![[TAS] SMG2: Freezy Flake GS2 in 37.95s @PurpleSunTAS](https://i.ytimg.com/vi/j_J_UWSEdpM/maxresdefault.jpg)
Uploaded February 2023 | Updated September 2026, 2 weeks ago
Green Stars TAS WIP
37.95s IGT - 60:59.31 real-time split (1:34.05 segment)
Freezy GS1/2 were two of the few stars I hadn't tested prior to starting this TAS, as star order didn't seem like it'd matter (more about that later).
The first launch star is a balance of optimisation for the earliest statue break (not waiting so long for the launch star to activate) while still being able to get to the launch star on its first active frame.
While other routes looked cooler (e.g. getting turned backwards in mid-air to have more speed), the fastest by 29 frames was how RTA runners do it (used Pottoww's IL as a blueprint: youtu.be/QWP7saYcBaw?t=10).
An instawalk from the launch star is common here, but interestingly, entering the launch star while turning ends up getting landstun, and entering while standing completely still gets faster and earlier acceleration.
1-flower cycle was attempted but failed (youtu.be/6GEkxIaaQ-k), missing out by around 0.6s.
With certain movement, conical slopes (e.g. the mountain, some crystals) can cause some very buggy fast slides for a frame or two, leading to glitchy slope jumps.
The consequence of slope jumps being patched in SMG2 is that they now stay at a constant speed and trajectory.
It is difficult to control the trajectory, and the jump in the 1-flower attempt has a much more wayward and slow setup after getting the 2nd flower.
As with the Tall Trunk slide, you don't lose any time with whatever movement you do provided you don't jump a lot.
However, exiting the tunnel without jumping is a little slower and doesn't let you spin/throw fireballs. Also, the gravity you leave the tunnel at will have an affect on your angle, sometimes changing angle mid-air.
Backwards exit with a mid-air fireball was 35 frames faster than the RTA method: forward exit, jumping on the lower brick block and fireballing around the corner.
Star order almost mattered at the end. The fastest regular route (by 6 frames) was to long jump onto the flying ? block, and its cycle changes with cutscene skips.
But another glitchy slope jump was found on the Goomba statue, getting a 67hspd jump that barely misses the lava geyser. Unfortunately this kind of jump couldn't make it onto safe ground, but the lava boost and movement on the sloped boundary was 20 frames faster than holding it back onto the ice platform and long jumping.
#supermariogalaxy2 #toolassistedspeedrun
Green Stars TAS WIP
37.95s IGT - 60:59.31 real-time split (1:34.05 segment)
Freezy GS1/2 were two of the few stars I hadn't tested prior to starting this TAS, as star order didn't seem like it'd matter (more about that later).
The first launch star is a balance of optimisation for the earliest statue break (not waiting so long for the launch star to activate) while still being able to get to the launch star on its first active frame.
While other routes looked cooler (e.g. getting turned backwards in mid-air to have more speed), the fastest by 29 frames was how RTA runners do it (used Pottoww's IL as a blueprint: youtu.be/QWP7saYcBaw?t=10).
An instawalk from the launch star is common here, but interestingly, entering the launch star while turning ends up getting landstun, and entering while standing completely still gets faster and earlier acceleration.
1-flower cycle was attempted but failed (youtu.be/6GEkxIaaQ-k), missing out by around 0.6s.
With certain movement, conical slopes (e.g. the mountain, some crystals) can cause some very buggy fast slides for a frame or two, leading to glitchy slope jumps.
The consequence of slope jumps being patched in SMG2 is that they now stay at a constant speed and trajectory.
It is difficult to control the trajectory, and the jump in the 1-flower attempt has a much more wayward and slow setup after getting the 2nd flower.
As with the Tall Trunk slide, you don't lose any time with whatever movement you do provided you don't jump a lot.
However, exiting the tunnel without jumping is a little slower and doesn't let you spin/throw fireballs. Also, the gravity you leave the tunnel at will have an affect on your angle, sometimes changing angle mid-air.
Backwards exit with a mid-air fireball was 35 frames faster than the RTA method: forward exit, jumping on the lower brick block and fireballing around the corner.
Star order almost mattered at the end. The fastest regular route (by 6 frames) was to long jump onto the flying ? block, and its cycle changes with cutscene skips.
But another glitchy slope jump was found on the Goomba statue, getting a 67hspd jump that barely misses the lava geyser. Unfortunately this kind of jump couldn't make it onto safe ground, but the lava boost and movement on the sloped boundary was 20 frames faster than holding it back onto the ice platform and long jumping.
#supermariogalaxy2 #toolassistedspeedrun
![[TAS] SMG2: Freezy Flake GS2 in 37.95s](https://i.ytimg.com/vi/j_J_UWSEdpM/mqdefault.jpg "[TAS] SMG2: Freezy Flake GS2 in 37.95s
Green Stars TAS WIP
37.95s IGT - 60:59.31 real-time split (1:34.05 segment)
Freezy GS1/2 were two of the few stars I hadnt tested prior to starting this TAS, as star order didnt seem like itd matter (more about that later).
The first launch star is a balance of optimisation for the earliest statue break (not waiting so long for the launch star to activate) while still being able to get to the launch star on its first active frame.
While other routes looked cooler (e.g. getting turned backwards in mid-air to have more speed), the fastest by 29 frames was how RTA runners do it (used Pottowws IL as a blueprint: https://youtu.be/QWP7saYcBaw?t=10).
An instawalk from the launch star is common here, but interestingly, entering the launch star while turning ends up getting landstun, and entering while standing completely still gets faster and earlier acceleration.
1-flower cycle was attempted but failed (https://youtu.be/6GEkxIaaQ-k), missing out by around 0.6s.
With certain movement, conical slopes (e.g. the mountain, some crystals) can cause some very buggy fast slides for a frame or two, leading to glitchy slope jumps.
The consequence of slope jumps being patched in SMG2 is that they now stay at a constant speed and trajectory.
It is difficult to control the trajectory, and the jump in the 1-flower attempt has a much more wayward and slow setup after getting the 2nd flower.
As with the Tall Trunk slide, you dont lose any time with whatever movement you do provided you dont jump a lot.
However, exiting the tunnel without jumping is a little slower and doesnt let you spin/throw fireballs. Also, the gravity you leave the tunnel at will have an affect on your angle, sometimes changing angle mid-air.
Backwards exit with a mid-air fireball was 35 frames faster than the RTA method: forward exit, jumping on the lower brick block and fireballing around the corner.
Star order almost mattered at the end. The fastest regular route (by 6 frames) was to long jump onto the flying ? block, and its cycle changes with cutscene skips.
But another glitchy slope jump was found on the Goomba statue, getting a 67hspd jump that barely misses the lava geyser. Unfortunately this kind of jump couldnt make it onto safe ground, but the lava boost and movement on the sloped boundary was 20 frames faster than holding it back onto the ice platform and long jumping.
#supermariogalaxy2 #toolassistedspeedrun")
![[TAS] Super Mario Galaxy 2: Beat Block Green Star 2 in 22.16s](https://i.ytimg.com/vi/jkGB2xNf978/mqdefault.jpg "[TAS] Super Mario Galaxy 2: Beat Block Green Star 2 in 22.16s
Luigi eeeeeyahs to the beat to collect a Green Star in 22.16 seconds [IGT, 1330 frames]
I started on Green Star 1 as I was testing both stars for cutscene effects. Havent actually done a cutscene run with either yet, so Ill update this with results later.
This isnt fully optimised, but its certainly fast. At the start, theres a good use of a short spin that doesnt lose too much speed so that the next long jump can happen earlier than normal.
I did have to slow the penultimate long jump on the green block cycle at 00:25 to be able to land for the last long jump in that section.
The second long jump up the steps was a frame-perfect long jump on the very edge, which let me conserve running speed when I landed, which was useful in getting a faster line-up to long jump to the corner of the last green block.
I landed awkwardly on this, which caused Luigis backflip to be stuck in a walking pose.
#SuperMarioGalaxy2 #ToolAssistedSpeedrun")
![[TAS] Super Mario Galaxy 2: Tall Trunk Green Star 1 in 10.35s](https://i.ytimg.com/vi/jraRdYBEKbc/mqdefault.jpg "[TAS] Super Mario Galaxy 2: Tall Trunk Green Star 1 in 10.35s
Luigi keeps kicking a tree to collect a Green Star in 10.35 seconds [IGT, 621 frames].
UPDATE: This has been improved to an 8.83: https://youtu.be/qshjDR8ch4M
At the start, I jump backwards off the tree stump to get in the position to pull off a tilted triple jump. While uploading this, I found that I couldve easily saved 9 frames by just changing the angle of the jump so I didnt have to jump as high to get over the edge, but oh well, add that improvement to the full game run.
The tilted triple jump lands on the planets hard curve in such a way that Luigi keeps the gravitational tilt of the rounded half while running onto the flat half, so hes almost running on his head. This causes a jump to go much higher than normal, and this triple jump spin wallkick spin can reach a high ledge.
This saves ~3s over triple jumping straight to the lower ledge on the left and getting to the higher ledge from there, and saves ~3.5s over the IL route that gets on the lower ledge and long jumps around the left side.
Almost every wallkick in this used neutral frame wallkick strain; if you go neutral on the joystick for even just a single frame after wallkicking, you can get significantly harder strain which can help go further sideways or backwards.
The following backflip was aimed sideways so that I could hold up and get lots of strain to make the angle of the wallkick much more sideways than normal.
I changed angle when landing to be flush with the tree when backflipping, since I didnt want to strain very hard for an angled wallkick here (but I did still go for a soft strain to avoid spinbonking against the ledge above).
On the last ledge, I performed a Z sideflip - before landing on the last ledge, I held up slanting left (opposite the facing angle), pressed and held Z (above the floor so it didnt do a ground pound) and then pressed A. A sideflip here was better than a backflip as it jumps 1 frame earlier.
This tilted triple jump route is ~0.6s slower than the IL route for Green Star 2 and the regular star.
Interestingly, it is a little bit faster to do GS2 first in a full game run, because the angle you land in after the cutscenes is really bad for setting up the tilted triple jump for GS1 - its the complete opposite angle to the optimal backwards jump.
#SuperMarioGalaxy2 #ToolAssistedSpeedrun")
![[TAS] SMG2: Cosmic Cove Green Stars](https://i.ytimg.com/vi/jzX9etyV_Gw/mqdefault.jpg "[TAS] SMG2: Cosmic Cove Green Stars
Green Stars TAS WIP
Green Star 1 - 20.58 IGT - 38:03.86 real-time split (1:16.51 segment)
For a more in-depth description, see the individual video here: https://youtu.be/XMDBlu3_u48
Brief notes:
- The start doesnt have any jumps in order to preserve hard landing state, which allows
the penguin launches to have uncapped horizontal speed.
- The launches work by clipping into the penguin, and then getting clipped out while spinning. If the clip-out reaches the surface of water, you get a spinny dolphin jump with all the clipping speed.
GS2/GS3 star order used to matter before new GS3 strats, as the cycles of the moving water blocks are different between cutscenes and no cutscenes. Neither of the two have water blocks sitting in the way of the shore launch, so there is no difference now.
Green Star 2 - 1:04.71 IGT - 40:03.76 real-time split (1:59.90 segment)
Ice-skating mechanics:
- Skating coasts at 18hspd, and spinning adds a boost of 9.25 to whatever the current speed is, but doesnt boost if current speed is above 18.75
- Long jumps have a fixed speed, while standard jumps are based on ground speed, while adding a small boost on top. Ground speed is conserved through jumps.
- There are two different deceleration rates of spin-skates. No full list of what causes what, but the lesser rate kicks in when youve been on ground for a short time. This is why I wait to jump at the start.
- Also for some reason uncrouch spin gained more speed than a regular spin for that initial one.
In the cave, 2P brick break + backwards slope jump + landing on the slope
just before entering water (so that holding Z will give a water long jump) ended up 0.4s faster than the nearest challenger (backwards sliding).
In the last section, double jump would be faster but this would prevent being able to long jump later on (even if Z is held during the jump), so a long jump spin is much better.
For the long jump, you can barely clip into the corner of the ceiling, and this causes a weird push out of the waters surface that lets you instantly long jump. This saved 0.8s over the normal water lj method, and 0.9s over a really cool backwards long jump.
For the star grab, a similar corner clip wasnt useful as water walls dont act like surfaces. Auto-ledgegrab is faster by 8 frames over backwards spinning out of the water.
Green Star 3 - 1:11.00 IGT - 41:58.10 real-time split (1:54.34 segment)
I was able to compare to a 2020 IL TAS I did with this one: https://youtu.be/FB2KuYX9OZo Saved 0.53s in the first section,
and another 0.91s in the cave.
The corner clip somehow conserves double jump state, giving a triple jump out of water. This saves 0.3s over the strat done in 2020 (jump & spin onto the curved ground, backflip into the water).
The Goomba launch works the same as penguin launches.
To get it high enough for the clip-out to reach the surface, some A Button Challenge tech came in handy: Goomba raising. When stunned by a spin or starbit, Goombas recoil up into the air. Spamming them with starbits can gradually lift them up.
While this causes a lot of hitlag, it was faster than the strat used in 2020, and also meant the Paragoomba could be 2P frozen higher up to make the next trick quicker.
The only downside to the launch is that when it hits the water above, Luigi immediately dives down. Fortunately not too much for 2P to let itself go of the Paragoomba.
Because Goombas immediately explode upon touching water, a Goomba launch isnt possible with the Paragoomba (it only clips Luigi down), but it can clip Luigi into the wall, where a shore launch is possible.
This is enough to reach the star; holding neutral during the launch allows for a spin once Luigi has reached the peak.
#supermariogalaxy2 #toolassistedspeedrun")
![[TAS] Bonsai Barber Perfect Day in 1:53.36](https://i.ytimg.com/vi/kwX_I9qlvIM/mqdefault.jpg "[TAS] Bonsai Barber Perfect Day in 1:53.36
Bonsai Barber is a WiiWare barber-sim made by Zoonami, a team led by the director of Goldeneye N64. Inspired by the Japanese art of bonsai and the barbershop experience, it involves giving charming anthropomorphic plants a fresh trim, with styles ranging from simple shapes to bizarre structures.
The game has a real-time mechanic where you can only cut up to 5 plants per real-world day.
This TAS aims to make an entertaining and fast run to complete the Perfect Day achievement on the first day. Due to bottlenecks in the games reward system, absolute speed isnt of the essence, so theres enough time to mess around with painting, birds, and apples.
10060 rerecords
april fools, lads, the joke is that i spent time on this instead of finishing galaxy")
![[TAS] SMG2: Flip-Swap Green Star 1 in 0.5xA Presses [Luigi]](https://i.ytimg.com/vi/l9m2kxQx2MI/mqdefault.jpg "[TAS] SMG2: Flip-Swap Green Star 1 in 0.5xA Presses [Luigi]
Goomba manipulation and triple bouncing into the star all while holding A.
This isnt a new strat but I made this as there wasnt yet a video of the star in 0.5xA. Its a difficult strat so having a video for proof is helpful.
This wasnt the fastest or cleanest completion but it worked so thats all that matters.
P2 was just chillin, ready for speed tests.
Aiming the starbit and timing the spin to stop the Goomba from falling is the most precise part, but there is also some precision to the triple bounce, as you have to specifically land on the middle of the Goomba - the sides make the bounce a bit too low.
#SuperMarioGalaxy2 #AButtonChallenge")
![[TAS] SMG2: Rolling Coaster Green Stars](https://i.ytimg.com/vi/laX3SZgt9A0/mqdefault.jpg "[TAS] SMG2: Rolling Coaster Green Stars
Green Stars TAS WIP
Star order doesnt matter here - the position of enemies doesnt change enough to make differences for either star.
Green Star 2 - 46.31 IGT - 2:25:22.90 real-time split (1:42.44 segment)
This is an improvement of 5.59s over my 2020 TAS, despite cutscenes: https://youtu.be/5zBcZT8DuUc
The main improvement is the spiral skip found by Sheep: https://youtu.be/XHtPWADnpds?feature=shared
This first bonks off the pillar and then jumps the frame after to keep high speed (around 85hspd) that is enough to get over the death plane and land at the bottom of the spiral.
The landing has to be precise and land on the rail along the edge in order to come to a complete stop, otherwise Luigi would roll down and off the other side.
A couple of smaller improvements:
- 0.26s at the start by jumping off the sloped edge of the platform to gain speed, rather than a chain of jumps (bunnyhopping, or bouncy-balling I guess) on the fences
- 0.55s plus higher speed for the skip by bunnyhopping from near the end of the rainbow halfpipe slide to maintain the higher speed from that slide (rolling on flat ground decelerates faster, jumping maintained the downhill speed)
Green Star 1 - 21.13s IGT - 2:26:27.36 real-time split (1:04.46 segment)
This is an improvement of 1.82s over my 2020 TAS: https://youtu.be/D5V8ibOL_tc?feature=shared
The main improvement is the same bunnyhopping technique, but this time it saved even more time (0.94s) to start much earlier and set up landing on the sloped side of the halfpipe slide right at the end, as this gets a much faster jump and leftwards trajectory through the Bob-ombs. This set-up didnt work out the opposite way for Green Star 2.
#supermariogalaxy2 #toolassistedspeedrun")
![[TAS] SMG2: Fleet Glide Green Stars](https://i.ytimg.com/vi/mEP0JCh_u28/mqdefault.jpg "[TAS] SMG2: Fleet Glide Green Stars
Green Stars TAS WIP
Star order has no real effect - the Thwomps may be a couple of frames different but nothing else.
Green Star 2 - 54.06 IGT - 1:49:23.00 real-time split (1:49.85 segment)
Compared to my 2020 TAS (https://youtu.be/DA8 G1KBlo), the flying section was improved by 2 frames (42.23 on timer vs 42.26).
I struggled to keep up with the 2020 TAS without copying inputs, and even adjustments to try to improve only got to being 0.2 units ahead by the tunnel entrance (Fluzzard Hspeed is generally 47 so thats a very tiny difference).
The main issue is there is a lot of downhill - the holes at the start, the tunnel, especially the Banzai Bill part, so alternating between boost and lift to get more distance like in Wild Glide is not as beneficial here as it also ends up higher, so you lose that extra distance with sharper downtilts later.
The framesaves came right at the end of the tunnel, where the bottleneck is no longer downhill but an uptilt right at the end.
A light uptilt followed by 1-frame sharp tilt under the ceiling saved a frame on that last stretch of tunnel compared to the gentle curve of tilt from before, and optimising that + general better movement to the star saved another.
Green Star 1 - 14.18 IGT - 1:50:19.86 real-time split (56.86 segment)
IL TAS is 13.98 (https://youtu.be/vDHeMTBJWAo).
I would usually delay the cutscene skip to optimise in-game time without losing real time, however, theres a quirk with the Jibberjays dialogue where - after skipping cutscenes - it has some cooldown or inactive state before the dialogue box opens, so you could press A 0.3s later than possible and still be as fast as pressing A as early as possible.
Skipping cutscenes earlier than I usually would manages to cut down a bit (but not all) of the waiting time, so its faster RTA by 13 frames.
Because this quirk gives freedom for suboptimal movement, it means there is space for another optimisation quirk - camera cutscene movement. For dialogue or other mini-cutscenes, the camera needs to reach a certain point to rest before the box appears or the thing happens, so frames can be saved by getting the camera closer to its point before activating the scene/dialogue.
If heading straight to the Jibberjay, the camera really slowly moves to its point, so instead I run and slide around the right, saving 7 frames on camera movement (which seems to not be limited by the cooldown quirk I guess?).
Also 7 frames difference was bigger than the IL of 6 frames difference. Maybe 1 more frame to save in IL!
As nemi43 found, tilting up and slightly to the side grabs the star 1f earlier because hitboxes.
#supermariogalaxy2 #toolassistedspeedrun")
![[TAS] SMG2: Stone Cyclone Green Stars](https://i.ytimg.com/vi/mIa6TaLJo2E/mqdefault.jpg "[TAS] SMG2: Stone Cyclone Green Stars
Green Stars TAS WIP
Star order matters for Tox Box cycles. It is faster to do GS1 first by 1.33s.
Green Star 1 - 8.55 IGT - 2:29:48.26 real-time split (1:04.66 segment)
The cutscene cycle strat is slower IGT than the IL strat (here TASed by Sheep: https://youtu.be/nGh7FFfgDbU ) but the cycles make the IL strat too slow - you barely miss making it onto the first Tox Box in time.
As for the Tox Box used, if I had tried to slope jump on the first flip, I would have gone sideways with the momentum from the prior backflip (which is done for the old GS2 strat). So I landed further down onto the edge to be able to get a regular jump to line up an extremely fast fling into the star.
Green Star 2 - 10.92 IGT - 2:30:42.56 real-time split (54.30s segment)
This uses the massive Tox Box fling found by Sheep: https://youtu.be/nGh7FFfgDbU?t=38
In my previous Stone Cyclone TAS from 2021, I added a clip of almost getting sent straight into the star from this Tox Box: https://youtu.be/2R9-WyeAbMY?t=44
Sheep then made the strat work out by redirecting the jump to the right and jumping fast enough from high enough up the Tox Box to scrape over the platform and ground pound onto the star.
While it saves 3.25s in-game time when delaying cutscene skips, I still have to wait for the Tox Box, so this only saves 0.38s in real time compared to the old GS2 strat you can see in my 2021 TAS.
#supermariogalaxy2 #toolassistedspeedrun")
![[TAS] Super Mario Galaxy 2: Honeybloom Green Star 2 in 59.31s](https://i.ytimg.com/vi/mKKOJihre2Y/mqdefault.jpg "[TAS] Super Mario Galaxy 2: Honeybloom Green Star 2 in 59.31s
Luigi and P2 fly through the bees garden to collect a Green Star in 59.31 seconds [IGT, 3559 frames].
This was optimised for real time comparison rather than in-game time, so 0.18s IGT was lost by skipping the flying cutscene.
Honeybloom GS2 joins the sub 1 club!
The first major change from the route is landing early from the swing and backflipping up the wall, instead of landing later and bouncing off the Goomba from a long jump. This saved a few frames.
The fastest method from this point was doing one wallkick and getting a long jump, spinning to the wall and wallkick spinning off the wall to the ledge with the Bee shroom, and backflipping up. This was faster than the usual double wallkick, no long jump, sideflip, backflip method by 0.5s.
The next section looks clunky, but is faster than several other routes I tested. After the first long jump, I held neutral for half of the single jump in order to avoid the Bee shroom, with the neutral landing still keeping most of the momentum for the double jump onto the bee.
The triple jump is the reason why this ended up faster, being able to get straight to the first cloud without going to the first flower.
I first tried doing a sideflip from the cloud and going through the flower to get warped up onto the second cloud, then backwards jumping to the checkpoint ledge.
However, I then tried jumping onto the flower and doing a sideflip from there and seeing if a spin could make it that way onto the ledge. It was just a bit short.
But by freezing the copter enemy (could also work with starbit stuns) I could use a weak spin to change how Luigi would collide with the enemy, including changing how the push effect would work - without the weak spin, Luigi would have done a somersault animation and got pushed to the left, but with the spin, he instead just floated up off the enemy and continued going right.
This saved 10 frames over the original sideflip-flower warp-backwards jump strat.
Long jumping on that ledge was faster than running and turning by a few frames, despite getting stuck on the wall for so long.
Crouch spin knockback was a few frames faster for grabbing the vine than just dropping off.
With the Bee suit, going upwards reduces horizontal speed to 8 units/frame, but while dropping, hspd goes up to 10.25 u/f.
With this in mind, mashing A to slowly hover down is the best horizontal tactic for Bee flight.
Hovering down past the first cloud was a bit faster than going up onto it and continuing from there, by only 5 frames.
On the cloud I landed on, I waited for it to bounce up to get more height in the jump, which was a bit faster.
For the pole, I spun just before grabbing to get a faster turn on it.
Sideflipping up the first ledge was faster than backflipping by 1 frame, due to being able to execute it 1 frame earlier.
Sideflipping up the second ledge was faster by a few more frames as backflipping needed to land further in on the first ledge in order to work, meaning it would have to land later, whereas sideflipping just needed to land on the edge.
Long jumping into the Bee shroom is 2f faster than jumping or running.
Having P2 bring a coin to refill some of the Fly meter saved 6 frames of the refill time spent on the swings log.
Jumping and starting the final flight with neutral joystick was faster as it gains a bit more height than holding a direction, and height was the key factor to grabbing the star early.
Spinning into the star saved 5 frames.
#SuperMarioGalaxy2 #ToolAssistedSpeedrun")
