This is a warpless tool-assisted speedrun of Superfast Mario Bros, a sped up romhack of Super Mario Bros., and serves as a companion branch to #4247: Denial140, negative_seven & chatterbox's NES Superfast Mario Bros. "warps" in 01:09.07.

Game objectives

  • Emulator used: FCEUX 2.6.3
  • Fastest Completion
  • Forgoes warps
  • Contains speed/entertainment tradeoffs

General Comments

Having watched Denial140 and negative seven's warps run many years ago, I became enamored and fascinated by the romhack. Looking around in the submissions area for runs of this game, I noticed that nobody has done a warpless run, so I decided to do one myself.
A few techniques I have to mention here. The big one is that the lag reduction trick described in the other run doesn't really work. When you let go of the Right button for a frame while running or press left then hold right while in midair, what the game would do half of the time is just move the lag frame to another place. The other half of the time, the lag frame may be gone, but there will be no net gain in frames saved due to the very brief loss of speed. Now there could be a way to actually save frames by letting go of the Right button at certain frames, but that would require extensive trial and error that I don't really have the patience to do manually. Because of all that, I just let lag frames happen if they aren't caused by an action I took myself.
There is another application for this kind of input technique, however. Thanks to the romhack struggling to keep up with rendering everything on screen, certain objects may not have proper collision detection, the big ones being the flagpole and the axe at the end of the castle stages; Mario may pass through the flagpole half the time and the axe can only be triggered if Mario touches it at a certain angle. By slowing down Mario by pressing Left for a frame and then holding down Right right after, Mario can properly touch the those objects and properly trigger the end level sequence.
One question you may ask is "Why did you grab the Super Mushroom at any point?" I actually did that in an earlier iteration of the run but to my horror, Big Mario actually causes more lag frames to suddenly appear than if I were to use Small Mario. I suppose that's due to Big Mario's animation cycle, but I could be wrong about that. Another question you may ask is "Why didn't you collect every coin you saw?" Also to my horror, I found that by collecting coins at certain sections of certain levels, lag frames may occur as a result. And if you collect over 99 coins and reset the coin counter, the game really starts lagging at that point.

Notable Stages

World 1-2: What Denial140 and negative seven did in their run at the beginning of this level immediately jump in midair and run across the top of the level. I deemed that unentertaining and decided to properly go through the level, stomping enemies and all. I didn't test if I lost time doing this, though.
World 1-3: This is the first of many levels where I have to slow down and move back for a frame in order to be in a better position to jump across the platforms while maintaining top speed.
World 2-1: If it weren't for the walljump glitch, I would have to use the trampoline to get over the wall at the end. Wallclipping doesn't work here due to how fast Mario movies.
World 2-2: There is one big thing that can make this level and other underwater levels very frustrating. The spawn points of the Cheep-Cheeps throughout the level is determined by the RNG. Should you get very unlucky, a Cheep-Cheep can actually spawn right in front of the exit pipe, which is exactly what happened to me in a previous iteration of this run.
World 4-1: Surpisingly, Lakitu doesn't appear to be affected by the romhack, so Mario outruns him and spends most of the level without him. This applies to every other level with Lakitu in it.
World 4-2: Another level where I forgo taking the roof route in favor of going through the underground portion for entertainment purposes. Unlike 1-2, however, I have to slow down for a better position to jump on or I'll hit a wall and lose even more time. And even if I did take the roof section, I have to get down there eventually and the way the last part of the level is designed forces me to slow down as there isn't enough space for speedy Mario to get to the warp pipe and maintain constant max speed.
World 4-4: What would normally be done in the maze-like section in the middle of this level is that Mario would clip through the block and bypass the maze. However, Mario can't wallclip in this romhack so I have to go through the maze conventionally. This level is also the first instance where touching the axe doesn't cause the bridge to open up and drop the fake Bowser. I don't know why this happens and I'm not sure if this can be done on command without losing time.
World 5-4: That huge firebar at the beginning is a big speed killer. Maybe with better RNG, the firebar could be in a more favorable position. The platform near the axe at the end is also a time killer as it serves as a barrier in the case of this romhack and leaves little space for Mario to get to the axe, forcing me to very briefly hover over the axe so Mario can touch it. This platform situation also happens in number of other castle levels.
World 6-2: Funny thing about the Starman powerup. Not only does the Starman move as fast as you, the powerup only lasts for half a second. The staircase at the end has a hidden tangible block that forced me to move back for a frame so I can get over it.
World 6-3: I encountered some hidden barriers in this level. Whether it's due to the glitched pulley platforms or a hidden block that suddenly became tangible, I don't know.
World 6-4: Here, I noticed that Bowser's flames can only affect Mario at certain positions. A good chunk of the time, Mario can pass through them unharmed.
World 7-2: The two Bloopers at the top are bastards here. They will always sweep that portion of the level, preventing me from going near the top. I actually have to slow down for a frame at the beginning because of them.
World 7-4: Here, I'm late to the party where people know that the loop triggers can only be activated while Mario is walking on the ground. This limits the amount of space Mario can move in while maintaining top speed. The fake Bowser in this level decides to leave to his home planet, which I guess explains why the bridge didn't open when Mario touched the axe.
World 8-1: I tested this level without using the Starman and as it turns out, it's faster to use it to kill the Goombas and the Piranha Plant you run into as that eliminates lag frames. Looks like this proves that the loading of multiple enemies on screen as explained in the "warps" submission does cause lag frames.
World 8-2: The big gapped staircase in the beginning required me to slow down for a frame as Mario can't make it over while maintaining top speed.
World 8-4: The first warp pipe appeared to be blocked by a Piranha Plant. Luckily, if I just hover over it just slightly to the right, it will despawn and I can safely enter the pipe. The third warp pipe was also blocked and the Piranha Plant in it won't despawn by hovering over it, but also lucky for me, I can enter it the moment I get into the right position and Mario won't be hurt by the Piranha Plant. I couldn't find a better way to end the underwater section of the level without hugging the warp pipe at the end for a few frames thanks to the position of the last firebar. Input ends when Mario makes his last jump.

Possible Improvements

  • Application of the frame rule? Being fast is tantamount in this run and Mario can't slow down unless he really has to. Also, with how fast things move in this run, I'm not sure if any difference will be made.
  • Slowing down could actually remove lag frames and save time, but I think a script would have to be devised in order to figure out which places are best to do so and I'm not good enough for that.
  • Late into the creation of this run, I learned while playing the later castle levels that you can save a frame by having Mario run for a frame, then jumping right after. I didn't want to go back and redo the earlier castle levels to incorporate this trick as this might affect the RNG and a romhack like this is very sensitive to the effects of the RNG. I might have to redo large parts of the run just to save like four frames.

Special Thanks

  • To Mizumaririn for helping me find the RAM address that keeps track of the looping sections of the game.

feos: Claiming for judging.
feos: Delaying, hoping acmlm will send this hack to RHDN (as we asked). If it doesn't happen in like a month, we'll see what we can do.
feos: It's finally accepted to RHDN (and other such hacks are now potentially RHDN-submittable since Acmlm attached the MIT license to them), accepting.

EZGames69: Processing...


TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 15583
Location: 127.0.0.1
This topic is for the purpose of discussing #7899: despoa's NES Superfast Mario Bros. "warpless" in 04:01.54
MrTASer
He/Him
Player (231)
Joined: 7/3/2021
Posts: 147
Location: Udaipur, Rajasthan, India, God'sBeautifulWorld
Ohh man, this was much more ridiculous than my SUPER SONIC CHEATS and faster too...https://youtu.be/cPFyOX-lDyk
My TASing channel ; My MUSIC channel ; My SoundCloud ; My HomePage ; Music Composer ; Mr.TASer#5922 VCop 2 TAS completed, but in a completely different way ; And SMB - GAGOTO (WL) TAS COMPLETED TILL W-4 ; Some more...
Banned User
Joined: 4/1/2016
Posts: 295
Location: Cornelia Castle
Hi. I have a couple things to say about this TAS. The previous TAS was rejected for hack quality. I understand why this decision was made. Although this was a very nice run, I have a feeling like the same decision will come through with this submission. I am still giving this a yes vote, though, as it beats all existing warpless TASes of this hack, and I still enjoyed this run.
DJ Incendration Believe in Michael Girard and every speedrunner and TASer!
Fortranm
He/Him
Editor, Experienced player (878)
Joined: 10/19/2013
Posts: 1121
https://www.romhacking.net/hacks/1720/
This is inspired by the Fast Foe Comet.
Interesting. I originally thought it's inspired by bootleg hacks like this, but I guess the speed here doesn't even compare? :P Link to video
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
I was surprised by how obscure this hack is. Google only finds tasvideos and some video hosting sites, they all link directly to the IPS file, acmlm's site doesn't have an index page since 2009, speedrun.com or any other communities don't know about this hack. It's probably obvious why nobody is interested in this hack outsite TASing, because it's mostly unplayable in real time, unless you've practiced a lot. But doesn't that make it even more worthy of a TAS? I think this hack is hilarious and it does provide some new challenges in the SMB1 universe. But we have this rule and I don't know what to do:
ROM hacks must not be overly obscure. If a hack is known to GoodTools, ROMhacking.net, SMW Central, or some other well-known database, that usually means its quality is decent and it won't be completely lost in the future.
  • There must be a publicly and readily available download link to the hack's patch file on the web. It should be easily found through Google. Having to join an online community (for example on Discord) to get the file is not allowed.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Patashu
He/Him
Joined: 10/2/2005
Posts: 4043
Bootstrap paradox? - We can become the reason why it's a well known hack.
My Chiptune music, made in Famitracker: http://soundcloud.com/patashu My twitch. I stream mostly shmups & rhythm games http://twitch.tv/patashu My youtube, again shmups and rhythm games and misc stuff: http://youtube.com/user/patashu
Expert player (2567)
Joined: 12/23/2007
Posts: 830
This looks like a good movie for demonstration, and can certainly entertain some people, but I have concerns about having movies like this published on TASVideos, so I'm going to post different opinions. To me, this looks like playing a game with super game-breaking cheat. I know technically it's not cheat, but it feels a lot like cheat. I'm not sure what the actual goal is. If the goal is "completing a game that looks like Super Mario Bros. in the fastest time", then what if someone else makes a even faster version "Hyperfast Mario Bros." one day? Technically I think if someone improves the code further, it can make Mario move faster and further reduce lags and waiting time. Will that obsolete this run?
Recent projects: SMB warpless TAS (2018), SMB warpless walkathon (2019), SMB something never done before (2019), Extra Mario Bros. (best ending) (2020).
Fortranm
He/Him
Editor, Experienced player (878)
Joined: 10/19/2013
Posts: 1121
HappyLee wrote:
I'm not sure what the actual goal is. If the goal is "completing a game that looks like Super Mario Bros. in the fastest time", then what if someone else makes a even faster version "Hyperfast Mario Bros." one day? Technically I think if someone improves the code further, it can make Mario move faster and further reduce lags and waiting time. Will that obsolete this run?
I wonder how a TAS of that bootleg hack from that multicart whose rom can be relatively easily found that I posted earlier would look like. If the result is somewhat similar, I wonder if that can be considered a case of notable/commonly seen bootleg hack like Street Fighter II': Champion Edition, which didn't even need an ips patch included in the package of all accepted hacks on the site from a while back.
Editor, Publisher, Player (47)
Joined: 10/15/2021
Posts: 376
HappyLee wrote:
what if someone else makes a even faster version "Hyperfast Mario Bros." one day? Technically I think if someone improves the code further, it can make Mario move faster and further reduce lags and waiting time. Will that obsolete this run?
Probably, probably not. Depends on how accessible the hack is. Also, I don't think it's possible to go any faster than this without it crashing in five seconds.
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
HappyLee wrote:
Technically I think if someone improves the code further, it can make Mario move faster and further reduce lags and waiting time. Will that obsolete this run?
Sounds possible, though such a hack would raise the same questions as this one - about accessibility and preservation.
Fortranm wrote:
I wonder how a TAS of that bootleg hack from that multicart whose rom can be relatively easily found that I posted earlier would look like. If the result is somewhat similar [...]
I don't see any similarities.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Banned User
Joined: 4/1/2016
Posts: 295
Location: Cornelia Castle
The hack is actually available on Acmlm's website, if you click the link in his video description. http://acmlm.rustedlogic.net/ips/SMB1fast.ips
DJ Incendration Believe in Michael Girard and every speedrunner and TASer!
Banned User
Joined: 1/6/2023
Posts: 263
Can't you just continually do simple ROM editing until you get the speed so fast that it is unplayable & unTASable? I don't think this hits that wall and as such is kind of pointless. You should hack into Super Mario Bros. and increase the speed to that level. Right now it's just arbitrary on what some guy made it. Picked a random number and went with it to create the ROM.
Published TASes: #1, #2, #3, #4, #5, #6, #7, #8, #9, #10, #11, #12 Please consider voting for me as Rookie TASer Of 2023 - Voting is in December 2023 My rule is quality TASes over quantity TASes... unless I'm bored.
Banned User
Joined: 1/6/2023
Posts: 263
Err, I mean the byte before that wall, obviously.
Published TASes: #1, #2, #3, #4, #5, #6, #7, #8, #9, #10, #11, #12 Please consider voting for me as Rookie TASer Of 2023 - Voting is in December 2023 My rule is quality TASes over quantity TASes... unless I'm bored.
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
The speed here is actually maximum. Go any further and you get an overflow, because the speed address is used a signed byte (sign bit meaning the direction). Same with various delays - they were all set to just 1 frame. Here's info from Acmlm, edited by me to look like an RHDN submission, which they rejected because I'm not the author.
The framerule was changed from 21 frames to 1 frame, instant acceleration to 0x7F speed, various delays cut out or speed values increased, nearly instant level clear. It's about as fast as acmlm could make it go, and was meant mostly for a silly TAS, but turns out to be actually beatable too!
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Banned User
Joined: 1/6/2023
Posts: 263
Ahh. Only other thing I could think would be editing how many pixels Mario travels forward per frame. (Almost like frame skip) I stand corrected then.
Published TASes: #1, #2, #3, #4, #5, #6, #7, #8, #9, #10, #11, #12 Please consider voting for me as Rookie TASer Of 2023 - Voting is in December 2023 My rule is quality TASes over quantity TASes... unless I'm bored.
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
OtakuTAS wrote:
Ahh. Only other thing I could think would be editing how many pixels Mario travels forward per frame.
Yeah that's called speed.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
negative_seven
She/Her
Active player (429)
Joined: 11/25/2012
Posts: 103
Location: Europe
The movement could be made faster, but then why stop there? There's really not much limit as to what you can speed up, and eventually you'd just be ending the game as soon as it starts. The hack strikes a decent balance as-is, I think. Maybe the choices Acmlm made were arbitrary, but they make for a fresh experience and at this point the hack is over a decade old and much better known on the internet than even some of the games we have published runs of on the site.
Post subject: Movie published
TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 15583
Location: 127.0.0.1
This movie has been published. The posts before this message apply to the submission, and posts after this message apply to the published movie. ---- [5111] NES Superfast Mario Bros. "warpless" by despoa in 04:01.54