If Esports Are the Sports of Video Games, This is the Parkour

How video game speedruns challenge the notion of play and video games

Games Done Quick Credit: Games Done Quick

When you think of video games on a historical scale, you might think of them as a relatively modern invention – the same way one might think of humanity on the Earth’s timescale – a blip at the end of millennia of media. The first video game recognized as such was Tennis for Two in 1958, and every video game humanity has known was created since.

But games aren’t an isolated medium, as they use influences from a huge combination of pre-existing arts and sciences. In Indie Game: The Movie, game developer Phil Fish excitedly states that games are ‘the sum total of every expressive medium of all time, made interactive’. There’s something bigger in that statement than the idea that games are a cool mix of all sorts of things: thousands of years of knowledge of color, architecture, composition, movement, animation, narrative, and sound come together to make the games we play on computers that are the result of thousands of years of math, physics, and chemistry.

One medium that is an intrinsic part of games is likely older than all of those, and it’s the one that is easiest to overlook: play. Play is most likely the oldest form of expression humans used. It is generally thought that when our species was taking their first steps, our curiosity tought us what was safe and warned others as to what wasn’t, and our playfulness allowed us to develop and practice nonexistent skills. Evolutionary, the humans with a healthy dose of playfulness had a distinct advantage: climbing a tree for no reason could mean you were more likely to evade an ferocious predator when you needed to, and throwing stones at a tree just to see if you could hit it meant you were more likely to yield food from whatever primitive weapons you had for hunting when the time came.

Over the centuries, play has taken many forms, both structured and unstructured. It has evolved into structured expressions of play such as sports, board games, and video games, but also continues to exist in that unstructured expression that you can so clearly see in children, rules and context springing into and out of existence just to facilitate fantasy.

Former Independent Games Festival chairman Brandon Boyer in 2011 referred to playfulness when he touted Little Big Planet as ‘an undistilled, unbridled expression of childhood naivité and creativity, it’s that thing we hear when a pre-schooler talks to us about the glorious dreamscape they spend every waking moment in, the thing we realize we also had, and then lost decades back and we’ll never regain, and that’s a tragedy’. His sentence has stuck with me ever since, as our contemporary society leaves little room for unstructured playfulness in adults, and only a little more for structured play. The structured play that is generally accepted are sports, card-games, and board-games. As a species we’re still struggling to think of video games as valuable structured play.

Ludos – the technical term for play – might be incredibly old as a form, it is still rather poorly understood from a theoretical level. The science of play deals with the same condescending responses that players of games experience – both from within academic circles and from within the games community at large. Suffice to say, I believe there’s tremendous value in understanding the medium of games beyond the practical.

One of the founders of modern cultural history, Johan Huizinga’s seminal 1944 book ‘Homo Ludens’ describes play in one short paragraph:

“Summing up the formal characteristic of play, we might call it a free activity standing quite consciously outside 'ordinary' life as being 'not serious' but at the same time absorbing the player intensely and utterly. It is an activity connected with no material interest, and no profit can be gained by it. It proceeds within its own proper boundaries of time and space according to fixed rules and in an orderly manner. It promotes the formation of social groupings that tend to surround themselves with secrecy and to stress the difference from the common world by disguise or other means.”

You’d think that with such academic consideration of play in the early days of games as a medium, the cycle between creators and academics would be entirely smoothed out by now. The reality is that that cycle remains somewhat muddy in games, but the general idea is that creators create, and academics discuss. The work of academics then feeds into new games being created, as the academics create understanding and terminology that helps us understand specific opportunities, challenges, and obstacles in game development. Sometimes, those developments happen smoothly, and sometimes, they’re surrounded by endless controversy and conflict.

An example that’ll make every game developer and academic groan, and as such is probably the least controversial example to discuss, is ludonarrative dissonance. The term was introduced by Clint Hocking, at the time a game designer at Ubisoft, and denotes a conflict between the play and the narrative context”. Examples of ludonarrative dissonance include a muscular supersoldier that leaps from a giant tank in a cutscene not being able to climb a knee-high obstacle during gameplay, a protagonist struggling with the emotional repercussions of killing a non-descript character after having spent the past hours gracefully murdering hundreds of unnamed hostiles. One of my personal favorites is the trope in games that allow you to breathe away bullet wounds in cover, where the player character has to limp to an emotional moment because they got stabbed at the end of a boss-fight. For a brief moment in 2012 and 2013, the term was everywhere as we tried to find a way to resolve that dissonance, and after that, as a medium, we concluded that ludonarrative dissonance isn’t that much of a problem as long as the internal consistency of the game world wasn’t compromised, and the players’ suspension of disbelief wasn’t affected.

You could argue that therefore, ludonarrative dissonance was a useless discussion and a waste of time, but I would posit that on the contrary, knowing that this effect it is not a problem is valuable. Having words to discuss it means we can take conscious choices about how to apply it, or how to use it to subvert players’ expectations. It also means that we can now recognize it faster, and in more diverse contexts.

One of the most amusing and extreme public examples of “ludonarrative dissonance” happens to be a twice-yearly charity drive called Games Done Quick. For the days-long live broadcast, they ask “speedrunners” to complete video games as fast as they, often breaking the structures, narratives, and level designs of the games being played in the process in spectacular and unexpected ways. If eSports are the sports of the games industry - highly structured plays with rigid rules and rewards, speedrunning is the freestyle parkour of the medium - often free of reward, aimed at self-improvement, and aimed at breaking rules. Speedrunning is obviously far less glamorous, and a bit more streetwise - participants practice games for months – or even years – to learn the skills and strategies to beat the game just half a second faster than the world records set by others. Most popular games and many less popular games have thriving communities of devoted runners, each a tournament of their own, that have learned how to run the game in as a little time as possible. Games Done Quick brings out the who’s who of the speedrunning community, and is equal parts charity drive, celebration, and community function.

In some games, a successful run means being able to expertly execute commands and inputs, being able to click through dialogue on exactly the right moment, making every jump perfectly, and landing each shot without fail. Sometimes, a successful run requires expert memorization of a specific route through the game, getting each power-up in the correct sequence, and tackling enemies before they even appear on the screen. In other games, a successful run is built upon glitches, knowing how to manipulate the game’s memory management to skip entire sections of the game, exploiting bugs to gain infinite speed, or forcing your way through small mistakes in the levels’ architecture so that one can take a shortcut, or simply walk around all the obstacles and dangers. In most games, a run’s success depends on a combination of all the above.

To allow meaningful competition in games, different categories were created to compete in. Most games have both a 100% and an Any% category that denotes completion requirements – in the 100% category players have to achieve and collect everything possible in a game, while win Any%, they could technically run straight for the final boss. Categories can include glitchless runs, which often get introduced when glitches take a severity that otherwise precludes meaningful competition or shows of skill, as some glitches can allow players to skip straight to the credits sequence.

For Games Done Quick, often there will be additional challenges or categories. Some speedrunners show off their skills by playing their game of choice blindfolded, some play the game on custom controllers or dance-mats, and some players will create contraptions that allow them to control multiple games, and the clock runs until all games are beat. There are even Tool-Assisted Speedruns in which computers are programmed to help beat the game as fast as possible, or to even fully control the game, in a battle of code versus code.

You might think of speed running as competitive, but the scene is both incredibly competitive and collaborative. Runners might compete for the best time for those few minutes or hours the race is on, but in general they’re working together to find new ways to speed up their community times. New strategies are theorized, discussed, tested, and spread through the community as they are found. When a new game is released, the fastest time often rapidly decreases as the community for that game finds new strategies, routes, skips, and glitches – sometimes an improvement of seconds, while at other times the strategy can save minutes, or even hours. Every now and then, a new strategy will force the creation of a new game-specific category – the recent Nintendo release of The Legend of Zelda: Breath of the Wild quickly led to the creation of a No Amiibo-category, a category barring the use of a plastic toy that Nintendo sells that -when held against the game console- gives the player access to a free horse to traverse the plains of the game that much faster.

Some skips and strategies take years to develop, spending months in theoretical discussion before being verified by a runner. In some cases, new ways to save significant amounts of time are discovered decades after a game has been released, and the game suddenly opens up entirely new ways of saving even more time. In some cases, the world record can have been moving by seconds for years, only to suddenly go back to a rapid decrease of minutes for weeks as new ways of applying the new trick are found.

Watching a speedrun, independent of your relation to gaming, is fascinating. Runners during Games Done Quick often keep the utmost level of concentration on their face as they flawlessly manipulate their character to pixel-precise locations, for a frame-perfect sequence of inputs, just to stave off six seconds off of their two-hour long world record attempt. As the timer ticks up the seconds, fellow expert-runners of the game will usually sit on the couch behind the runner, explaining the tricks and strats that the player is manipulating as they happen, and discussing the risk-assessments in favor or against certain risky strategies the runner might be making during their attempts. The couch often offers a play-by-play commentary of the actions and context of the run, emphasizing subtle displays of extreme skill, and discussing how the runner will try to get as good a time as possible. Occasionally, the runner will ask for silence, because they have to focus, often to execute a highly complex and timing-dependent sequence of inputs with a 0.03 second precision for each press. At the end of the run, they will state ‘time’ – the magical word that indicates completion of the speedrun, and the moment the clock on the screen stops counting.

As a game developer myself, speedrunning is a fascinating phenomenon. Games are created with a purpose, and an intent, and with the explicit purpose of letting a player immerse themselves into the game. It is our goal to create a piece of media that allows the player to suspend the disbelief, and engage in the structured rules of play we have created to experience a narrative arc or various levels of challenge. Specific regions of the game are painstakingly created with certain colour palettes and auditory cues to evoke specific emotions, skirmishes and obstacles are set up to offer a pleasant ramp in difficulty and challenge, and mechanics are tweaked to be as dependable and accurate as possible.

Speedrunners, instead, reject all of those efforts, emotional draws, and narrative. To them, the videogame is not the game, and they are not playing. To a speedrunner, the videogame is a machine, while the attempt to beat it as fast as possible is the actual game. At no point do they allow themselves to be convinced the game is anything more than a program, a machine that responds in predictable and exploitable ways. They are manipulating more than playing.

As a community, speedrunners spend hundreds of hours figuring out how to speed up their runs, and in doing so try to deconstruct the inner workings of the game. Speedrunners therefore often discover programming tricks developers have used in their efforts, and try to find ways to abuse those tricks. One such example is ‘RNG Abuse’. Random Number Generation, refers to a game containing random elements that unfold differently any time. Most games feature randomness of some sort, to determine the way the artificial intelligence behaves, what kind of enemies you encounter, or what sort of items you acquire after finishing a battle.

Randomness, as such, can have a make-or-break effect for a good speedrun. The games community at large has many (more or less) affectionate names for ‘hoping for good random’, but for speedrunners, any control over this randomness can be critical to getting the game beaten faster.

One thing you learn as a programmer is that computers can’t actually generate random numbers. The best we can do is to find a value in the computers’ memory that we can assume is different every time, and use that as a seed. That number is then run through a series of mathematical formulas called functions, and the result is what we call a pseudo-random number. Most games then use the resulting number as the new seed when required to generate a new random value, and since those varying resulting values are used in all sorts of different places in the game, users can generally not find or see a pattern in what is effectively a pre-determined sequence. That leaves one challenge to the programmers: to find a starting value that we can trust to be different seed every time when we first need a random number.

In the popular adventure game series Pokémon, the programmers used the game device’s clock to initialize the random seed, as they could be certain that the clock would be a different value every time the game would get booted up. Pokémon speedrunners eventually realized this, and created an entire science around “RNG Abuse”, finding certain times and dates you could set your devices’ clock to prior to starting the game. Making sure you set the time to a certain moment, and then booting up the game at exactly the right moment would ‘seed’ the game in a specific way, effectively controlling the randomness in the game. This way, players can manipulate the randomness of the game to receive better Pokémon creatures, and to ensure encounters occur in the most advantageous way for their runs.

As a programmer, there is something lovely in seeing your work deconstructed like a puzzle you never intended. When programmers write the tens or hundreds of thousands of lines of code that make up a game, they’re thinking about structures, techniques, and architecture that balance the need between keeping the code readable for humans to work on, and performing well. When the game gets finalized for distribution, computer programs optimize that human-readable code into something only readable for computers to get that tiny bit of extra performance. At that point, the game becomes a black box that no one can look into, and speed runners pry into that black box with their curiosity, theories, and skills of deduction. Sometimes, they uncover enormous parts of the intent and thinking of the programmers, which – in a way, means they get to know the people making the game more personally than normal players do.

The same goes for the artists that create the three-dimensional worlds of many modern games. In most games, the world you see is nothing but a big façade, a word I choose intentionally as a huge majority of art assets in a game are constructed in such a way that only details that could be visible are created. This means that most buildings in a game are, in a way, simply empty cubes that look good from the outside. When you’d get inside that cube, the cube would be entirely invisible, as computers don’t have any information to process. That means that looking from the outside-in, you’d see a building, but looking from the inside-out, it’d appears as if no building exists. Players are obviously not meant to ever find a way into those cubes.

The structures on the outer areas of a level are generally not even cubes, but resemble the famous fake house façades at Leinster Garden in that beyond the façade, there’s actually nothing. If a player were to somehow get behind those, they would fall forever, unless the developer implemented a fail-safe to place the player back into a ‘safe location’, or an invisible ‘floor’ that the player could continue to stand on. As such, developers make every effort to make sure there’s no way to get behind those outer edges.

That also means those locations are a prime target for speedrunners. They use every mode of movement available to them to find ways to ‘clip’ through those art assets and get on the other side of them. In certain games, coming at a specific spot of a specific wall in a specific way will cause the code to allow you to wiggle your way through the wall. In certain games, being able to string together a certain complex sequence of jumps and precise movements allows you to jump over those walls by climbing your character higher than the programmers expected.

If and when you make it behind those walls, given there is a way to keep you from falling forever, you can effectively walk straight past any obstacles and enemies the game placed in the intended side of the façade. In trying to find ways to move through a level faster, speed runners both explore and reveal the pragmatic and artistic work of the many artists that work on these games.

For the game designers, the people who establish the rules of a game, and tweak the numbers to feel as good as possible, speedruns that exploit a minimal amount of glitches or bugs might be the most interesting. These runners simply use the settings and rules that are available in the game to the most extreme extent. I didn’t finish Super Mario Bros, the 1985 platformer that turn Mario into a world-famous character, until I was in my early twenties. As a kid, I had spent hundreds of hours playing the game, finding some of the later levels impossible to clear. Almost two decades after I first played the game, I found a copy of the game that I managed to clear, realizing that the game is only about two hours long.

In October of last year, speedrunner “Darbin” broke the previous Super Mario Bros world record of 4 minute, 56 seconds, and 87 milliseconds by 0.3 seconds after almost 28,000 attempts. Super Mario Bros’ has been ‘solved’ to such a point that the fastest time is usually stuck for months, until someone figures out a way to be just a few frames faster. Almost the entire run consists simply out of running and jumping with extreme accuracy, and given the decades the community has had to find the optimal route, there’s almost no room for improvement left.

As a game creator, watching speedrunners treat a game with respect, as a dance partner, or a beloved vehicle, there is a unparalleled beauty in a game designed to be one game, being played as another game. When speedrunners treat a game as something to be subjugated and defeated, it feels downright ugly to me. Some speedrunners will refer to glitches found after a decade of intentionally trying to break the game as ‘lazy work by the developers’, as if anything could withstand years of continuous and intentional challenge. The best speed runners understand that their game of speed running lives in symbiosis with the game they’re playing.

Whatever tools or rules they use, speedrunning adheres to the highest ideals of play – it is voluntary, intrinsically motivated, and enjoyable – but it has very little to do with video games per sé. It can be taken seriously inside the ‘magic circle’ in which the speedrunning community exists, but is completely frivolous and non-consequential outside of it. It has more in common with video games than it is a subset of video games, because in a way, the speedrunners themselves have become game creators - designers of their own games.

Rami Ismail is the Business & Development Guy at Vlambeer, a Dutch independent game studio known best for Nuclear Throne, Ridiculous Fishing, LUFTRAUSERS, Super Crate Box, GUN GODZ and Serious Sam: The Random Encounter.