Link to videoMovie file
Rerecords: 7530 (on this level)
This is .71, human world record is .55.
We just finished the holy trinity: these three levels are among the 15 hardest in the run, no doubt.
I may go and redo 1-4 but it's so frustating I don't even know if I'll go back.
Nice job. I wonder if that’s any proof that the TAS syncs on console, or if replays don’t require much in the way of syncing.
What I did was just save replay in emulator, export backup memory, then exported that on my microSD, got the microSD into my flash card (ACE3ds) and booted up the game with the exported memory from the flash card.
Replays aren't different from normal gameplay; technically, you could sync the entire TAS on console; and that would be a good way to test DS console verification, since Touch&Roll is the most complex game I know that has absolutely 0 RNG or outside influcences (ng+ saves don't affect gameplay, for example.)
We're moving on onto the worst fking cycles in game history Fluctuant Travelator. This is still world 1, and this is probably the second last annoying and hard level that will get in our way. It's not Curvature hard, but it's definitely up there.
Here is an overview of the level (the actual WR is 54.06:)
https://www.youtube.com/watch?v=n_jCHLSa1ZQ
As you can see, the human record has to wait for the right cycle. I'm currently trying to skip waiting for the cycle: although it is frustating as hell, it works.
There's one particular reason this level is annoying: it's the ONLY SINGLE LEVEL IN THE ENTIRE GAME which has a starting platform you cannot abuse to gain speed off. You see the red border on the starting platform? There's an invisible wall ON TOP of it, and I still have no idea why. As such, it is impossible to boost of it (but possible to corner boost of it, since it is a wall).
Link to videoMovie file
Rerecords: 1351
Shoutouts to cycles for making this level a pain.
Note that I take a wierd angle to hit the goal at the end, bouncing off the left goal pillar; it isn't faster than going straight through the goal (I tied), so I just did this for extra style points.
https://youtu.be/KTb4FLrEgfU?t=5m19s
Swathe is up next, and oh boy. Collision galore. (Go to 5:19 in the video for the WR, by Silver)
This should take about the same time as Fluctuant Travelator to make; expect it relatively soon.
This should take about the same time as Fluctuant Travelator to make; expect it relatively soon.
This was a terrible thing to say. Swathe is infamous for being a copy of B4 in Super Monkey Ball, except all the edges are bigger, which leads to completely different and extremely tough strats. I haven't figured out a good way to tie the human record atm thanks to stupid cycles.. so don't expect it out soon.
This is the only level in the way of World 1 (along with Spiral Staircase, which is really hard). Hoping to finish this in two or three days.
Link to video
Just before we hit 30 000 rerecords, Swathe (finally) decides to cooperate. This new start has an insane boost on the left of the start platform. The bounce will have to be lowered a bit (for faster speed) but it's really looking solid right now.
I'm going to throw this out there: I am desesperatly in need of help (mostly for bruteforcing stylus inputs and strategies.) If you are interested in the run and game in any way, please contact me on the forums or on the IRC channel of TASvideos.
Link to video
There's definitely something here.. but we're talking about Swathe. If this trick could be implemented, it would definitely speed up the roll on the bridge to a 85+ speed and save quite a few frames.
I'm currently heavily investigating (AKA trying random stuff)the goal clip because it manages to defy every single ounce of good collision and logic this game has.
If you are interested in seeing what's going on, here is a frame by frame breakdown of the clip, up to the frame where we goal (and end with a .73.) Note that we are holding (255,0) aka up-right the entire time.
Album for your viewing pleasure
help
Small post for small progress: This TAS is going to be restarted (for the levels that matter, which are Progenitor and Curvature). The reason being is that boosting is horrible on Progenitor and with some tweaking, I'm sure I can get the Fladdermus bounce, which would save close to 45 frames (which is huge!). Just bruteforcing the inputs for tweaking will be the hardest part, as just optimising tweaking will take about 2k rerecords; but the hardest part will be adapting tweaking to have a position to the right when entering the position for the Fladdermus bounce to be executed, and I'm going to spend a lot of time there (up to a month if needed, though it shouldn't take more than around two weeks) because I want Progenitor to have the Fladdermus bounce.
For Curvature.. Many Super Monkey Ball TASers cannot support the thought of having an unoptimised TAS versus human records, even if the human record is the most retarded and luckiest collision in existence. This is why I will fully restart Curvature, engineer strats for it, and break the hell out of this piece of garbage that's called a level. The hardest by far is optimising falling: it's 56 frames you have to independatantly optimise, and just for that, I'm considering around 230 'viable' inputs per frame, which will lead.. to a lot of bruteforcing, pain and tears. Going to devote up to three months on this one (we've alderaly spent two, and I honestly don't want to stretch the pain), but after that, I'll just let the 3 frame slower version slide (for now..)
Note that I will spend a LOT of time on making sure I get a perfect down to input boosting. Boosting determines everything in this game, and it shapes up your whole strat: it's actually more important than the rest of the level. Curvature has perfect boosting, Progenitor I'll play around a lot with.
I've found a lot of useful RAM adresses, notably Aiai's facing position and his 'inertia' (aka direction he is going to be ejected to for around 30 frames before being locked to that direction). This adress is active only in collision and when boosting, and is the reason boosting exists.
The branch name for this run is going to be "All Worlds, 230 exits". The game doens't have a defined ending (it gives the same at the end of each world); as such, a normal completition of the game is finishing all 12 worlds. The main problem is that the 12th world is locked behind a grindwall of 2006 bananas (stage collectibles). Thankfully, we can get 1900 by just redoing the first 11 worlds and collecting all bananas, leaving just a very small grinding portion.
Essentialy, this run is - almonst - a 100%; it collects all bananas (except World 12 bananas); and plays through the entire game, though the Touch&Roll community AKA me and Silverboxer doens't see it as a 100%; it doens't fill the requirements for it. Feel free to discuss about it.
One last note: I have 230 levels to play through. Have fun seeing this TAS in five years!
Project Masochism has begun.
Dang, that sounds insane lol. I've always thought Super Monkey Ball games looked absurd to TAS so I tip my hat to you sir.
I'll be following your every WIP as I have been for the last 4 months and I look forward to your progress with this game. :) Good luck with Progenitor and Curvature!
I honestly think this could become a Star worthy TAS.
We're starting on Fluctuant Travelator. The idea for this new strat is to clip to the left side of the platforms for the first 'wave', to avoid having to slow down as much. This pushes you up continously.
But before, the wierdest clip of all time.
https://youtu.be/rhV41kChfxk
Whoa! I haven't been working on this for a while (other than the reason that TASing this game sucks); I currrently have two short other projects running: a secret GBA game and Altaïr's Chronicles. Progress will be slower while these two are running, and I expect being able to be able to get back on this by full speed once these two are done, which shouldn't be long.
What is left of the current Fluctuant Travelator strat? First off, this is TAS-only, since it doens't wait for the cycle like the human record does. Only two people have attempted to skip the cycle and succeded: they both have a time .5 slower to the current TAS.
As such, I have to create ridiculous strats pretty much on the go, with my only reference being year-old human records and Silver's records.. since some of these strats wildly vary, this is incredibly hard in some occasions (see: Curvature, Anaconda)
The reason this stage is ridiculous is because the red platform.. actually has an extra invisible barrier on top of it. Have absolutely no idea why anyone thought this was a good idea, but it does make our life a lot harder..
There are a lot of things I want to investigate:
- boosting left for one extra frame. The entire strategy resides around being able to clip off the right border with a really hard angle and be able to get enough speed to skip the cycle (but we still have to slow down in some occasions.) While boosting any more extra left would give an angle too wide and ultimately waste time, I've found that boosting left for one extra frames gives you just enough speed so that your angle isn't too wide, and may give me some extra speed upon exiting the platform. It won't save one frame by itself, but the increased speed throughout may save one frame or even lead to a clip on the first series of platforms.
- getting a clip off the first series of platforms without waiting. The human records wait for the cycle then get a clip because the platforms are well positioned, and I do it in the TAS at the second to last series of platforms, which saves one frame. Hovewer, a clip that could be gotten earlier could completely turn the level on its head and save a ton of frames.
- clipping off the far left of right of the platforms. I'm inspiring myself from the M1 strat on SMB1 here, but the far sides of the platforms actually have different collision, and I'm almonst sure a clip off them is doable. The edges are very steep and can make for something interesting.
A lot to try, a lot to route. I'm more optimistic on this level, as it's relatively fun to TAS and doens't feel like you're slowly killing yourself.
Congratulations! Seemed like quite the task, hope the rest of the levels go smoother.
Hooray! Hoo-zzz..
help