This beats the current published run by 57 frames (0.948s). The whole timesave comes from the use of a newly discovered pause buffer strategy in the Ultimate Race.

Objective:

The objective of this TAS is to complete all the levels on single player mode as quickly as possible. It makes use of skips to save time in both Silly and Ultimate races. No glitches were used.

The "Kukie Skip"/"0 Cycle" strategy.

Backstory:

So in November 24th of this year, I was practicing the Ultimate Race on stream so then I could do any% speedruns. Then I had the realization that I was really close from skipping a cycle in the last bit of the level. Later in that same day I went to test out possible ways where I could theoretically pull of the cycle skip. I found out that it's normally not possible to do that, tho I had the idea of pause buffering so then I could in theory delay the cycle to give me enough window to squeeze through and save time. It just so happens that it works.

How it works:

By carefully timing the pause and unpause inputs, the game will not advance the state of the bridge momentarily. If carefully timed, there is more than enough time to get to the finish line earlier than it was intended.
Tho, there are some limitations to this mechanic:
  • Good timing is required to have a good enough window, so you can't just do it every frame.
  • You can't make the game not advance the cycles of platforms that are right in front of each other. That means that you can only delay the platform you're on, the one platform that comes after the one that's in front of you, or the platform that stands after a checkpoint platform.
As of now, the Bridge in the Ultimate Race is the only place we know where this can be used to save time. You can't skip cycles in places like the lifting bridge in the Intermediate Race or the Catapult in the Aerial Race. The most you can do is just delay the cycle by 2 frames on the Intermediate Bridge, which is not enough since pause buffering your way to the bridge will be way slower than just waiting for it to go down.

Observations and conclusions:

Most of the gameplay in this TAS is from Aglar's 2:43.01 submission, so I consider him an author of this TAS as well. That's due both to me using his TAS as a base to look for things, as well as to how this game works since all you can do is move around (either not move, move slowly or move quickly). That makes it so if I was to change the gameplay it would not save time at all and it would not be any more entertaining, plus it barely would look any different. Tho I was able to change some of the gameplay in both Intermediate and Ultimate races.

Can any time be saved anywhere else?

As of now, no one knows. There's a strategy in the last bit of the Intermediate Race where you can fall from the carpet directly to the finish line platform, tho it doesn't trigger the level end because it's way too offscreen. So it cannot be used to save time in the TAS unless if some way of doing it offscreen is found. It took a good bit for this to get improved again, but in the end it did get improved. So we never know what to expect, only time will tell.

Special thanks:

Thanks to the Marble Madness speedrunning community for basically teaching me how this game works and how to speedrun it. And thanks to the authors of the previous submissions for finding all the other timesaves.

Samsara: Losing my marbles, gaining a submission to judge.
Samsara: Nice improvement! Really clever thinking to pull something like that off in a game that felt like it couldn't be optimized any further. Accepting as an improvement to the published run!
feos: Pub.


TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 15583
Location: 127.0.0.1
This topic is for the purpose of discussing #6945: Aglar & LeKukie's NES Marble Madness in 02:42.07
Challenger
He/Him
Skilled player (1689)
Joined: 2/23/2016
Posts: 1061
Nice improvement. Yes vote.
My homepage --Currently not much motived for TASing as before...-- But I'm still working.
tormented
He/Him
Editor, Player (227)
Joined: 9/25/2015
Posts: 37
Nice to see a new trick pop up in a short and already well-TASed game. Yes vote. Given that a large chunk of this run is the exact same as Aglar's, I think it would be fair to credit him as an author as well.
Elipsis
He/Him
Joined: 11/11/2014
Posts: 7
Location: Pittsburgh
Definite "yes" vote. I can't fault the author for not improving any other levels, I'm pretty sure that Agular made them optimal already... and given the game's 16-frame frame-rule, single frame improvements aren't going to matter anyway. I know the author tried to use the "hot drop" on Intermediate race for awhile during a draft of this TAS, but unfortunately it only worked with the camera hack enabled.
LeKukie
He/Him
Player (87)
Joined: 11/11/2020
Posts: 4
Location: Brazil
tormented wrote:
Nice to see a new trick pop up in a short and already well-TASed game. Yes vote. Given that a large chunk of this run is the exact same as Aglar's, I think it would be fair to credit him as an author as well.
I'll add a bit where I talk about aglar's TAS saying that he is an author, since I can't reupload the .fm2 with the metadata changed (as far as I can tell). If at some point the moderators change the metadata on the file itself i'll remove that bit since there's already a paragraph talking about how this submission majorly relying on Aglar's submission.
Spikestuff
They/Them
Editor, Publisher, Expert player (2643)
Joined: 10/12/2011
Posts: 6438
Location: The land down under.
If you want to place Aglar as the author just go to the Player Field in the submission edit and just do LeKukie & Aglar.
WebNations/Sabih wrote:
+fsvgm777 never censoring anything.
Disables Comments and Ratings for the YouTube account. Something better for yourself and also others.
Emulator Coder, Judge, Experienced player (729)
Joined: 2/26/2020
Posts: 783
Location: California
Most of the gameplay in this TAS is from Aglar's 2:43.01 submission.
You should probably add Aglar as a co-author then (you can edit your submission to do so, on the player field). Anyways, nice job, yes vote.
LeKukie
He/Him
Player (87)
Joined: 11/11/2020
Posts: 4
Location: Brazil
CasualPokePlayer wrote:
Most of the gameplay in this TAS is from Aglar's 2:43.01 submission.
You should probably add Aglar as a co-author then (you can edit your submission to do so, on the player field). Anyways, nice job, yes vote.
Oh nice, thanks. I guess I'll keep the bit I said I was gonna remove since it's better for it to be there anyways.
Evan0512
He/Him
Joined: 10/18/2017
Posts: 171
Location: San Francisco
No vote. I looked at this video. The movie stopped at least 60 frames before it reached a goal line of the ultimate race. There were some "negatives" I knew about this:
    On the beginners race, you bump the black marble for at least 3-5 seconds, as I said it's not quick. On the ultimate race, there is a pause button in the platform, as it appears within 60 frames. I don't think the pause button will help. The screen hides the marble for at least 20% of the run. I don't know the region differences of the game.
I think a 2:36 or 2:37 run is good, with several issues.
Planned runs: Marble Madness (Arcade, Genesis J, GG/SMS) Proposed: Ecco the Dolphin (Genesis, GG/SMS, CD: regular, camera freeze) Ecco: The Tides of Time (Genesis, GG/SMS, CD; normal mode) Mario Kart DS (all cups, all missions) In progress: Grand Poo World 3 (all exits hard mode)
Spikestuff
They/Them
Editor, Publisher, Expert player (2643)
Joined: 10/12/2011
Posts: 6438
Location: The land down under.
Evan. "Did you find the movie entertaining?" is what's being asked not "is it optimal?". Also movies ending input early is a pretty normal thing here... what are you on about?
WebNations/Sabih wrote:
+fsvgm777 never censoring anything.
Disables Comments and Ratings for the YouTube account. Something better for yourself and also others.
Evan0512
He/Him
Joined: 10/18/2017
Posts: 171
Location: San Francisco
Spikestuff wrote:
Evan. "Did you find the movie entertaining?" is what's being asked not "is it optimal?". Also movies ending input early is a pretty normal thing here... what are you on about?
I found that it is somewhat entertaining.
Planned runs: Marble Madness (Arcade, Genesis J, GG/SMS) Proposed: Ecco the Dolphin (Genesis, GG/SMS, CD: regular, camera freeze) Ecco: The Tides of Time (Genesis, GG/SMS, CD; normal mode) Mario Kart DS (all cups, all missions) In progress: Grand Poo World 3 (all exits hard mode)
LeKukie
He/Him
Player (87)
Joined: 11/11/2020
Posts: 4
Location: Brazil
Evan0512 wrote:
No vote. I looked at this video. The movie stopped at least 60 frames before it reached a goal line of the ultimate race. There were some "negatives" I knew about this:
    On the beginners race, you bump the black marble for at least 3-5 seconds, as I said it's not quick. On the ultimate race, there is a pause button in the platform, as it appears within 60 frames. I don't think the pause button will help. The screen hides the marble for at least 20% of the run. I don't know the region differences of the game.
I think a 2:36 or 2:37 run is good, with several issues.
I stopped the movie 37-38 frames before the Marble crosses the finish line. The previous submission also stops the movie a couple frames before hitting the finish line, so I opted to do the same. Bumping the black marble loses no time since you gotta wait for the bridge at the middle of the level to go down anyways, so I used that knowledge to mess with the black marble at the start. That's what I changed over the previous TAS in that level pretty much, since I knew the beginning of that level is the one where I can do something. I did it with the objective of it being entertaining (At least I find the way I bounce on the marble funny). Plus if I was to just wait at the bridge it would be WAY more boring that what I did. This TAS ties Aglar's TAS in every single level, and it saves time in Ultimate. The pause buffer is pretty much the ONE reason the TAS got improved in the first place. Each pause is 44 frames long and that is 57 frames faster than waiting for the normal cycle. It's not possible to do the skip without pause buffering as I stated in the submission. About the screen hiding the marble 20% of the time in the run: I unfortunately can't do anything about that. The game was programmed that way, and there's no known glitch or exploit to circumvent that behaviour. If this was an actual big deal this game wouldn't even have been submitted here. I'll probably make a re-encode of this movie using the Camera Hack luascript so then whoever wants to understand whats happening can watch that.
Joined: 6/4/2009
Posts: 893
Evan0512 wrote:
I looked at this video. The movie stopped at least 60 frames before it reached a goal line of the ultimate race.
hi, here at tasvideo, we don't use RTA timing for our submissions but "from power up to last input"; so it's pretty common to stop the timer the moment the last luck manip is done (eg :http://tasvideos.org/1899M.html) that's also why the time here is 2:42 ( from powerup) while the WR is 2:40 (from the first race start) as for race to race comparaison with the WR they are on par (at least for the 2 first levels) but knowing aglar (the orignal submiter this improvement is based on) those are optimals, but of course and as always, feel free to try to improve this submission if you can. anyway, nice improvement, yes vote
Alyosha
He/Him
Editor, Emulator Coder, Expert player (3822)
Joined: 11/30/2014
Posts: 2832
Location: US
Too bad this trick is only applicable at the end, but an improvement is an improvement, yes vote. Nice job once again on finding this.
Lord_Tom
He/Him
Expert player (3144)
Joined: 5/25/2007
Posts: 399
Location: New England
Yes vote; I love a good pause glitch and liked tangling with the enemy marble. I would be interested in a camera hack encode.
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. ---- [4349] NES Marble Madness by Aglar & LeKukie in 02:42.07