Speedrun Science

James Parkinson
James Parkinson
Speedrun Science

For over 30 years, people have been in pursuit of the perfect run in Super Mario Bros. That is, completing the game in the fastest time possible, also known as a speedrun. Now in 2021, the current world record holder may have reached the human limit, coming within milliseconds of the fastest time achievable by a computer. By analysing what’s technically possible, speedrunners are able to find the exploits, refine their approach and practice relentlessly, to push the absolute limits. This is the science and art of the speedrun.


JAMES PARKINSON: For over 30 years, people have been in pursuit of the perfect run in Super Mario Bros. That is, completing the game in the fastest time possible, also known as a speedrun. In the 2000s, the record of 5 minutes and 8 seconds stood for years, as players struggled to beat the game any quicker. By the 2010s, speedrunners began to exploit glitches and tricks in order to cut down their times - beating the game by any means necessary became the standard. Over the following decade, players started to gradually improve, posting sub-5 minute times. It took years just to shave a second off previous records. Then in February 2018, streamer and YouTuber Kosmic set a new world record of 4 minutes, 56 seconds, and 462 milliseconds.

JAMES PARKINSON: The rankings are in a constant state of flux, as players around the world submit their best times, beating other speedrunners by fractions of milliseconds. Kosmic’s record didn’t stand for long though. He broke it himself by becoming the first person to beat Super Mario Bros. in 4 minutes and 55 seconds, a record he achieved in September of 2018.  

JAMES PARKINSON: Kosmic’s complete time was 4 minutes, 55 seconds and 913 milliseconds. And for the following two years, it was considered impossible to break the 55 second barrier. The top 10 on the leaderboard was full of records within the 4:55 range and Kosmic’s own eventual personal best came in at 4 minutes, 55 seconds and 646 milliseconds, submitted in January of 2020. But then in April of 2021, a speedrunner by the name of Niftski achieved a new world record of 4 minutes, 54 seconds and 948 milliseconds.

JAMES PARKINSON: At the time of recording, this record still stands. That could change at any moment, but the difference now is that there really is only milliseconds left to slice off this time. That’s because it’s the closest a human has come to the theoretical limit of 4:54.26 - that’s the fastest time achievable by a computer.

NERISSA HART: So there are these things that people will do called TAS, or Tool Assisted Speedruns, where they'll use a computer to essentially log every single perfect input that you would need in order to get the best time possible for a game. Now, they don't exist for every game, because they take a really long time to put together. But there are a lot of games where you can look up Tool Assisted Speedruns for more popular games, and just see how insane it would be if someone was on god level and could do every single thing in the game perfectly.

ERIC KOZIEL: The entire point of a Tool Assisted Speedrun or a TAS is that you're able to go through, piece-by-piece, inspect the game’s running state and other elements, so that you can identify what is perfect. Without the limitation of human, physical abilities, what is actually possible in this game? What is the exact, optimal way to get through this? And what starts to happen is that people find creative solutions to do the same things, just in a more human, approachable manner.

JAMES PARKINSON: By analysing what’s technically possible, speedrunners are able to find the exploits, refine their approach and practice relentlessly to push the absolute limits, blurring the lines between humans and computers. In this manner, speedrunning is a very precise science, but it’s also an artform.

JAMES PARKINSON: I’m James Parkinson. From Lawson Media, this is Gameplay, stories about video games and the virtual worlds that power culture and community.

JAMES PARKINSON: Like a lot of things within the games community, speedrunning goes back to the very early days of video games. The drive to get the high score and top the leaderboard on your favourite arcade game is central to what made many of those games fun and competitive. But firstly, how do you define a speedrun?

ERIC KOZIEL: There's a lot of different levels to that question. And actually, this is something that I tried to define in my book.

JAMES PARKINSON: This is Eric Koziel, the author of Speedrun Science.

ERIC KOZIEL: Speedrunning is trying to complete a goal within a video game in the shortest amount of human computer interactions possible. And what that really means is, there's kind of this feedback loop going on, of you the player, you're pressing buttons that cause something to happen inside the game, the game outputs it to the screen, light travels into your eyes, you see what happens, you react and you keep going. So that's, in a more technical way, that's what a speedrun is. Really though, it's just taking a fixed game, a medium that you want to try and attach a goal to, and then you're going to perform that goal as quickly as possible. And that lends itself to not only just how well you were actually able to pull it off, but what strategies you're able to come up with along the way, how well you understand what's going on in the game, and how what you're doing to effect it might come back at you. So really, it is just going through any goal within a game that has a definable start, and a finish, in the shortest amount of interactions, or in measurable format, the number of frames possible.

JAMES PARKINSON: We’ll talk about frames shortly, because that’s a key component to understanding and mastering a speedrun. When this all started though, it emerged very organically.

ERIC KOZIEL: Speedrunning is, in many ways, just a divergence of sport. People care about being able to speed run well, in the same reasoning, that they care about how fast a human can run 100 meters, there's just something about that level of performance and optimisation for a human to perform. And that that is an important distinction here, what they can do to be able to pull off these incredible feats, and then to drive it even further.

ERIC KOZIEL: There's this kind of a recipe of sorts that allowed it to really flourish. And it really relied a lot on the technology to come forward, well enough, so that different elements in this recipe could really come together. Earliest games and things during the arcade era, they had, predominantly they were focusing on score. So you're going through and you're trying to hit more platforms, kill more enemies, anything that's going to make that score counter go up. And that's primarily how people competed. Time errors, and timing in general is something that came along a little bit further on towards, say, the Atari era. So Pitfall might be an early example of this, where not only could you go through, try to collect everything to maximise your score. But the game also had a rudimentary but perfectly functional timer. So once people were able to go through and identify, “Okay, here's where I can get all these keys and pickups”, all the gold bars, I believe. It's not just a question of, “can it be done?”, it becomes a much better competition of, “Who can do it faster?”. But still, that's just kind of from the game side, something that promotes trying to go through faster. And even through the 80s, this wasn't necessarily a particularly popular design standpoint for many games, at least not for their main focus.

JAMES PARKINSON: That began to change though, heading into the 90s.

ERIC KOZIEL: [Game started to get more diverse. You have a lot more kind of experimental, quirky things that various developers were trying out. But the community and communication aspect was not there. You might have more games that are open to being time-able, whether by something in-game or just a stopwatch, but you didn't have the community, you couldn't easily share that performance that you're going through with somebody else. Or similarly, watch somebody else's and learn your own set of tricks from that. But there are a couple of key things that started to kind of plant the seed in other people's minds in the early to mid 90s. The first of which, I'll actually point to as the Nintendo World Championships.

ERIC KOZIEL: If you're not familiar, this is a competition that took place, I believe in the late 80s, early 90s, where Nintendo actually had a number of venues set up all around the US, possibly, I don't recall if there were any in an international space as well. But they had developed a special game set, that was actually a combination of three games.

JAMES PARKINSON: This was an exclusive NES cartridge that Nintendo produced just for the competition. It included Super Mario Bros., Rad Racer, and Tetris.

ERIC KOZIEL: This was a timed session. So players would come in, they would have, I believe, some six minutes and 28 seconds  - correct me on that some other time.

JAMES PARKINSON: Pretty close - it was 6 minutes and 21 seconds.

ERIC KOZIEL: But they go through, they'd start out. And after you complete the goal in Mario of collecting 100 coins, you'd immediately warp into the next game. And after you've completed, I think, two laps in that game, you move onto the third game, which is Tetris. And you use up all of your remaining time to just build up your score in Tetris. On paper, this was all a scoring event. You kind of have a multiplier associated with your points in Mario, your points in the racing game, and then finally, your points in Tetris. But it's all done by a timer. So what became much more important is to figure out, where can you score the most points the quickest. And for many of the competitors that turned out to be Tetris. There was a very high multiplier on that, so they had to not only just play all three of these games well, they had to go through and quickly determine, “alright, I need to actually just get through Mario as quickly as possible. So where's my most efficient way to get 100 coins?”, finish the racing game without any sort of hiccups or crashes, and then just go ham on Tetris. Try to get as many Tetris’ as possible, blow up your score counter, and then that would maximise your score towards the end. So that was the end strategy and it kind of became like a speedrun, even though that wasn't the intention, just because there was this hard limit timer associated with it.

JAMES PARKINSON: Also during the 90s, magazines like Nintendo Power would set challenges for players, particularly for games that had built-in timers, like Super Metroid. You’d take a photo with your old film camera as evidence and send it in by mail. The magazine would then publish their top times in the next issue, and for those that made it, it was instant glory and bragging rights among your friends, with your name in print. So already, players were thinking about games in a different way. But it would still be some time before the technology caught up. The biggest challenge being that it was hard to actually share your accomplishments and provide proof of your speedrun. Even if you had an old video camera and made a VHS tape of your playthrough, you couldn’t really share it with anyone. There is one game though that made this process easier.

ERIC KOZIEL: Doom was pivotal for speedrunning in many different ways. The main one that I'm going to discuss here, though, is that from the get go, Doom allowed players to record replays of their performances. And a replay is not a video, a replay is a recording of the inputs that the player used during their defined play session. So this would be every input and how long and various other parameters of what was going on in the stage, that basically form all of what a player did, and if played back would perfectly recreate that same performance. The big catch here is that this is a significantly smaller type of file. And Doom is also immensely popular. So this whole aspect of a replay, even in the early ages of the internet, where not everybody has a big - a lot of bandwidth to be able to share these kinds of broader bigger videos, you could pretty easily share a replay of you doing something, and then other people could easily identify it and say, “Oh wow, yeah that was a great run, you had a neat strategy there”, and, “I want to try that”. You finally had a community and an ability to share these speedruns, essentially.

JAMES PARKINSON: If you can’t share and compare your times with other speedrunners, then you’re only competing against yourself. But the community aspect really started to grow, once video capture became more accessible to consumers, and we had the beginning of streaming.

ERIC KOZIEL: Streaming has helped speed runs in a lot of ways. The main one, at least as far as I see, is that it enables people to still pursue their personal projects, their goals. In the earlier days, kind of everybody who was doing a speedrun referred to it as their “project”, as something that they had a goal in mind something that they were pursuing, and they would keep it up until they got that particular time or the set of tricks that they really wanted to see in a finished run. What streaming allows is that whole process, going through the project, to not be a solitary one. So you'd be able to have people along for the ride, they basically be part of that experience, encouraging you, offering their own insight, wherever they might have it. And really turned it from a solitary but enjoyable hobby, into one that you could actually share with others, whether they were helping you or simply just rooting for you. And that started off in the late 2000s, where streaming platforms like Twitch - well, at that time it might have even been JustinTV - came about and actually gave people a reason, some platform to actually go and put their streams out.

JAMES PARKINSON: You also had new software for accurately timing your runs, and websites like speedrun.com, which launched in 2014. It became not only a hub for the community but also the destination for online leaderboards. Fast forward to now, and speedrunning is thriving. As we heard at the top of the episode, retro games are still incredibly popular among the community, but even newer titles like Hades are very appealing for speedruns. But whatever game you choose to speedrun, you better like it because you’re going to be playing it a lot.

NERISSA HART: I would say it's thousands and thousands of hours, considering for a while, when I first started, I actually would do at least one run of the game every day, because I just wanted to get my times to a better and better place.

JAMES PARKINSON: That’s Nerissa Hart.

NERISSA HART: Or Ruby Hart on the internet. I am a 28 year old gender fluid actor, a theatre artist, a streamer, and primarily a speedrunner. So the main game I speedrun is Shadow of the Colossus, but I'm actually learning some other speedruns as well.

JAMES PARKINSON: Nerissa says they got into speedrunning by accident, just because of their pure enjoyment for Shadow of the Colossus.

NERISSA HART: So I fell in love with this game, just the way that the storytelling is done in such a unique way, the music of it is absolutely gorgeous. It's unlike anything I've ever heard. And when I got to college, I met one of my best friends in the world, Ian. And we bonded over the fact that we liked a lot of the same video games, and SOTC was one of them. So we made this yearly tradition of passing the controller back and forth and playing through the game. So we would kind of switch who was playing which Colossus. And then as time went on, it ended up turning into full on, us having two monitors and two PS3s, side-by-side, doing full races that would last several hours, a bunch of our friends would pile into a dorm room together. And it was just a fantastic time, we had a lot of fun with it.. Fast forward, I got so good at the game that my roommates who are all gamers of some sort - they basically looked at me and went “Um, Nerissa, why have you never considered speed running? You've gotten really good at this”. And I didn't think that my times would match up, I didn't think I was on a, you know, pro level because I didn't know any of the speedrun strats. But when we pulled up speedrun.com and looked at the times on the chart, I would have been ninth in the world with the time that I got that day.

JAMES PARKINSON: After the break, the skills and techniques involved in speedrunning, and what it takes to be one of the best.


JAMES PARKINSON: Speedrunning is inherently competitive, but unlike most competitions cheating is encouraged. If you’re submitting a time to official leaderboards, and you’ve used certain techniques or “strats” to improve your time, there are dedicated categories which your run will fall under. Here’s Nerissa Hart.

NERISSA HART: Yeah, so “strat” means strategy. So basically, what you typically denote as a strat is something that you don't use as a regular player. So if you're doing a run, basically using the tricks, or the puzzle solving that the designers meant for you to use, we call those intended runs. And there are actually a bunch of speedrun games that will have category extensions. And basically, that means it's not one of the regular categories that you typically submit for with speedrunning. So it's like a specialised category. But when you're using strats, so speedrun strats, that's the term that we use for things that have a lot more specificity to it, that are meant specifically to cut off that time.

JAMES PARKINSON: Strats are also very specific to each game. For example, two important strats in Super Mario Bros. are the wall clip and the flagpole glitch. The wall clip is a glitch that allows Mario to basically run through a wall, and the flagpole glitch skips the animation that lowers the flag at the end of each level. Doing this saves a little bit of time, which all adds up at the end of a run. And this is also where the importance of frames comes in. Here’s Eric.

ERIC KOZIEL: So you'll often see speedrunners talk about, “Oh, we found a new strategy that saved us five frames, twenty frames”. Anytime that they bring up a frame, they're talking about a single display, a single image output by the console or the computer, whatever you happen to be playing on. And that is kind of like the minimum measurable time frame for when a goal is started or finished. But it also happens to be the minimum unit that a game typically operates on user inputs. So many retro games, and most modern games try to shoot for something in the realm of 60 frames per second, you might also see it as 60Hz. And that just that just says that the game console or computer is producing one image in every 60th of a second or 60 images in one second. But that also means that that's kind of the minimum amount of time that a change in inputs is going to result in something happening in the game. Many older consoles only look for input once a frame. So that's kind of the player-to-computer interaction frequency as well. And because of that, there's also some quirkiness that can happen. Many times there will be a particular trick or interaction in the game, that is something that must be performed within one frame, or just a couple frames, and only have a very small window. And what that means is that the player reactions or their motions need to be within that margin of just one or two frames, however many is necessary for that particular trick. And that directly equates to the amount of time that they have to be able to perform it. So if it's a one frame trick, that means that they need to hit that button in exactly 1/60th of a second at a particular time point, in order to get that interaction to occur. So frames overall are very important from the input handling side, but are also like I said, the ultimate kind of tiebreaker between two equivalent runs. So this is something that you see compared only at the games that are run to the most optimised levels, where there's very little margin of error, or any opportunities for just the difference in a human pressing the starting and stopping of the timer to allow for any additional error. So they count the frames, as in how many video frames were produced, from a known start point to a known endpoint. And they use that as the final determination of, “Is this run faster or is that run faster?”.

JAMES PARKINSON: Putting together a good run though, is about much more than just knowing which strats to use. The best speedrunners will approach a game in a very analytical way.

ERIC KOZIEL: So I think every player has their own way of familiarising themselves with the games, and really assessing what they want to do and what they want to get out of it.

JAMES PARKINSON: For Nerissa, it starts with a casual run through the game, to get a feel for all the mechanics and how the game functions.

NERISSA HART: And then from there, I will look at what videos exist. So a lot of times, if a game has been speedrun quite a bit, by different runners, like multiple different runners, they'll create video tutorials or document guides, usually in Google Docs or something like that, that they'll put on a Discord, that walks you through their terminology. So when you're looking through their guides, if there's a word that they use, and you don't understand it, but it's really common, they'll often have a glossary of terms, so that you can understand what they're referring to. Looking at the video tutorials that walk you through the strats that they use step-by-step, because a lot of times that will break down the bigger strategies that you have to use throughout the entire game, as well as the ones that are more specific, moment to moment. So it could be small things like holding the D-pad down so that you can make sure that your character jumps in a specific direction, even if the camera isn't pointing that way, to sometimes in games, if you whistle or let out a yell, that makes your character move just a little bit faster than normal. So you constantly have to be yelling or whistling to keep your character moving at top speed.

JAMES PARKINSON: As Eric outlines in his book, this is the first step of approaching a speedrun, which he calls the investigation phase.

ERIC KOZIEL: You're just learning. You want to try to find everything that's under the hood of that game that might be relevant, whether you know it will or not, in the present state of things. So this, for the most part, is just playing the game, making observations, seeing what you can do. Anytime you notice something a little off, like spend a little bit more time, try to do things that the developer probably didn't expect you to do, just push the game kind of to the edge of its limits, and see what gives. There's a lot more technical approaches that you can go through this too, depending on your skill set, and really, how deep in the weeds you want to get. But this is the stage where you can also start to understand, “what actually is the game?”. And what I mean by this is, every game is a software program. There's a bunch of math going on to determine whether or not you came into contact with that enemy. Labelling what the contents of your inventory are, anything that you can think of, is controlled by some mathematical elements. And just because it's a piece of software, I will argue that it's never going to be free from flaws or perfect. So a lot of times, it's kind of realising that, “Yes, my motion here is controlled by a special algorithm, a formula that describes my speed as a function of position and prior states, what form of movement I'm going, what the state of the controller is”, and so on. But also things like, “This enemy has 10 hit points. Alright, that enemy has 12 hit points”. Getting into a full kind of data logging mode is critical, in my opinion, to being able to make the best decisions about how to dispatch them, get out of their way, decide whether they're worth spending time on at all. And it's really just building up a huge toolbox of knowledge that you're going to try to work with in the later stages.

NERISSA HART: And then once I'm done with that step of just looking through the tutorials, pausing and playing while I'm going through the game, and just trying to optimise them, then typically I'll try and go through it on my own to see if I can remember it, what I do well, what I do poorly. And then I'll try and pull someone in who was way better at the game than me, to let me know what I'm doing wrong. So I know that's actually one thing I'm still doing with Shadow. There are a bunch of speed runners who I really respect and I'm friends with, who I'll regularly do races with, or they'll analyse my runs and give me feedback. So when I'm new to a speedrun of a game, I try and do the same thing, of just asking someone who is much better at it than I. “What advice do you have? Where can I improve? What phase do you think I'm in? Do you think I can work on these more advanced strats now or should I stick with the basics?”.

JAMES PARKINSON: The next phase is something you’ll hear speedrunners talk about quite a lot, which is routing.

ERIC KOZIEL: Routing is the process of going through and actually developing a plan based on the knowledge that you have. And this is an iterative process, trying to go through, pick out, “alright, I know this room, I'm going to have these resources available to me, how can I actually get through this, and what's quicker than that?”. And routing is a lot more complex and interesting than a lot of people might think, just because you're turning the game at that point into a puzzle. You kind of know the rules of the game, because that's what you learned in investigation. And now you really want to go through and figure out, “What is actually going to get me an optimal outcome for the purposes of speed””. Because it's not just, “Okay, this room, I can do this, and it's the fastest. And then in this room, I'm going to do this”. If you have a limited resource pool, you have to decide, “oh well, I've only got these three bombs. And I need to spend one of them in this room. And then in this next one, if I spend two, I can get through in five seconds, if I don't spend any I can get through in 10”. That's a routing situation, you need to decide whether using up those remaining two bombs is going to be worth it to save that five seconds difference, or whether they'll save more time using them somewhere else. And extrapolated out across an entire game, this can be very challenging, especially in a game such as an RPG, where a decision you make early on about what equipment you take, or what enemies you choose to fight, might play off a huge outcome in three, four or five hours from that point, where it suddenly makes a really big deal. And being able to capture that, think about it and decide what is optimal, is a key part of actually coming up with a good run. But more than anything, routing is where you develop your route. This is your script, what you actually want to do, when you get to doing your full runs. And the full runs are part of the execution phase. This is where you go when, yep, you have all of your information, you've used it to come up with your plan of action, your script that you're going to follow, as you go through, trying to get this optimal time, because you believe that that script that you came up with, is going to be the optimal way to play through the game, according to your goals and your rules.

JAMES PARKINSON: From there, it’s a lot of practice, and ensuring your chosen route is actually going to be achievable on a consistent basis. It’s also building up your muscle memory and honing all of those skills that you’ve worked into your runs.

JAMES PARKINSON: This constant process of refining and optimising comes down to human skill and determination. But some speedrunners also owe a lot to tool assisted speedruns for revealing new tricks and setting the bar for the ultimate time to beat. Strats like the flagpole glitch in Super Mario Bros. were once considered a TAS-only ability, something that only a computer could pull off in a tool-assisted environment. But as people began to break those tricks down and find ways to incorporate them into their runs, the gap between humans and computers, at least for some games, began to close.

ERIC KOZIEL: And that erosion, as people started to understand, find ways to do things - they're not necessarily getting better hands, they're still humans, it's the understanding that changed, as well as the ability to set up something that would otherwise not be possible or would be too inconsistent, to just to try without any means of configuring it. And I've seen that in many, many different games, where something might be known for a while, but it was widely considered to be too difficult, too inconsistent or otherwise. And then the major changes and updates and the new discoveries that come along, are all about those same things, those same tricks, those same skips, but somebody had found a new way to do it, or they came up with a more consistent approach. And that's really what has helped, allow this progression to where we are today. And I expect that'll continue as well. And it's really interesting for me to watch, as people gradually become more and more computer-like.

JAMES PARKINSON: You can of course take a more casual approach to speedrunning if you like, there are no rules there. Eric says it should always remain a fun activity. But you can also go in the opposite direction and be as detailed and meticulous as you want, which is why Eric compares speedruns to a scientific exercise.

ERIC KOZIEL: You can kind of consider every speedrun to be an experiment, an experiment to show what the believed best possible way forward through this game is, for a human. And sometimes those experiments will be successful, sometimes they won't. But in any case, when they're accomplished, you look back, you assess the results, what allowed them to do so well, where might they be able to improve further, that's where the science comes in. And that's where it's particularly valuable to approach it as a scientific process, where you break it down, find out where the weaknesses are, find out where you can further push it and excel, and really make the most of it.

JAMES PARKINSON: While the speedrunning community existed prior to streaming, platforms like Twitch and YouTube have certainly made it more popular. Charity events and speedrun marathons like Games Done Quick bring together players from around the world, raising money for a good cause, and strengthening the speedrun community in the process. From what used to be a solitary activity, speedrunners can now share their runs with the world, which has brought it much closer to an artform.

JONATHAN HAY: So a lot of the way that academics have discussed it to date is kind of in terms of looking at how the speedrunners kind of do something different to what the developer intended. I think that's kind of important. But I think it's important to realise that it's maybe not so much about what the developer intended, and kind of circumventing that, as a kind of practice that exists for its own means, you know, it is a form of art itself. It's not trying to be transgressive.

JAMES PARKINSON: This is Jonathan Hay. They’re a visiting lecturer at the University of Chester, and in 2020, they published an article titled: Fully Optimised: The (Post) Human Art of Speedrunning. Jonathan argues that the streaming era has seen speedruns evolve into a kind of performance art.

JONATHAN HAY: The idea of the post-human is the idea that technologies are this massive part of our lives. So really, applying this to speedrunning, it seems that speedrunning is kind of this dramatisation of that, where people are making speedrunning part of their everyday lives. And they're constantly doing the same runs again and again, and again, interacting with these gaming technologies, and learning to get better and better at the game. [12:25] And suddenly, you've just got this almost merger of the two, where they're becoming so close to the game, after doing it so many times, that they're really kind of showing that there is so little distinction between themselves and the technologies they're using. And they're showing how, by doing this speedrunning process, this process of optimisation, that humanity and technology are kind of converging to a large degree. So you've got this really artistic practice where it's basically performance art. You know, they're doing stuff in the game, and operating in tandem with the game to such a level that they almost become one and the same really.

JAMES PARKINSON: Speedrunners push themselves to their limits, often in front of passionate fans, which Jonathan calls an “artistic pursuit of perfection”.

JONATHAN HAY: That's a Friedrich Nietzsche quote. So Nietzsche talks about art as the production of perfection. And I think for me, this kind of comes down to the sense that speedrunning is about getting this perfect, in a lot of cases world record run. You know, some players only work for personal bests - and I only work for personal bests myself, because I’m nowhere near good enough to get world records. But it's all about, for the individual player, working towards what is the most optimal run that they're able to perform. So it's producing this perfect run, where, you know, every aspect of it is as good as they consider it can be. And I feel the performance aspect of it, you know, where they’re doing all of these runs - hundreds and thousands, quite often, of failed runs, it’s all part of the performance art aspect of it.

JONATHAN HAY: Where it's not so much just about the completed runs, the most optimal ones, the ones that succeed, but also about the runs where, you know, you fail on the first jump, or you fail on the final jump, you know, Bowser spits fire in the wrong direction, etc. and you don't quite complete it. You know, it's as much about those failed runs, because it's all this performance sort of art that kind of draws the audience in.

JAMES PARKINSON: Even though he has a very methodic approach to speedrunning, Eric can’t deny its artistic merits.

ERIC KOZIEL: It is definitely an art, just because it is something that individuals can put their personal flair into. Whether it's they decided to come up with a slightly different route, or they handle certain parts of executing different stages. There's finesse and a lot of creative personality allowed in speedruns that players are able to showcase all the time.

JAMES PARKINSON: As an actor, Nerissa Hart is very familiar with the expressive nature of performance. And through the practice of speedrunning, they’ve found a kind of fulfilment and self-improvement that continues to challenge and inspire them.

NERISSA HART: It started off just as a, “Let me see if I can”, you know, “let me see if I can get a time on the chart, why not? My roommates say I can, so let's do it”. And now I've become so deeply invested in the personal growth aspect of it. Every time that you're looking at doing a speed run, you're trying to find ways to make yourself better. So weirdly, finding all of the different things I can optimise, that I can make better, the things I can practice, kind of scratches that itch in my head of knowing that I'm constantly working towards a goal to make myself better. And I love that personal improvement aspect of it. And as well as the competitiveness, there is this friendly competition within the community to not only help each other improve, but to constantly try and beat each other's time. And that's something I really love. So that all mixed with just being in love with the game, constantly motivates me.

JAMES PARKINSON: Many thanks to Eric Koziel, Nerissa Hart and Jonathan Hay. As usual, there’s so much more to speedrunning that we couldn’t fit into the episode, but we have links to references and further reading on our website, including Jonathan’s article. Just follow the link in the episode description in your podcast app. Eric’s book is called Speedrun Science: A Long Guide to Short Playthroughs, and we have a direct link to that as well.


JAMES PARKINSON: Gameplay is a production of Lawson Media. This episode was written and produced by me, James Parkinson.

JAMES PARKINSON: The Gameplay theme was composed by Breakmaster Cylinder, and our artwork is by Keegan Sanford. Additional music this episode from Epidemic Sound and Breakmaster Cylinder.

JAMES PARKINSON: You can follow the show on Twitter, Instagram and Facebook at ‘gameplaypodcast’. And you can join our community on Discord. Gameplay is an entirely independent show, so if you like what you hear, please tell your friends about the podcast and help us to spread the word. For all the links, transcripts and more, head to our website, gameplay.co. Until next time, thanks for listening.


Speedrun Science: A Long Guide to Short Playthroughs by Eric Koziel

‘Perfect’ Super Mario Bros. speedrun beat after two years

How a speedrunner broke Super Mario Bros.’ biggest barrier

World record Super Mario Bros. run of 4:54 achieved, finally breaking the 4:55 barrier

Niftski Breaks The Super Mario Bros. Speedrun Record With A Perfect Run

YouTuber Breaks Super Mario Bros. Speedrun World Record

Super Mario Bros. Theoretical Limit reduced to 4:54.26 (-1 frame)

Summoning Salt: Super Mario Bros: The Human Limit

The 33-Year Quest for the Perfect Run of 'Super Mario Bros.' Is Almost Over

Speedrun.com - SMB1 Rankings

TAS videos - SMB1

Fully Optimized: The (Post)human Art of Speedrunning

Mitch Fowler and the brain-bending world of 'speedrunning'

Automating Speedrun Routing: Overview and Vision

Great! Next, complete checkout for full access to Gameplay
Welcome back! You've successfully signed in
You've successfully subscribed to Gameplay
Success! Your account is fully activated, you now have access to all content
Success! Your billing info has been updated
Your billing was not updated