Sheep
Skipping a part of Rolling Coaster in Super Mario Galaxy 2 (Rolling Coaster Skip)
updated
This is ~1s slower than Luigi's TAS and a bit less than 3s faster than current IL (ILs have a ~12s time advantage over cutscenes runs). This is also 16.16s faster than previous TAS, done by HIME.
This level follows mostly what azulon did in his Luigi's TAS: youtube.com/watch?v=WLn61dSCwMY. That same TAS follows what usernameispassword did in his own TAS, without the improvement azulon found before Bowser and without my own strat, the quicker hits for Bowser (youtube.com/watch?v=6VMwtusRyK4).
The start was really tricky. I got really fast slope jumps, but everything would miss either the checkpoint, either the death area (dying by the black hole is slower and even with a lot of speed, not getting the Death Area on top of me would be also slower). There is possibility to get sloped really quick from the pipe at the left, then doing maybe a really fast triple jump, but that would probably be slower than this - hard to tell, but I couldn't get it to work. I tried getting the Checkpoint Area from below me as well, but reaching the pipe is impossible or I would die before grabbing the Checkpoint Area.
What follows after I respawn, until Bowser, is mostly what azulon/username did in their TAS. The improvement found by azulon was used here too, but it's after the 2D area.
For the 2D area, it was possible to get a big momentum from the platform under me, but it was ~1s slower, sadly. Entering the top area this way makes the backward momentum possible - otherwise, it doesn't work. I keep crouch activated when I land to keep my speed and get an instant long-jump.
The area after is where I lose ~1s over azulon's TAS, as Mario needs to get on the fire throwers to clip back to the fence. Not going that route was always slower and while it was possible to go on the fire throwers by clipping a bit inside them, it takes a bit too much time and was, as such, at least 20f slower. The way to go to Bowser is also hard, because with Mario, you are required to get a better angle than Luigi can take. I tied three times, but this attempt had -1 Coin, and while I'm pretty sure we don't care about that, I decided to keep it.
The Bowser fight is just my strat - a way to hit Bowser's tail way earlier (it has no name, so we call Sheep's Bowser Strat lol). While it is possible to do it while making Bowser stuck on crystal and the green things, it is overall just better to do it this way - I also make sure that no crystal or green platform is broken, as it would be losing 2f for no reason.
Lastly, for the first hit, I lost sometimes 1f if Bowser went to the right or the left. I guess he lands 1f slower...? Not sure.
As for improvement, it is possible to be faster with Bowser if he lands on a crystal after you spin him (should work on any phase). It requires an insane manipulation with this strat, so I was not up for the challenge, even tho I did try.
Again a Mario TAS because I was bored. I might TAS Good Egg 3 to see what time Mario could get (kinda), but I probably won't TAS too much of Mario levels.
I wanted to TAS something today, so I decided to do a level no one ever did - Gold Leaf S.
The 2P slope jump at the start is faster than anything else, but I need to do a long-jump on the slope at first. I wish I could keep speed from it, but sadly, it wouldn't keep the speed.
A route by doing only long-jumps and doing one on the mushroom was 1f faster than being on the water, keeping crouch on to go far, and long-jumping inside the bubble. Sadly...
I do another strat that I thought about, and that's to activate the bell really early. Even though I enter the bubble later, it doesn't matter, because I'm so fast normally that I had to slow down in the bubble, like RTA does. Here, I am able to keep my speed (except a bit at the start, but that's normal), which is faster by almost 100f (90f is 1 second and a half).
I was able to save 3f by going inside the bubble earlier. I might try to improve it on the side, sub 28 has to be possible!
I'm not done yet with Kitchen (Dome 3), but I'm making some progress. After what I'm doing right now (a level), it will be only 2 Galaxies left before moving to the Bedroom (Dome 4), which I'm hype about!
This is not a "level showcase" like I do when I'm done with a Dome, though. I have NO idea what level I will showcase, as I don't want to really show anything lol. I might only do a "half-level showcase" instead, but I'm not sure yet - I have my idea for what I want to show, though ^^ I will see with the two other levels I'm doing.
But you can take this, as I did a video in the past showing that I couldn't have added Fiery Dino in another level - we foudn the issue and here he is! The gimmick is a bit different than what I would have done for the other one, but it's still really good in my opinion ^^ (Video in question: youtube.com/watch?v=98nJ-Lg0dPs)
I continue working on Kitchen! See you, hopefully, soon enough for more Red Sun! :3
The route was changed a liiittle bit compared to what it's often done. I have nothing to really say though, but most of the things were thought about - I should be able to improve this in the future.
It might be possible to skip the Life Shroom, which could save ~2s - this would make this strat way more "viable" (as a time save). Sadly, I have no proof of that working.
I did that strat pretty long ago, a bit after Sheep Boost, so almost 1 year ago. I did this RTA and I even skipped a day of uni for making this (lol). This was just crazy work.
The wall jump on the rotating platform is really tricky as you can't go too high with how many death areas there are. The issue with going too down is that you get taken by the gravity right under you, of the drill planet.
There is an insanely big gravity that puts you towards the Drill Planet, which is impossible to escape (its radius is like 10 000 000), so no matter what I could try, I couldn't escape it.
There is the possibility to go even faster by going down really quick after taking lava damage, but you die after a bit because you are in the void while being on fire - that means that sadly, you can't get a GGen Skip. Any other possibilites are either ruled out, either tried. It is possible to get kinda low, but even on fire, you will get grabbed by that gigantic gravity and even if I didn't die, I would get trapped.
This is the best I could do. I am sure you can improve this strat, but I doubt it will be ever fast enough.
I was messing around while I was bored/tired tonight and I decided to check Bigmouth, a level azulon had an idea with: going on the top so you could take the shell from there and not to have to deal with the brick blocks. Sadly, you need to go insanely high and while it's possible with slope jumps, I don't know how fast you could do that (after some tries, it seems too slow). After checking Whitehole, I saw that one Water Area extended quick down, but that you could probably reach it without touching the Death Area. After few tries, I was able to finally get it! This is BOTF - Bottom Of The Fish. It saves 2~3s from what I can tell compared to normal way, but it could be more - hard to tell. It's pretty simple and the surface swims are pretty fast - swimming also is pretty fast overall.
I have to slow down at the top because if I continued, the shell would break. Not going far enough, though, means that after the jump with the shell, the leap in the air, I would go back to the same Water Area and it would be too slow.
The ending is annoying as throwing the shell earlier makes it homing to either the coach Penguin, either the koopa that's walking. I tried some sorts of ways to make the cutscene with the toad stops, but nothing, sadly, worked.
Backflip saved 4f I think?
1:06.XX easily reachable - 1:05.XX would be harder. If there is a way to use the slopes on the side and get an insane 2P slope jump (I can promise, I tried, but nothing goes far enough), a time close to 1:00.XX is reachable - a sub 1 would be kinda insane.
Again, this is more of a LOTAD - a "not so optimized" TAS.
- Egg Skip can be a tad faster.
- King Kaliente might not be as good as my Boss Blitz C TAS.
- Nothing to note on Major Burrows.
- Bouldy's RNG could have been better in first phase, as the second rock is the bomb-boo and not first one.
- Nothing to note on Fiery Dino.
Overall really good, but I am sure a 2:35.XX is reachable!!
While this one technically has not a lot to say, I still can say this much:
Skipping the cutscene is a bit weird as the IGT starts the moment you spawn. It means instantly skipping the cutscene like you would do in SMG1 is not as good in SMG2 - you need to delay it by a bit. Normally, you can wait a good amount of time before doing it, but here, it was different - there is a cycle, which means the moment you need to skip the cutscene is precise. Every 1f you wait to skip the cutscene isn't 1f you lose in the overall the cycle - it means I could always reach the ending the same frame even if I delayed skipping the cutscene by, like, 5f. Though, after a bit, you start losing frames as there is a "point" where skipping the cutscene doesn't change anything: you can start running at the same moment. I delayed the cycle the exact frame where I could start running instantly. Crouching to long-jump 1f faster (which is done a lot of times) was 1f slower here sadly.
The cycle of the chomp is, from my tries, unskippable. As close as I was, even with doing everything perfectly, I just couldn't skip it. You can reach the big lava wall on the left from the starting platform, so no lava boost can be done there. The "lava wall" that on the ground is just to be pretty, as the hitbox of the lava is flat - you can't lava boost there. As such, reaching the last platform in time is not possible - I'm at least 10f too slow for a perfect cycle (just an estimation, maybe I'm actually just 5f too slow, idk). I tried not skipping the intro, but it would genuinely do nothing - to save some time in the video, I decided then to skip it the first frame I could.
The most insane and important part is here. You see, the moment you freeze the Golden Chomp actually saves or loses time. While it may sound crazy, this IL actually saves 3f from what it should be ONLY because I 2P freeze the chomp a certain frame - or at least, a certain point on the railings. For some reason, around that curved path, if you freeze it a certain point (seems like around the middle, just a tad before?), it will without any reason be 3f faster. This has nothing to do with 2P freezing it a shorter amount of time - or if it does, then where the chomp is is all that matters. After all, freezing it while being further or closer is the same: as long as I freeze it the same exact frame, at the same exact part of the path, it will save 3f. If I freeze it 2f later, I lose those 3f. It is possible to win 2f or 1f a bit further on the path, and then I could lose 2f if I freeze it really close to the platform. While there are some theories, nothing really explains this. username theorises that the chomp being on curved path makes it gain a bit more speed when it's unfrozen - this would make sense, but the fact that only one frame it works and that I can still win 1 or 2f at some other parts that aren't curves raises questions. I have no idea how this works.
Whatever - those 3f allowed a 46.7X - I started with a high 46.8X and I didn't think I could save time. Good surprise! While there are potential lags with the enemies being on screen, this doesn't affect IGT - same with star bits.
Finally, the end was done just for a little flex - this didn't save or lose time.
This level was done only because I didn't like having a 51.28 on the SMG2 TAS sheet. This is an IL and can not be beaten in a TAS level that imitates a 120/242 or an Any% run.
While I'm more TASing and working on SMG1, I might TAS some levels in G2 again just to improve some levels or make our sheet even "prettier".
Decided to chill and to TAS while chilling so I decided to TAS a chill star. I optimized the ending as good as I could - I saved ~0.7s in that last section alone.
The problem with the start is that two frame perfect long-jumps back to back weren't enough - I sadly would be few units away from the slide, meaning I would get on the edge of the collision in front of me - either I would let myself get sloped (yes, it's a slope.) and it would be slow, either I would do another long-jump which was too slow as well. I had to delay my long-jump by few frames and barely make it on the slide. After a sloped jump that allows to go down without sliding, you can enter the water.
There is a trick where you can enter the water even faster, but it would require you having a spin - which I didn't have due to the sloped jump. Not doing the slope jump would be slower no matter how I want to see this, so... There was a lot of ways to enter the water, but doing two spins like that was finally the fastest by 5~10f than any other thing I would do - if there is an improvement somwhere, it's probably there.
Anything in the water was... well, it's just you riding a shell. Just move the less possible and that's it. I tried using some parts of the map to see if I could gain a bit of speed forward on them, but nothing seemed relevant - still, maybe you can.
The ending is, like I said, where I optimized the most. I tried for 20 more minutes, but I could NOT beat my time. I could tie it, but not beat. When you arrive at the end of the water, multiple things:
- Being higher was always slower. Being the lowest possible was the best, even if too low could lead to bonks or being sloped (which I will talk about instantly).
- About that "slope", you can actually get sloped if you hold a bit back. It's nothing great, even if it first saved me 0.13s. Sadly, it makes you stay in the air longer - while it could have been great with a good angle as I could throw the shell and get to the star, I just couldn't do it from where I standed - maybe from the other part, on the right, and not on the left? The Star is behind the wall so I doubt it could be ever be faster or useful.
- Holding towards the Launch Star would make me leave the water by more than 6f than holding another direction - holding up was also useful, so I held up-right here. When the camera switch (or like, 2f after), you need to start holding down and a bit right to get to a right position. You aren't "stun-locked" on the floor, making you walk fast enough to be... well, not slow. I had to cut the corners like I could to get to the star.
- You can technically grab the star 1 to 2f earlier if it turns a bit (as all stars do) - I do not know if this could be manipulated in any way as it is not shown in the intro cutscene or fly-in.
Way more to say about this than I thought. This will be beaten by PurpleSun whenever the latter will do this level anyway, but if we can't TAS levels that were already TASed before, then we would run out of stuff soon enough.
If anyone dares to say "D-Does that mean that you will TAS a 242-" I swear to god I'm deleting your comment (me when I'm ragey).
This was requested by Pablo (@pablodiaz-violin9513) to see if this could be used in Any%. As I was doing nothing today, I thought it could be a good idea!
From what I tried, it is impossible to do a frame perfect jump upon landing. No other TASes is able to do it neither, so it is what it is (I'm pretty sure all levels in Sea Slide don't let you fpj). I was able to get a star bit on top of the middle egg planet a moment, but I overwrote that save state and I couldn't do it anymore. With that star bit, it might be possible to not have to do a spin cancel long-jump, which could save 1-2f. I have no idea how I achieved it, but it is possible by agitating your cursors on the grass on that planet - I tried for 5 more minutes but just couldn't do it.
Going in the water the quickest by going on the right was ~30f slower than this attempt. I was able to surface swim to the right of the stone gate, but it lost ~20f for some reason. On the land, I was able to improve by 24f the path to the Luma with a good 2PLJ - it conserves your speed, making the jump after quick enough to be better than anything else. Doing only one long-jump and 2PJing mid-air to get on the platform where the Luma is was ~30f slower as well.
For the Luma, I went the earliest I could to the cannon - maybe you could get there even earlier, but that seemed almost impossible.. I wonder what would happen if you could? It wouldn't cancel the Luma cutscenes, but it could be funny to see.
The Hurry-Scurry planet had tto be redone mid-way as I messed up my routing. While some frames could be saved there, I cut the corners as good as I could. From the cannon to the Star Grab, I'm faster by 6,483s over the current best Luigi IL WR (by Payduck).
I wasn't planning to TAS this level, but as I want to work with RTA runners/routers, and that Pablo asked me, I thought it would be great to do it while I was doing nothing (even if I need to mod and write and draw and... Oh... Oh god...).
Enjoy this one ^^ ~3h to TAS it fully.
Long-jumping at the start is faster than anything else by at least 10f - that's all I really have to say about frames and stuff. Though, it is possible to press Z and go down really quick when the boss starts, which I do.
This is technically a LOTAD I did to check if there was any improvement possible to make Bonefin an Any% candidate.
The new strat here is to actually take the shell the latest possible after the first hit, but not actually getting it right away - you want it to follow you, take another shell and hit Bonefin with that same shell. While it actually doesn't save time, the timesave is right after: as I took that shell in front of me for the second hit and I hit Bonefin a third time with the shell I took from the first Phase, I am able to go back to that same shell under me and throw it on my left to hit Bonefin really early. With that, I'm able to gain ~4s over the fight itself in Tibs' current IL WR (with Mario).
This is almost faster than current Luigi IL WR on this level, but I'm slower by a bit less than 1 second. As I have cutscenes to imitate a Speedrun, I lose ~12s over any IL.
Probably a TAS only strat, but it's hard to tell sometimes. Maybe you can do this but slower for RTA? idk
This was hard. Clipping is extremely precise, as it's probably the only way to go inside the cage - spinning to join the ground or wall-jumping inside the cage are both impossible deeds. With that, I saw that you could get higher with a bonk (I knew that since, like, even before strat-hunting, but still) and thought about getting inside the cage that way.
And there you go, it is possible!!
I would guess that this saves ~9s. It also means Sky Station 2, for a 120 TAS, would be rerouted to Luigi. I tried this already when I started my 120 TAS for SMG2, but I never got it - this is a good timesave for the TAS, but I'm more glad to have finally done it than anything.
Finally!
This could be improved by 2 - 3s with better movement and not being a scaredy-cat at the end; I failed three attempts there, so I got scared. it is possible to 2PLJ to the Sling Star with Spring Mario, but I didn't feel like doing that.
Uses my strat that saves ~5s on the second planet by doing a triple jump and then a backflip - requires a Wiimote spin so you can reach the platform on top without the need to wall jump.
Nothing else to say - not taking the Spring Mushroom is just too slow.
Previous WR was by Sp4ce ( • *WR* Super Mario Galaxy - Watch Your Step ... ) who is ~6s slower than this.
Decided to TAS this level after watching PurpleSun's TAS on it (Mario): youtube.com/watch?v=_Ivqaw5BtdI
I decided to TAS this because I knew I could do it faster, even tho I'm Luigi and should be slower - though, until the end, I'm faster by, idk, 0.5s? I then lose some of this advantage by the last long-jump, the one to the star, as Mario as the advantage there.
Nothing I can say that PurpleSun didn't say. The reason I am faster is due to the first hit being way faster - not because PurpleSun didn't do it frame perfectly, but because Major Burrows is faster for me than him.
If this worked, it wouldn't save more than 5 seconds. From what I can see here, it probably saves... idk, 3s? 2s? Still, timesave is timesave :)
It's called "Pamsy Skip" because it's my girlfriend who thought about this (and her username is Pamsy)!
The TAS starts the frame I press A on "Play This File" and ends the first frame the cutscene of the last Star I get starts. This TAS was done from power start, but in the mod itself, the files were already there - this is why they are all there. Unlike my 120 Stars TAS for SMG2, this one was done without any editing, but only dumping frames/audio.
I started this TAS after seeing people asking me again and again for a SMG1 TAS. As I didn't want to simply do a SMG1 TAS, I decided to mod this one as it changed enough the game to be able to feel like something else. I was wondering about TASing something like Blue Eclipse or Cosmos Collapse, but the problem with such mods and TASing is that they often require a certain path (Blue Eclipse even more), meaning it wasn't that interesting to do. While it could still be great to see, I am not planning to ever TAS such a mod.
There aren't too much to say about the TAS otherwise. It is optimized, but you can probably improve some stuff there and there for sure: there are maybe minutes of timesaves, but I didn't want to optimize this to an extent where I would take years to do. Still, it is really optimized and I went for every strat I could except the ones that I couldn't do. Some strats aren't Mario doable neither, like you probably can guess. I am not going to explain every tricks I did, but if you are wondering about some stuff, feel free to ask.
For the route, my only true miss would be the Secret I get first in Honeyhive, losing 0.25s. I didn't want to redo it, so I kept it, but it's the only true route thingie I did badly. There is a possibility I could have do some levels first to save time, but I don't know which ones it would be. I prioritized staying in Domes to not have to go back to them and doing Lumas later on to save the most time possible. Star bits were taken whereI thought it would be worth it and Grand Finale was done before Bowser3 due to the whole credits being there - this saves a lot of time, like you can guess.
As you can see, the whole Star Festival is skipped because of these files. Not really anything I could do about it if I didn't want to simply long-jump to a cutscene, skip that cutscene and then get another unskippable cutscene. While it could devalidate this TAS, I couldn't care less: the Star Festival is not really anything anyone cares and more cutscenes than there already is in this game would have just added... nothing. It would have been boring, actually, even if the Star Festival Skip I found is a cool strat!
While this TAS can be improved, I still tried and tried a lot. A lot can be done with Luigi that you can not with Mario, sadly, but as I started from Power On, I didn't have a choice (plus a Mario TAS is nice). 2P was chosen to be more interesting, even if the mod wasn't intended to be played with it.
I didn't find a lot during this TAS, but I still found some improvements there and there. Some stuff can be taken in RTA, but with how much people in the SMG speedrun community seems to care about all of this, I think we will have to wait that they find it themselves or that someone watching this actually try them out and gets a record (IL) with them. I don't think I found a lot, though.
I do not have any special thanks this time, as I did mostly everything alone. Still:
- Thanks to azulon for checking that Freezeflame Skip. He proved it possible after I was done with Freezeflame, even if it was Luigi 1P, and not Mario 2P - I still have a lot of faith on that skip working for Mario, just would be harder.
- Thanks to Pamsy, my girlfriend, for her support as always. She watches the TAS by my side and it makes me pretty happy. Her support is why I can continue working on stuff like this while not sinking into thinking nobody would care.
- Laurozz, TGW and Melyssa for their work on such a fun mod. I would advice anyone to try it out! :3
I know a lot of y'all wanted a SMG1 TAS, not a mod one, but this might be almost more interesting for people who knows extremely well the speedruns of that game. Do not ask me too much about a SMG1 TAS, please - I might delete such comments because it starts to bore me.
I can't offer the file of the TAS as it might not work for other people. Nothing was edited (thank god, actually). ~50% of the TAS is cutscenes, but you will get the timestamps in the comments.
Thanks for watching!!
#supermariogalaxy #greenstars #femboy
Edit: Thanks for the 5 likes guys :heart: I never got this many likes before
Edit2: 10 likes!! The creator might see it, thanks!!!!!
It is possible to do this strat for every single Bowser phases, but only Bowser1 Phase 2+3 (from what I tried) can be really RTA doable. First Phase of this fight is giga tricky because the camera is different and you need to go to that planet, so you don't have the same liberty on movements.
This is RTA possible and you would save 1 to 2s with this strat, but it's really risky and easy to lose time over it.
He should get the skip himself pretty soon with this, but I wanted to show it because I take pride in my findings and I won't try to get the skip myself anyway; the fact alone that I can prove it and show it is all I care about.
I will update my comment whenever Azulon gets it, which can be right after this video or a bit later on.
With a good to perfect angle, ig this can be slower by like 10s or more, but still faster than doing normal Freezeflame.
O Freunde, nicht diese Töne!
Sondern laßt uns angenehmere anstimmen
Und freudenvollere
Freude!
Freude!
Freude, schöner Götterfunken
Tochter aus Elysium
Wir betreten feuertrunken
Himmlische, dein Heiligtum!
Deine Zauber binden wieder
Was die Mode streng geteilt
Alle Menschen werden Brüder
Wo dein sanfter Flügel weilt
Wem der große Wurf gelungen
Eines Freundes Freund zu sein
Wer ein holdes Weib errungen
Mische seinen Jubel ein!
Ja, wer auch nur eine Seele
Sein nennt auf dem Erdenrund!
Und wer's nie gekonnt, der stehle
Weinend sich aus diesem Bund!
Freude trinken alle Wesen
An den Brüsten der Natur
Alle Guten, alle Bösen
Folgen ihrer Rosenspur
Küsse gab sie uns und Reben
Einen Freund, geprüft im Tod
Wollust ward dem Wurm gegeben
Und der Cherub steht vor Gott
Froh, wie seine Sonnen fliegen
Durch des Himmels prächt'gen Plan
Laufet, Brüder, eure Bahn
Freudig, wie ein Held zum Siegen
Seid umschlungen, Millionen!
Diesen Kuß der ganzen Welt!
Brüder, über'm Sternenzelt
Muß ein lieber Vater wohnen
Ihr stürzt nieder, Millionen?
Ahnest du den Schöpfer, Welt?
Such' ihn über'm Sternenzelt!
Über Sternen muß er wohnen
(youtube.com/watch?v=Jg3sEE18WsE&list=PLuCfyLNwCvPlM63uXENNdvM_V9fnAjz04)
We hereby notify you that your account will be marked as "inactive" in 7 days, as you have not logged in for the last 12 months. Once your account is inactive, your nickname can be claimed by active users. You can reactivate your account by simply logging into your FACEIT Account.
Why are we doing this?
We want to ensure there is equal opportunity for our existing user base who have taken a break from FACEIT as well as our new player base who wish to diversify themselves.
What happens if I don't log in?
Your nickname will be released back into our unclaimed nickname database and you will be provided a generic FACEIT Nickname: e.g. yellow_12345.
What happens to my personal data?
Nothing. You will not lose any other personal information or game-related stats. All your information is still available in our systems and in your account. The only change is that your FACEIT account will have a generic FACEIT Nickname.
Can I change my new nickname?
Yes, you can change your generic nickname immediately from the point of login. This can be done one time free of charge.
Please refer to our updated Terms and Conditions that became effective on May 3rd, 2024, as well as our FAQ.
Thanks,
The FACEIT team
It makes resetting horrible as you need to do Kamella manip. This is for TAS so it is what it is.
I have way too much to say about this, so I will keep it short.
The Pull Star strat is not mine, but MATLF's one. It's known since a while, but it was always proven to be slower... Well, here, I actually land faster than Xander's TAS - though, due to my position on that pull star planet, I am then slower by the end due to Xander landing further on the last planet, making him go into the Sling Star ~1s faster than me (a bit less). I tried to do the normal strat to see if I could be faster than this, but after a bit, it didn't really work out too well AND this strat would still be around the same time - I then decided to keep this because it took me 4 hours to get it. Well, to be fair, I tried other strats before.
I did mega triple jumps in all directions, I tried leaving by the left, I tried a lot of different things, but nothing works. I was able to actually leave the gravity, but I'm instantly getting back into the one of the cylindrical planet, making anything impossible. I was finally able to get this.
That Sling Star planet has a lot of different ways to do it, but it was the fastest I could do it. It's where I lose my advantage like I said, but I can't do anything about it. I passed too much time on the other strat to just give it up, so... I am ~40f (or more?) faster than when I tried the other strat, so either I suck, either Luigi is just slower for that.
Finally, the last planet. I tried a route proposed by Azulon, but it was more than 100f slower despite how optimized I was able to do it. Sadge :( The whole planet is faster than Xander's TAS because I can get the third Silver Star with a 2Plj that keeps my speed and let's me get it without needing a double jump or a spin. 1st Silver Star to 2nd could have got a 2Plj, but it wasn't faster or better from my tries - maybe an improvement there anyway.
Finally, for the last Silver Star, I got two tries tied at first, but the other was 2f slower despite being a bit "further" than the other attempt. No idea how it's possible.
Some levels are with Azulon, but this one was mostly me. Thanks, though, for his ideas with the backflip on the Sling Star planet, even if it finally led to be slower than my best attempt.
Pretty proud about this one. It could be improved by few frames there and there and maybe this strat is simply too slow compared to the "normal" one, but it is what it is - with only 2 GPs, it has to be the best way to do it, at least.
One of my most optimized level to date. (For G1 at least)
(Don't take this seriously.)
IL record is 2:21.13 - my TAS is 01.65 slower, but I have cutscenes, meaning I lose more than 10 seconds over it.
For the first section, I tried going on top of the volcano with the new strat username did for Melty Molten C, but it was more than 1s slower than this attempt - that volcano strat was still faster than the best IL section. lol
For the Chips planet, I tried to get them the opposite way, but it was way slower and harder - led to nothing, even if it can be tried again with better optimization. The bonk I do is required to get on the platform, so I had to do it.
I would love to have more to say, but... That's about it. Frame advance was used in the fight.
I tried this few months ago and Duck absolutely hates this A press - glad to have been able to help and make it possible to totally save it!
For Mario, it's maybe a 0.5xA, due to the air acceleration and stuff, but honestly, it might be a 0xA.
I wanted to save an A press since so long, so I'm really happy to have finally done it.
This is a LOTAD, as I didn't want ot try hard the ending. GG!
I win time before the Hungry Luma Launch Star, only to lose most of the advantage on that last planet, due to Mario long-jumps being close to the ground (cycles and stuff like that make that last planet not really better with Luigi from what I tried. Could be wrong).
From landing to the Hungry Luma LS, the ground-pound is what saves the most time.
PurpleSun spawns the Power Star before me, but as I'm closer than the Star than him, I'm able to grab it 0.09s faster.
Previous time was 1:09.51 - I had to rework the ending.
Sub 1:09 is clearly possible, but I'm not 100% sure how to improve things with cycles.
This one was done for Ghost - I wasn't planning to TAS Toy Time S, but oh well, there you go!
'Super Kaizo Galaxy: Red Sun' is the prequel of 'Blue Eclipse', the first mod of the Eclipse Trilogy. This mod, Red Sun, follows the story of the previous Savior, "Fiel", after that Bowser kidnapps the Princess.
Red Sun is a mod that is Luigi only and can not be completed without playing 2P. Be it by yourself with a second remote, with keyboard and mouse or any way possible, or by playing this mod with a friend, your lover or a member of your family, 2P is required to be able to even pass the first level. Even though this mod calms down in difficulty, this mod is still extremely difficult and requires good knowledge and good skills to be able to beat it. Game Over'ing is impossible without actually doing it on purpose and save states are allowed as long as you are having fun - whenever the mod, do not start calling people using them cheaters or "lame", thank you.
I am not able to give you a clear date for this mod, but I am trying to make Red Sun be out for the 1 year of Blue Eclipse, so end of October this year (2025). I can not promise that it will happen, but I will try, that, I can promise.
Like the trailer says, a lot will be changed, same for the galaxy order. This mod will also be 100% beta-tested by "Cesar" (TheCesarMK), which was also my beta-tester in Blue Eclipse, even though he was not able to beta-test World 6 or S (no shame, I was just too hasty with Blue Eclipse). SuperAntonioGalaxy also beta-tested the whole World 1 and Tibs a bit - thanks to them, even tho I have some other beta-testers that either gave up, either just don't play anymore.
I want to be clear: Red Sun, while it's going to be really, really hard, is NOT going to try to be the hardest possible. I already nerfed a lot - I want to make the save stateless experience more fair than I did in Blue Eclipse. Will this mod be hard? I already answered it. Will this mod be the hardest of SMG1 yet? Yes, no doubt about it. But will it be fun?
My goal with Red Sun is to offer a good continuation of Blue Eclipse (even tho it's a prequel). It will focus a lot on atmosphere and story - almost every NPC has their text changed and gives lore and world building, as some explanations on the previous Savior, their disappearance, a bit more about Bowser and stuff I can not talk about...
In // of Red Sun, a short story that will do 20~30 Chapters that is a prequel to this story as well will be released little by little, starting by this week-end or start of next week.
People know Blue Eclipse mostly because it's extremely hard. My biggest sorrow about Blue Eclipse is that it tries to be atmospheric, gives a good story (even tho Red Sun/Green Moon are the ones who really enters that story) and be fun(ny) - and yet, people only see its difficulty. People think the only goal of Blue Eclipse is to make the other suffer. I am honestly sad about it, so I hope Red Sun can change that. More custom levels, more lore, more fun - this will be a good improvement over Blue Eclipse.
Lastly, I would thank PKSMG2 and Bu1626 for some of the changes (UI, Coins, Launch Stars paths, etc...) in my mod. Thanks to my betya-tester to play this mod or have tried, thanks to Jhay for some answers and thanks a TON to SuperHackio to have helped me with the reordering of Galaxies. There are more people to thanks, but they will be in the credits, don't worry - if you are not mentionned, you will be there anyway! :3
The first planet has nothing that is really new, except that the climb to the top is a bit faster while doing this strat twice. The start could be few frames faster as I think I did the 2P jump a tad too late, but it works. There is a potential zip jump from the bowser statue, but PurpleSun wasn't able to get it and me neither - I didn't try a lot, but as my Lua Script is not working, I wasn't going to try more than this. It seems not doable anyway, but who knows?
Anything after, until almost the end, is mostly like RTA. Really can't tell much about these ones.
The lava cylindrical planet with bullet bills and all that is maybe few frames faster or slower than the best time there is there - I tried the normal way, but I wasn't able to get it faster. I kept this way as it's kinda cool, but that's... all.
For the first Phase/Planet again Bowser, doing that strat where I hit Bowser in the air a second time was slower due to not having a good acceleration. This was not an issue for the first hit, making it faster (by almost nothing).
The second Planet is a bit better due to how quick I hit Bowser with the ballons with the help of 2P. This was close to perfect, if not perfect.
The last Planet is where ~5s is saved: a new TAS strat I found myself, while fighting him.
Bowser is programmed to always go the opposite way of you until he hit the floor - when he touches the ground, you can spin him, but he is also go in a straight line until you meet him, where he stops and go the opposite way after a bit. Though, you can manipulate that to trick the game into making Bowser go a certain way and follow that way - he will still try to go away from you, it's not me breaking the game's code. Though, as I change directions a lot and that I'm already close to him, it looks like I'm going the same direction as him. He is still way faster, meaning I have to long-jump at a certain time to be able to get pushed by him - he then pushes me with him, as I'm under/in his collision/hitbox, letting me be close enough to spin him. I do that twice, saving ~5s. This is really, really hard and requires a precision that RTA can't replicate: with save states and slowing down the game, I can manipulate it, which RTA can not do.
There is a potential other strat between the ice and desert planet, but I wasn't able to get it (getting into the beam at the end of it without touching the ground of the other planet).
With the potential zip jump, maybe you can have a sub 4:50, only if it works. The fire/bullet bill planet is clearly not optimal, so you can be faster there too.
Overall well optimized from me, but nothing insanely optimized as well. Still a really good time and happy to FINALLY find a strat that is 1P usable... Well, for TAS, at least.
Yeah, yeah, I finished the Wool Speedrun not a long time ago and I am already TASing a SMG1 level - what is wrong with me? I just wasn't able to focus on a lot and after seeing username's TAS on Star Reactor, I wanted to do this. I took a level not TASed yet, neither M or L, and decided to go with this Ghostly.
This one is pretty well optimized, I can't lie. There are few frames there and there that you can save, I am sure, but I doubt you can make this time that much lower without finding a new trick.
I didn't really time the difference between simply walking or doing that fpj to spin at the start, but I doubt the walking would be faster - or at least, the time would be around the same. If I lose 3 - 4f there, then I'm just stupid, but it looks cool, r-right...?
More seriously - the next session... I tried a lot. Going from bones to bones seemed at first pretty fast, but it was way slower than I thought - more than 2-3s slower than best attempt. The problem is that even tho you can join two bones with a 2P long-jump, meaning no spin, the last one was few pixels too far/high to let me do it with a normal spin after, meaning I had to use a 2P jump - it was too slow. Even without that, it would have been too slow. Using the Rainbow Star was ~50f slower than this attempt, even though I was able to join the bone without a spin, seeming like a good option. Sadly, taking the Rainbow Star and its cutscene just are too slow for what it is.
I didn't really try what is best after the spider bounce: 2P jumping into the Sling Pod (like I did) for 2P jumping to reach the wall the highest possible? From my little tries, it seemed way more clean and better to do it like this, as I could actually reach the Sling Pod. You can do it without 2P too, but I am sure it would have been slower. Going towards the Launch Star and wall jumping (from the Sling Pod) instead of taking the Sling Star was not possible and even if it was, it would have been ~40f slower.
The next section doesn't really have a lot. Backflip was faster but idk by how much and instead of going top or bottom on that rotating planet with holes, I long-jumped from the middle, saving a good amount of time. The bouncey planet could be improved, but I wasn't able to do better - good chances I would have bonked against the hole planet thingie if I tried to get an instant Launch Star entry. Doing a long-jump might also be a good shot, but when timing with the best Luigi IL (not taking in account the fact that it doesn't get fpj), it seemd around the same time, even tho I'm still few frames faster. Again, might be improvable, but it was the best I could get.
Bouncing on that shop Luma was way slower, no matter how you do it. More than 2s at least.
Bouldergeist itself was full of RNG manipulation. The only thing that matters with the strat I do, which I call Triple Hits, is having the first rock on the second sets of rocks, Phase 1 or 2, be a Bomb Boo and that you already have 2 Boomb Boos spawned. This require a "perfect RNG", in a sense. Every weird movement I do is not random, but actually to manipulate that RNG - this was a bit annoying, but I made it. The last hit could have been 1f faster, as I'm not 100% sure I threw the star bit the perfect frame, BUT it should have been. You can improve that boss fight by having the Bomb Boos being at the perfect position to get the hits the fastest possible, but this would require a lot of manips and would been few frames faster at best - this is still really, really clean.
Funnily enough, I am 0.2s straight away from having the same time as username's TAS that was released today (well, yesterday for me, but it was like 5 hours ago). Absolutely coincidence, but a fun one!
Welcome to the "Wool Speedrun", a 120 Stars TAS for Super Mario Galaxy 2. The timer is from the first frame of clicking on the Yoshi icon to the first frame the Grand Star animation shows when beating Bowser a second time.
The "120 Stars" category is doing the 120 'normal' Power Stars and beating Bowser again, unlocking the Green Stars as such. The reason I don't TAS the Green Stars is due to motivation, time, and because "PurpleSun" does such a TAS already.
When starting my TAS, I got issues instantly because I didn't start from a totally new file, making it desynced instantly. Then, it started desyncing for no reason. Due to that bug, I had to record every level and editing the together - the time of my TAS seems to be actually 5:08:28, but due to some frames maybe dropped during the editing and some maybe lost during the levels where you see "Playback", I decided to reduce the timer by 1 second - actually, it's more probably the time is actually 5:08:26, but I felt more fair to say it was 27.
Sorry for any audio bugs sometimes presents at the start or during some of the videos with "Playback" on, same for the Yoshi flashes that happen due to the version of my Dolphin.
This TAS started as simply a fun thing, something that would not be too serious, as seen with my World 1: it kinda blows. It's nothing great, it's seconds and seconds slower even than RTA. While I started this as a fun thing and an excuse to strat hunt on every level, I started to focus more and more with the start of World 2 - Puzzle Plank S to be more precise. Even then, I continued doing a bit whatever and some levels, even with Luigi, I didn't want to bother and didn't optimize a lot.
While 5:08:XX is good, even in English, you can make this time at least a 5:03:XX simply by optimizing everything. New strats were found during my TAS and I found new ones while I was TASing (like the Lava Lair 1 zip jump). With way better optimization overall and new strats, I am sure a sub 5 is possible - a better route would save time as well.
This doesn't mean this TAS is a bad or that I shouldn't have make it - a lot can be used by RTA and it's what jda7 and myself call a "template run": it's meant to be a great ground to start the foundation of a better, more optimized TAS in the future (that I will help with). Does that mean the Wool Speedrun is mid? Bad? I wouldn't go that far and calling it a LOTAD anymore sounds wrong, but yes, this TAS is easily improvement by minutes. Star Bits management. Routing. When we switch to Luigi. New or other strats not done. A lot more.
I am still proud of this journey. I had a lot of fun with this TAS, despite how often my Dolphin Lua would crash, sometimes every 5 minutes. Mole Skip also made me think about giving up, because of how bad World 1 is. But it was fun. It was good.
The other reason I did this TAS is to finally give a 2P and a 120 Stars one to this game that deserves it. I don't care how bad this TAS will be compared to the next one. I don't care how many levels will be beaten by RTA. I don't care about all of that. No, I care about people being able to see how crazy SMG2 can get. No, how crazy SMG in general can get.
I have nothing else to add. I doubt a lot of people will read this, but I like writing like that. Anyway, thanks. For more stuff:
- Thanks to @jda7 for helping with the route (we didn't think about Beat Block and Wild Glid L issues, which is why the route is not perfect). He also helped a lot with explaining techs or helping with strats (mostly routes in Purple Coins stars or Romp ones - for example, the Flash Black one, for the ending, a lot was with his help). Also great morale support and friend overall!
- Thanks to @NolifeVV for helping with some stuff, but also checking every level in advance and reacting. He also showed some strats like the SLJ in Lava Lair and made them RTA viable (tweaking them a bit).
- Thanks to @azuulon for the thumbnail.
- Thanks for the SMG community and SMG TASing community (and servers) for the help and the documentation.
* Also, thank you @PurpleSunTAS, @usernameispassword4023, @EpicRandomGamingChannel, @/BillyWAR, @/MayuTheMayu and other TASers/people with some help on the side or offering them.
Finally, thank you to my lovely and great girlfriend Pamsy for being there overall. While not really of any help with this TAS, I know continuing trying stuff like this, even when it gets no attention, is worth it, because I always have someone that will wait for me and want me to feel happy in my life. Thank you, chou.
#supermariogalaxy2 #TAS #toolassistedspeedrun #sheep #woolspeedrun #120stars #SMG2 #supermariogalaxy #record #toolassisted #femboy #toletole #mariogalaxy #SMG #hastag #blueeclipse
This one is weird. I did this while my Wool Speedrun video was exporting on my computer and due to my computer being surcharged, my TAS got a bug and totally desynced, meaning I wasn't able to go back to it. I am saying it because, idk how, it was around 40f faster - yes, almost 1 full second ahead of this. It means this time, with my previous TAS, would probably be a 1:49.XX. I have no idea what I did to win 40f there, but I did - I tried every possibility, but nothing worked.
The bounce on the first enemy is maybe not optimal and maybe even slower, but to my tests, it was the fastest I could do - keeping A pressed was slower than not doing it by at least 4f.
The place where I get faster than RTA is the 2P long-jump, letting me land without a spin needed, saving a good little amount of time.
usernameispassword proposed that I'd long-jump from the tornadoes, but all I was able to get was 2P mid-air jumps or single jumps. Sadly, it doesn't seem to be possibled, compared to pull stars - I am not the best at 2P in SMG1, so it's not declared impossible yet, but username himself seemed to not beleive in it anymore.
The start is where most of the optimization can come from another TAS - the rest was really clean.
The 2D section was 20f faster than previous attempts, but with username's ideas, it was more than 40f faster than previous attempts. On those wheels, it's possible to get an insane amount of speed, like CK showed, but themself have no idea what they did and nobody really knows - here is a link to their clip: https://x.com/MrCloudKirby/status/1887061180885164085/video/1.
I was not able to replicate such a thing despite so many attempts. I kinda understand a part of it with the slope, but getting sloped seems impossible (it is because CK got it, but how? You probably need other factors before, like where you jump and everything).
Why going longer backwards could have been good, the Fire Flower spawning made it impossible - a perfectly timed spin allows me to bounce from the first Piranha, then letting me bounce on all of them even tho it was really precise to get the second one to work. The bounce on the piranha on the ceiling was not planned, but without it, this would not have worked. 18f faster than not doing it and bouncing off only one.
The tower was really clean. The 2P freezing on the enemy is useless, I did it only if I could use that enemy, but of course, no. I wanted to join the Sling Star on the side, but I was not able to get that much distance - even if it was possible, good luck-
Well, anyway, I was writing that when I had a 1:52.86. Then, I wondered if joining that Sling Star from the top would lead to a fast time... and holy did it.
It first saved 99f over normal way, then saved saved 9 more frames with a better entry to the Sling Star. You also *need* the backward momentum from the tornado going towards the right to get in the other one. Maybe it's possible to go 3 - 4f earlier, but eh.
Using that 2P jump then spin puts me in the right place of the wall to start wall jumping without the need to ledge grab, saving a little amount that I can't refuse.
The ending can be done with a slope jump that goes backward quick enough to not require the instant crystal break, but it wasn't necessarly faster - here, I get an instant grab on the star with a good instant crystal break, saving even more time from my previous attempt.
This TAS was kinda annoying to work with due to so many desyncs from nowhere - I had to record segments because I was NOT redoing the whole star again. Luckily, keeping my frames on screen led to the TAS not desyncing (for whatever goddamn reason) and I was able to record it.
This is the last TAS before the Wool Speedrun is uploaded. It might be also the last TAS until a bit, as I want to focus on other stuff, but who knows?
I don't really have too much to say or add over what could have been said before. I did the star chips my way (with a route someone showed me, I "improved" it/put my own in it) and the ending has my strat, which saves few frames only over others, so it's nothing too great, but time save is time save.
With the MTJ, it would be a 1:37.XX. Still, sub 1:40 is nice!
The improvements are as such:
- The cannon shot actually gets to the lever faster (by few frames).
- The spider bounce is faster than doing it the normal way (I wasn't able to improve it, but should be possible).
- New strat in the fight where you shoot the shell BEFORE the cutscene starts (second phase).
Improvements:
- The lever can be hit a bit before, by at least 3f.
- Killing the spider lost 2f due to hitlag.
- No cutscenes.
Did this level for Azulon once again.
First SMG1 TAS (outside of the strat hunting for King Kaliente) - this was Azulon's choice, so I went with it.
This is a 7.57s improvement over previous Luigi TAS (which was 1P): youtube.com/watch?v=39KnC0eQtw4.
It's also faster than PurpleSun Mario's TAS by 5.32s: youtube.com/watch?v=MhaBPBFn77M
The start has nothing new and it's maybe the part where some frames should be improvable.
The second section countains the Launch Star clip (Screw Skip), which requires you to go from the pink platform to the orange one while crouching, then spinning to get pulled by that Launch Star. It saves a good amount, but every TAS does it, so nothing new.
What is "new" is the triple bounce on that mecha-bowser that usernameispassword found some time ago. It was "weirdly" the hardest part of this TAS for me, as the Hole-In-One didn't take too long. It totally skips the Launch Star and saves 3~4s (could be more around 2.5s with really good optimization with the path to the LS) over just taking the Launch Star.
Lastly, the boss itself. The Claw Hand Jump (Hole-In-One) saved 189f over climbing the boss from the left side (with the help of 2P). Maybe a better route would make this skip save more around 2.5s as well, but my climb was pretty clean. Sadly, I wasn't able to get quicker to the claw and skip a cycle, but maybe doable with cutscenes skip at the start of the level (this TAS kinda replicated how the level would look in an Any%/120 run).
Nothing really to add, some frames saved there and there to better movement. Doing a bounce on the mecha-bower after the Screw Skip was a bit more than 10f slower than not doing one. The backflip saves around 10f as well over the triple jump. Going on those ?Blocks is around the same time, but I consider it slower by at least 6f (from what I timed, could be more due to a bad wall jump).
Next one should be Airship Armada!! o7
Everything up to the Goomba is probably a bit slower than him - I am then faster due to Goomba shootings.
This can be beat by RTA (2P) with Chinese/Japanese language, as both of them saves more than 1s over English, but I doubt my time is beatable by RTA otherwise.
I am putting this one public because why not. This might be the last one I upload before I finish my TAS and upload the whole thing! :3
The Mole Skip uses what we call Edge running - it's something I also found (youtube.com/watch?v=p99FIl2Xhp8). It's running on the slope to gain more speed... Though, this is not the best way to do Edge run, as this zip jump led to a 74.564hspd. At first, I thought I would be out of the slope, making the jump a 0.001hspd the frame after (happened to me MULTIPLE TIMES btw.), but I was lucky (and talented #swag) and I got this. The angle was always close to perfect - no, it was perfect.
Ground-pounding earlier was possible, but you would need to walk to the center of the cage, making it slower. It is maybe possible to do this 1f faster, maybe 2f, but hard to say.
This make GS3 an insta grab, btw, saving around 22s over normal way (for TAS) - this Skip is around 21s faster than TAS, while Pumpkin Skip is a bit less than 29 - those two zip jumps saves a bit less than 50 seconds.
I also have an idea to make ICS - Improved Crab Skip. It's still unsure if it's possible or not, but it would save a bit less than 1 second, or a little more - hard to say as it's still not proven to be possible and needs good RNG.
Whatever - Mole Skip is finally proven possible. This save spins and time.
I lose already around 1 second by not playing Luigi (not that I have a choice, it's how the route is) because I don't have the backward momentum. I do win a lot of time by optimizing some stuff... mostly the ending.
The ending took less time than excepted. There are a lot of i-frames on the topman when you hit it, making it impossible to shoot a star bit or 2P it whenever you want. In a way, that was for the best, as star bits are pretty much needed for my TAS.
I was able to 2P a mini topman to 2P the topman itself, saving 1 star bit.
While I take some time to get to the Gearmo, I had to slow down at the start to be able to shoot a star bit to the Topman.
Another improvement was before, when I use a Z cancel at the start of the second planet, which I found in Space Storm 1.
As I was uploading my 1:18.78, I noticed an easy improvement. While it led to -3 Star Bits in this run, it's worth as it's a 0.87s improvement - that was the long-jump I do after slowing down. In my 1:18.XX, I had to spin on top of the void, losing time to get to the Gearmo.
While I have no doubt you can improve this, it's still a damn clean Space Storm L. Almost done with the Any% part of my TAS.
* My Egg Skip was a bit slower than it should be.
* Maybe you can enter the Launch Star quicker.
* Better ending I guess.
This is with the new found improvement for the Double Coconut Freeze strat.
This is a LOTAD, but I did it RTA too, just was a bit slower. You can probably make this even faster, but you get the general idea
Was just to showcase something quick. Exact time difference would be actually 4.9s, instead of 4.86s.
I already proposed this idea around 2 months ago, but nobody really cared that much. Not a problem, but it's a ~5 seconds time save with possible improvement.
You can also be around 1~2s faster if you do the strat we do for Boss Blitz, when you freeze the coconut later on.
Overall, just a strat you should do.
I timed it with Tib's IL because it's the current IL WR. If there is a faster King Kaliente currently, then maybe it saves less than 4s, but uuh?
Improvements:
- Better start overall. The neutral jump to the Launch Star is the only reason I'm faster than Gory.
- Yoshi is with backward momentum instead.
- CM flutter was way better/faster.
- The Star Chips section was kinda the same, but with Star Bits.
- The ending is improved pretty much anywhere. The Goomba Bounce is way better and allows for a 1:19.XX. The blimp fruits grabs are also better.
*/ Going Coinless is faster for the whole run, but slower for the IGT. I got greedy for the IGT time, so I took them, and it could let me have the 100 for GGen anyway. It's around 0.4s slower iirc (maybe a bit more).
Overall, very happy about this one. I thought I would get a 1:20.XX, but it seems 1:19.XX is possible!
Even with every optimizations possible, idk if 1:18.XX is possible. Could be, but damn.
Let me take care of you, shake shake, with my knack for lovin’
If you won’t cry, it’ll dry up, let’s let it pour
Please say yes? I wanna lick away our spices, drink up the tears from your dissected heart
Rolling Masterjank, great galaxy, thank you #Nintendo for #SuperMarioGalaxy2
For the rules of my TAS:
- Starts from a new Save File. Bank Toad won't have any Star Bits (he should have 100). [Because of a mistake, he has currently 9999 - they will be removed between Puzzle Plank S and Hightail Falls 1.]
- 2P is allowed; everything is played on the Vanilla game.
- Any strat is allowed.
- This will be played on the English version of the game. While it loses at least 20 seconds over Chinese (and a bit less for Japenese), I decided to go with English.
Nothing really important to note for these rules - it's just a 100% SMG2 TAS.
- The way this TAS is played is mostly me playing on Slow Down, which is why I could miss some Star Bits that a "optimized TAS" won't. TASInputs will be rarely used except for infinite flutter angles or some other things (like the Fire Flower minigame). For Frame Advance, could be use to compare strats or for some other textboxes.
[For more casual people (that don't play the game often or don't know too much about speedruns):
- Coins are avoided because it gives a textbox after the star that loses a bit of time. If coins are unavoidable, the less possible will be taken. 2P can help with taking a coin.
- Star Bits are farmed/taken for Hungry Lumas (one per World, unlocks a new Galaxy per world) and because the Perfect Run requires 9999 Star Bits at the end.
- The reason you can the camera going first person and leaving that first person instantly after some launch stars is because of a trick called "Camera cancel" that allows you to skip being "staggered" from landing from a Launch Star. I won't explain every trick, but this one is going to be used pretty much every level, so I think it's a good call to explain that.
- While there is nothing weird for this part of the run, note that the route for a 100% won't be simply doing every Star you can one by one. There are a lot of reasons on why we skip stars, like cutscenes being together or Comet spawns.
With that, enjoy this TAS! Super Mario Galaxy 2 is often considered more boring to watch because it's a "long-jumping simulator", but there is so much to see in this game. I want people to see how fun it could be to watch SMG1/2 speedruns/TASes, so I hope to be able to make you interested in it.]
Thanks to:
@jda7 for the reroute and helping me overall with possible strats and helping me overall. While I am alone doing this TAS, he should be considered a huge part in it.
@NolifeVV for watching my levels in advance and agreeing to help on some parts if needed.
Mayu for the overall support (also forced me to improve Spin-Dig 2 by 9f).
Pamsyy for being always there for me. While not really anything for this TAS, I will alwasy thank her for everything she is doing for me~.
(For the final product, the "PLAYBACK" that you see on top will either be hid by the name of the level I am on, either I will keep them as it won't ever be seen once again in the future. I had a bug in my TAS, forcing me to use my OBS recordings instead of Dolphin Dump Frames/Audio. I am deeply sorry for the lack of quality - I do feel ashamed about that.)
If you are interested into seeing progress being made and everything related to the levels: docs.google.com/spreadsheets/d/1lejEzaOUbVYO12PXxwRoUBcYSJ4BbWZlLeYAimNOd4E/edit?gid=0#gid=0
You could spoil yourself some strats used or the IGTs though, so do as you wish.
Thanks for your attention and I can't wait to release this TAS when it's fully done. Some levels... are going to be a hard grind.
Thanks for Mayu for looking at this and giving ideas/having fun watching me TAS. This is not optimized like I wished, but it's pretty good. Sub 1:20 is of course possible.
This one was done because I decided to TAS one Star (Green Star 1 - Egg Skip) and they decided another Star that I should TAS.
jda7 proposed the mega triple jump that he does for Green Star 2 at the start, but for Mario, it's ~20f slower than what I do here. Maybe a certain angle let's me go on the planks instantly, as the problem is that the wall jump makes it slower, while more "stylish". The climb is pretty clean, but I could have been a bit faster - for this reason alone, the start should be able to lower by at least 40f with TAS. Probably more, but hard for me to tell... We will need to wait for Shadow64's TAS.
2P is pretty useful to break the egg - going with a long-jump to a backflip seems better at first glance, but I can almost directly do a backflip, saving a bit of time. This is too risky and saves so little for RTA that it's probably not needed. Breaking Yoshi's egg to win a bit of time is useful in this case, and while facing the other direction to start Yoshi's run instantly seems to be the best, I think what I do is almost the same time. Maybe a bit slower, but it didn't seem so.
The CM was a pain in the ass to take and I think I still am slower than Gory's newest record. IDK how he does it.
You can eat the spiky plant instead of using 2P - I don't think it changes anything. The only reason I did this is just because I do play with 2P - it deserves to be used a bit, right? :3 Nothing to say about this part, otherwise - it's pretty much what RTA does.
For the last planet, impossible to do a land cancel or make of any use of the camera. Long-jump to a spin/ground-pound (frame perfect) is probably the best... period. I can do a land cancel at the end and with two backflips after, while having breaking the egg, I take control of Yoshi pretty quickly.
I don't think I played the best with the berries, but it wasn't too bad. A "strange" time save I got is at 1:25, where it actually gives me a slight height boost. Nothing insane, as it saves like... not even 5f, imo. It's nothing huge, but I take it.
Using the goomba to boost very high is where I actually win a lot of time over Gory/Hime's TAS and makes this Star easily faster with 2P for both RTA and TAS. Of course, hitting that RTA is... good luck. It's a bit "RNG" where you will go, after all. I am pretty sure I can improve that ending and save few frames, maybe going to 1:21.XX, but I am very happy with this so far.
1:22.26 is still my PB. Sad that I lost it, but 24f is not enough for me to complain to much.
This is not a TAS - I didn't use anything else than Recording and Save States. Everything was done by me, without slowing the game, using TAS inputs or even frame advances - I didn't make of any use of my Lua Script neither, while I could. :3
This took a bit longer than excepted because of something called me fucking it up. It was one of the most fun thing I ever done, though.
This is NOT a TAS made to see how quick you can do the level spinless. The boss was very bad, if it was the case. I just wanted to have fun.
I was able to reach almost 400 speed while playing around with the hammers. Wild stuff.
Nothing to note, otherwise. It is possible to go the other way in a more simple manner by just waiting that he slams his hammer and walk on him, but this way was more flashy and fun.
Not having the ability to spin meant some of my flutters weren't able to go very far or they weren't very good... I am also not too great at them. Maybe that, at 00:28, I could have gone to the ? block instantly while using the hammer bro's bounce, but when I tried my speed stopped before - I still think it's possible, but again, this isn't me trying to do the quickest I can this level - I am just having fun, honestly. :3
Could do stuff like this a bit more often just for the fun of it. We will see...?
The way this works is... uuuh, weird? The area (or wtvr it is) that says that you are done with the race is reached, so you are removed from the "slide" - with that, you are not affected by the slide itself, but by the normal gravity. This leads to this - it looks like a boost, but it's not. Because you are not in the "slide" anymore, you gain a new spin, your ground-pound, etc... as well - it would be slow to use a second spin, sadly.
You can reach the top of the tree with this, but it would be slower, straight up.
Impossible for me to say if this has potential for either TAS or RTA. I doubt for RTA, even if it's not really hard to pull off, but for TAS? It would be my 3rd strat that's few frames (or like 0.2s) slower than RTA, but improved enough with TAS to be useful/used (like both Green Stars in Rightside-Down).
I have NO IDEA if I found this or if it was already tried/found before, but if so, sorry: I found it myself by messing around.
After few strats I tried but were already found before, glad to find something that seems new. The only way, right now, I can see this being faster than RTA would be to land just next to the NPC that gives you the star - even better if you can land on his right for a good star grab.
Sadly, my zip jump idea in the start of the second planet is slower by almost nothing. Nothing in World 2 from me (for Green Stars) and for Tall Trunk, nothing yet. :(
My Egg Skip was close to perfect - the only part I could improve is the ground-pound timing, but for RTA, I think it was close to perfect. The Egg Skip itself was also close to TAS-like (thanks to PurpleSun for telling me how he did it himself).
The small lags during King Kaliente is either my PC, either me PUTTING a save state (I didn't use them) in case I failed the fight again and wanted to see what record I could if I didn't fail like a shithead. I think verifiers will be able to see that I didn't use any, but still.
Thank you to youtube.com/watch?v=ICEzGr7x3t4 this song who was with me when I got the IL Record.
Green Star 2 next? Clueless
The movement to go to that moving platform is pretty much the same as PurpleSun current TAS. You use the momentum of that moving platform to do a simple jump, leading to a long-jump where you don't need to stop holding right and where you can keep A pressed.
You need to then spin (once only) to get on the collision OoB/clipping inside the planet, just do a normal long-jump and after a bit, throw another fire ball to stay on the collision (otherwise, you will fall). As long as you don't go up too quick, you won't get pulled by the gravity and you can fall into the star.
This is faster overall because you don't need to be "pushed" by the 2D Area, which is pretty slow.
Pretty happy about this find. Nothing to say about it otherwise, though. Maybe it will be used for ILs; maybe for TASes, this will be few frames faster. Hard to tell.
All I have to say is that you need to do the first two fire balls while holding right, and then switch to up-right. Maybe the angle can be a bit more right or up if you have a nunchuk/a controller (I play on keyboard&mouse, after all).
The triple jump is not really too hard/bad, as you just need to make sure you keep a bit of speed, do your second jump far enough, but not too close, and have your triple jump be on the edge of the box. It's not really the hardest part, but you need to have jumped before.
The part in the air, with the fire balls, is harder than it looks. Took me some time (around 5 - 10 minutes) to finally do it once (but after, it took me like 50 seconds to record this lol).
I thought this strat would be what the TAS does, but I am happy that it's not - I am planning to go through almost all Green Stars to see if I can find any new strat. Well, at least, that could have more chances to be RTA viable? I don't think I will surpass PurpleSun.
This strat is pretty risky and saves around 1.5 seconds over current RTA strat, while losing few frames (and I mean it) over TAS strat - could be worth it, but the risk is there. You wouldn't lose a lot, but still.


