It wouldn't be a SMB speedrun video if it didn't have the bus analogy.
Jokes aside, Summoning Salt is a great Youtube channel, and it makes me feel emotionally invested in speedruns of games I've never heard of before. Great storytelling. Not to mention that speedruns (and the glitches people find to get world record times) are pretty interesting from a computer science standpoint.
Came here to add comments about Summoning Salt as well. I don't follow speedruns at all; but, I really love his work and I absolutely echo the sentiment that he will get you emotionally invested in his content.
As an interesting aside, not only does he make great documentary style videos about Speed running, he also happens to be the current (as of June 2021) world record holder for Mike Tyson's Punch-Out!
I do wonder if the “check for win condition after every X number of frames” is for performance reasons.
Since I believe it’s every 21 frames that means you could perform 20 other functions dispersed across the 20 other frames, but if the system doesn’t need to do anything in that time then I don’t know why you wouldn’t just check every frame.
I found it to be very similar to science at large - the idea that there’s a community nibbling away at the edges of the unoknown, working in collaboration (and some competition) to maximize something.
This is a great video, but it glosses over some of the technical aspects because it assumes the viewer already understands them. This video provides a basic gestalt of the tech used for these runs: https://www.youtube.com/watch?v=_FQJEzJ_cQw
It‘s probably way too late to change the rules now, but even though it‘s totally legal according to the rules on the leaderboards, using a keyboard seems like quite a difference from using a controller, and from the end of the video, sounds like it does make certain moves easier.
It's complicated. Keyboards are almost certainly 'better' in the abstract if they're good enough (low activation threshold, not a lot of bus latency on whatever you're using to connect it, etc) because it's easier to press complex sequences of buttons when you're not holding the thing you're pressing them on ('fat thumb') or you can choose the layout of the relevant keys (vs. 'claw' or 'piano' gripping a controller).
But there are advantages to controllers too -- it tends to be easier to press two buttons on the same frame for eg.
And the thing is, it's not like there's only one controller you can perfectly standardize on. Even if you require only an 'original controller from the console manufacturer', well does that mean you have to be using the original rubber contact pads? Those are mostly all dead by now. Are all replacement rubbers equal? What about putting tape in the middle of the dpad to make diagonals more consistent?
And then it's also an accessibility issue. A lot of people can't use the original controllers. A lot of people who are otherwise very good at video games. Many of them can't use them because they played so many video games on them and they've gotten RSIs.
All that said there are usually limits. Almost no games allow macros (exception: Celeste allows a macro for a dash technique) or any other "press one button for multiple inputs" kinda stuff. Very few games allow turbo (exceptions: some RPGs do because mashing through cutscenes is very bad for your hands).
A lot of games do require controllers that are at least no more capable than one that came with the console though. But then there are things like one-handed controllers that are no more capable but could potentially change what's easy/hard.
So.. yeah it's complicated. It's just down to the game and its 'community' to decide what makes sense. That's largely how the whole speedrunning world works these days.
But no it's not too late to change the rules. Speedrun leaderboard rules change all the time to accommodate new information, techniques, and glitches. If people who run SMB feel it's unfair, it might change.
From what I've seen, speedrunning communities tend to be pragmatic about that: If something like this becomes a concern, the leaderboards are typically just split and a new category added. (Although I guess for a console game there is a fundamental question if keyboard is an extra category or gets allowed for the "main" one)
For some games there are tons of categories for different variations or restrictions. (It can even happen that when a new fundamental glitch is found, a category "Game X without this one specific glitch" is added if people enjoy running the game without it anyways)
The video mentions that moves that can only be done by pressing right and left at the same time aren't allowed, as that can't be done on a controller. I wonder if keyboard input is tolerated but only if it blocks impossible combinations, or if it's just on the honor system not to do those moves?
So I'm going to use the GameCube controller with the keyboard in the middle or I'll make a custom controller board or use a high end fighting games pad.
Though an original NES controller speedrun coming close should definitely be a legit accomplishment.
I enjoyed the video, but I have to disagree. For me, the storytelling was the worst part. I feel like they said “the problem seemed insurmountable, and the community needed a little help. [x months/weeks/years] later [clever person] decided to give it a go” about 4 times.
I suppose, from a certain perspective, repetition can make good storytelling, and the actual real life events of course restrict the storyline to an extent, but they could have at least phrased it differently
Highly recommend going through Summoning Salt’s back catalogue as he is a great storyteller. He has a way of getting me interested even in games I’ve never played, like Ninja Gaiden [1].
I think the epitome of this for me was his vid on Mario Kart 64 and the A1A trying to stop a prolific speedrunner from claiming 32/32 simultaneous track records [2].
My teenage son is currently 223rd in the world in Super Mario Bros with a sub-5 personal best. He's explained just bits and pieces of the SMB speedrun subculture to me (frame rules, etc.) and it's just fascinating.
At under 5 minutes, it seems like it's the ideal length for practicing as well - if you screw up a full run, you at most have lost 5 minutes of time, you just start over again.
No-warp runs which go about 20 minutes are way more painful when you make a dumb mistake in 8-4.
> how did they figure out the theoretical best time in the first place
A rarely used tool is brute-force searching for ideal inputs by making a computer play the game, trying all possible inputs. In theory, this process could find the ideal set of inputs for any game, but since the space of all possible inputs grows exponentially with the length of the sequence, this is only viable for optimizing very small portions of the speed run. Instead, a heuristic algorithm can be used. Although such an approach does not guarantee a perfect solution, it can prove very effective for solving simple puzzle games.
They don't know the theoretical best time but they know what is possible using tool-assisted speedruns (TAS) using the current set of known tricks. So it's possible there are still tricks to discover that will improve the speed floor, but it's been a long time since anything new was found.
> - how did they figure out the theoretical best time in the first place?
Depends on the game...
With mario, if you know the max acceleration and speed of mario, and the length of the level, you can calculate the theoretical minimum needed to come to the other end, especially if no glitches are allowed. If you have to backtrack a bit, you can calculate the optimal path to pass through. Glitches (eg. passing through walls) make this harder.
With games like Zelda, breath of the wild, it's pretty much impossible to calculate exactly, except for taking the current records, finding flaws, errors and/or adding new optimizations, and subtracting the "lost time" from the record.... until a new, better strategy comes out.
> how did they figure out the theoretical best time in the first place?
Generally by trial and error and with a TAS, which is a "Tool Assisted Speedrun". Some people determined what all of the most optimal segments were, then programmed a computer to push the buttons at exactly the right time (down to frame perfect). So theoretically the current best time could be a local maximum, but with how long the game has been out for that seems unlikely.
This also led to a kind of funny thing where some runs were considered "TAS Only", only to be later achieved by a human in a real run.
[+] [-] oceliker|4 years ago|reply
Jokes aside, Summoning Salt is a great Youtube channel, and it makes me feel emotionally invested in speedruns of games I've never heard of before. Great storytelling. Not to mention that speedruns (and the glitches people find to get world record times) are pretty interesting from a computer science standpoint.
[+] [-] bcrescimanno|4 years ago|reply
As an interesting aside, not only does he make great documentary style videos about Speed running, he also happens to be the current (as of June 2021) world record holder for Mike Tyson's Punch-Out!
https://www.speedrun.com/user/Summoningsalt
[+] [-] Pulcinella|4 years ago|reply
Since I believe it’s every 21 frames that means you could perform 20 other functions dispersed across the 20 other frames, but if the system doesn’t need to do anything in that time then I don’t know why you wouldn’t just check every frame.
[+] [-] seventytwo|4 years ago|reply
[+] [-] Agentlien|4 years ago|reply
I don't follow the speedrunning scene, but I always get really excited when Summoning Salt releases a new video.
[+] [-] Afforess|4 years ago|reply
[+] [-] css|4 years ago|reply
[+] [-] thenoblesunfish|4 years ago|reply
[+] [-] stormbrew|4 years ago|reply
But there are advantages to controllers too -- it tends to be easier to press two buttons on the same frame for eg.
And the thing is, it's not like there's only one controller you can perfectly standardize on. Even if you require only an 'original controller from the console manufacturer', well does that mean you have to be using the original rubber contact pads? Those are mostly all dead by now. Are all replacement rubbers equal? What about putting tape in the middle of the dpad to make diagonals more consistent?
And then it's also an accessibility issue. A lot of people can't use the original controllers. A lot of people who are otherwise very good at video games. Many of them can't use them because they played so many video games on them and they've gotten RSIs.
All that said there are usually limits. Almost no games allow macros (exception: Celeste allows a macro for a dash technique) or any other "press one button for multiple inputs" kinda stuff. Very few games allow turbo (exceptions: some RPGs do because mashing through cutscenes is very bad for your hands).
A lot of games do require controllers that are at least no more capable than one that came with the console though. But then there are things like one-handed controllers that are no more capable but could potentially change what's easy/hard.
So.. yeah it's complicated. It's just down to the game and its 'community' to decide what makes sense. That's largely how the whole speedrunning world works these days.
But no it's not too late to change the rules. Speedrun leaderboard rules change all the time to accommodate new information, techniques, and glitches. If people who run SMB feel it's unfair, it might change.
[+] [-] detaro|4 years ago|reply
For some games there are tons of categories for different variations or restrictions. (It can even happen that when a new fundamental glitch is found, a category "Game X without this one specific glitch" is added if people enjoy running the game without it anyways)
[+] [-] elihu|4 years ago|reply
[+] [-] undersuit|4 years ago|reply
Though an original NES controller speedrun coming close should definitely be a legit accomplishment.
[+] [-] ranie93|4 years ago|reply
[+] [-] xiphias2|4 years ago|reply
[+] [-] permo-w|4 years ago|reply
I suppose, from a certain perspective, repetition can make good storytelling, and the actual real life events of course restrict the storyline to an extent, but they could have at least phrased it differently
[+] [-] joshschreuder|4 years ago|reply
I think the epitome of this for me was his vid on Mario Kart 64 and the A1A trying to stop a prolific speedrunner from claiming 32/32 simultaneous track records [2].
[1] https://youtu.be/7u1tVD7UEqw
[2] https://youtu.be/D6cpa-TvKn8
[+] [-] epaga|4 years ago|reply
At under 5 minutes, it seems like it's the ideal length for practicing as well - if you screw up a full run, you at most have lost 5 minutes of time, you just start over again.
No-warp runs which go about 20 minutes are way more painful when you make a dumb mistake in 8-4.
[+] [-] amirmasoudabdol|4 years ago|reply
[+] [-] disambiguation|4 years ago|reply
some thoughts
- given the timeline of progress, i guess we might see THE perfect run in the next 12 months or so
- how did they figure out the theoretical best time in the first place?
- is machine learning being applied to this problem yet? might be able to discover some more time saves
[+] [-] tantalor|4 years ago|reply
A rarely used tool is brute-force searching for ideal inputs by making a computer play the game, trying all possible inputs. In theory, this process could find the ideal set of inputs for any game, but since the space of all possible inputs grows exponentially with the length of the sequence, this is only viable for optimizing very small portions of the speed run. Instead, a heuristic algorithm can be used. Although such an approach does not guarantee a perfect solution, it can prove very effective for solving simple puzzle games.
https://en.wikipedia.org/wiki/Tool-assisted_speedrun#Method
[+] [-] readams|4 years ago|reply
[+] [-] ajsnigrutin|4 years ago|reply
Depends on the game...
With mario, if you know the max acceleration and speed of mario, and the length of the level, you can calculate the theoretical minimum needed to come to the other end, especially if no glitches are allowed. If you have to backtrack a bit, you can calculate the optimal path to pass through. Glitches (eg. passing through walls) make this harder.
With games like Zelda, breath of the wild, it's pretty much impossible to calculate exactly, except for taking the current records, finding flaws, errors and/or adding new optimizations, and subtracting the "lost time" from the record.... until a new, better strategy comes out.
[+] [-] fuzzer37|4 years ago|reply
Generally by trial and error and with a TAS, which is a "Tool Assisted Speedrun". Some people determined what all of the most optimal segments were, then programmed a computer to push the buttons at exactly the right time (down to frame perfect). So theoretically the current best time could be a local maximum, but with how long the game has been out for that seems unlikely.
This also led to a kind of funny thing where some runs were considered "TAS Only", only to be later achieved by a human in a real run.
[+] [-] ChrisArchitect|4 years ago|reply
[+] [-] jordigh|4 years ago|reply
[+] [-] andrewmcwatters|4 years ago|reply
[+] [-] anorphirith|4 years ago|reply
[+] [-] ChrisArchitect|4 years ago|reply
The vid title or description could use a one liner about what it's actually about.