• Emulator used: DeSmuME v0.9.7 (Make sure Advanced Bus-Level Timing is off before watching.)
  • Goal: any%
  • Takes damage to save time
  • Uses death as a shortcut
This is an improvement of 2002 frames (~33.366... seconds).
New Super Mario Bros. was released in 2006 for the Nintendo DS. The story has been the same every time for the past 25 years: rescue some princess who can't defend for herself. So you're playing as this fat plumber, right? Because everybody else is too lazy to save her, it's up to this guy. Once again, he's gotta move through eight themed worlds to fight the evil king Bowser and save the princess. He really doesn't care about her, though - he's only in it for some cake afterward. (No point in doing it for fame anymore, anyway.) So he effortlessly skips almost every world and gets there without the need of some huge fungi, apparently exclusive to this story. Rushing through micro-sized, beating up some Bowser minions without a thought or care, and boosting around with a cool shell he found lying around, he does so without fail... well, almost.

Addresses

nitsuja's Lua script was very helpful with watching X and Y values. You can find it here: http://pastebin.com/gz975ErQ
  • X Speed: 021B6A5C - Fixed Point, 4 bytes
  • X Position: 0208B70C - Fixed Point, 4 bytes
  • X Camera: 020CA2EC - Unsigned, 2 bytes
  • Y Speed: 021B6A61 - Signed, 1 byte
  • RNG 1: 02088A68 - Unsigned, 2 bytes
  • RNG 2: 02088A6A - Unsigned, 2 bytes
Note: This submission will refer to speed by nitsuja's Lua script, since that's what was used most of the time.

Luck Manipulation

The starting RNG depends on the date and time you start the movie.
The RNG changes whenever certain sound effects are played, whether by Mario or the background music itself (though this depends on the level). Luck can always be manipulated by either double jumping, throwing something, or wall jumping. We sadly do not know exactly what the numbers in the address mean yet, so luck was usually manipulated by changing the amount of double jumps and wall jumps.
We currently know that the RNG affects:
  • The Red ? Block and Hammer Bros. movements on the world map
  • The movement of the Scuttlebugs
  • The volcano rocks in World 8-8
  • When the final cutscene in World 8-Castle 2 (Bowser Jr. throwing Dry Bowser into the pot) activates

Glitches, Tricks, and Notes

  • Outside of manipulation, this game is very edit-friendly. We've often rearranged parts of the movie file to compare attempts and such.
  • While in the Blue Shell, as long as you don't hold left or right and keep jumping, you won't lose speed.
  • Luigi does not affect luck at all. (You can watch this TAS with him by simply adding L and R to frame 298.)
  • With the Blue Shell, and at full speed while out of the shell, running down a hill for 2 frames and going into the shell while jumping will keep your speed slightly above maximum running speed (3087, compared to 3072) for as long as you stay in the shell.
  • For luck manipulation: To lose the least amount of speed when doing repeated double jumps, cancel a triple jump by landing on a double jump with no input for 1 frame.
  • You can super corner boost with the Blue Shell by coming out of it, just like "un-ducking" with big Mario.
    • If you come out before hitting the block first, you will perform a midair jump.
  • Ducking before falling down a cliff will make you fall quicker.
  • You can start duck jumping/get into the Blue Shell without losing speed by running off a cliff and then starting the action.
    • With the Blue Shell, try to jump and land on the very edge of the cliff.
      • After doing so, if you get out of your shell early, you can perform a midair jump. (A silent double jump will not change the RNG.)
    • As with the shell, you can perform a silent double/triple "fall" by "un-ducking" in midair with small/big/Fire Mario.
  • A midair double or triple jump can be done by duck jumping and hitting a ceiling (to "un-duck") nearly right after jumping. (Possibly linked to the previous note?)
  • When wall jumping, you can save a frame by performing the wall jump a little away from the wall.
    • If you push against a wall until you're able to wall jump and let go or move away from the wall, you can "store" a wall jump for about half a second.
      • You can wall jump off of nothing - as long as you're not close to a floor, you can wall jump at any time for that half of a second.
    • You can shoot a fireball to gain speed faster after wall jumping.
  • Being able to "catch" a wall for a vertical wall jump with Mini Mario seems to depend on the speed and frame you approach it.
  • There is odd behavior with trying to enter pipes - sometimes, you will be stalled before being able to enter. This stall can usually be reduced by holding forward in input longer, changing the state of your animation (usually by double/triple jumping), or running along the ground into the pipe.
  • Run down hills to the point where you need to slide to gain more speed. This is because sliding takes longer to gain speed.
    • Hold left or right while sliding (keeping down held, too). Holding left (or right) will help you gain more speed.
    • When you slide off a hill and start falling, keep holding forward to keep your speed slightly higher than usual (3096, compared to 3072) until you land.
  • When hitting a button to raise the water level, as Mini Mario, swim above the water and fall back into it. You will suddenly jump back up to the surface and can then walk on it.
  • Ground pounding off of a Broozer as Mini Mario will bounce you very high.
  • Use screen scrolling to your advantage. Time activating moving platforms so you can run across them in the direction you want to.
  • In World 8-6, there are some areas where a spring pad is next to a wall. If you jump between that space and the pad and push against the wall, you'll get a vertical boost.
  • You can "freeze" your animation by rapid-firing in the opposite direction you're facing. (This looks very silly when doing it after a triple jump with Mario's head toward the ground.)
  • The hanging metal platforms in World 8-5 have odd behavior. Running while they're tilted against you will sometimes make you run faster.
    • The unstable rocks act normally, however. Try to land on approximately the middle to make them tilt in favor of your running.
  • You can repeatedly jump off of cliffs (like in World 8-7) to keep your speed. (You can even keep jumping after landing on solid ground to stay facing forward as long as you don't stop.)
  • While in water currents (currents from underwater pipes, like in World 8-3), your speed cycles. You can use this cycle to your advantage.
    • Here's two pattern examples. Make sure to focus on the "xvel" value instead of the "speed" value in nitsuja's script.
    • (> = hold forward (or backward - it doesn't matter in a current), (nothing) = no input; quotes mean repeat what's in the quotes)
      • For all forms except Shell Mario:
        • 1152 (Start)
        • > 1244 1336 1428 1520 1612 1704 1796 1888 1980 > 2060 > 2044
        • 2136 2132 2128
        • 2124 > 2108 > 2092 > 2076 > 2060 > 2044 >
        • 2124 > 2108 > 2092 > 2076 > 2060 > 2044
        • 2136 2132 2128
        • 2124 > 2108 > 2092 > 2076 > 2060 > 2044 >
        • "2092 > 2076 > 2060 > 2044 >" (2 times)
        • (Release the rest of the time until back to 1152.)
      • For Shell Mario:
        • 1728 (Start)
        • > 1808 1888 1968 2048
        • "2032 2112 2096 >" (7 times)
        • "2032 2080 2064 2048 2032 >" (2 times)
        • 2080 2064 2048 2032 2048
        • (Release the rest of the time until back to 1728.)
    • What patterns you mainly want to aim for repeating the most are:
      • All forms except Shell Mario:
        • Major: 2124, 2108, 2092, 2076, 2060, 2044
        • Minor: 2092, 2076, 2060, 2044
      • Shell Mario:
        • Major: 2112, 2096, 2032
        • Minor: 2080, 2064, 2048, 2032, 2032
    • Of course, this will still depend on your X position when you reach the current. Sometimes, interrupting the "major" patterns can save time overall.
    • What is important is that you repeat the minor cycle on the very last frame you're still in the current. (You can tell when that is by checking to see if the next frame will not allow you to reset the loop.)

Level-By-Level Notes

Intro:

17 frames saved because of frames mysteriously being added to the previous movie's beginning.
13 frames mysteriously saved, likely from using a different emulator.
  • 30 frames ahead.

World 1-1:

1 frame saved from performing more corner boosts.
21 frames saved from using the Starman to run faster.
  • 52 frames ahead.

1-1 to 1-2:

No frames were saved.
  • 52 frames ahead.

World 1-2:

Room 1
7 frames saved from catching the ledge.
16 frames saved from super corner boosting with the Blue Shell and better speed management in the narrow passage.
8 frames saved from triple jumping off of the teeter-totter.
Room 2
3 frames saved from jumping to turn around after coming out of the pipe.
11 frames saved from wall jumping earlier.
  • 97 frames ahead.

1-2 to 1-Tower:

7 frames saved from better map manipulation.
  • 104 frames ahead.

World 1-Tower:

12 frames saved from getting through another passage earlier.
  • 116 frames ahead.

1-Tower to 1-Cannon:

17 frames saved from better map manipulation.
  • 133 frames ahead.

World 1-Cannon:

No frames were saved.
  • 133 frames ahead.

1-Cannon to 5-1:

1 frame saved from choosing "No" when asked to save by pressing right and A on the same frame.
  • 134 frames ahead.

World 5-1:

17 frames saved from not having to wait to manipulate luck.
  • 151 frames ahead.

5-1 to 5-2:

5 frames lost to manipulate the red ? block to move to World 5-B.
  • 146 frames ahead.

World 5-2:

Room 2
18 frames saved from super corner boosts.
1 frame saved around the multiple pipes.
3 frames saved upon the pipe entry.
Room 3
8 frames saved at the flagpole by making a full turn.
  • 176 frames ahead.

5-2 to 5-3:

33 frames saved from better map manipulation.
  • 209 frames ahead.

World 5-3:

11 frames saved from super corner boosting.
1 frame saved from keeping speed in the narrow passage by repeatedly jumping.
  • 221 frames ahead.

5-3 to 5-Ghost House:

38 frames saved from not having to watch the red ? block move.
  • 259 frames ahead.

World 5-Ghost House:

Room 1
1 frame saved from an earlier wall jump.
1 frame saved from better speed management at the Hammer Bro.
54 frames saved from ground pounding off of the Broozer and wall jumping to the next Broozer.
(133 frames saved total.) About 14 frames saved from better turning at the staircases. About 119 frames saved from ground pounding the Broozers and "activating" the Broozer next to the secret door earlier.
Room 2
45 frames saved in the elevator room from better usage of the vertical wall jump and taking a straight route upward.
  • 493 frames ahead.

5-Ghost House to 5-Cannon:

12 frames saved from better map manipulation.
  • 505 frames ahead.

World 5-Cannon:

No frames were saved.
  • 505 frames ahead.

5-Cannon to 8-1:

1 frame saved from choosing "No" in 1 frame.
  • 506 frames ahead.

World 8-1:

2 frames saved from corner boosting.
  • 508 frames ahead.

8-1 to 8-2:

No frames were saved.
  • 508 frames ahead.

World 8-2:

Room 2
8 frames saved from optimizing the water current.
25 frames saved from going above the rising water to jump and enter the pipe earlier.
Room 3
1 frame saved from a better wall jump. (2 frames saved from it, but 1 lost to positioning before entering the pipe at the end due to the wall jump.)
Room 4
11 frames saved from wall jumping.
Room 5
6 frames saved from avoiding pipe entry delay.
Room 6
2 frames saved from alternating A and B to swim up faster. 11 frames saved from avoiding pipe entry delay.
Room 7
24 frames saved from wall jumping up instead of using the rising water.
Room 8
2 frames lost to avoid a time of 255.
  • 594 frames ahead.

8-2 to 8-Tower 1:

No frames were saved.
  • 594 frames ahead.

World 8-Tower 1:

Room 1
82 frames saved from activating two blocks to move more in cooperation.
Boss
4 frames saved from a better fighting strategy.
4 frames saved from turning toward the front before the cutscene activates.
1 frame lost from having to wait for the timer to tally for another frame.
  • 683 frames ahead.

8-Tower 1 to 8-3:

33 frames saved from better map manipulation.
1 frame saved from choosing "No" in 1 frame.
  • 717 frames ahead.

World 8-3:

Room 1
6 frames saved from not being stopped by the ledge when falling to the pipe.
10 frames saved from avoiding pipe entry delay.
Room 2
27 frames saved from better cutting of corners.
42 frames saved from a combination of corner boosts and optimizing water currents.
2 frames saved from entering the pipe at a specific X position.
Room 3
1 frame saved from starting the wall jump a bit farther from the wall.
  • 805 frames ahead.

8-3 to 8-4:

33 frames saved from better map manipulation.
  • 838 frames ahead.

World 8-4:

2 frames saved from a super corner boost.
28 frames lost to pick up a Koopa shell.
1 frame lost from being unable to super corner boost.
28 frames lost to get the Starman.
77 frames saved from a combination of using the Starman to run faster and scrolling the screen to have the platforms move in favor.
  • 860 frames ahead.

8-4 to 8-Castle 1:

No frames were saved.
  • 860 frames ahead.

World 8-Castle 1:

Room 1
1 frame saved from getting on a moving platform earlier.
2 frames saved from running on the platform longer.
1 frame saved from a better super corner boost.
Boss
2 frames saved from a super corner boost.
5 frames saved from landing at a specific position on the button to activate Mario's "victory" cutscene earlier.
  • 871 frames ahead.

8-Castle 1 to 8-5:

1 frame saved from choosing "No" in 1 frame.
  • 872 frames ahead.

World 8-5:

Room 1
4 frames saved from super corner boosting.
4 frames saved from slope optimization of platforms.
2 frames saved from entering the pipe at a specific X position.
Room 2
1 frame saved from an earlier wall jump.
  • 883 frames ahead.

8-5 to 8-6:

6 frames saved from better map manipulation.
  • 889 frames ahead.

World 8-6:

1 frame saved from an earlier wall jump.
3 frames saved from a better jump to the platform.
19 frames saved from wall jumping to gain speed faster.
8 frames saved from better cutting of corners.
17 frames saved from stepping on the flying Koopa earlier.
36 frames saved from a different route up the Spiny area.
108 frames saved from wall jumping to the top of the room and removing the fadeout delay.
  • 1081 frames ahead.

8-6 to 8-7:

5 frames saved from better map manipulation.
  • 1086 frames ahead.

World 8-7:

72 frames saved from a super corner boost, and repeatedly jumping on the cliffs to keep speed.
19 frames saved from turning at the Starman earlier.
12 frames saved from running on floors and platforms more (using screen scrolling to manipulate platform timing).
6 frames saved from ducking to land on the ground near the pipe earlier.
  • 1195 frames ahead.

8-7 to 8-8:

3 frames saved from better map manipulation.
  • 1198 frames ahead.

World 8-8:

2 frames saved from a super corner boost.
30 frames saved from turning at the invisible block earlier and wall jumping off of the block.
3 frames saved from ducking to land earlier.
  • 1233 frames ahead.

8-8 to 8-Tower 2:

1 frame saved from better map manipulation.
  • 1234 frames ahead.

World 8-Tower 2:

Room 1
683 frames saved from a new strategy by dying after activating the midpoint (approximate - this is impossible to compare directly due to map manipulation).
42 frames lost from getting a Mushroom.
42 frames lost from getting a Fire Flower.
3 frames saved from jumping to the final platform near the boss door earlier.
Boss
2 frames saved from a better fighting strategy.
4 frames saved from turning toward the front before the cutscene activates.
20 frames lost from having to wait for the timer to tally longer.
  • 1822 frames ahead.

8-Tower 2 to 8-Castle 2:

4 frames saved from better map manipulation.
1 frame saved from choosing "No" in 1 frame.
  • 1827 frames ahead.

World 8-Castle 2:

Room 2
18 frames saved by wall jumping to gain speed faster.
Room 3
7 frames saved from minimal slowing down at the horizontal fire shooter.
1 frame saved from super corner boosting.
Room 4
29 frames saved from a different strategy.
Room 5
8 frames saved from wall jumping to gain speed faster.
Room 6
43 frames lost from getting hit.
134 frames saved from a different strategy.
27 frames saved from wall jumping to gain speed faster.
Room 9
3 frames saved from corner boosting off of more places.
Boss
1 frame saved from corner boosting.
10 frames lost to aim for the goal of ending the game earlier rather than the input (4 frames faster in hitting the button).
  • 2002 frames in total.

Y05H1's comments

Opening comments

If it wasn't for my friend mindnomad, this TAS never would've actually started! He's a fanatic of New Super Mario Bros., and it was his encouraging that got me to try again at this game (I made an attempt to run this game a while ago, but it didn't turn out very well). After making some progress, we posted a WIP on TASVideos, and soon enough, we were all working to create this TAS. Regarding the level of optimization here, I'd say it's fairly nice, but I think it just needs that "push" to make it a run that will be very hard to beat.

1-1

An interesting level to start the TAS. This level has a 2-year history for mindnomad and I. Both before we even met and when we planned to make a TAS of this game together, this level was redone about 10 times (done throughout the period from repeated loss of motivation) due to our lack of knowledge on luck.

1-2

This was a fairly fun level to rush through. Learning about the super corner boost with the Blue Shell here made this level (and all the others in general) harder to optimize in terms of trying to manage speed. It's still pretty simple, though.
The pipe at the end gave us a bit of "delay" trouble, but luckily, terrotim was able to completely get rid of it.

1-T

It was disappointing to know we couldn't get the Blue Shell here instead - we were unable to make it through the moving blocks when we tried. It would've been a nice timesaver. I was happy, though, that we were able to save enough time to reach 465 as soon as possible. Unless another timesaver is found to reach 467, this level should be perfect. (This is the first of a few levels that have seemed to reach the time limit.)

5-1

I had fun with this one! terrotim had the idea of messing around as much as possible. So, we did so by repeating double and triple jumps (and cancelling them) while occasionally facing backward. Admittedly, maybe more could've been done with entertainment, but as it is, it seems okay. (I did want to face backward coolly without doing anything for the exact amount of frames we had to wait at the end, but I couldn't do so without messing up the beginning we have...)
Even with super corner boosting and the above-speed-shell trick, we couldn't pass 366 - we were off by 10 frames.

5-2

Shown here are the first few uses of the full-speed-off-ledge trick with the shell. The part with the many pipes required quite a bit of testing, since there are a lot of ways to go about it. Near the pipe at the end, I had an idea to break through the mass of bricks, but that was slower. (Ohhh, it looked much cooler, though!)

5-3

This level and its RNG... I had to edit this level so much to get the manipulation right. (No idea why this level was so difficult...) The other half of the level with the Snailicorns was more complex than 5-2's pipe part. A random midair jump is thrown toward the end. Oh, and yay, just one frame off from missing 268.
(Then, to find out that this level had to be redone because it had a timesaver... Attempting for remanipulation stopped me early at 5-1.)

5-Ghost House

How absolutely confusing the intro is. You hold B/A for just two more frames, and the Broozer above the beginning changes its path - slightly. That slight change leaves you with a loss of 2 frames from being unable to get enough height after ground pounding off of it. The part right after wasn't much better, either!
Well, anyway, the second room of the level had much time spent on it. I tried many tests with Y speed - whether a certain amount of frames holding B/A would be optimal, and if not, trying with a different amount of frames. Bouncing off of the Splunkin in midair near the end of the room looks cool.

8-1

A rather straightforward level. Corner boosts seem difficult to do as Mini Mario. (Well, at least to me.)

8-2

Finding the water surface boosting glitch surprised me. It just sort of... happened. I like the bonus room - lasting 2 seconds of game time! The wall jump to the pipe gave us delay, but I was surprised to have been able to completely get rid of it by holding forward against the wall. Pipe entry delay is such an odd thing...
With all the improvements throughout, we were able to finish the level with 2 frames to spare (to avoid 255).

8-Tower 1

At first I was overwhelmed by the thought of trying to manipulate timing of all the blocks ahead, but it turned out that the two big blocks near the end couldn't be manipulated, which made it so that I actually didn't even have to manipulate that much. That is, only in terms of the blocks themselves - yay, wall jump spam! terrotim mentioned that we needed a ton of RNG advancing to manipulate a Blue Shell in 8-3, so I had to perform as many wall jumps and double jumps as I could to make it... and I did, with 2 chances of advancing to spare.

8-3

This level was a bit of trouble because of trying to find out the fastest speed cycles for every water current, and because of trying to optimize the corners at the beginning. We made multiple attempts to pass corners by pressing and letting go of forward at random times. (We don't have any real way of calculating an optimal way to pass them, sadly.)
The pipe at the end was difficult, too - not only did we have entry delay, but X positioning had to be perfect to reach the point to start the fadeout transition as early as possible.

8-4

Taking a hit here looked "unclean" (I personally wanted it to look more obvious that we wanted to take damage). So does the Koopa shell grabbing, but there's nothing that can be done. It was surprising to find that using the shell to get the Starman was faster (without, we were about 10 frames slower).
Lots of screen scrolling toward the end to manipulate platform movement.

8-Castle 1

When I attempted this level, I was unable to do the big jump past the overhanging wall to the platform. (Hmm, not much to say...)

8-5

The way this level turned out makes me very pleased. terrotim found out that we coincidentally end up jumping in tune to the first part of the beginning loop of the song. I added in some screen scrolls to make it even more "musical." (I do wonder if these will even be noticed...) Aside from that, it, in general, looks polished as a run.

8-6

Vertical levels always seem to be the hardest to speedrun...
terrotim came up with the nice route for this difficult level. The boosting glitch with the spring pad and the wall has interesting properties and is confusing to pull off. At the last minute, mindnomad discovered that we could simply wall jump up to the last room. This luckily saved us even more frames thanks to better map manipulation!

8-7

mindnomad and terrotim found out about keeping speed on the cliffs, as well as staying face-forward. It looks a lot different seeing Mario triple jump "at the screen" while moving forward. Mario coolly gets past the Fire Bro here with a little screen scrolling, although it was actually unnecessary.

8-8

This level surprisingly seemed easy to do, regardless of hearing a bit about trying to manipulate rocks to not drop in the wrong places. terrotim did a great job at the invisible block, including wall jumping off of it.

8-Tower 2

This was quite a fun level! I was making this level as mindnomad was giving some entertainment suggestions to me, like having that jumping "seizure" in the random small hole 3/4ths up the tower. Thanks to terrotim for finding out about freezing Mario's animation after a triple jump. In the new edit, I added in having Mario's head in the ground for a bit like terrotim did in his attempt. I'm still not fully satisfied with my attempt at entertainment, however.
Saving time by dying in the level was a nice discovery - it's the biggest improvement of the whole run, after all.

8-Castle 2

The last level. I actually don't have much to say here. In the room with the giant Thwomp, we tried a different strategy by revealing the switch from above and performing a wall jump after flipping the room (by storing a wall jump off of the wall right before it flips). The wall jump off of nothing looked cool; too bad it was slower.

terrotim's comments

1-1

At first, I was surprised that using star power was actually able to be faster. The wall jump off the star block certainly helped this.

1-2

I really like the triple jump off the second teeter totter into the secret route. The Blue Shell corner boosts are shown to add extra height to jumps, which could be useful in the 100% run.

1-T

We actually saved enough time that we had to stall a little to avoid fireworks. Quite a few later levels needed tiny bits of stalling

5-1

Syncing actions to the music is always amusing to me. And Y05H1 did well in that.

5-2

Having Mario invincible while in a shell proves to be most useful in this level for getting certain boosts and avoiding certain enemies.

5-3

Enemy positions start becoming a problem from here on out. In this case, it wasn't too big, but the camera needed to be altered a bit to put the enemies in spots where they can slow Mario down the least.

5-G

I spent so much time on this level. So many possible approaches to this level because of the uncooperative Broozers roaming the level. Manipulating Broozer actions and positions became troublesome. But I really like how this turned out. Plenty of frames saved in this level. Also, the time saver in the elevator Y05H1 discovered is a pleasant surprise.

8-1

Using a crowber to corner boost was pretty nice. Other than that, not much else was done to improve the level.

8-2

Two more tricks here that surprised me when the level was completed. I must commend Y05H1 for finding these tricks and making the run of this level so fast that he actually needed to slow down a little in the end.

8-T1

Making the stone blocks cooperate with Mario was key to saving frames here.

8-3

The part before the swimming section is another part of this run I enjoy; so much is strung together smoothly in those couple of seconds. Interesting to note that the currents from the pipes underwater are higher than you think (making it faster to approach the pipes from the top).

8-4

Camera manipulation again is important in this level. Getting the enemies and platforms in just the right place took a bit of time. To be honest, duck jumping under a moving stone block was just a random idea that happened to worked.

8-C1

A few frames saved here. Not much else different here.

8-5

Who knew running up would make Mario run faster? I certainly didn't. But nevertheless, some frames were saved as a result. There is quite a bit of syncing to music here as well.

8-6

Another level I personally spent a lot of time on. There are not as much possibilities here as in 5-G, but there was still so many different approaches here, being a vertical level. There is also an interesting vertical boost that lets Mario reach springboards quicker.

8-7

Having Mario facing sideways while moving forward is a very nice touch to this level's run. There are a lot of tricks that can be done from not pressing any directions and this was one of the more memorable ones.

8-8

Making the most out of the star power's speed boost was key to this level's optimization. Wall jumping off the block with the star power is used again here.

8-T2

First, we find out that dying is indeed faster. Then it became a matter of "get to the midway point as fast as possible and die". After that, its pretty much a freerun up until near the end.

8-C2

One of my favorite levels to TAS. The level has Mario move in every direction. Saving frames in this level was actually very enjoyable and rather simple to do because each section is divided by many doors.

Known Improvements

In World 1-2, if we had more knowledge on catching walls and vertical wall jumping, a frame could probably be saved at the corner catch near the beginning by not holding back as much.
In World 5-3, 2 frames can be saved by using the Blue Shell trick for the first half of the level to stay higher than maximum running speed. This was found too late, however, and we were unable to edit the improvement in without also fixing the manipulation.
In World 5-Ghost House, we found other routes through the first room that were always ruined by the movement of the Broozers. They seem to be affected by how close you are to them. It is likely that we're missing certain paths through them to make them move in the desired way.
In World 8-Castle 1 and on, we are not sure if map manipulation is the fastest. We often counted how long timing takes by how many "flap" sounds the red ? block makes. However, when off-screen, no sounds are made. We thought this meant perfect manipulation, but we came across a few occurrences where we somehow saved more frames (even though there was no sound difference) that doubts this. (The transition in World 8-Tower 2 (after dying), however, is more likely a place we can save time, since there's a flap sound that plays there.)
(?) There are a few extra lag frames throughout the run. We have no idea how to prevent these without simply waiting. From testing, gaining them seems to depend on the frame you enter levels. For example, in World 1-2, if we entered a frame later, we wouldn't get that extra lag frame. Changing luck would not eliminate it. So, if we were to save another frame earlier in the run, we may be able to avoid this lag frame. None of them actually lost any frames, so it's unlikely that this matters.

Closing Notes

We're glad that we finally finished this TAS. It took us nearly a year to complete, starting from when we all joined together (but a lot of the time was given to stalling; in honesty, this run probably could've been created in half the time it took). We all are satisfied with the result.
Anyway, we hope you all enjoy!

Mukki: Judging...
Mukki: Accepting as an improvement to the currently published run.
adelikat: Setting to delayed pending a new improvement
adelikat: New movie updated, improves 8-6 and has a few frames of improvement in map movements between later levels as a result. Total of 118 frames were improved.
Aktan: Processing...


TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 15566
Location: 127.0.0.1
Editor, Skilled player (1536)
Joined: 7/9/2010
Posts: 1319
well, on the intro I think this is random, maybe you can manipulate this and save more frames
Favorite animal: STOCK Gt(ROSA)26Sortm1.1(rtTA,EGFP)Nagy Grm7Tg(SMN2)89Ahmb Smn1tm1Msd Tg(SMN2*delta7)4299Ahmb Tg(tetO-SMN2,-luc)#aAhmb/J YouTube Twitch
Editor, Skilled player (1439)
Joined: 3/31/2010
Posts: 2108
A highly anticipated 4-author submission for a popular game? Oh no, you didn't! Anyway, I watched it and I especially enjoyed the use of damage. Yes vote.
TheKDX7
He/Him
Player (118)
Joined: 7/9/2011
Posts: 392
Location: Switzerland
The pure entertainment, good work the guy votes yes
Joined: 3/10/2011
Posts: 136
Location: Inside the FCEUX emulator
Extreme epicness of Mario continues!!!! Yes vote.
Blah Blah Blah... Why I'm writing this?
Warepire
He/Him
Editor
Joined: 3/2/2010
Posts: 2178
Location: A little to the left of nowhere (Sweden)
Sweet improvement, good job. Yes vote.
Synahel
She/Her
Former player
Joined: 1/19/2011
Posts: 260
Location: France
Amazing. Sheer awesomeness. Thanks a lot guys for this.
Joined: 7/7/2011
Posts: 140
Location: Germany
Even my sister liked it a lot YES vote
How should I know what I think before I read what I post?
mklip2001
He/Him
Editor
Joined: 6/23/2009
Posts: 2227
Location: Georgia, USA
Very nice! 8-T2 was incredible to watch. I'd recommend that the screenshot for the movie use a frame from that stage when Mario's animation is frozen, such as when he's barely in front of the rolling spike ball (that part reminded me of "haters gonna hate"). Out of question, what did adelikat do in this movie? I see he's not in the comments now, and usually his runs have a link to a repository of savestates and such. Excellent work. Yes vote. P.S. Since the publications for this movie have been alternating between Mario and Luigi, maybe this should be published as a Luigi movie for the fun of it.
Used to be a frequent submissions commenter. My new computer has had some issues running emulators, so I've been here more sporadically. Still haven't gotten around to actually TASing yet... I was going to improve Kid Dracula for GB. It seems I was beaten to it, though, with a recent awesome run by Hetfield90 and StarvinStruthers. (http://tasvideos.org/2928M.html.) Thanks to goofydylan8 for running Gargoyle's Quest 2 because I mentioned the game! (http://tasvideos.org/2001M.html) Thanks to feos and MESHUGGAH for taking up runs of Duck Tales 2 because of my old signature! Thanks also to Samsara for finishing a Treasure Master run. From the submission comments:
Shoutouts and thanks to mklip2001 for arguably being the nicest and most supportive person on the forums.
Joined: 6/4/2009
Posts: 570
Location: 33°07'41"S, 160°42'04"W
No matter how much I listen to the NSMB main theme, it's one of the best songs ever. Oh, the TAS, right. That's good too. Very good. Gotta low how Mario jumps around. Easy yes vote.
PCachu
He/Him
Joined: 10/1/2009
Posts: 166
This is the Epic Hitbox Abuse run. Most notably that bit in World 5 where Mario hops up into a warp pipe while there's still a Piranha Plant in it.
Editor, Experienced player (931)
Joined: 7/20/2011
Posts: 345
I've never played this game, but that doesn't stop me from seeing this movie's brilliance.
Current thoughts: Hachiemon (J) for GBA.
NitroGenesis
He/Him
Editor, Experienced player (556)
Joined: 12/24/2009
Posts: 1873
Hey, whoever keeps adding any% as a branch can stop, it's generally useless unless a 100% is made. I've removed it.
YoungJ1997lol wrote:
Normally i would say Yes, but thennI thought "its not the same hack" so ill stick with meh.
Editor, Experienced player (570)
Joined: 11/8/2010
Posts: 4036
Wow! Very nice improvement! I was surprised to see the part where they used death to save time. And I noticed the jumping syncing with the music. Easy yes vote.
Joined: 11/3/2010
Posts: 22
Yes, yes, absolutely yes. Incredible run, loved the playaround in the moving platform levels. Without a doubt, worthy of the site, a no-hesitation yes vote from me.
N._Harmonik
She/Her
Joined: 8/27/2006
Posts: 502
Location: Canada
Impressive as usual. But is anyone working on a warpless run?
Why, oh, why do I even <i>try</i> to understand my own species?
Joined: 5/17/2007
Posts: 393
Location: Sweden
Nice seeing something "new, and 30 sec faster too
"No love for the game gear"
Joined: 9/20/2010
Posts: 80
N. Harmonik wrote:
Impressive as usual. But is anyone working on a warpless run?
funny, i actually wanted to do one, but i really havent done much with this game. i may make one in the future.[/off topic] i know i havent said something here but as i said on the vid, fantastic!
Current Projects: SPW3 any% (World 1 Finished 8/29), SMW 100% Glitched (Making a Path), SMW any% (Glitched?) (Testing)
Joined: 2/20/2010
Posts: 209
Location: I'm in space
Clearly another troll submission.
Oh, play it cool. Play it cool. Here come the space cops.
Patryk1023
He/Him
Joined: 3/1/2011
Posts: 288
Location: Inside out house.
Eeeeeeeeee... What to say... Oh well, that's DS, Mario is too epic and has too much submissions! But this is too good :D Yes vote.
<Nach> scrimpy is fretty with her sunglasses on I'm here. never visible.
Joined: 12/19/2010
Posts: 126
nice run, maybe it is possible to do this game under 20 minutes
live life
Joined: 5/2/2006
Posts: 1020
Location: Boulder, CO
Nice! big yes vote here
Has never colored a dinosaur.
Joined: 4/25/2004
Posts: 498
*throws another yes vote onto the stack* Nicely played, guys. That death skip in 8-T2 was interesting, hehe. :p
Joined: 7/5/2011
Posts: 13
Location: Mississippi
A definite yes. Pretty slick moves throughout the run. Also love what you guys did in the moving platform stage.
Former player
Joined: 4/7/2010
Posts: 181
Location: Taguig City, Philippines
Great Improvement guys... And great entertainment also. I VOTE YES!