Why Tool-Assisted Speedruns Have a Right to Exist


A tool-assisted speedrun of a video game (TAS for short) means that the player edited together a video of them beating a game with the use of scripted inputs and an emulator, which is a computer program that runs a video game console such as a Super Nintendo on your computer.

Using tools – as opposed to playing the original game on the original hardware – has elicited some controversy among the gaming community. When a video surfaced in 2003 of player Morimoto beating Super Mario Bros. 3 in an incredible eleven minutes, many people assumed that he had done so on a regular console and controller. After discovering that the run was tool-assisted, the video experienced some backlash.

Since then, TAS videos have remained contentious, with many people scoffing at their existence. While most of this disdain stems from the accidental mislabeling of videos to withhold the fact that they are tool-assisted, I think a deeper cause for this indignant reaction lies in a general misunderstanding.

As humans, we pretty much have the whole “art” thing on lock. Having an excess of free time in lieu of searching for food every day probably helps. Sure, some other animals have been shown to make art, but they usually do so only for survival or because a human trained them.

Maybe this monopoly on art is why we humans are so quick to call out someone who’s “doing art wrong,” especially if a computer is involved. We are so narcissistic as a species that we assume that any sort of digital assistance in art signifies a betrayal, or an admittance of the artist’s incompetence. Our disdain is especially perplexing considering we’ve become comfortable with machines doing all our math and most of our driving for the past thirty-plus years.

Still, despite no one being bothered at how deceptively scripted a show like Top Gear is, whenever someone posts a TAS video of a game, cries immediately rally of “FAKE!” or “CHEATING!”. These commenters fail to grasp that the people who create these videos are not disillusioned. Contributors to TAS forums such as Masterjun and nitsuja never claim to be experts at playing video games; they’re just experts at the game’s code.


No, I’m not talking about coding in general either. These players never go in and edit the ROM code of the game directly, like genuine cheating devices such as the infamous GameShark and the earlier Game Genie do. Instead, TAS runners meticulously examine the code of a game for exploits. The only difference between them and unassisted speed-runners is that they don’t rule out the exploits that require inhuman timing.

TAS players also only use the given controller inputs. Sure, they might slow down the game to allow pixel-precise scripting of these inputs, but when played back at normal speed, the game is still only being influenced by the controller.

This fact can be proven by watching the games being played on the original console using a computer wired to the traditional controller inputs. The scripted moves, created entirely by the user, are played back through the controller via signals that skip pushing the button manually.

Yes, they are using a computer to play the game back for them, but they aren’t doing so because they can’t otherwise beat the game. They use tools because they are curious how far they can push the boundaries of the game’s program, and they also want to share the immense entertainment value that the results can provide.

If further proof for legitimacy is needed, look no further than these videos that dispel common myths:


Myth #1 – TAS players are lazy


Using a computer aid doesn’t mean any less work. If anything, TAS players spend far more time researching and play-testing theories that could lead to relevant tricks and exploits. Some of these feats shave off time, but some just look cool.

The above Yoshi’s Island video demonstrates three years of hard work recording and editing play sequences so that the timing is not only perfect, but the play style is also incredibly entertaining to watch. Every action had to be repeated endlessly so as to generate the illusion of super-human skill when edited together.

The makers of this video also had to input all the controls manually at first and then adjust the time values to make sure nothing ever got out of sync.


Myth #2 – Abusing glitches isn’t fair

If you’re talking about playing a game as it was intended to be played, then yes, one should never use glitches if they want the full satisfaction of legitimately beating a game.

However, keep in mind that games are developed by programmers as a series of preset challenges. Jumping through their hoops is great, and makes for a multi-billion dollar industry, but going beyond those boundaries can yield a treasure trove of interesting scenarios. People have been sequence-breaking Super Metroid on the original console for years, as seen on #4 of my previous article.

What hasn’t been duplicated without TAS yet is beating all the bosses in the opposite order they’re intended. The above video accomplishes quite the feat – taking on the strongest boss first with the worst equipment.

Rather than focus on merely duplicating the normal play style but with some help, TAS players boldly seek out ludicrous challenges and then meet them head-on, like beating four Mega Man games at once with one controller.


Myth #3 – TAS ruins the human element

The above video of Super Mario Bros. 3 being glitched via a specific set of inputs isn’t terribly special. Game-crashing glitches like pressing up and down at the same time while on ladders in the Japanese version of Kirby Super Star have been known to sometimes lead to game-end screens.

What makes this particular TAS special is that it can reproduced by a human being on a normal console. Thanks to the tireless research of TAS contributors, the SMB3 pipe glitch was made known to the world and was later duplicated by Twitch player Jokaah. So if you really want to watch someone perform the same trick slightly slower and with a couple of foul-ups, that video is available.


Myth #4 – Skipping most of the game makes it boring

The Mario video above would beg to differ. Beating a notably long game in five minutes is both weird and impressive by itself, but watching TAS contributors Mijitsu and Swordless Link hand the game its own ass is both awe-inspiring and hilarious.

Other players have accomplished even more jaw-dropping feats, like reaching Gannon in Ocarina of Time as Young Link. When someone can take a game that’s been played to death and still leaves you wondering “What’s going to happen next?” he or she deserves some props.


Myth #5 – A TAS makes the game too easy

Like the Super Metroid video, this Castlevania video sets out to challenge the player with an unconventional limitation. Player Grincevent painstakingly researched and worked out a run that didn’t destroy a single object or enemy in Dracula’s castle, with the exception of the bosses.

While he was allowed to repeat individual sections endlessly until he got it right, the madness-inducing frustration involved in dodging everything perfectly is astounding. The only greater feat I’ve seen was Youtube user NotEntirelySure’s unassisted 500 point Super Mario Bros. Run.


Myth #6 – Doing a TAS is ultimately pointless

Playing video games in general can be deemed pointless, but for the people who insist that using a computer to assist you with a game is even more pointless, I would say that the discovery involved makes it well worth it.

The above video demonstrates a game-ending glitch similar to the pipe one in SMB3, but instead of ending the game, the player inputs just the right controls to manipulate code using only the controller port.

By executing a precise sequence of button presses before activating the glitch, TAS player Masterjun was able to incorporate lines of arbitrary “junk” code into the Super Mario World game’s RAM and run mini-games using existing assets.

Masterjun has thus walked beyond the boundaries of playing a game in this instance and has managed to create something entirely new. The game was even able to played once the AI-controlled input was removed and a regular controller was plugged in.




TAS videos don’t take away from the healthy gaming community that already exists. No one is deleting legitimate speedruns off the internet, and no one is going to come over and smash your precious vintage game collection like so much James Rolfe. Instead, TAS players take their love of these old games and try to transform the experience.

Also, using a computer to input controls faster than a human could is not an irredeemable sin. People don’t expect Smooth McGroove to sing all the vocal tracks at once, talented as he is.

Rather than reject  TAS videos, recognize them for the problem-solving and entertainment value that they hold. A TAS video can help renew interest in a neglected classic or just pass the time by watching a favorite game played through in mind-boggling ways.

So, instead of judging, keep TAS videos in a separate mental category than game hacks or human-controlled single-run playthroughs, and give them the respect that they deserve.

Similar Posts


  1. I think viewing this debate in terms of art is a “backwards logic” sort of mistake. ANYTHING can be viewed firstly as art and then be conveniently boxed up to fit that label.

    I believe the real issue here is perception of “human achievement vs machine achievement”… and has nothing at all to do with art. That’s not to say that in the end, their respective achievements can’t still be “recognized” as art though, thus validating “tool-assisted speed runs”, at least minimally, as art.

    To help demonstrate, I’ll borrow from the popular quote… “guns don’t kill people… people kill people”, and re-purpose it…

    …”Games don’t play games… people play games.” (again… not about art… but rather inherent capability)

    What a game is “mechanically capable of” (being finished in time x for example), is quite a different thing than a living/intelligent creature then demonstrating what ‘they’ can do.

    Ultimately, I think the dissatisfaction simply stems from this particular expectation of capability. So it comes down to mislabeling I would say… specifically, the “purpose” of the video… to see what the game is mechanically capable of… or to see what a human can do with that capability?

  2. This sounds like watching a video and expecting to see a man tight rope walk across the grand canyon. Only to learn that he tight rope walked across his back yard a bunch of times and edited in the footage of the grand canyon and cut out all of the times he fell.

  3. TAS runs aren’t contentious within the speedrunning community, and that’s all that really matters in this discussion. Speedrunners have many different categories for records, like any% and 100% runs, and TAS runs are just another category in the list. The simple solution for someone who doesn’t like TAS runs would be to watch a different category.

    Random people complaining about YouTube speedruns would be like someone who doesn’t watch football saying that the defense shouldn’t be allowed to blitz because it’s not fair to the QB.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.