were still present in this new version of the game. From one account of the event, a runner successfully executed the super swim glitch and actually crashed the game demo as a result (Gulyas, 2013).

It is not clear whether a Nintendo representative at the event took note of the glitch, whether the multiple online accounts[9] of the event drew too much attention, or indeed whether the development team simply had not had time to patch out super swimming for the demo build of the game, but whatever the reason, when Wind Waker HD was released the glitch that c super swimming had been patched out. As a result of this, the route through Wind Waker HD differed considerably from the original speedrun and many in the community were unhappy with the decision to remove what was perceived as a glitch so obscure that it could only matter to speedrunners from the game.

Nintendo’s internal reasoning for removing the glitch has never been made public, although some motivations can be guessed. Super swimming represents a diversion of the virtual space put together by the Wind Waker development team. Sailing and harnessing the power of the wind to control one’s navigation of the Great Sea are arguably the core mechanics that were introduced to distinguish the game from other Zelda titles. Link even has a fatigue meter associated with swimming so that players could not undertake the sequence breaking task of swimming from island to island without a boat. In a Wind Waker speedrun, all of these measures to ensure a particular narrative experience are thwarted and the accident of the super swim is used to re-curate the game in a different light. As the speedrunning museum of accidents is p through videos and livestreams, even non-runners are exposed to the programming oversights that exist below the apparent polish of a finished, triple-A title.

For Nintendo, this likely raises at least one of two concerns:

In the first place, although runners tend to maintain a sense of respect for game makers despite finding errors in their works, the existence and dissemination of an accident like the super swimming glitch arguably carries with it an implication of a job poorly done on the part of Nintendo. It recalls something that Tanja Sihvonen speaks of in relation to developers and modding communities, namely that “the game industry fervently aims at staying in control of its business interests” by curtailing how games are played and tinkered with (Sihvonen, 2011: 75). Just as game companies seek to control how (if at all) their games can be modified, so too do they sometimes seek to put restraints on how play itself is undertaken. The parallel is particularly apt if one considers both speedrunning and modding as curatorial practices. Indeed, if speedrunners can be said to re-curate a game through its play experience, then I would argue that mods re-curate a game through the code itself.

Second, although perhaps inseparably, the transformation of the Wind Waker’s narrative into one that involves Link racing across the ocean while flailing wildly, can potentially be said to do violence to or ‘ruin’ the game – not, on the level of the narrative as guaranteed by the code, but rather on the level of the implicit rules, or the immersion players create for themselves. The prospect of one’s game being exposed to Virilio’s violence of speed in a forum as large as the speedrunning museum of accidents does not necessarily appeal to game makers, triple-A or otherwise. Whatever the reasons, the example of the Wind Waker HD points to some potential tensions that underlie the developer-runner dynamic. However, not all game makers are so quick to curtail deviant play of this nature.

Compromise is a more recent development in the relation between runners and developers that seems to occur exclusively in the case of smaller, independent (‘indie’) game companies. It refers to a middle-ground approach to glitches and the speedrunning community whereby some exploits are left untouched. Because the tricks executed by runners often require a very specific series of inputs or only function in particular locations, some developers decide that the glitch is so unlikely to be encountered by regular players that it may as well remain in the code for what it gives to the runners that engage with their title. Other examples of compromise include games that feature ‘speedrun modes’ that incorporate a reliable in-game timer, further incentivizing people to try and play through as fast as possible. To better illustrate this, I will look at one brief example – A Hat in Time (Gears for Breakfast, 2017).

A Hat in Time is a 3D adventure-platformer game that consists largely of jumping and running across various environments collecting items to advance the story. While the game was being Kickstarted in 2013, an early build was given to a well-known member of the speedrunning community, Narcissa Wright to play on her stream. Although giving an early build of one’s game to popular Twitch personalities is a common practice for advertising one’s game, the fact that Narcissa was a speedrunner provided some specific benefits.

Since she would be interacting with the game in non-intuitive ways, the likelihood of Narcissa finding mistakes in the game’s coding to allow for sequence breaks was high. Beyond being free advertising for the development team during a crucial period in their game’s lifecycle, the decision to offer the game to a runner also gave Gears for Breakfast access to someone who was effectively a seasoned play tester. Gears For Breakfast’s policy from this period into their beta release has been that any bugs discovered by speedrunners or otherwise will be fixed. In return, however, they “will not fix an exploit after release, if it has speedrunning value and it is unlikely to be reducing the experience of a regular player” (A Hat in Time Blog).

From this, one can see several issues that exist even when developers and runners seek to collaborate in some way. There is a clear sense of there being ‘good’ and ‘bad’ glitches, that varies from title to title. In general, ease of execution, coupled with a major sequence break means that even indie developers will want to patch out the bug. As Gears for Breakfast explains in a blog post about the Hat in Time beta:

If you find some wicked way of getting into the sky or under the ground and it requires a crazy combination of button presses, then we have no reason to remove that alternative experience from the game. But if there is a chance it might happen on accident to a regular player, then we will have no choice but to patch it. I hope this is understandable. (A Hat in Time Blog)

As speedrunning grows in popularity, some developers are now slowly beginning to think of them as ‘alternate experiences’ as I have been considering them, rather than insults or implications of a job poorly done.

A Hat in Time an ‘indie’ game – it is made by a group of people that are not directly affiliated with a triple-A studio. Their budget and their audience are smaller than a game like Wind Waker HD. The development team used this to their advantage and was able to directly engage with speedrunners in a way that was mutually beneficial: speedrunners found a new titles to engage with and enjoy without fear of their practice being curtailed by patch notes while the developer found an audience to play and share their game. Apparently, speedrunning has come to occupy a significant enough place in game culture that independent game developers wish to cater to them on some level. However, when a developer’s efforts to harness the speedrunning demographic are too pronounced, the result is not nearly as positive for either game player or game maker.

Endorsement refers to a relatively uncommon developer response to speedrunning whereby individuals strive to design a game with the community explicitly in mind. As is the case with compromise, endorsement only occurs at a very grassroots level, with individuals usually posting on the speedrun subreddit or other community forums to pitch their game idea and ask for input or testers. This type of developer response is quite uncommon for two reasons – on the one hand, designing a game with such a niche demographic in mind means that games in this category seldom gain much of an audience and, on the other, in almost all cases these games do not actually generate interest from speedrunners and so do not get made. As a result, there are not any particularly poignant examples of endorsement to present here, nor does it seem likely that there may ever be one.

For indeed, it would be impossible to design a game with intended glitches because glitches, by definition, are never intentional. A developer may intentionally leave a glitch unpatched as was shown in the examples of developer compromise, but to code in a deliberate error is something that arguably cannot be done. If speedrunning, by its nature, is a re-curation of a developer-intended play experience, then any attempt on the part of a developer to anticipate and design the re-curation of a game that never existed would be unable to escape from being an original play experience that could then inevitably be altered by runners.

In other words, even if, one day, someone made a game that deliberately has glitches left in for speedrunners to use, runners would still undoubtedly play it in a manner that differs from the developer-intended path. Until someone codes a gamification of the meta-game of routing a speedrun, however, this claim cannot be verified. Instead, I must be satisfied with the statement that efforts to design a game specifically to be speedrun have, to this point been poorly executed and poorly received. I do not believe that these attempts at game design are damaging to the community or the practice, however. The very notion that games are being designed with speedrunners in mind at all is quite the positive thought with which to conclude the arguments of this paper.[10]


Conclusions – enclosure and the velvet rope

We have now seen that speedrunning is a museum of accidents, a re-curation of games through how they are played. The notions of re-curation and curatorial play have been explored with reference to museology and game studies and it has ultimately been shown that, while not all play is curatorial (or re-curatorial), the specific character of the speedrunning museum of accidents as a collaborative project to exhibit the accident distinguishes it from other forms of play by fitting into this category. This, in turn, has been used to account for the various ways that game developers have responded to their games being speedrun. Some have found the re-curations through speedrun play to be too divergent from their vision of the game’s narrative and have therefore patched out glitches while others have taken a more measured approach to whether or not a glitch ought to be removed. Others still have been so inspired by the play exhibited in speedruns that they have sought to design games specifically with the community and play practice in mind, although these efforts have thusfar been failures due to the inability to program ‘intentional accidents’. All the same, the range of responses that I have discussed are firm indicators not only of some of the tensions between game makers and game players, but also the unprecedented significance speedrunning has come to hold for games and game culture.

In Players Unleashed! Sihvonen discusses, among other things, how, in an attempt to incentivize the release of mods for The Sims (Maxis, 2000) while still maintaining some creative control over what gets released, the developers created “The Sims Carnival […] a modding toolkit that is extended to promote creativity” (Sihvonen, 2011: 74). Maxis realized that a community of players that interact with and re-curate their software is a demographic worth befriending and rather than antagonizing. And while there is certainly a discussion to be had about enclosure in relation to such practices (for indeed, the mods that are released are now entirely on Maxis’ terms), for our purposes it can be said that it is arguably better for both parties if the developers of The Sims stimulate the modding community rather than issuing them cease and desist orders.

In the same vein, as I have been researching speedrunning, even triple-A companies have begun to acknowledge speedruns and the speedrunning community. As a throwback to the early days of competitive gaming contests and a celebration of the original event’s 25th anniversary, The Nintendo World Championship was certainly a milestone event for Nintendo and their entire playerbase. Still, few, I imagine, felt the momentousness of the occasion more than speedrunners. For indeed, six of the eight invited players were high-profile members of the community and several of the rounds of competition involved completing parts of various Nintendo titles as quickly as possible. Whether this is an indication of Nintendo warming to the idea of their games being speedrun or whether it was simply a move to maximize viewership for their event remains unclear and only in time will we know whether the community will come to view such public recognition from major developers as a boon or a hindrance. For the present, however, my research on the subject is concluded.



ARINÇ, C. (2013), « Curating Ghostly Objects: Counter-Memories in Cinematic Space », in J.P. Martinon (ed.) The Curatorial: A Philosophy of Curating, London, Bloomsbury Academic, 183-95.

BAINBRIDGE, W. A. and Bainbridge, W.S. (2007), “Creative Uses of Software Errors: Glitches and Cheats”, Social Science Computer Review vol. 25, no. 1. 2007.

BISMARCK, B. V., Schafaff, J., and Weski, T. (2012), Cultures of the Curatorial. Berlin, Sternberg.

CONSALVO, M. (2007), Cheating: Gaining Advantage in Videogames, Cambridge, MIT Press.

DE CERTEAU, M. (1980/1984), The Practice of Everyday Life, Berkeley: U of California Press.

GOOD, O. (2013), « Here’s Why Sailing Took So Damn Long in Legend of Zelda: Wind Waker », Kotaku. <>

GULYAS, D. (2013), « The Wind Waker HD Still Has Super Swimming. » Oprainfall. <>

LEMIEUX, P. (2016) “Losing Time: Microtemporal Histories of Speedrunning”, SCMS 2016, March 30 – April 3, Atlanta, United States of America.

NOKE0 (2013), “Superswimming is here to stay”, June 12, GameFAQs. <>

NEWMAN, J. (2005), “Playing (with) videogames”, Convergence vol. 11, no. 1.

NEWMAN, J. (2008), Playing with Videogames, London: Routledge.

NEWMAN, J. (2012), Best Before: Videogames, Supersession and Obsolescence, Milton Park, Abingdon, Oxon: Routledge.

RAQS MEDIA COLLECTIVE (2013), “On the Curatorial, From the Trapeze”, in J.P. Martnon (ed) The Curatorial: A Philosophy of Curating, London: Bloomsbury Academic, 2013, 17-23.

RUSSEL, C. (2014), “Speedrunning as Media Archaeology”, 1er symposium de la série Histoire du jeu, June 27-28, Montreal, Canada.

SALEN, K. (2002) “Quake! DOOM! Sims! Transoforming Play: Family Albums and Monster Movies” Walker Art Center.

SCULLY-BLAKER, R. (2014) “A Practiced Practice: Speedrunning Through Space With de Certeau and Virilio” Game Studies Vol 14, No 1, <>

SCULLY-BLAKER, R, BEGY, J, CONSALVO, M. & GANZON, S. C. (2017), “Playing along and playing for: Livestreaming from tandem play to performance” HICSS-50, January 4-7, Waikoloa, Hawaii.

SIHVONEN, T. (2011), Players Unleashed!: Modding the Sims and the Culture of Gaming. Amsterdam: Amsterdam UP.

SQUIRE, K. and JENKINS, H. (2002), “The Art of Contested Spaces”, <>

ŠVELCH, J. (2014), “Comedy of contingency: making physical humor in video game spaces”, International Journal of Communication Vol. 8, p. 2530-2552.

WATSON, N. (2016) “Procedural Trivia: How Players Decode Minecraft”. CGSA 2016, June 1-3, Calgary, Canada.

VIRILIO, P. (1986 [1977]), “The Museum of Accidents”, Art Press No 102.

VIRILIO, P. (2009 [1980]), The Aesthetics of Disappearance, Los Angeles, CA: Semiotext.



[1] Scully-Blaker 2014, Scully-Blaker et al. 2017

[2] An important glitch in both The Legend of Zelda: Ocarina of Time (Nintendo 1998) and Majora’s Mask (2000) that tricks the game into thinking that Link’s sword is constantly swinging, bestowing other odd properties to Link that lead to other glitches.

[3] I am here thinking of Russel (2014) and Lemieux (2016), for example.

[4] Some of which are, in fact, on display in museums.

[5] The example of Atari’s E.T. may strike one as strange given that it is not explicitly tied to glitches, however, as the emblem of the Crash of 1983, an event treated by most game histories with as much gravity as the stock market crash, it is by far the best example of a romanticized relic of the ‘dark days’ of video games. The example of old video game magazines, another common game history artifact that sometimes showed pictures of glitches, could serve just as well here.

[6] Beyond Newman (2005), (2008), other academic treatments of glitches include Consalvo (2007), Bainbridge and Bainbridge (2007), and Švelch (2014), among many others.

[7] Watson (2016)

[8] Particularly those working in the genre of memory-film.

[9] Beyond Gulyas’ article, there were numerous forum posts on the subject including Noke0 (2013).

[10] It may be objected that games such as Super Meat Boy (Team Meat 2010), which include skippable cutscenes, the ability to quickly reset or retry, and a reliable in-game timer might contradict this. However, game does not become a ‘good speedrun’ simply through having these gameplay mechanics, nor are most developers who make such games likely to have made them ‘exclusively with speedrunning in mind’.