Name: SMGalaxy - Matter Splatter Mansion - last room strategies @Joselle_Astrid
Uploaded: 2026-09-16T08:17:00-05:00
Duration: 8 min 17 s
Description: SMGalaxy - Matter Splatter Mansion - last room strategies

SMGalaxy - Matter Splatter Mansion - last room strategies

Joselle Astrid 2026-09-16 | SMGalaxy - Matter Splatter Mansion - last room strategies

F-Zero GX - CPSO - 046415 with Fat Shark (Max Speed non-SA)

Joselle Astrid 2022-01-05 | 17"762 - 14"298 - 14"355, 2021.01.03.
I started playing this course to secure a Max Speed PB in 2021, which I did, thus keeping my streak of at least one Max Speed PB per year since 2012. I then kept playing to get a run I was satisfied with, which I did here!

PB progression:
2014.06.24 - 0'48"084 - youtube.com/watch?v=Y2wtYwCn3WI
2021.12.28 - 0'47"792 - youtube.com/watch?v=4t2DvVB5Ct4
2021.12.31 - 0'46"971 - youtube.com/watch?v=N0Or5Wz2buQ
2022.01.03 - 0'46"415 - This run

My best splits are 17"341 - 13"930 - 14"112, though there's a risk/reward factor with the jumps which makes it tough to optimize all the way. I figured a reasonable goal for me would be 17"7 - 14"3 - 14"3, and that's almost exactly what I got.

My jump strat of choice was the MTSISB (MTS into shift boost) jump. mtthew was the first to show this technique for the CPSO jump in his 48"290 Black Bull run (youtube.com/watch?v=JCuAX5NLBMY). Then in the past year or so, KAZOON (youtube.com/channel/UCJ7rQzTrQ_porN7oAujGtNw) put a lot of work into further developing this MTSISB jump - making it work for Fat Shark, Blue Falcon, Space Angler, and just about every other MTS-capable machine.

This jump is very flexible in a sense, since it works for a wide range of approach speeds. However, you do have to adjust your MTS timing and positioning accordingly. I think in general, if your approach speed is faster, you want to be positioned more to the right before MTSing, and you'd generally do the MTS earlier.

Using KAZOON's 45"454 run as a guide, a good target for Fat Shark is to be on the right side, with speed in the mid-1800s, when starting the MTS. The two most common ways to fail are:

- 1. Your machine starts lifting off before you can MTS, and thus you fly off way too far right to your doom.
- 2. You overcompensate the timing/positioning and fail to lift off entirely. This probably means you MTSed too early or too far left, given your approach speed.

Trying for faster approach speeds and earlier MTSes may help to get even faster jumps, but at the cost of consistency. The sweet spot between cases 1 and 2 above seems to be pretty small when your MTS-start speed is in the 1900s or faster. Also, when you jump earlier, you may start to have an issue of flying too far to the right after the MTSISB.

The thing is, Fat Shark 'naturally' reaches the MTS-start point at over 2000 km/h with good driving, so I have to tap the brakes (the X button on my input display) to get into the 1800s. Now, I'm usually concentrating on the dash plates and not looking at the speedometer here. So I have to get a feel for how much I have to brake (possibly a little more on boost laps), as well as when to MTS.

Sometimes I get in a rhythm, and sometimes I don't get any good laps for 10 minutes. Still, when I had tried this jump at high 1900s / low 2000s, my success droughts were more like 40 minutes instead of 10 minutes - so the braking really helped me.

About the dash plates strategy: KAZOON got 45"454 by skipping the 3rd and 6th dash plates, but still using small quick turns to preserve speed (at least on Lap 1). From my own testing, I also found this to be a good strategy. I wasn't getting to the jump point any sooner by hitting the 3rd and 6th dash plates. Hitting only the 3rd might've even been a couple frames slower. Hitting all the dash plates may help to reach the jump point with higher speeds, but it doesn't matter for this particular jump strat since I'm braking anyway.

Course: Casino Palace: Split Oval
Machine: Fat Shark
Setting: 100% (all the way to the max speed side)
Rules: 100% settings only, no snaking
Other category details: No side attacks

Date: 2021.01.03 22:12 UTC-0800
Played on: NA version, run with Wii RVL-001 (NA) using Wii disc channel
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS (480p, 59.94 FPS), with NintendoSpy input display
Editing: VapourSynth (upscaling), aacgain (audio normalization), ffmpeg (audio sync)

Celeste Classic - All strawberries in 3:31

Joselle Astrid 2021-11-28 | I had fun playing through Celeste earlier this year, and wanted to show off some parts of the game. Well, or parts of the classic PICO-8 game in this case. I haven't looked up speedrun strats, just seeing what I can figure out.

3 deaths: 1400m, 2400m, and 2500m, but all not very costly. At the end of 1500m, I missed the dash and awkwardly stopped, since my brain had moved on to the next screen. The rest went pretty much to my plans. 2600m is the hardest screen for me, so it was pretty nice to not die there (or at any point later).

I know of a better 2100m strat, but couldn't do it very consistently, so I went the long way.

Date: 2021.11.27 03:27 UTC-0800
Played on: Celeste, Steam version 1.4.0.0 (Windows 10), PICO-8 mode
Capture software: OBS with x264, capturing game at 720p60 + Open Joystick Display
Controls: iBuffalo BSGP801

Pikmin 3 Deluxe - Any% - 43:15

Joselle Astrid 2021-06-14 | I got this run in April, and initially just uploaded to my Side B channel since I was expecting to improve again soon - but I ended up putting off further practice/runs for a while. So now I'm re-uploading this to my main channel for better visibility (and to improve the audio sync a bit). I want to get back to this category around August or so, but I do feel pretty good about the level I've demonstrated so far.

I learned most of the strats and mechanics from Claris, who got the first sub-40 and sub-39 Any% runs back in December and is still ranked 2nd for the category. She is also running the game for SGDQ 2021 (early July), and I'll be on commentary!

For anyone else interested in learning this speedrun, I've written everything I know about the strats here: github.com/JoselleAstrid/game-notes/blob/main/Pikmin-3-Deluxe/Any-percent_Normal.md

As for this run: Other than maybe Day 6, none of the losses stand out that much from the others. From here I'd want to just polish most of the strats overall, and get more experience with different situations so they don't catch me off guard as often.

Twitch highlight - twitch.tv/videos/978505902

Day 1 - I should review the best strats for the sheargrubs and 5 pellet.

Day 2 - Pretty good for me, though there's decent room to clean up my play. There are so many little optimizations to think about in this day.

Day 3 - Good. Always easy to lose a lot of time here, especially in the beginning cave.

Day 4 - Stalled for several seconds at the end when deciding whether to wait for the orange to get completely beamed in. I wasn't sure if I had to wait for it, but turns out I did have to (unlike the Whiptongue Bulborb). Would've been better to charge Pikmin onto the orange before charging onto the bulborb.

Day 5 - Orange Bulborb got close to passing the gate. That was close! My cursor wasn't aimed correctly to whistle off the Winged Pikmin (I can't put the cursor "on" the gate, I should point at the ground below instead). Louie carry after the boss needs work; I could only get 7 Pikmin on this time, and I'd rarely been getting more than 8. At the Louie clip, I could have whistled the Pikmin off of Louie sooner, but at least I didn't need multiple attempts.

Day 6 - I forgot to make Brittany 'Go here' to the Blue Onion, so that was the biggest single mistake in the run (15-20 seconds). I'm still kind of sloppy at the end of the day. The 3 Rock Pikmin that got left behind were all around the Aristocrab, I think.

Day 7 - The bomb rocks got Plasm Wraith down to lower health than usual; I don't remember seeing it that low before. Usually the Pikmin have to attack it for a few more seconds to finish it off. Maybe the bomb rocks were perfectly timed with the blob drops?

Category: Any% (beat the game), Normal difficulty

Date: 2021.04.06 23:48 UTC-0700
Played on: Version 1.0 (digital), English language, using joycons
Capture hardware: Elgato HD60S
Capture/recording software: OBS, x264
Editing: Avidemux (trim), aacgain (audio normalization), ffmpeg (audio sync correction)

Super Mario Galaxy - Any%, 1P Luigi - 2:35:49 (2011/12/28)

Joselle Astrid 2021-05-31 | Hi! I've recently come out as a trans woman. I now go by the username Joselle Astrid (or just Joselle), and my pronouns are she/her. I used to go by yoshifan28.

To prove it's still me, here's a vintage Galaxy run! I recently went digging for any runs I still had on DVD from 2011, the year that I started attempting full speedruns of this game. I didn't think much of chronicling my runs at the time, but Galaxy's quite a prominent game for speedruns by now, so it's good to try and preserve some of the history behind it.

Full runs I've saved from 2011:
2:40:10 - April - youtu.be/YGD6bgmz1NY
2:38:09 - October - youtu.be/sVrZQM73wrA
2:35:49 - December - (this video)

I can confirm this was my last PB run before my 2:35:20 at AGDQ 2012.

After this run, I demonstrated 6 non-Any% stars (included in this video): Dreadnought 1 and 2, Melty Molten 1 and 3, Toy Time P and 3. These were my candidates for stars to show off as a bonus after the GDQ run, and I ended up picking three (DN2, MM3, TTP).

This is highlighted on my Twitch as "2:35:14", but the highlight is broken. Part 1 is 1h53m long, and "part 2" is just the first part of part 1. Still, you can check the highlight to verify that the early game is the same: twitch.tv/videos/45590557
I was capturing with a modest laptop at low quality, which is why I was recording on DVD as well as streaming.

I believe the old "2:35:14" timing (which I also mentioned at the end of the GDQ run) was due to two things. First, I used to consider the start of timing to be the start of Luigi's control in Gateway, rather than file select, which meant I was starting about 27 seconds later. Second, I believe I timed the run by taking the number of DVD-recorded frames and dividing by 60, rather than by 59.94, which meant the timing was 8-9 seconds too fast.

This was cleaner than my 2:38:09 to start, though I still had my share of mistakes, like losing 45 seconds getting up the tree in Gold Leaf 3. Compared to the 2:38:09, my Kitchen and Bedroom routes were rearranged slightly. I think I moved the Freezeflame stars earlier to ease the star bit collection task before Sling Pod. I'm not sure why I switched up the Kitchen, maybe I thought it was a bit simpler to remember if I did all of Beach Bowl at once.

Route:
Gateway, Good Egg 1 2 3, Honeyhive 1 2 3, Flipswitch, Robot Reactor
Space Junk 1 2 3, Rolling Green, Battlerock 1, Space Junk C, Star Reactor
Beach Bowl 1 2 3 H C, Ghostly 1, Sweet Sweet, Honeyhive C, Good Egg H, Bubble Breeze, Ghostly 2 H, Airship Armada
Gusty Garden 1, Freezeflame 1 2 C 3 H, Gusty Garden 2 H 3, Buoy Base 1 H, Battlerock H, Hurry-Scurry, Sling Pod, Honeyclimb, Gusty Garden C, Dark Matter Plant
Gold Leaf 1 2 H 3 C, Toy Time 1 H 2 C, Sea Slide 1 C 2, Drip Drop, Honeyhive H, Galaxy Reactor

Any% speedrun - Start at the beginning of Luigi's quest, Finish by completing Bowser's Galaxy Reactor
Luigi, 1P, 4:3 aspect ratio
Time from file select (frame 1753) to Grand Star cutscene (frame 562183): 2:35:49.84

Date: 2011.12.28 UTC-0800 (Based on Twitch highlight date, Wii menu date before the run, and practice file date after the run. My Wii Menu time was not guaranteed to be accurate, but probably within a couple hours.)
Played on: NA version, NA Wii RVL-001 + Wii disc channel
Capture hardware: DVD recorder, LP quality with a DVD-RW disc
Editing: Anri-chan 3.3, IQ, with .avs edited to upscale to 720p

Super Mario Galaxy - Any%, 1P Luigi - 2:38:09 (2011/10/08)

Joselle Astrid 2021-05-31 | Hi! I've recently come out as a trans woman. I now go by the username Joselle Astrid (or just Joselle), and my pronouns are she/her. I used to go by yoshifan28.

To prove it's still me, here's a vintage Galaxy run! I recently went digging for any runs I still had on DVD from 2011, the year that I started attempting full speedruns of this game. I didn't think much of chronicling my runs at the time, but Galaxy's quite a prominent game for speedruns by now, so it's good to try and preserve some of the history behind it.

Full runs I've saved from 2011:
2:40:10 - April - youtu.be/YGD6bgmz1NY
2:38:09 - October - (this video)
2:35:49 - December - youtu.be/o10PNHMAp3Q

This seems to be the only run I have saved between my 2:40:10 in April and my 2:35:49 in December 2011. I know my full runs (as opposed to practice sessions) were very infrequent, and I also know I wasn't focused on full runs from April to July (when I finally finished my segmented run). Overall, if I had any more missing PB history in this time period, I doubt it was more than a couple of PBs missed.

According to the Galaxy WR history spreadsheet, this had apparently been highlighted on my Twitch as "2:37:33" on 2011/10/08, but the highlight went missing on Twitch at some point.

I was capturing with a modest laptop at low quality, which is why I was recording on DVD as well as streaming. However, the DVD recording is quite low quality too, since it was set to EP quality for some reason. This was much worse than LP (the next quality level up) because instead of interlacing 2 consecutive frames together, the recorder would just blend the 2 frames together, making it impossible to separate the 2 frames to get 60 FPS video. As such, the video is very blurry and artifacted and it can't be helped. Take care not to watch too closely and for too long at a time. Sorry!

I believe the old "2:37:33" timing was due to two things. First, I used to consider the start of timing to be the start of Luigi's control in Gateway, rather than file select, which meant I was starting about 27 seconds later. Second, I believe I timed the run by taking the number of DVD-recorded frames and dividing by 30, rather than by 29.97, which meant the timing was 8-9 seconds too fast.

It's evident that I wasn't in the habit of resetting runs, as I lost over 40 seconds in Gateway Galaxy, and a fair bit more time in the rest of the Terrace. The best part of the run? Not dying on Conquering the Summit (1:53:10). It would've been faster to just fall and respawn, but where's the fun in that?

Route:
Gateway, Good Egg 1 2 3, Honeyhive 1 2 3, Flipswitch, Robot Reactor
Space Junk 1 2 3, Rolling Green, Battlerock 1, Space Junk C, Star Reactor
Beach Bowl 1 2 3 H, Ghostly 1, Sweet Sweet, Good Egg H, Honeyhive C, Beach Bowl C, Bubble Breeze, Ghostly 2 H, Airship Armada
Gusty Garden 1 2 H 3, Buoy Base 1 H, Battlerock H, Hurry-Scurry, Sling Pod, Gusty Garden C, Freezeflame 1 C 2 3 H, Honeyclimb, Dark Matter Plant
Gold Leaf 1 2 H 3 C, Toy Time 1 H 2 C, Sea Slide 1 C 2, Drip Drop, Honeyhive H, Galaxy Reactor

Any% speedrun - Start at the beginning of Luigi's quest, Finish by completing Bowser's Galaxy Reactor
Luigi, 1P, 4:3 aspect ratio
Time from file select (frame 947) to Grand Star cutscene (frame 285354): 2:38:09.71
(+/- 0.02 due to being 30 FPS)

Date: 2011.10.08 UTC-0700 (Based on former Twitch highlight date, Wii Menu date before the run, and game file date after the run. My Wii Menu time was not guaranteed to be accurate, but probably within a couple hours.)
Played on: NA version, NA Wii RVL-001 + Wii disc channel
Capture hardware: DVD recorder, EP quality with a DVD-RW disc
Editing: Anri-chan 3.3, IQ, with .avs edited to upscale to 720p without deinterlacing

F-Zero GX Unleashed 4.0 - Chapter 7 (TA) - 123713 with Mighty Hurricane (80% non-SA)

Joselle Astrid 2021-01-06 | When building F-Zero GX Unleashed, there's an option to replace some of the regular Time Attack courses with Story courses - specifically the ones from Chapters 3, 7, 8, and 9. Here's Chapter 7, which is also being dubbed "Mute City: Championship Circuit" for Unleashed's purposes (although that name doesn't show in-game).

There are a couple of tricky narrow turns, so I wanted a machine that was relatively easy to handle, and I also wanted something that could get strong mine boosts. So, Mighty Hurricane came to mind. It doesn't have the best body stat for taking mine damage, although it's not as bad as advertised either (it really should be C or D, not E).

I accidentally ran through a bit of lava before the Lap 2 refiller, and also probably should've boosted later so I could stay in the refiller longer. Without those errors, I could've had an extra boost for Lap 3. As for Lap 3... how's that for an explosive finish? I have no idea if getting those kinds of speeds is consistent, but my success rate on it stands at 1 out of 1.

I was racing the Expert non-SA staff ghost, but turns out I also surpassed the Grandmaster non-SA staff ghost (1'24"109) with this run: youtube.com/watch?v=Hto59nu1uvU&list=PLmh9KAbzYQuhWpycVAwjDFwrl2PawzwVQ&index=43
Didn't expect that, so I didn't think to save a ghost of my run! Oh well.

There are various other GX mods that can be applied which I haven't shown in these Unleashed videos, such as alternate track textures and music replacements. You can find more on GX modding here: gamebanana.com/games/5899

Course: Chapter 7 - "Mute City: Championship Circuit"
Machine: Mighty Hurricane
Setting: 80%
Rules: Expanded Max Speed - 80%-100% settings
Other category details: No side attacks

Date: 2021.01.01 20:52 UTC-0800
Played on: Unleashed version 4.0, run with Wii RVL-001 (NA) using Nintendont (modded to tolerate GC stick drift) + Sony SF-32UY SD card
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling, crossfade), ffmpeg (audio normalization)

F-Zero GX Unleashed 4.0 - ASD - 057269 with Silver Rat (100% non-SA)

Joselle Astrid 2021-01-06 | Silver Rat was really lacking redeeming qualities in GX, so in Unleashed, it was upgraded from having a near-nonexistent MTS to a very good one. I'm only showing some of its potential here, as I didn't practice enough to maximize the MTS speed at each turn, and I also didn't go for the explosive finish strat at the last hairpin. I thought the loop strategy on boost laps was pretty decent though.

This beats the Expert non-SA staff ghost of 0'59"041: youtube.com/watch?v=T8HWnBNyWWk&list=PLmh9KAbzYQuiMD3pvFhGkowPhiBOHvsoX&index=24
However, I'll need to raise my level to match the Grandmaster non-SA staff ghost of 0'53"410... coincidentally done with Silver Rat (at 80%): youtube.com/watch?v=QV9Co5m-vCU

Course: Aeropolis: Screw Drive
Machine: Silver Rat
Setting: 100%
Rules: Expanded Max Speed - 80%-100% settings
Other category details: No side attacks

Date: 2021.01.01 20:22 UTC-0800
Played on: Unleashed version 4.0, run with Wii RVL-001 (NA) using Nintendont (modded to tolerate GC stick drift) + Sony SF-32UY SD card
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling, crossfade), ffmpeg (audio normalization)

F-Zero GX Unleashed 4.0 - LTR - 251225 with Space Angler (80% non-SA)

Joselle Astrid 2021-01-06 | Space Angler was a contender for the slowest machine in GX, but it's a viable and unique pick in GX Unleashed, boasting one of the highest top speeds in the game. AKC advised me that 80% settings would be better than 90-100, though, sacrificing a bit of raw max speed for better accel. I took it for a spin on a long course, and it seemed pretty effective.

I didn't grind this run quite enough to beat the Expert non-SA staff ghost (2'50"911). My shaky driving at Lap 3's first refiller was the most costly mistake. I was real close at the end, and my railslide wasn't half bad, but the ghost kept ahead by railsiding from an earlier spot! youtube.com/watch?v=qJASboR07YE&list=PLmh9KAbzYQuiMD3pvFhGkowPhiBOHvsoX&index=27
It's refreshing to challenge staff ghosts that use advanced techs like that, though.

It looks like I haven't resolved my recording freeze issue yet (1:43), sorry about that.

Course: Lightning: Thunder Road
Machine: Space Angler
Setting: 80%
Rules: Expanded Max Speed - 80%-100% settings
Other category details: No side attacks

Date: 2021.01.01 19:56 UTC-0800
Played on: Unleashed version 4.0, run with Wii RVL-001 (NA) using Nintendont (modded to tolerate GC stick drift) + Sony SF-32UY SD card
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling, crossfade), ffmpeg (audio normalization)

F-Zero GX Unleashed 4.0 - BBDH - 056291 with Super Piranha (100% non-SA)

Joselle Astrid 2021-01-06 | I enjoyed this machine/track combination in GX, so I thought I'd give it a go in GX Unleashed as well. As in GX, Super Piranha's trademark is its overkill steering ability, and it's ever-tempting to show that off by cutting the corners close. It's a bit trickier when the machine's got a much better booster in Unleashed, though!

Wall hits always feel pretty costly on short courses like this, and I had a couple of them here. I'd already spent about 30 minutes on this though, and decided to move on for the time being. Besides, I had already passed the Expert non-SA staff ghost of 0'57"149: youtube.com/watch?v=cY-SiyxO_nw&list=PLmh9KAbzYQuiMD3pvFhGkowPhiBOHvsoX&index=6
And was still quite a ways off of the Grandmaster non-SA staff ghost, which gets 0'50"331: youtube.com/watch?v=GIHKzZfcGaA&list=PLmh9KAbzYQuhWpycVAwjDFwrl2PawzwVQ&index=13

Course: Big Blue: Drift Highway
Machine: Super Piranha
Setting: 100%
Rules: Expanded Max Speed - 80%-100% settings
Other category details: No side attacks

Date: 2021.01.01 19:21 UTC-0800
Played on: Unleashed version 4.0, run with Wii RVL-001 (NA) using Nintendont (modded to tolerate GC stick drift) + Sony SF-32UY SD card
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling, crossfade), ffmpeg (audio normalization)

F-Zero GX Unleashed 4.0 - MCTR - 053100 with Elegance Liberty (100% non-SA)

Joselle Astrid 2021-01-06 | F-Zero GX Unleashed has a Climax edition, which replaces 5 of the machines with ones that only appeared in F-Zero: GP Legend and F-Zero Climax. The Climax edition is here: gamebanana.com/gamefiles/9106

GP Legend and Climax are Game Boy Advance games, so the machine designs don't have official 3D models. However, CH-MCL managed to model these machine designs in GX's style. It's not a perfect replacement, since the GX/AX character models are still used, for example. But still pretty impressive modding work!

This machine is Elegance Liberty, and it's normally yellow but I was using a color swap of it. As you can see, in Unleashed, it's got MTS power to spare. I'm not used to controlling an MTS this strong, so I'm sure I could be strategizing or optimizing it better. I also played it simple at the end instead of going for an explosive finish. No unintentional wall hits in this run though, and got an MTS into shift boost at the start of Lap 3, so it's a respectable result for me.

I was racing the Expert non-SA staff ghost, which gets 0'54"751 with Wonder Wasp: youtube.com/watch?v=W3ybspdlFZA&list=PLmh9KAbzYQuiMD3pvFhGkowPhiBOHvsoX&index=1

Course: Mute City: Twist Road
Machine: Elegance Liberty
Setting: 100%
Rules: Expanded Max Speed - 80%-100% settings
Other category details: No side attacks

Date: 2021.01.01 18:49 UTC-0800
Played on: Unleashed version 4.0, run with Wii RVL-001 (NA) using Nintendont (modded to tolerate GC stick drift) + Sony SF-32UY SD card
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling, crossfade), ffmpeg (audio normalization)

F-Zero GX Unleashed 4.0 - PTAD - 158352 with Green Panther (80% non-SA)

Joselle Astrid 2021-01-05 | F-Zero GX Unleashed is a GX mod by AKC12. It rebalances the machines' stats so that they can all be competitive for time attack - and it does so by greatly buffing most machines. It also has multiple sets of staff ghosts to provide plenty of challenge. A major update, version 4.0, was released a couple of weeks ago. You can find it here: gamebanana.com/gamefiles/5792

Unleashed has 3 levels of staff ghosts now: Standard, Expert, and Grandmaster. Also, there are side attack and non-side attack variants for a few courses in Expert, and for a majority of courses in Grandmaster.

I'd like to get more into Unleashed later this year, but for now I tried it out for about 4 hours to see what it had to offer. I chose to race some Expert non-SA staff ghosts, and so far they've been a pretty nice challenge level for me. I'll be uploading several runs which I got when trying things out, starting with this one. The runs aren't super polished, but the main goal is to showcase the mod a little bit.

This ghost here gets 2'01"037 with King Meteor: youtube.com/watch?v=7L2qq1EiOn4&list=PLmh9KAbzYQuiMD3pvFhGkowPhiBOHvsoX&index=7
It took about 25 minutes to beat, using what I thought was a good machine for the track (I needed time to adjust to the handling though).

I was basically following the same strat as my GX Fat Shark run. Green Panther is sort of comparable to Fat Shark in this mod in terms of weight, accel, MTS ability, and quick turn feel. Well, the quick turns are a bit more sluggish, maybe like GX King Neptune if I remember correctly. I didn't entirely trust my ability to handle it, so I used safer, less optimal boost timing near the end of the lap. Otherwise this went to plan, at least a couple mistakes in there but nothing that costly.

For Unleashed time attack competition, AKC is accepting machine settings from 80% to 100%. Note that this is different from the unofficial category "Expanded Max Speed" for the original GX game ( fzerocentral.org/viewtopic.php?t=14364 ), which ranges from 70% to 100%. So 80-100 is Unleashed's version of Expanded Max Speed. The general idea behind this category is to encourage settings experimentation and help machines which are more limited at 100%, while not allowing settings low enough to enable snaking.

I used 80% for this run. On the machine settings screen, you can adjust the setting by increments of 1% with the D-pad, and you hear a sound effect when you reach a multiple of 10%.

Course: Port Town: Aero Dive
Machine: Green Panther
Setting: 80%
Rules: Expanded Max Speed - 80%-100% settings
Other category details: No side attacks

Date: 2020.12.31 19:52 UTC-0800
Played on: Unleashed version 4.0, run with Wii RVL-001 (NA) using Nintendont (modded to tolerate GC stick drift) + Sony SF-32UY SD card
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling, crossfade), ffmpeg (audio normalization)

F-Zero GX - SOSS - 140920 with Gallant Star-G4 (Max Speed non-SA)

Joselle Astrid 2020-12-29 | 40"698 - 29"036 - 31"186. Improved from 1'42"840 (-1"920, youtube.com/watch?v=KjgcMMRAdBg). This continues my streak of at least one Max Speed improvement per year, since 2012!

Unfortunately, my recording has a 0.5-second freeze coming out of the S-turns on Lap 1. It's not a really critical part of the run, but it's still annoying. I played the instant replay after getting the run, and the recording of that has no freezes, so I've included the instant replay at the end.

It's been a year since I've done any GX time attack, and longer than that since I've used Gallant Star-G4 specifically. Considering that, I'm pleased about getting this kind of run in less than 2 hours. It does help that Surface Slide without side attacks isn't very demanding, but still, I expected to be less precise and consistent overall.

Best splits, without the MTS into shift boost at the last turn, are 40"423 - 29"0xx - 31"1xx. So, not much to complain about there. I've gotten the MTSISB twice on Lap 2, but I'd probably need some iterations of experimenting and studying video in order to really get a knack for it.

My results on the Lap 1 tunnel MTS vary by a few tenths, and this run's result was about average. Occasionally, I tap the inside wall during the MTS in a way that saves a couple more tenths, but I haven't been able to do it intentionally. Generally if I tap the wall, I'm losing time.

On the Lap 2 tunnel MTS, I was able to cut to the inside more than usual here, and got a fairly clean exit. On Lap 3, I was sometimes having trouble carrying enough speed from the tunnel to the 360-degree left MTS. So I tried delaying my boost after entering the refiller, with the idea of helping the MTS speed and the speed afterward. However, I think the real issue was that I was initiating the Lap 3 tunnel MTS later than needed, for some reason.

Course: Sand Ocean: Surface Slide
Machine: Gallant Star-G4 (Dread Hammer, Maximum Star, Titan-G4)
Setting: 100% (all the way to the max speed side)
Rules: 100% settings only, no snaking
Other category details: No side attacks

Date: 2020.12.28 19:32 UTC-0800
Played on: NA version, run with Wii RVL-001 (NA) using Wii disc channel
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling, crossfade), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - LTR - 201770 with Heat Grampus

Joselle Astrid 2020-12-27 | This is a 2"812 improvement over the time I got back in June (2'04"582): youtu.be/gBTnbrffLLA

Lap 1 was sloppy at the first tunnel and final twist, but the bottleneck before the second jump salvaged it more or less. The close call at the Lap 2 wavy refiller made me kinda high-strung for Lap 3, but it worked out. Both boost laps were a few tenths off of my best lap (38"716), so can't complain much there.

The final twist could've been better on all 3 laps really, though some more than others; I guess Lap 2 lost the most time there. I was (mostly) handling the twist better in my 2'04 run. So I should focus a little more on that part next time.

Other than that, I really didn't have any strat changes from my 2'04. I was just in control a bit better overall. That might be attributable to the new Gamecube controller I got recently. Or, maybe I was more alert due to the cake I had eaten before playing... let's go with both.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Lightning: Thunder Road
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.12.25 19:50 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont (modded to tolerate GC stick drift), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (SSBU edition)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - GPS - 235788 with Heat Grampus

Joselle Astrid 2020-12-27 | This is a 8"986 improvement over my 2017 time with Dark Grampus (2'44"774), the last time I wrote a detailed description for this course + category:
youtu.be/CoIg9xt--UM

Lap 1 is my best opener. Lap 2... started well, but got pretty messy, losing like 3 seconds. Lap 3 got the spiral skip, and was pretty good other than crashing after the skip. I have yet to get two spiral skips in the same run, but that would obviously cut another chunk of time if I could manage it.

There was an unfortunate 0.3s freeze in the video at the final refiller. Fortunately there wasn't too much action at that moment, but I'll still want to get that figured out before too long (I've also noticed other freezes in my recent recordings). I got a new laptop recently, so there's a good chance that's related...

What's different from 2'44? Well, spiral skip is the most obvious thing, although getting it once like this "only" saves about 3 seconds. The rest was better driving generally, Heat Grampus being a bit faster than Dark Grampus, and a few smaller strat tweaks.

At the start of boost laps, I used to kind of allow myself to get tossed into the air as the pipe started to turn, similar to Lap 3 of GX Max Speed runs. Now I steer along with the pipe turn and stay grounded, which lets me retain more speed. Lap 2 went particularly well here, as I started the next T-drift with a lot of speed.

I've been getting the spiral skip reasonably often, like 20% of the time. Getting enough speed isn't that hard really; my approach seems to work fairly well. Navigating the turns at 2500 km/h just before the skip is the tough part. As you can see from my inputs, I just use stick steering to get through the turns, then at a certain point I attempt a right strafe without touching anything else. If I don't crash into any walls, and my positioning and speed are good, then I can go into the air neatly and skip the spiral turn.

Here's a clip showing a spiral skip with a somewhat lower speed (21xx km/h at liftoff): youtu.be/osnV85NAiqY

It seems that getting the spiral skip like this could be doable on Lap 1, too, but I haven't managed it. Either way, the hope is to stay on the right side when the track dips down. In Laps 2 and 3, I've gotten the most success with the spiral skip on that side. And in Lap 1, the right side tends to give me a cleaner 'fall' down the slope compared to the center or left.

If there's a part of the strategy to try and polish more, it's probably the cylinder. Nothing about the cylinder geometry totally prevents me from trying to go 3000 km/h, it's just a matter of keeping up with the curves; and presumably there are easier and harder ways of doing that.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Green Plant: Spiral
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.12.25 19:37 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont (modded to tolerate GC stick drift), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (SSBU edition)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - PTCW - 121796 with Heat Grampus

Joselle Astrid 2020-12-26 | This is a 1"436 improvement over my time from the FZC marathon last month (1'23"232):
youtu.be/TRjcZtlpjwQ?t=608

Lap 1 is my second-best opener, Lap 2 is my best boost lap, and Lap 3 was shakier but decent enough. One of the best parts here was the speed at the turn before the branches on Laps 1 and 2. The part before the first refiller on Lap 2 was also very clean.

Erm... yeah. Honestly, I haven't learned much, if anything, about this course since my 1'25"562 PB (strategy comments there still apply: youtube.com/watch?v=6h8N3ymL2ag ). Since then, all of my three improvements have happened on one-off attempts, and my actual grind sessions never got anything better than 1'27. So I guess this course is all mental for me now?

For some reason, the video (or game) froze up for a second as the end splits were appearing. Unfortunate, although also fortunate that it didn't happen during the run itself.

If you figure out what the camera was trying to do during the post-finish landing, let me know, because I have no idea myself.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Port Town: Cylinder Wave
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.12.24 18:50 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont (modded to tolerate GC stick drift), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (SSBU edition)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - ASD - 118944 with Heat Grampus

Joselle Astrid 2020-12-26 | This is a 1"269 improvement over my previous time from May (1'20"213):
youtu.be/_u66ZG9yZaI

Very solid run for me. Happy with all the splits except Lap 3, which wasn't too slow either. I've only gotten two laps under 12"5 before, so having three 12"5xx laps to finish the run just makes this feel so... optimized.

That said, laps around 12"0 should be feasible with a shift boost at the end. The T-drift WR is 1'14"214 with Aerial Whale-EX.

Most of my strategy is the same as the 1'20, except when I'm coming out of the last hairpin turn. I used to keep more to the right, just stopping all steering for 0.3s to make sure I was stable enough to start another T-drift. Now, I strafe farther to the left coming out of the turn, getting to a part of the track with a gentler slope. It's a slightly more indirect line, but I can start the next T-drift almost immediately without getting tossed around.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Aeropolis: Screw Drive
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.12.24 18:46 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont (modded to tolerate GC stick drift), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (SSBU edition)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - All Tracks T-Drift with commentary - F-Zero 30th anniversary marathon

Joselle Astrid 2020-12-20 | Here's something a bit different. Kind of a late follow-up, but I decided to re-upload the AX run I did for F-Zero Central's mini-marathon, held a month ago to celebrate F-Zero's 30th anniversary.

The run's also here on FZC's channel, but in 480p30 due to being a direct, non-upscaled Twitch export: youtube.com/watch?v=R4swqTXFuW8
Here's the Twitch vod: twitch.tv/videos/811818761

I don't generally seek out marathon events these days, especially big events, but this was a pretty modest community event where we wanted to showcase as many F-Zero games as possible, and I had played AX recently. So I gave it a shot.

So yeah, I'm not a big commentator to begin with, and I also don't normally commentate this coherently. I prepared pretty much all of it in advance, since I wanted to explain as much as I could about the game in a short time, giving all the context for a GX player to understand what they're watching here. Actually, I was largely reading off of a slideshow (an advantage of remote events!). Even so, my voice was a bit shakier than I'd like it to be, which can happen to me if I'm even the slightest bit nervous - oh well.

Still, I'm glad I got to explain pretty much everything I wanted to, which isn't a given in a timed run like this. I was advancing through menus pretty slowly even if I wasn't talking about the menus themselves, so that was a bit awkward, but I think that did work out well timing-wise. Overall, if you're wondering how I'd plan things out if I ran AX at GDQ, this would be about it.

Commentary:
0:00 - Intro, T-drift basics, MCSO
2:46 - GX Action Replay code which loads AX leftovers, Nintendont, ASD
5:40 - Controls, GX/AX control differences, T-drift controls, OSMS
8:49 - More GX/AX mechanics differences, PTCW
11:56 - License cards, pilot points, memory cards, LTR
16:25 - AX arcade cabinets and locations, GPS

Track times:
MCSO: 1'04"822
ASD: 1'22"978
OSMS: 1'35"439
PTCW: 1'23"232 (Personal best!... with Lap 1 being the fastest lap)
LTR: 2'25"904 (At least Lap 1 was the silly kind of bad)
GPS: 2'45"923

More complete explanation of the AX-exclusive T-drift mechanic: youtube.com/watch?v=TC07KEB7K0g

Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.11.21
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (purple)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with LiveSplit and NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - PTCW - 123968 with Heat Grampus

Joselle Astrid 2020-11-19 | Thought I was going to upload Spiral next, huh? Well, I have yet to write comments for that - but in the meantime, I accidentally improved by 1"594 in Cylinder Wave while de-rusting for the F-Zero 30th anniversary marathon.
Event schedule (I'm running AX All Tracks): horaro.org/fzc-fzero-30th/schedule
Event stream channel: twitch.tv/fzerocentral

If you're wondering what's with the LiveSplit display on the left: I was doing an All Tracks runthrough without really worrying about menu speed. So I was just using the splits to remember my in-game times, while having real time at the bottom for reference.

Lap 1 was decent; a little slow before the first energy refill and at the end (T-drift took a second to kick in). Lap 2 was good. Lap 3 was very good; I went for a faster T-drift setup at the very end (at the risk of screwing up the entire thin stretch) and it paid off, saving half a second.
My best lap is 25"671: youtu.be/LrTfQd1tjgE

For strategy notes, my 1'25"562 comments still apply: youtube.com/watch?v=6h8N3ymL2ag

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Port Town: Cylinder Wave
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.11.18 22:47 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (purple, clear back)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with LiveSplit and NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - LTR - 204582 with Heat Grampus

Joselle Astrid 2020-11-15 | Done on June 19. Previously uploaded to my Side B channel with short comments.
Lap 1 is my best (but nothing too crazy). Lap 2 was a little slow early on, but fairly good. Lap 3 had a close call at the first jump, though I certainly lost time there and at the end.

This is just a 2"775 improvement over my 2'07"357 from 2017, done with Dark Grampus: youtu.be/grlVv3jSPqA
2'07 was a pretty strong record for me at the time, since it was the first time I got sub 2'10. At my current level though, this 2'04 is a little underwhelming. I thought I had put enough effort into grinding this track for now, though, so I decided to settle.

This is the second-easiest AX track to pick up for T-drift beginners (after MCSO). Yet, this is probably the hardest track for me to grind at this point because... it's almost TOO fast. If I practice the track for more than 20 minutes, I usually lose enough focus that I start playing badly, especially on the thin stretches. The effort required on the thin parts also tenses up my hands, so my hands get tired more quickly here than in GPS (the longest track).

I ended up making a few consistency adjustments to make life a little easier, at the expense of some potential. Just up to the first checkpoint gate ("time extension!"), I already end three T-drifts a little earlier than necessary: the first T-drift (when starting laps 2 and 3), the T-drift before the first energy refill, and the T-drift coming out of the first thin tunnel.

The zigzag road leading up to the second jump is another good example. You can do a right T-drift through the zigzag, and it works out beautifully if you can keep the speed (youtu.be/vicJ_iOtrG4). But there's a big risk of a head-on collision at the zigzag, and due to the extra speed, a big risk of pinballing off the walls if you mess up at the thin section. So in this run, I just steer normally through the zigzag to stay at more manageable speeds.

Finally, I also MT more than I need to at the end twist, though it's less because of the speed and more because the terrain is tricky. (Touching the rails is especially harmful here because you tend to get 'sticky' rail grinds rather than bouncing clean off.) As I'm exiting the twist, I try to start with a right T-drift if I can, so that I only need to switch directions once at most to complete the lap.

Here's a 2'02"480 by FighBat with Heat Grampus: youtube.com/watch?v=Ovw1Xa3gFJU

Here's a video of YUTA03xs and SHINTA posting 41 openers and a 36 lap, using machines made with the Mars -EX booster: youtube.com/watch?v=skGuP-GJVaE

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Lightning: Thunder Road
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.06.19 18:07 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - PTCW, Lap - 025671 with Heat Grampus

Joselle Astrid 2020-11-07 | Done on June 7. Generally I'm not uploading my best laps on this channel, but I'm making an exception for this one because it came out of the blue. My second best lap is 26"453, I've gotten three 26s, and even 27 is kind of uncommon for me.

This was fairly good up to the first energy refill, very good up to and including the branches, and outstanding after the branches. I kept quite a high speed through the upward twist. I then popped into the air briefly before the final T-drift, but it didn't lose much time (if any) because I wanted to slow down to line up the drift anyway.

Here's my second best lap, which entered the first energy refill 0"400-0"500 faster than this lap: youtu.be/oXowuBQtlOk

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Port Town: Cylinder Wave
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.06.07 15:06 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - PTCW - 125562 with Heat Grampus

Joselle Astrid 2020-11-06 | Done on June 7. Previously uploaded to my Side B channel with short comments.
A pretty consistent and clean run; no standout parts, but not many mistakes either. On Lap 3 I lost speed at the first left, and on Laps 2 and 3 I played a little on the safe side before the branches.

This is just a 2"018 improvement over my 1'27"580 from 2017, done with Dark Grampus: youtu.be/2h9-z8aVbAU
It was likely my strongest time from back then, though. So it's still nice that I was able to improve it by a couple seconds.

In this track, I'm constantly trying to answer a risk/reward question: how long can I build speed with T-drift before the next part of the track is unmanageable without catching air? The vast majority of the time, slowing down to not catch air is better than going fast and catching air. Going too far to the side, or going to the side too suddenly will make me catch air (or even fall off course).

The last thin stretch is another risk/reward question where the sooner I start my T-drift, the harder it will be to line it up properly and keep it under control. Catching air on the thin part is quite dangerous, since I can spin out of control and it's too thin to start another T-drift. Another concern here is that I have to turn slightly right (against my left T-drift) to keep a good line during the thin pipe, so I might regain grip if I'm not careful. To prevent this, I have to make sure I turn lightly enough, or turn with enough speed (a boost can help with this).

The main changes from my 1'27 Dark Grampus strat are at the first left turn, and the upward twist after the branches. For both parts, I now opt to MT instead of pressing the accelerator multiple times. Honestly, if either of those changes are improvements, then they're pretty slight. I think the first left is now more consistent for me, but it's still kind of tricky, and the old strat could possibly have a bit more potential. For the upward twist, I think there's now a little more potential, but a little more variance with results depending on my speed going in.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Port Town: Cylinder Wave
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.06.07 14:57 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - OSMS - 129601 with Heat Grampus

Joselle Astrid 2020-10-29 | With 21"033 lap PB. Done on June 20 in an All Tracks run (10'19"312: twitch.tv/videos/656827231). I've been pretty lazy about making 'formal' uploads of my AX times this year, but I'm getting back to it now.

Since I was in the middle of a run, I skipped past the results screen quickly, which makes for a slightly awkward IL video. But the lap splits were 25"074 - 21"033 - 21"494 - 22"000, and max speed was 3623 km/h.

In terms of potential, my best opener is 23"827 (and still improvable: youtu.be/LKy7vhSCwl4). Lap 2 was my best lap, but Lap 3 was actually 0"650 ahead of that when exiting the tunnel! So if we figure that 23"5 + 20"4 + 20"4 + 20"4 could happen, then the T-drift WR of 1'26"225 is certainly in the realm of possibility. Easier said than done of course.

Taking a decent line through the first ice patch isn't a big problem compared to Fat Shark. However, any number of things can happen with the bumps/dips after that ice, since the steering controls aren't very precise. On boost laps, I boost-MT here to keep up speed as much as possible before getting back to driftable terrain.

Next, ideally I'll keep a long T-drift going until the dip after the second ice. But I need a good approach line, and I can't go too fast, otherwise I'll pop into the air before the dip. If the drift is going well, I generally want to slow myself down a bit, such as by doing small steering adjustments throughout the drift. If my initial drifting line is poor, like in Lap 1, I need to correct and start another drift. Doing two shorter T-drifts intentionally is a decent consistency method, but doing a single long T-drift seems to be fastest if you can get it right.

Maximizing speed before the tunnel exit is quite important, because there isn't much opportunity to gain speed from the last part of the tunnel to the jump out of it. The tunnel exit itself can be tricky at high speeds. It's easier to take the right turn wide instead of trying to cut to the inside.

The right branch seems to be fast enough that it's better to take it on every lap. The branch entry can be tricky though; the switch from left to right T-drift is just something I need a good feel for. Sometimes I'm handling it well, other times I'm bouncing off the walls a lot.

If I'm going fast at the end of the right branch's right turn, it seems best to MT until the last turn of the lap, like on Lap 2. If I'm not going very fast, it seems better to T-drift the following left turn briefly, like on Laps 1 and 4.

At the beginning of the tunnel on boost laps, staying around the center of the track seems to reduce the chances of catching air multiple times, thus helping speed. To do so, I try to angle slightly left as I go off the ledge.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Outer Space: Meteor Stream
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.06.20 15:17 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with LiveSplit and NintendoSpy input display
Editing: VapourSynth (upscaling), ffmpeg (audio normalization)

F-Zero AX (Nintendont) - ASD - 120213 with Heat Grampus

Joselle Astrid 2020-08-03 | This is a 7"212 improvement over my 2017 time with Dark Grampus (1'27"425), the last time I wrote a detailed description for this course + category:
youtu.be/PEffivUAOHc

First lap was pretty sloppy, but the boost laps were really consistent. I messed up the final lap's loop by going hard on the steering too early, but got some time back at the end with a clean shift boost.

I now MT (release accel) to keep up speed in the first left turn, but on boost laps, it's fairly easy to go too fast to handle when doing this. So it's a matter of slowing yourself down just enough to manage the turn. There are probably a few ways to approach this, but what I do is switch from right drift to left drift just before entering the turn, and then release accel from there.

After the loop, I used to do a long MT over both road bumps. Now I keep left to avoid the first bump, and then start MT just before the second bump. Once my machine is stable, I start drifting again. In this run I seemed to always stabilize pretty quickly, but sometimes the machine wobbles around for a half second or so.

For the last hairpin, the first accel tap helps the machine steer more sharply. Then I press and hold the accel without drifting, with the aim of slowing and stabilizing enough to prevent catching air. If the machine does get airborne here, it can get tossed around for a half second or so, which means I have to wait (at relatively low speed) before starting the next drift. Sometimes it's a simple hop and landing, like in the final lap, but that's not guaranteed by any means. It's not uncommon for me to strafe off course here due to misjudging my stability.

That second boost after the last hairpin was added during this play session - Faizer suggested it while I was streaming. With ideal steering it doesn't seem to help much, but the boost does seem to paper over any speed loss from larger steering adjustments. Despite the added boost, I still had plenty of energy throughout this race.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Aeropolis: Screw Drive
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.05.25 23:34 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - MCSO - 102643 with Heat Grampus

Joselle Astrid 2020-07-30 | I've been lazy about uploading my recent AX course times to my main channel, but I'm getting around to catching up now.

This is a 4"203 improvement over my 2017 time with Heat Grampus (1'06"846):
youtu.be/8AnSsBgN-Uo

My basic strategy on the straightaways is to not steer unless I really have to. No steering toward or away from dash plates, only away from walls (and maybe into the refiller if I'm out of energy). In particular, I leave it up to chance to hit the first dash plate of the lap and the dash plates alongside the refiller.

Why? AX steering is fairly sensitive with the Gamecube control stick, so oversteering on high-speed straights is a real issue. Thus, less unnecessary steering means fewer wall hits. Also, anything other than the tiniest steering adjustments (again, tricky with GC stick) will drop my T-drift speed. In comparison with these things, the dash plates don't matter quite as much.

Depending on which dash plates I'm going to hit, I try to adjust my boost timing accordingly so that I boost as late as possible before each turn. I misjudged this at the start of Laps 2 and 7, and just didn't get to boost in those cases. It didn't make that much of a difference though, I think.

The dynamic in this course could change significantly if you're a skilled player on an arcade cabinet, since the wheel could allow you to make tiny steering adjustments reliably.

Finally, the bump before the last straight can interfere with restarting the T-drift there. To avoid getting tossed by the bump, I try to use a fairly small steering movement when restarting drift, and I just aim to nick the left side of the final dash plate, if anything. I might miss the dash plate, but it's not that big a deal if I can just use a boost in its place.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Mute City: Sonic Oval
Machine: Heat Grampus (Splash Whale, Heat Snake, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.05.28 22:28 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - All Tracks T-Drift in 1017061 (15:44.29 RTA) with Heat Grampus

Joselle Astrid 2020-06-22 | In-game time: 10'17"061
Real time: 15:44.29
Twitch recording: twitch.tv/videos/657800297

Course times:
MCSO 1'06"060 (+3"417 on personal best)
ASD 1'21"772 (+1"559)
OSMS 1'35"051 (+5"450)
PTCW 1'28"961 (+3"399)
LTR 2'04"629 (+0"047)
GPS 2'40"588 (+2"020)
Total 10'17"061 (+15"892 versus PBs)

MCSO - Okay I guess. This actually was my best Lap 1, but I made mistakes later on; the Lap 7 swerving in particular was unnecessary. Happens sometimes with the sensitive steering though.

ASD - Pretty clean. A little slow in the middle of Lap 1, and a couple of wall hits later, but not a big deal.

OSMS - So-so. Lap 1 started all right, but I had trouble starting a T-drift after the tunnel. Lap 2's tunnel exit was slow.

PTCW - Two significant mistakes in the first half, but good otherwise. It can be easy to fall off in this course.

LTR - Quite a solid time despite a few collisions, the biggest being in Lap 2. Lap 3 is my first sub-39 lap!

GPS - My best Lap 1, and my second best course time (best has a spiral skip). I tried to lift off for the skip in Lap 2, but I either needed better positioning or more speed.

Overall, a pretty satisfying run, especially since the finish was strong. As indicated by the splits, I had a previous run that was really strong until Spiral, where I got an off-course in Lap 3 and finished with 2'49. I'm glad I was able to beat that run here (by 2 seconds of in-game time). If I kept up my current level, I could probably cut a few more seconds with more attempts, but I'll settle with this for now.

For real time, I start timing on pressing A at the settings screen before Sonic Oval; James McCloud's pose changes abruptly at this point. I end timing upon crossing the finish line in Spiral. The LiveSplit timer on the video is just an approximation, since I have to control it manually while playing. Also note that the game feed may be slightly behind other elements of the video.

The load times don't vary significantly with my setup, but they do seem to vary enough to throw off my splits a bit. Enabling license card functionality slows down the loads and menus, but I need it enabled to use custom machines (Nintendont doesn't have Gamecube memory card support). Custom machines are much better for T-drift.

In-game time (IGT) is the primary timing method for AX runs. However, the time card after the race is rather difficult to read (even with frame advance) if you press A as quickly as possible. So I wait a moment after the race to make sure the course time is clearly visible before pressing A. In other words, I sacrifice a bit of real time to make IGT easier to track.

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Machine: Heat Grampus
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2020.06.21 15:58 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with LiveSplit and NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

Highly Responsive to Prayers - Lunatic, Makai route - 11,881,815

Joselle Astrid 2020-03-13 | Score: 11,881,815
1 credit clear, 18 misses
Lunatic difficulty, 魔界ルート Makai route

(Note: The screen flashing effect when using a bomb is somewhat intense. The first bomb is at 10:00.)

I was interested in trying more Touhou games, so I started playing this one, and ended up sticking with it until Lunatic difficulty. Initially, I practiced some stages for score because I wanted more extra lives to get 1 credit clears. Eventually, I thought I could probably learn a full route's worth of scoring, and this is the result. This isn't a super-competitive score, but I think it's a decent effort considering that I didn't look up other people's strategies yet.

This game is unlike others in the Touhou series, playing more similarly to Arkanoid than to shmups. It also has big differences from traditional 'block breaker' games:

- You control a character (Reimu), not a paddle. You can deflect the yin-yang orb by attacking it, but you need to time attacks well: if the orb hits you, you lose a life.
- There are lots of bullets to dodge.
- The orb is allowed to touch the ground, but doing so resets your card chain in non-boss stages. Getting big card chains is the game's biggest scoring mechanic.

Hitting with the top of a slide attack makes the orb follow a fixed trajectory and distance. So if I keep using slides, connecting with the top (not the side) of each slide, I can form consistent strategies. The bullets complicate things, though. The bullets that come out of cards are generally aimed towards you, but their exact angle is random. The bullets also move faster or slower depending on how well you've been surviving recently.

I missed around 1 million points from dropping chains in Stages 2, 12, and 18. But there's more potential improvement than that. The second best source of points are the P items that drop from every 10th card. If you keep collecting these P items without missing any, they eventually give 65k points each. However, trying to get P items along with the card chains adds a whole other level of complexity, so I've basically ignored them for now.

Stages 1-5 (0:05)

I don't have a consistent strategy for Stages 1 and 8, since the teleporters shoot the orb at random directions/speeds.

I mis-aimed a slide and dropped the chain on Stage 2, but I only had a few cards left, so it wasn't too costly.

3 shows a couple of special moves that I use occasionally. The spin kick is useful to deflect bullets shortly after a slide. The flying kick acts like a slide, but imparts more horizontal speed; here, I use it to clear the left cards while avoiding the teleporter.

Stages 6-10 (2:32)

Stage 7 is where things get tricky. I was trying to spin-kick after the 3rd slide, but I didn't slide early enough to make it come out in time. Dying changes the timing of the turrets' shots, which means I have to improvise somewhat.

In 9, sometimes I wait at the center for the turrets to shoot, so that the bullets don't corner me later. It's hard to remember when I should wait, though.

I finished off YuugenMagan more quickly than usual. It's always preferred to spend less time dodging the missiles and star bullets.

Stages 11-15 (6:12)

On Stage 11, I stand close to the bottom-center cards when they're about to shoot, since their bullets may aim almost horizontally if I'm too far away.

I was trying to spin kick after the first slide in 12, but I don't think that works with this bullet speed. Then I dropped the chain. I had trouble remembering how to play the stage 'normally' afterward.

In 13, I did a slide in the wrong direction (at 7:57), but then I did another slide toward the wall to get the orb in the same position as before. It's a handy trick sometimes.

14 went really well, despite me forgetting to do a strat that would result in fewer close calls (start with a right slide, not left).

Elis was pretty cooperative with her attack choices until near the end. Then again, most of her late-phase attacks are pretty brutal.

Stages 16-20 (10:29)

The center wall in Stage 17 forces me to run back and forth a lot, which makes dodging difficult. I was in the zone at the start (and also got a bit lucky), but made mistakes later.

Stage 18 had a good start, but unfortunately I got pressured too far left and dropped the chain halfway through.

The Sariel fight... wasn't great, but at least I got through. I think I can handle the last phase without bombs, but with 0 lives, I didn't want to risk it.

Date: 2020.03.10 23:17 UTC-0700
Game: 東方靈異伝　～ Highly Responsive to Prayers (PC-98). "Touhou Reiiden", Game 1 in the Touhou series.
Game version: 1.10, with 'Heaven'/'Hell' route menu. Not sure if it's an unofficial translation or not.
Platform: Neko Project II fmgen (190506), CPU setting at x48
Controller: iBuffalo BSGP801

F-Zero GX - GPMR - 121369 with Fat Shark (Max Speed non-SA)

Joselle Astrid 2019-12-27 | 30"753 - 26"281 - 24"335. Improved from 1'22"836 (-1"467, youtube.com/watch?v=ZoLVKswh4bg). This continues my streak of at least one Max Speed improvement per year, since 2012!

The weaker lap in this run was Lap 2, since the speed after the first MTS cancel wasn't that great, and it would've been better to not strafe as much at the end of the lap (I was even unintentionally holding L a bit after the refiller). Lap 1 and 3 were close to my best, though Lap 1 could add an MTS at the end, and the mine boosts on Lap 3 could've been better. Still, it's always nice to get the big boosts in Lap 1 and then not miss the boost-lap refillers.

At my skill level, results in this course still heavily depend on the triple mine boost to shift boost on Lap 1. Anything less and I'll finish the lap with 33 or higher. I'm not sure if a certain set of inputs makes it easier to get these boosts or not, but I do have the controller inputs shown in this run. Naegleria also has a PB video with inputs, so that's another possible reference. At any rate, getting the boosts is just half the battle. I harnessed them well in 2 attempts, but in maybe 6 other attempts I hit the wall or flew off the track.

My old 1'22 run's biggest flaw was how it handled the bumpy right turn near the end of the lap, on every lap. That run lost speed from the bump in Laps 1 and 3, and hit the walls in Lap 2 due to an unexpected shift boost. In this new 1'21 run, I avoided the bump by going around the left side of it. Hugging the right side of the track would also work.

It's funny that I wasn't really paying attention to the bump back when I got 1'22, because now I find that getting tossed by the bump like that is non-trivial to do intentionally. I also tried to practice the shift boost on that bump, but I couldn't get it consistently (youtu.be/dOUTWmxRTkU).

I wasn't getting appreciable consistency on the MTS into shift boost (MTSISB) after the first straightaway, and I was already having trouble picking up the refill without that extra speed, so I didn't go for it. Relatedly though, I once accidentally got an MTSISB later in that same left turn. It's probably even harder to use since it's so close to the refiller, but here's what it looks like for the record: youtu.be/VAgxzsuZUV0

One other point of interest: GX has two general ways of going about the quick turn technique. For as long as I've played GX, I've stuck with the "press L+R, then release one" method. This method gives me good control with consecutive small quick turns. But it makes it awkward to alternate between stick-only steering adjustments and small quick turns, since I have to press L+R before each quick turn, and release both for basic steering. So I've gotten interested in trying the "lightly tap one of L/R, then press the other" method. It's tough to learn new muscle memory for such a fundamental technique, so for now I'm just trying it in low-speed situations on Lap 1 (including in this run).

Course: Green Plant: Mobius Ring
Machine: Fat Shark
Setting: 100% (all the way to the max speed side)
Rules: 100% settings only, no snaking
Other category details: No side attacks

Date: 2019.12.19 23:36 UTC-0800
Played on: NA version, run with Wii RVL-001 (NA) using Wii disc channel
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - GPS - 259338 with Fat Shark

Joselle Astrid 2019-11-25 | You might be able to guess what my goal time was. Lap 2 didn't keep pace, but I played Lap 3 very well and scored a personal-best lap. I was worried that I'd choked sub-3 after the ice turn, but I managed to recover decently and didn't lose too much time.

Best splits are 1'00"716 - 57"779 - 57"188 = 2'55"683. I can see potential for 55 laps and sub-2'50 if all went to plan, even without any spiral skip strats. But sub-3 was already pretty tough. The course has a lot of sloped turns, so it's rather unforgiving for T-drifting, especially with Fat Shark's handling.

The pipe at the start gives an opportunity to build lots of speed, which is good since I can't build much speed on the steep upslope afterward. Results may vary, though. In this run it was pretty much how I wanted, exiting the pipe at 2000-2400 km/h on every lap. Sometimes I'm well below 2000; other times I get a shift boost to 3000+, which makes it quite difficult to survive the upslope (but I've done it a couple of times: youtu.be/cETNNNvoAwI).

For the sloped turns, I need to start on the right side before the left turn, or I might fly off (youtu.be/G2SCtnABG5Q). The right turn was a bit deceptive for me at first, as I used to turn too late all the time. Once I get to the last left turn, I can T-drift to gain speed again. I think you can also gain speed at the start of the first left turn, but I didn't need it in this case.

For the cylinder, it could be worth it to drift the entire first turn instead of getting the cylinder refiller, if you could play until the next refiller with very few mistakes. However, I got better results by driving through the cylinder refiller, so I needed that extra room for error.

The left turn before the right-turn refiller can be tricky to T-drift through, because it's such a gentle turn. You have to enter the turn with some speed and then tap the steering gently, but not so gently that you lose your drift. Boosting can make it a little more consistent, but make sure that you can control the next turn at the speed you're going. Whatever you do, don't hit the right wall here, or you'll grind against it and lose all your speed.

A spiral skip (youtu.be/NlSGQZp5RKs) could save a few seconds per lap, but there isn't a consistent skip strat for AX Fat Shark yet, and it's kind of a pain to practice the skip. It didn't seem worth the effort for me at this point, so I just tried to get a shift boost going down the steep downslope. The spiral turn itself is pretty awkward due to the slope features, so I got better results here with MT instead of T-drifting.

The ice turn makes things a bit ugly for Fat Shark again. Generally, Fat Shark kind of wipes out and loses speed after drifting in one direction for too long, so I go for 2-3 taps on the outer wall to kind of reset the grip. I shouldn't hit the wall once the track straightens out though, otherwise Fat Shark starts to bounce around like an air-hockey puck.

The left turn after the ice is relatively flat, so I can T-drift on it. But the turn is very tight, so I had best results by tapping the outer wall intentionally, just like the last hairpin in Screw Drive. The last couple of turns can be a bit deceptive. I have a habit of going too fast after the refiller and hitting the left wall hard, so I recently added a bit of MT to prevent that. There could be a slightly better strat for this part though.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Green Plant: Spiral
Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.11.23 22:33 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - PTCW - 134041 with Fat Shark

Joselle Astrid 2019-11-19 | Similar to my custom-machine runs here, I initially disliked running this course with Fat Shark and didn't think I'd put much effort into it. But I did in the end, and got a much better time than I expected.

Best splits are 31"907 - 30"334 - 30"205, so this run was very close considering how chaotic the course can be. There are a lot of turns where things can get messy, starting from the first left turn and the thin section afterward. But there are two main parts that tend to make or break a lap for me.

The first part is the stretch from the first energy refill to the start of the branching paths. The gentle curves during and after the refill are the best place for Fat Shark to gain serious speed. But if I'm too far left or just going a little bit too fast, I'll pop up and possibly get tossed around the cylinder to lose all that speed.

Then for the long right turn, my plan is to turn with accel off until I'm getting too close to the left edge, at which point I tap accel + brake to adjust my angle further inward. It all happens quickly, so it's quite easy to misjudge something and understeer or oversteer. It's common to get tossed completely off-course as a result. On Lap 3, I wanted to increase my survival chances, so I went for a bit less speed than usual.

The second crucial part is the last stretch. It's thin and tricky to line up with, but I also can't gain much speed on straights with Fat Shark, so I want to enter this part at high speed if I can. I eventually settled on this long right drift into a short MT left turn. It's not particularly easy to hit the final dash plate this way, but that's not too big of a deal on boost laps. The dash plate does help on Lap 1, but funnily enough, I missed it here and still scored my best Lap 1 time.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Port Town: Cylinder Wave
Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.11.18 19:43 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - LTR - 222979 with Fat Shark

Joselle Astrid 2019-11-15 | Non-custom record at this time of writing. Improved from 2'28"640 (-5"661) which I got in my All Tracks run last month. This was a near-sum-of-best-splits run that came pretty suddenly. I got a bit shaky around the middle of Lap 3, and definitely got nervous when I saw my high speed going over the bumpy refiller, but fortunately held it together.

The last time I talked about this track + machine combo in-depth was in my 2'39"668 video (-16"689): youtu.be/WVu8tW6aYH4

Since then, I've improved my strategies for some of the turns. Most notably, at the sharp left turn near the end, I strafe out wide to enter the turn, then use accel+brake to restart my drift in the middle of the turn. My overall skill with Fat Shark has also improved a lot, so I'm pinballing off the walls less often now, and I've got the confidence to do things like boost-MT into straight stretches.

This 2'22 run in particular had virtually no pinballing, and that's the main reason why it's such a big PB. Most of the bottlenecks still have room for improvement though. The first bottleneck in Lap 1 (0:19) was pretty mediocre, probably because the speeds I got beforehand were faster than I'm used to. Lap 3 (1:55) is a good example of that bottleneck done right.

For the second bottleneck (0:28, 1:16, 2:02), not hitting walls would let me go off the jump around 2000 km/h on boost laps, instead of 1675-1850 like I got in this run. Actually, I've even shift-boosted off the end of the jump, yielding a 23xx km/h glide (youtu.be/NZEUJGxJ8NE). With stuff like that, I bet Fat Shark can cut 2'10 at least, but it would take quite some finesse to pull off.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Lightning: Thunder Road
Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.11.09 19:35 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - All Tracks T-Drift in 1209933 (16:04.09 RTA) with Fat Shark

Joselle Astrid 2019-10-17 | Real time: 16:04.09
In-game time: 12'09"933
Twitch recording: twitch.tv/videos/495752521

Individual level (IL) times:
MCSO 1'19"418 (+2"060 on personal best)
ASD 1'39"556 (+0"105)
OSMS 1'47"729 (+4"284)
PTCW 1'43"013 (+3"015)
LTR 2'28"640 (new PB)
GPS 3'11"577 (+6"314)
Total 12'09"933 (+15"778 versus PBs)

For real time, I start timing on pressing A at the settings screen before Sonic Oval; Don Genie's pose changes abruptly at this point. I end timing upon crossing the finish line in Spiral. The LiveSplit timer on the video is just an approximation, since I have to control it manually while playing. Also note that the game feed may be slightly behind other elements of the video.

The load times don't seem to vary significantly with my setup, but I haven't tested for slight variations. I turned off license card functionality in the service menu to greatly speed up the loads and menus.

The in-game times are a bit tricky to catch; you'll have to frame advance (the . key in YouTube) from the fade-to-white after 'FINISH'. The in-game time can be seen briefly on the rankings screen.

As for the gameplay, this run felt quite solid! OSMS and GPS were fine, MCSO and PTCW were good, while ASD and LTR were PB level. GPS almost turned sideways near the end (literally), but I handled it well, all things considered. There's a lot that can go wrong in this category, so this is better consistency than I expected.

At this point, the best PTCW and GPS ILs I have on this channel are outdated, though I've uploaded them to my Side B channel (youtu.be/CNtWLjyQeDI and youtu.be/C1wLOBr6XyM). I'm still going to try and polish those IL times - once I think I'm done, I'll upload the runs here with more detailed comments.

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.10.16 22:39 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with LiveSplit and NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - OSMS - 143445 with Fat Shark

Joselle Astrid 2019-04-07 | Non-custom record at this time of writing. The best done on an arcade cabinet is 1'44"934 by Macabre with Mighty Typhoon. It was tough to beat; this is probably my most-played track with Fat Shark now.

Lap 1 (26"900): Should have MTed a bit longer right after the first ice. The tunnel end was quite good; did a couple of small drifts between dash plates, still didn't miss the next dash plate, and then got a good jump. The right path shift boost was rather lucky, and the end shift boost was solid. This was my best opener.

Lap 2 (25"767): Could have used another boost before the tunnel last dash plate. Then again, I could have run out of energy on Lap 3 as a result, so maybe it was a good call. Bit of slow wall scraping at the end, but otherwise fine.

Lap 3 (25"589): Got tossed to the tunnel ceiling, but it didn't lose too much time. Rest of the tunnel was good. I was too far right on the last turn to do anything useful (left is good for drifting, center for a shift boost).

Lap 4 (25"189): An even bigger tunnel jump, similar to the one in GX snaking. It's very hard to land this properly using a slippery machine in AX, but this result wasn't bad; I escaped the side of the tunnel pretty quickly. I took the left path for consistency; I think the right path is marginally faster for this lap, but it's easier to pinball badly there. I'm happy with averaging 25"5 for boost laps. My best lap is 24"350, but I've only gotten a few low-mid 24s.

AX Fat Shark is pretty slippery, and extra care is needed with pipe tunnel physics. It's easy to drift too far to one side of the tunnel where the curvature is tighter; and when that happens, I can easily get 'caught' on the side for a couple of seconds, flailing around 800 km/h or lower.

For the first ice patch, I can barrel into the left 'wall' at full speed and sometimes get good results, but it's hard to control exactly what happens. I throttled myself here to avoid the walls, thus increasing consistency. Specifically, I either try to pause my snaking for a moment before the ice patch, or I start MTing in the middle of the right T-drift. Even with throttling, speed and timing are pretty important for both ice turns.

If I'm close to the left wall as I turn left, I might get a big jump like the one in GX snaking WRs (got it partially in Lap 3, and a bigger one in Lap 4). It's a potential timesaver on Lap 1 especially, but it's tough to land correctly, especially with AX slipperiness. Often after landing, I end up losing all my speed at the tunnel wall, and sometimes I even end up facing backwards as a bonus. Overall, I don't go for this jump on purpose, but the left turn is somewhat tight to make at these speeds, so it just happens sometimes.

At the end of the tunnel, I approach the jumps by first angling a little left, then T-drifting lightly to the right until I pop into the air. If I don't angle left first, I tend to get a slower, more lopsided jump.

When taking the right branch, I throttle myself at the start, either by avoiding the dash plate or by stopping drifting for a bit. Then the goal is to T-drift right with the correct steering amount so that I touch the outer wall gently, without pinballing off both walls. If I don't get a shift boost after this turn, I follow up with a left T-drift which should tap the right wall lightly around the end of the branch, or possibly tap the left wall lightly instead. If I don't turn hard enough, I'll probably get a slow grind on the right wall at some point.

When taking the left branch, don't stray too far right at the first refiller, since the right wall is actually pretty close and it can easily make you grind to a halt. I always do a little left stride before entering that refiller, then a right stride as I begin to grab energy, and I'm usually safe that way. At the end of the branch, it's quite tough to avoid wall hits when going at a decent speed, so I just make sure I hit the right wall instead of the left wall. The results are usually acceptable that way.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Outer Space: Meteor Stream
Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.04.06 19:06 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - MCSO - 117358 with Fat Shark

Joselle Astrid 2019-03-24 | Non-custom record at this time of writing. The best done on an arcade cabinet is 1'18"660 by ルイマ (Ruima) with Fat Shark.

This run had one of my best starts, and all of the boost laps were at least decent. I actually skipped 1'18, with my 2nd best time being 1'19"400 (followed by about 5 other 1'19s). Theoretically there's still a lot of improvement to be had, to the tune of 10"8 + 9"5 + (8"7 * 6) = 1'12"500, though best splits get kind of unrealistic when there's 8 laps.

There are two major drifts on each lap. The first drift should ideally drive onto (or very close to) the dash plate before the refiller, and then get extended almost to the dash plates alongside the refiller (thanks to Valyssa and CGN for demonstrating this in their runs). However, aiming a little too far right can slow you down a fair bit, and aiming too far left cuts the drift short and reduces the speed gain. It's also tough to make adjustments while keeping the drift going, since you have to be very light on the steering in the first place.

However, it seems that timing a boost slightly before the refiller makes the too-far-left case more acceptable; I'm still able to reach 3000 km/h frequently this way. This gave me some leeway with aiming the first drift: I could either aim at the dash plate or somewhat left of it. Without this leeway, it would've been extremely hard for me to get sub-10 on every boost lap.

For the second drift, if you're positioned too far right (or angled too far right) as you start the drift, you'll get slowed down by the outer slope. It's something to watch out for at higher speeds in particular. I ended up watching for a little 'valley' (low point) on the left side of the road as a cue for starting my second drift.

At the end of the second drift, aiming for the right side lets you extend the drift longer. But the road there is a bit uneven, so approaching at a certain angle (more left-to-right) may make you pop into the air and regain grip. Avoiding the last dash plate also lets you boost-MT at the end of the drift, though I think boosting just before the dash plate would work fine too.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Mute City: Sonic Oval
Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.03.22 22:47 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - ASD - 139451 with Fat Shark

Joselle Astrid 2019-03-24 | Non-custom record at this time of writing. The best done on an arcade cabinet is 1'41"356 by MACABRE with Mighty Typhoon. Most of the laps had minor/medium mistakes, but overall this was a pretty consistent run without serious chain reactions of errors.

This short course has a few spots which are challenging to optimize:

1) The first turn. I settled on T-drifting with gentle steering, and tapping the outer wall in the process. After two wall taps, I restart the accel to straighten out and complete the turn. The speed after straightening out isn't that great, but I figured I could snake to get back up to speed reasonably quickly. And the accel restart makes my exit angle/speed quite consistent compared to alternatives.

If I'm going slow enough, I can use a normal drift turn (without strafing) to get through without wall hits, but the turn is long enough that Fat Shark will wipe out and lose speed by the end. A T-drift restart or two could solve that issue, but I was having trouble getting through the turn cleanly with drift restarts, since the track isn't that wide. That's why I settled on wall hits.

I've had faster results with a carefully angled railslide (youtu.be/-6s_8ZMvBQY) and a shift boost (youtu.be/oQOTfSvi-aE?t=102), but I couldn't seem to replicate either of those.

2) The loop entry. Snaking/drifting within the loop is ineffective, so having a good entry is important. I try to approach the loop generally the same way each time, with a left T-drift and then boost MT. But often there is a large speed loss as I start the boost MT, so my entry speed can be anywhere from 1000 km/h to 2200 km/h.

I tend to lose more speed when I angle too far to the left, or when I start MTing too late. By avoiding both of those things, it seems like I can avoid the 1000 km/h case most of the time, but results still vary more than I'd like.

Other than the loop entry, not tapping the wall within the loop itself is also important. I had mixed success with that in this run; probably the most improvable part for me overall.

3) The last turn. Like the first turn, I settled on using a wall tap or two. If the wall tap pushes me more to the left, I lose more speed, but I'm able to avoid popping into the air and the finish becomes easier to manage. If I end up more to the right, I'll generally have more speed (and a slightly more efficient driving path), but I'll also pop into the air and things get more unpredictable.

It's probably wiser to deliberately go for a wall tap that pushes me to the left (by aiming for a certain part of the outer wall), but I guess I felt like rolling the dice to try and get slightly faster lap times. It worked out OK in this run. When I'm pushed more to the right, sometimes Fat Shark needs a second to stabilize, or gets a shift boost, and I need to adjust my next drift to compensate.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Aeropolis: Screw Drive
Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.03.14 22:06 UTC-0700
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - GPS - 312966 with Fat Shark

Joselle Astrid 2019-02-22 | Currently, this is the best non-custom time for this course in AX. The best done on an arcade cabinet is 3'17"181 by P-MODEL with Groovy Taxi(!). From personal experience, real-life stamina is definitely an issue in this course when playing on cabinet, so I would say Nintendont has an advantage.

This is still a nice run for me though. Lap 1 was sloppy at the start, but the rest went rather smoothly, enough to be my best opener. Lap 2 could be better in some places, but was my best lap due to a lack of big mistakes. Lap 3 had the biggest mistakes, at the cylinder jump and the pre-ice spiral; but after those first two laps, I'll take it. I could have needed a lot more attempts to get a time like this.

It's possible to skip the pre-ice spiral with Fat Shark (youtu.be/NlSGQZp5RKs), but I don't know if it can be done consistently. If it can, then sub-3 would certainly be feasible, especially considering this run still has its share of minor/medium mistakes.

Also, I made some attempts at the start shortcut which is used in GX (going off the left side at the end of the pipe), but I couldn't get it with Fat Shark at all. I guess AX's slippery steering was making it harder for me. Perhaps it would be easier with a lighter machine. In any case, I don't think it would save a lot of time for Fat Shark, since it can get a lot of speed by drifting in the pipe.

As I've indicated in my arcade cabinet run of this course (as well as my T-drift tutorial), Fat Shark has some trouble with the tight turns here, requiring a lot of accel restarts in some places to compensate for its wide turning radius. However, Fat Shark can get back up to speed pretty quickly in many cases, so it's still a very viable choice as far as non-customs go.

I'm actually not sure how Groovy Taxi handles in this game. Perhaps something to try out later.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Course: Green Plant: Spiral
Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.02.21 18:25 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display

F-Zero AX (Nintendont) - GPS - Spiral skip with Fat Shark

Joselle Astrid 2019-02-22 | I wasn't even thinking of going for this, but it happened. I typically try to get a shift boost at the top of the hill, but without hitting the wall. A few times before, I had hit the wall and just flew out off-course to the right. This time, my angle allowed me to maintain a lot of height.

This lap was otherwise nothing special, so it was just 1'03"640 (best is 1'02"603 without the spiral skip).

Course: Green Plant: Spiral
Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)

Date: 2019.02.21 18:04 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.

F-Zero AX (Nintendont) - Lightning: Thunder Road - 239668 with Fat Shark

Joselle Astrid 2019-02-17 | At this time of writing, this is the best non-custom time for this course in AX. The best done on an arcade cabinet is 2'43"697 by るいま with Blue Falcon.

The record would probably be stronger if not for the difficulty of snaking on an arcade cabinet. Snaking is very beneficial for non-custom competition, because no non-custom machine can T-drift in a straight line.

My best splits are 55"243 + 50"056 + 49"527 = 2'34"826, and there's definitely still room for polishing on top of that. Mistakes can be pretty punishing though (head-on collisions, slow runs through bottlenecks), so I still get a lot of runs over 2'50.

Lap 1 here was the strongest lap relative to previous runs, though it had its ups and downs as well. The stretch from 12 seconds to 30 seconds (in-game time) was pretty much exactly what I wanted to do. The bottleneck leading up to the second jump was quite sloppy though; it's hard to recover there if I hit the wall early. The start of the wavy refiller wasn't great, but the shift boosts made up for it; and the last couple of turns were mediocre.

The part before the zigzag road (before the second jump) is a bit awkward. The left turn massively slows you down if you have the accelerator on while doing it, so I release accel there. I do need to slow down shortly after, though, or it'll be hard to avoid a head-on collision at the zigzag road. So I either scrape the wall there, or tap accel to slow down (like in Lap 3).

Before and after the wavy refiller, I intentionally hit the outer wall while T-drifting through the long turns, since Fat Shark just kind of wipes out if it attempts big turns normally. If done correctly (with a good entry angle and turning radius), I can exit with 2000 km/h or so. If I turn too sharply, I might hit the wall too hard and pinball into the other wall as I exit the turn. If my turn is too shallow, I might scrape the outer wall repeatedly and lose all my speed.

By the way, most of my recent videos had a frame drop issue; there were some stretches where one frame would get dropped every 5 or 6 frames. That should be fixed as of this video. After adding my Elgato as a Game Capture Device in OBS, I just had to go to Properties and enable buffering.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Machine: Fat Shark
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2019.02.15 17:33 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - Port Town: Cylinder Wave, Flying - 047695 with Combat Grampus

Joselle Astrid 2019-02-16 | I had an idea for improving my 54"553 with a new strategy, involving one respawn per lap instead of just one respawn on Lap 3. It worked out:

1. Drive off the starting platform. Fly to the side to avoid the cylinder's gravity.

2. Fly near the blue round building. This puts your checkpoint past the first turn. Try not to fly too high around here, or you could get stuck on another part of the track. But don't fly too low or you'll be considered off-course. Remember, longer strides are better for maintaining/gaining height.

3. While staying far away from the lap line (so that your checkpoint remains past the first turn), cross the lap line backwards. I fly to the tower structure to do this, because it's an easy landmark.

4. Dive down and let the respawn take you through the lap line. This should complete Lap 1, and you'll be respawned near the start of the cylinder.

5. Drive backwards off of the cylinder; make sure you become airborne before reaching the very beginning of the cylinder. The reason you need to do this is that the very beginning of the cylinder is actually behind the lap line. So if you just drive straight off of the cylinder without getting airborne early, your last landing point will be on Lap 1. The game doesn't let you finish Lap 2 unless your last landing point was on Lap 2.

Specifically, I first go to the underside of the cylinder, since that actually makes me start higher when I escape the cylinder's gravity. I also aim diagonally so I can build up a bit more speed; driving completely sideways on a cylinder is very slow. Then I strafe off of the sharp-curving part of the cylinder to get airborne.

6. Repeat steps 2-5 to complete Lap 2 and start Lap 3, then 2-4 again to complete Lap 3. Unless you are very good at step 5, your starting height here will be less predictable, making the flight a bit more difficult. Keep an eye on the cylinders around you to gauge your height, and make sure you fly cleanly under or over those cylinders instead of landing on them.

I think this is the best strategy for Combat Grampus, since the respawns help cut out some height-gaining time. But Maximum Dragon should fare better with the GX strat, since it can gain height more easily.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Machine: Combat Grampus (Splash Whale, Combat Cannon, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: No restrictions; checkpoint skipping allowed

Date: 2019.02.14 18:40 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - Port Town: Cylinder Wave, Flying - 054553 with Combat Grampus

Joselle Astrid 2019-02-01 | This is the GX strat performed in AX, except for the ending, where I took advantage of AX's respawns to make things slightly easier (not necessarily more optimal though).

1. Drive off the starting platform. Fly to the side to avoid the cylinder's gravity, while staying high enough to not be considered off-course.

2. Fly near the blue round building. This puts your checkpoint past the first turn. Try not to fly too high around here, or you could get stuck on another part of the track.

3. While staying far away from the lap line (so that your checkpoint remains past the first turn), cross the lap line backwards. I fly to the tower structure to do this, because it's an easy landmark.

4. Stay behind the lap line as you return to the starting platform. When you are close to the lap line again, land (because you have to land once per lap for the next lap to count) and cross the lap line forwards to complete the lap.

5. Repeat steps 1-4 to complete Lap 2.

6. For Lap 3, repeat steps 1-3, then just dive down and let the respawn take you through the lap line. (Alternatively, you could get close enough to the lap line and then cross the lap line forwards while airborne.)

Combat Grampus needs to gain some height to complete step 4, so it may save more time with a respawn-every-lap strategy. I tried this, but I wasn't able to complete Lap 2 this way. I couldn't figure out why, but I think I have a good guess now.

The respawn puts me at the start of the cylinder, and then what I would do was drive backwards off the start of the cylinder to get airborne again. But it's quite possible that the very beginning of the cylinder is actually behind the lap line, which means I would be getting airborne with my last landing point being in the previous lap. When my lap status is like this, the game wouldn't let me complete the next lap.

If, instead, I used the cylinder's awkward shape to get airborne before reaching the beginning of the cylinder, then I could keep my last landing point on the current lap, and the strategy should work. So I intend to try that when I get the chance.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Machine: Combat Grampus (Splash Whale, Combat Cannon, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: No restrictions; checkpoint skipping allowed

Date: 2019.01.25 18:33 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - Outer Space: Meteor Stream, Flying - 102390 with Combat Grampus

Joselle Astrid 2019-01-31 | This is the GX spaceflying strat done in AX. A breakdown, as I understand it (I might not have everything right):

1. Drive off of the starting platform while avoiding the pipe's gravity. This requires a fairly well-timed sharp turn.

2. Gain some height. I change from L to R once first, in order to stay out of the pipe's gravity. Combat Grampus needs a couple of spirals for height, but Maximum Dragon probably doesn't need to do this.

3. Fly backwards until you almost reach the final turn, then turn around while staying far to the side of the lap line. Don't fly too far backwards, and don't fly too high, or you might pop onto another part of the track (you can't see it until you land on it).

4. While you're still far to the side of the lap line, cross the lap line forwards, then cross the lap line backwards. Make sure you are above the lap line as you do this, too. Afterward, your checkpoint status should be at the last turn of Lap 1.

5. Approach the track again and land behind the lap line; this should snap your checkpoint forwards from the last turn to your current position. Now just complete the lap.

6. Repeat steps 1-5 for Lap 2, and repeat again for Lap 3.

You'll notice I end up spin-attacking a lot when attempting flying in AX. A spin happens when steering while holding L+R, so if I fail to release one strafe completely before pressing the other, then I spin. I guess I didn't have a huge incentive to be careful about spinning, since it doesn't really seem to affect height; I think it does slow down my flying a bit though.

So that's something to work on, along with incorporating more boost power. For now, I was just focused on completing 4 laps without major mistakes.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Machine: Combat Grampus (Splash Whale, Combat Cannon, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: No restrictions; checkpoint skipping allowed

Date: 2019.01.25 17:20 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, mkv 480p, with NintendoSpy input display
Editing: Avidemux (trim), AviUtl (upscale + audio normalization)

F-Zero AX (Nintendont) - Lightning: Thunder Road, Flying - 156881 with Combat Grampus

Joselle Astrid 2019-01-27 | FighBat came up with this flying shortcut in 2016: youtube.com/watch?v=hw1wVCoKWbM

When trying recently, I came up with a decent idea to apply the shortcut on all 3 laps, and this run demonstrates that idea. Here's a walkthrough of the strategy:

1. T-drift to the first jump and start flying.

2. Cross the lap line forwards after going off-course, but before the respawn ship picks you up. On Lap 1, I did this by landing at the end of the lap into a checkpoint-induced 'off course', then sliding across the lap line (unlike in GX, an AX 'off-course' landing is treated the same way as 'broken down', so you keep some momentum). This completes the lap, and then respawns you back at the first jump. But when you respawn, you are still on Lap 1, even though your latest lap is Lap 2.

3. Start flying again, and this time land in front of the lap line, at the beginning of Lap 1.

4. Drive across the lap line backwards. You might think this would put you at the end of Lap 0, since you were at the beginning of Lap 1. However, this actually puts you at the end of Lap 1, since your latest lap is Lap 2, and the game doesn't let you go 2 laps earlier than your latest lap.

5. Drive across the lap line forwards. You're now at the start of Lap 2.

6. Repeat steps 1-5 to complete Lap 2 and reach the start of Lap 3. (On Lap 2, I unintentionally completed the lap a different way: I fell into the abyss, but still crossed the lap line while off-course.)

7. Repeat steps 1-2 to complete Lap 3.

Laps 2 and 3 actually come close to my expected T-drift lap times (without flying). So, an alternative strategy which gets around the same time is to T-drift Laps 1 and 2 normally, and just do the shortcut (steps 1-2) on Lap 3.

Here I used Combat Grampus, which is the best spaceflying machine we've found so far which doesn't require AX memory card support (not available in Nintendont). Assuming spaceflying feels mostly the same as in GX (which I believe is true, after also trying Combat Grampus in GX), Maximum Dragon should be noticeably better for flying, and could achieve a better timesave with this strat. Combat Grampus needs longer strides to gain or maintain height in the air.

A couple of other ideas for more difficult strategies:

- Bounce out of the track sometime before the end of the first narrow passage, then fly behind the lap line and land there. I'm not 100% sure, but I think this would complete a lap during the respawn movement. If it works, it might not be usable 3 laps in a row, but it should lead to a faster overall strat.

- Copying GX's flying strat, which is very difficult due to needing several bounces on and off the track, as well as long, mostly-blind flying routes.

AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Explanation of the AX-exclusive T-drift mechanic (not present with the "AX in GX" code): youtube.com/watch?v=TC07KEB7K0g

Machine: Combat Grampus (Splash Whale, Combat Cannon, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: No restrictions; checkpoint skipping allowed

Date: 2019.01.25 16:44 UTC-0800
Game version: Revision D
Game configuration: Run with Wii RVL-001 (NA) + Nintendont v5.485 (Force Progressive: on), with game ISO stored on Sony SF-32UY SD card.
Controls: Gamecube controller (silver)
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS, 480p with NintendoSpy input display
Editing: Avidemux to re-encode audio with faac (this seemed to avoid audio distortion), then AviUtl (upscaling, audio normalizing)

F-Zero GX - Big Blue: Drift Highway - 057659 with Fat Shark

Joselle Astrid 2018-12-30 | 23"818 - 16"999 - 16"842. This continues my streak of at least one Max Speed improvement per year, since 2012. Procrastinating like last year, but I got around to it!

This run isn't a big improvement (-0"756), but I did fix the slowest part of my previous best. That run ran out of energy before the lap 3 chicane. This time, I didn't boost after the lap 2 end jump, thus saving a full boost to keep my speed up throughout lap 3.

Having to boost through the chicane a second time reduces my completion rate, but I managed anyway. It helps a bit to end the pre-chicane MTS close to the right wall so that I can get a better angle into the chicane. I still need good timing on the quick turns, though.

I actually still had about 1/4 of a boost left before the final dash plates, which could have translated into a full boost (stacked with the dash plates) if I'd activated it just before entering the refillers. So using that could have shaved some time. The 'railslide' could also be better, but it's always a bit awkward here with Fat Shark, since the dive takes you so close to the finish already.

My best opener is 23"420, and I got 23"4xx several times. I think the MTSes in this run's opener were good for my standards, but I seem to have missed the edge plate at 0:19, which I usually get.

MTS technique is really key for optimizing this course with Fat Shark. I think I've got it down about as well as I had in the past, which I can't really complain about. It's something to polish further if I put more time into Max Speed though.

I've overhauled my recording setup since my last video. Now I'm capturing using Wii component cables instead of S-video, and I'm converting that signal to HDMI for use with an HD capture card. There should be no more interlacing artifacts for games I can capture like this, as long as those games support progressive scan (as GX does).

I've also started using a GC controller input-display adapter (bought from pidgezero_one), which works with the NintendoSpy software to get the controller display shown on the video.

Finally, I'm using OBS to record instead of AmaRecTV, so the video will end up matching my Twitch.tv stream. I still generally prefer to have game audio only on my YouTube videos, so I'm using OBS's audio tracks to make sure my mic is only included on my Twitch stream.

My previous run: 0'58"415 (-0"756) with Fat Shark
http://youtu.be/ZvODt9UJAjM

Machine: Fat Shark
Setting: 100% (all the way to the max speed side)
Rules: 100% settings only, no snaking
Other categorizations: No side attacks

Date: 2018.12.29 19:01 UTC-0800
Played on: NA version, NA Wii RVL-001 + Wii disc channel
Capture hardware: Component + Portta component-HDMI converter + Elgato HD60S
Capture software: OBS (480p), with NintendoSpy v1.11.0 display
Editing: 720p upscale with AviUtl + x264guiEx 2.34

F-Zero GX - Casino Palace: Double Branches - 228591 with Fat Shark

Joselle Astrid 2017-12-31 | First GX personal best in 2017! My old time was 2'33"680: youtube.com/watch?v=zf1S1sWVJXE

Besides a personal best, my main goal was to get better at running this track smoothly, without worrying much about the edge shift boosts yet. My edge-SB fundamentals are relatively weak, and this isn't the easiest track to practice them on, from what I understand. Hopefully sometime in the future, I can devote the necessary time and focus to get good at edge SBs.

Lap 1: Had better speed than usual in the first half, but missed the dash plate before the final hill.
Lap 2: Pretty clean throughout. Could have used another boost after the first refiller, but often I'm a little short on energy in Lap 3, so it's hard to say.
Lap 3: Nicked the wall after the first MTS, and missed a dash plate at the first pair of "8" shaped branches. Both fairly minor though. Shaky at the second-half dash plates, but my speed there was fine.

Overall, this is about as clean as I can expect for now. This isn't an easy course to run perfectly. I'm a bit disappointed to not get a 45 Lap 3 even once, but not too big of a deal.

0:18 - Doing this shift boost close to the jump plate takes me significantly higher compared to shift boosting a little earlier.

0:47 - MTSing on this narrow stretch can be tricky. I hold the stick halfway left in advance of starting this MTS, to ensure my machine is unstable enough to change orientation quickly. Don't prolong the MTS too much so you can line up for the dash plate. Be sure to press L to help get a quick, smooth MTS exit. (My MTS here on Lap 1 was fine, but I missed the dash plate anyway.)

0:53 - This MTS into jump plate would save more time if I could get closer to horizontal, but I couldn't figure out how to do that. I think it helps if I quick-turn left earlier, but I often hit the wall when trying that followed by an immediate MTS.

1:06 - It can be easy to skid at this slope before the refiller. I think you need to follow the track line closely with a quick turn or two, but it's tricky to judge the turn. I mess it up more often when going slower than usual, actually.

1:12 - This track transfer sometimes allows me to stay completely level like this. Most of the time I get tilted a little bit instead, which makes the landing less clean. I'm guessing you want to be perfectly perpendicular to the track edge when strafing off, but I'm not sure.

1:28 - This MTS actually elicited a lot of wall hits for me. I suppose the braking makes the speed vary here a fair bit; it's hard to brake to an exact speed.

1:31 - These 6 dash plates can feel significantly tougher to hit in boost laps. I don't think your position when approaching the hill matters too much. The main thing to me was quick-turning far enough to the right before the first dash plate. The hill is a bit deceptive, and I used to not turn right far enough because of it.

1:39 - You can get a shift boost (or a double shift boost) at the end of this hill if you're going fast enough, but it requires you to be pretty risky, going just barely slower than the point where you would fly off the hill.

1:42 - An MTS into the jump plate can let you land right at the start of the refiller, but again, I couldn't seem to figure out this MTS well enough to land this consistently.

1:50 - I often misjudged this MTS on Lap 3 in particular. The boost timing is a bit different from Lap 2, at least with my strategy. The MTS itself is executed about the same, but the MTS exit needs to be a bit earlier, from what I can tell.

Machine: Fat Shark
Setting: 100% (all the way to the max speed side)
Rules: 100% settings only, no snaking
Other categorizations: No side attacks

Date: 2017.12.31 02:21 UTC-0800
Played on: NA version, NA Wii RVL-001 + Wii disc channel
Capture hardware: S-Video + GV-USB2
Capture software: AmaRecTV 3.10 with deinterlacing to 480p + x264vfw
Editing: AviUtl + x264guiEx 2.34

SMG1 - Sea Slide 2 - Luigi - 1:17.97 (43.60 penguin race)

Joselle Astrid 2017-11-09 | The main improvement here is spin-swimming faster in the eel tunnel by not holding A. For some reason, when you're not holding A during spin-swimming, your speed depends on your distance from the sea floor. The more distance, the more speed you get (up to a certain point; the speed cap seems to be close to the top of the eel tunnel). The fastest speed you can achieve with A released is faster than the speed you get while holding A.

My specific strategy is to release A starting from the bubble coin, keep A released through the eel tunnel while staying as high as I can, and start holding A when I get close to the "X" rocks.
In this run, I adjusted poorly after the twister boost and didn't stay as high as I could have, so I lost some time. It's possible to reach the second bubble coin before losing 2 health, with a bit of time to spare. Despite being suboptimal, I still gained 1.3 seconds over my old Luigi time (45.71 youtube.com/watch?v=aAzDpdgmRJM) due to A-release.

Mario can of course ignore the bubble coins, and might benefit a bit from A-release swimming near the surface to the right of the machine, since the sea floor is somewhat deep there. I haven't tested this enough to determine if it really helps, though.

After the first surface spin, I pound into the water, shake the Nunchuk once, and start Wiimote shaking right after the Nunchuk shake. In terms of when I want the game to register the inputs, I want the Wiimote shakes to register as soon as possible (optimally 1 frame) after the Nunchuk shake. But in practice I perform the shakes a little farther apart.
This strategy avoids the extra 14-frame delay between the 1st and 2nd spins which normally happens when you just use the Wiimote. It's kind of complicated to explain, but I should have a research video to fully explain this within a couple of weeks.

I still don't have a setup for optimally getting a surface spin from underwater. It's a 37% chance. When you're repeatedly Wiimote spinning underwater, a spin lasts 30 frames, and you want to reach the surface during the last 11 frames of a spin to start a surface spin.
If you want a safe, sub-optimal strategy for this kind of surface spin, just hold the stick Down all the way and spin repeatedly (at a flat water surface), and you should get it within 3 tries.

I think this is roughly the optimal place to start the last surface spin, just a little before the bricks. As far as I know, klar was the first to do it like this (46.40 youtube.com/watch?v=jlz808FcwwQ). A bit earlier and you go back underwater after one surface spin. A bit later and the "up-slope" of the water interferes with initiating a surface spin.
During the last surface spin, cutting to the right of the stone arch saves about 0.2s, and cutting to the right of the small rock after that (which I didn't do) would save about 0.1s. The leeway for not bonking might depend on your spin timing, but I'm not sure.

I didn't do any low-yield surface spins in this run (just out of laziness). There are a couple of surface spin strats available at the eel tunnels. One is to surface spin from the end of the right eel tunnel to the "X" rocks, which Vallu has done in a run (44.95 youtube.com/watch?v=pPxuafQQeww). Another idea is to surface spin from the start of the left eel tunnel to the "X" rocks, but I don't know if that is faster.
It's also possible to surface spin near the moving machine, but I'm not sure what's the best way to transition that surface spin back to normal spins.

For those unfamiliar: the surface spin at the start of the race works by shaking the Wiimote starting just after the "1". Yes, shaking before the start of the race is important, and the reason is that it avoids the 14-frame delay I mentioned earlier. Surface spin chaining only works when none of the spins are delayed.

Sea Slide Galaxy, Star 2 - Faster Than a Speeding Penguin
Luigi, 1P, skipping cutscenes
Time from star select (frame 208) to Star Get cutscene (frame 4882): 1:17.97
Time from first movement (frame 543) to Star Get cutscene: 1:12.38
Race time: 0:43.60
Spreadsheet of my level runs: docs.google.com/spreadsheets/d/1nS2sEv06TpbsJLWb7Jl7vBzIZfabCDiskY4cM_44VkM/edit?usp=sharing

Date: 2017.10.24 23:57 UTC-0700
Played on: JP version, NA Wii RVL-001 + Gecko OS v1.9.3.1 on Homebrew
Capture hardware: S-Video + GV-USB2, Line In audio
Capture software: AmarecTV 3.10 with deinterlacing to 480p + x264vfw
Editing: AviUtl + x264guiEx, upscaled to 720p

SMG1 - Sea Slide H - Luigi - 1:08.91

Joselle Astrid 2017-11-08 | This shows how the level would go in a 120 star run if you previously fed the Hungry Luma during Star 3.

I used a consistent setup to initiate the surface spin:
1. Enter the water without holding Z.
2. Hold Up on the stick and spin, so that you do a little spinning hop which goes out of water and then back in.
3. After re-entering the water, wait a little bit. Then start shaking the Wiimote. Your third spin should initiate a surface spin.

The "wait a little bit" on step 3 is the only vague part, but the timing is pretty lenient, which is why I consider this a consistent setup.

Make sure you are holding Up starting from step 2 until you get the surface spin, and keep holding Up to maximize your surface spin speed. And don't press or hold Z during any of these steps. By not pressing Z, your swimming gets pulled toward the water surface even when you're holding Up, and your horizontal direction is locked. Normally, this limited swimming state is annoying to deal with, but we take advantage of the limited movement for the surface spin setup.

When you land on the notes planet from the cannon, there's a 2 or 3 frame window to do a D-Pad Up landing cancel. Typically it's 2 frames if you're holding the stick, 3 frame if not holding the stick. It's nice if you do get it, since usually you exit first person view automatically due to the tile's vanishing movement. Unfortunately, I didn't get the cancel in this run.

I introduced the music note route in this with-Luma run: youtu.be/YGQZrHdajrk (1:50.64 with cutscenes and Luma)
In a comment, Really_Tall had a suggestion for optimizing the last note: "run over the tile that has the final note you collect when you go past that way, without collecting the note. Then you could ground pound as you get it instead of having to run past the edge of that tile."
I looked into this, but any path alteration I attempted felt like it added close to the cost of running one extra tile anyway. Maybe there's a good way to do it but I didn't figure it out.
However, I did improve the last note over my 1:50 Luma run: I turned to the side while grabbing the note, thus reaching the edge of the tile a bit more quickly compared to running all the way across the tile.

One other thing to explain at the end. After the black hole disappears, but before the star appears, there is 1 frame when you are able to move. If you have the stick released during this 1 frame, you will come to a stop and will have to re-accelerate after the cutscene. So, make sure you are holding the stick during that frame of movement. This also applies to Hurry-Scurry Galaxy (Mario/Luigi is typically off camera in that case, so it's even harder to notice the frame of movement). For my 1:50 Luma run, I didn't know about this frame of movement, which is why I jumped before the last note to keep momentum.

So yeah, I was actually holding onto this run for 3 months, and feel kind of bad about it since the surface spin setup is quite useful (for the comet star as well). I was looking into other surface spin setups, and that escalated into a full investigation of water spin mechanics, which I ended up getting burnt out on instead of finishing. I was also hoping to get the notes planet landing cancel in this IL, but I'll settle for talking about it in the description. The notes planet walking was pretty well optimized, so all things considered I'd have a tricky time beating this run by much.

As for that investigation, I'll continue working on it slowly. In the meantime, I think the strat in this video is still the best surface spin setup I have.

Sea Slide Galaxy, Hidden Star - Hurry, He's Hungry
Luigi, 1P, skipping cutscenes
Time from star select (frame 207) to Star Get cutscene (frame 4338): 1:08.91
Time from first movement (frame 541) to Star Get cutscene: 1:03.34
Spreadsheet of my level runs: docs.google.com/spreadsheets/d/1nS2sEv06TpbsJLWb7Jl7vBzIZfabCDiskY4cM_44VkM/edit?usp=sharing

Date: 2017.08.03 20:21 UTC-0700
Played on: JP version, NA Wii RVL-001 + Gecko OS v1.9.3.1 on Homebrew
Capture hardware: S-Video + GV-USB2, Line In audio
Capture software: AmarecTV 3.10 with deinterlacing to 480p + x264vfw
Editing: AviUtl + x264guiEx, upscaled to 720p

Super Mario Galaxy - Any%, 1P Luigi - 2:40:10 (2011, for JRDQ)

Joselle Astrid 2017-11-01 | Here's a bit of SMG speedrun history that I never got around to uploading. I did this run for Japan Relief Done Quick (JRDQ), an online speedrun marathon organized by members of Speed Demos Archive (SDA) in April 2011: http://gdqvods.com/event/jrdq

The idea of an online-only relay event was pretty new for speedrunners at the time. Even streaming itself was new territory for some runners like me. The event was also organized on very short notice, in order to be as significant as possible for the disaster relief cause.

At the time, I was still trying to finish the last part of my segmented SMG speedrun. But then this marathon started getting organized on the SDA forum, and I became very eager to participate in it. In fact, I got a capture card and stream set up so that I could participate.

By signing up, I gave myself 2 weeks to figure out streaming, derust the first half of the game (which I had done segments of over a year ago), and adapt my segmented strategies to single-segment as needed. I'd never done an SMG single-segment run or practiced for one, so my estimate was 3:20 to buffer for gameplay and setup uncertainties. Leading up to the marathon, I did get decent practice in, except I never managed to squeeze in a full practice run for some reason.

So yes, for those keeping score, things that would never be seen in a marathon today:
1. This marathon run was my first ever full SMG run.
2. I had no streaming experience except just before this online marathon!

There was moderate schedule competition (about 50% accept rate if I had to guess), so it's not like this was the only run to pick. I suppose this got accepted due to the popularity of the game in general, and possibly the very good reception that my segmented run was getting. In any case, the run happily did go smoothly enough and under estimate.

The VOD situation was another story. If UStream had a VOD solution at the time, it either wasn't utilized or wasn't adequate for the marathon somehow. So, VOD recording was up to the players instead of the organizers. For my part, my laptop wasn't too powerful, so I couldn't make a good quality recording while streaming with Adobe FMLE. In fact, I'm pretty sure I had to use Skype on an older laptop to reduce the load on my streaming laptop. To record the game in better quality, I split my game feed to my DVD recorder.

I think I made the right decision with my tech setup, as I was prioritizing the marathon stream going smoothly. However, I got lazy about uploading the VOD for a few reasons. There were only a handful of marathon runners uploading their VODs, and I thought mine wouldn't be much to write home about, given that I couldn't record commentary. Also, processing 3 hours of DVD recording in Anri-chan would take a while on my laptop.

From an SMG records standpoint, this was the best run I knew about at the time; but since it was my first run, I thought it would be very improvable. It was a while until I actually improved it in late 2011 (though I don't have stream VODs of this anymore), while preparing for my AGDQ 2012 run. At the event itself I got a personal best of 2'35, so at that point my PB finally had a recording, and I figured I was done.

6 years later, I've decided to upload this JRDQ run after all, for historical records. It's awkward though, because as far as the vast majority of today's SMG community is concerned, this purported 2011 run didn't exist until now. Since I took so many years to properly reveal this, I hesitate to tout it as a former WR. On the other hand, this WAS a live-streamed marathon run, and the first full SMG run for many of its viewers. So I hope you agree that I'd have to be pretty audacious to fake this video somehow.

Anyway, my takeaway here is that it's easier to upload runs as I achieve them; I'll try to be better about that. If you do accept this as a 2011 run, hopefully you'll find it interesting to see the strats and level of play. At the time, it's safe to say this run exceeded my expectations. I remember those Galaxy Reactor deaths pretty well though, heh.

Other notable records posted on Nicovideo around this time:
1P Mario 2:48:47 by konkuu, June 2011
2P Luigi 2:37:44 by NEG_taro, August 2011

Any% speedrun - Start at the beginning of Luigi's quest, Finish by completing Bowser's Galaxy Reactor
Luigi, 1P, 4:3 aspect ratio
Time from file select (frame 1960) to Grand Star cutscene (frame 577987): 2:40:10.05

Date: 2011.04.08 23:54 UTC-0700 (based on file date shown after the run)
Played on: NA version, NA Wii RVL-001 + Wii disc channel
Capture hardware: DVD recorder, LP quality with a DVD-RW disc
Editing: Anri-chan 3.3, IQ, with .avs edited to upscale to 720p

F-Zero GX - Aeropolis: Dragon Slope - 235150 with Fat Shark, no MT/SA

Joselle Astrid 2017-10-31 | From 2003. This is the last of my decent-quality GX recordings from 2003-2004. The others that I've already uploaded are Cylinder Wave 1'45, Slim-Line Slits 1'48, and Sonic Oval 0'43.

Momentum Turbo/Throttle (MT) is releasing the accelerator to maintain high speeds. At this point, either MT wasn't widely used by Max Speed players yet, or it was too new to me and I decided to stick with what I knew at the time. MT makes a huge, obvious difference in this course due to the big dives, so this is my best demonstration of the pre-MT era of Max Speed GX. (This definitely wasn't a WR at any point, but still.)

In general, I think I didn't use MT for most of 2003, then came back for a few MT and MTS improvements in 2004. The improvements included 2'23 here (Fat Shark, no SA), a decent update to PTAD, PTCW 1'45, and not-so-great Black Bull times in SOSS and LLC. I'm not sure where MCSO 0'43 goes on my timeline. It seems to be one of my latest 2003 times, and maybe even my first one with MT/MTS.

The rest of my 2003-2004 GX recordings (14 of them) are only available in really poor WMV quality, because I didn't save the runs on VHS or DVD in the long term. I did save replays, but unfortunately my memory card was stolen at some point. Anyway, putting the WMVs on YouTube will probably make them even harder to watch, so I've directly uploaded the files here: ln.sync.com/dl/63ca3dfe0/umdz6we7-3g9sjw8f-7y59iwn8-csjn2ryi

Machine: Fat Shark
Setting: 100% (all the way to the max speed side)
Rules: 100% settings only, no snaking
Other categorizations: No releasing the accelerator, no side attacks

Date: 2003.12.21 UTC-0800 or earlier (from Date Modified on my old .wmv file)
Played on: NA version, NA GC DOL-001 + Standard disc loader
Capture hardware: VHS, then later re-recorded on DVD
Editing: Anri-chan 3.3, XQ, with .avs edited to upscale to 720p

F-Zero AX - Credits

Joselle Astrid 2017-10-30 | I was practicing Meteor Stream in Race mode, and just felt like watching the victory lap cutscene after one of my runs. To my surprise, as the victory lap started, the staff credits started showing. I'm not sure if this AX staff roll is on YouTube anywhere, so I decided to upload it.

For those who don't know what the victory lap sequence looks like without the credits, here it is: youtu.be/pHOVzWtK_go

Normally the victory lap goes for about 17 seconds without any text on screen, but here it went 55 seconds along with the credits. And normally the podium scene fades out while the camera is still showing the winners, but this time the camera panned up to the sky like in GX's Grand Prix, while showing the copyright text.

I honestly have no idea what triggered the credits. The only information I can find about the credits is from this old 2channel thread: game4.5ch.net/test/read.cgi/arc/1066408295
"If you've gotten 1st place in Race mode on every track, and satisfied some other requirements, you can see the staff credits."

This was definitely not my first time winning Meteor Stream in Race mode, so I don't think I accomplished anything new like that on my license card. There were also other occasions recently where I decided to watch the victory lap cutscene, but no credits appeared. Also, while this is my first time seeing the credits, it's possible that I had satisfied the credits requirements before in the past, but decided to skip the victory lap on those occasions.

Here are the content differences from GX's credits:
- AX-only sections: Product Manager, Mechanical Engineers, Cabinet Designers
- AX adds "Powered by Hardware R&D" under "Thanks to"
- GX-only sections: Music Engineer, Theme Songs for the Characters, Movie Producer/Directors/Production, and the last five sections under Nintendo

Game configuration: Rev D or E AX iso (not sure which, but I don't think it was C) run with Nintendont v4.436, played using a Gamecube controller. Running Homebrew on a Wii RVL-001 + Sony SF-32UY SD card.

F-Zero AX - Port Town: Cylinder Wave - Bloopers and jank

Joselle Astrid 2017-10-29 | A single video of my personal best doesn't do this track justice.

Some of these clips are pretty common occurrences, but help complete the picture of things that can happen during runs. Other clips aren't so common.

The part at 1:18 caused the game to save a Speed record of 4260 km/h, a big improvement over my previous 35xx. Being able to go off-course in Time Attack has its advantages!

If you're wondering why I was in Race mode instead of Time Attack mode sometimes, I was using that to try and practice more efficiently. You can configure the game to have a lot more laps in Race mode.

Machine: Dark Grampus (Splash Whale, Dark Chaser, Euros-01)
Setting: 0% (all the way to the acceleration side)
Game configuration: Rev D AX iso run with Nintendont v4.441, played using a Gamecube controller. Running Homebrew on a Wii RVL-001 + Sony SF-32UY SD card.

F-Zero AX (Nintendont) - Port Town: Cylinder Wave - 127580 with Dark Grampus

Joselle Astrid 2017-10-28 | Although we don't have a 2003-2004 video to reference for Cylinder Wave, it feels like I've figured out this track better than any other so far. Lap 1 was my best (very smooth except before the branching paths), 2 was pretty solid, and 3 was a little shaky but got the job done.

The narrow section up to the first refiller was one of the toughest parts to get down. I needed a slow enough speed to manage the right turn without flying off, hence the MT. After that, as I approach the refiller, I can either stay on the same side of the cylinder or I can go to the other side. Lap 1 is a nice example of going to the other side. Lap 2 is a decent example of staying on the same side, although you can also see that the bump where I restart the T-Drift might make this the slower option.

The right turn before the branching paths is another spot where I could in theory be going faster, but I opt for lower speeds for better control. If you're going somewhat fast here (1700 or so), and you get into the air, you are practically guaranteed to go off course. The cylinder gravity seems less forgiving than in GX.

After the drop-off, we have that tricky corkscrew. It's hard to not slide like crazy to the side here, especially with speed, so I just do a short T-Drift beforehand and then tap A a few times to slow down and prevent sliding. It looks weird, and is likely suboptimal, but at least it gets me through consistently.

For the straightaway at the end, I generally start the T-Drift a little later than I could, just to make sure I can line it up well. It's essential to not pop into the air where the pipe is thin, because it's near impossible to restart your T-Drift there. Also, note that you need a small right turn before the last dash plate to maintain a good line for the remainder of the cylinder.

Differences between AX and GX mechanics: http://yoshifan28.blogspot.com/2017/04/f-zero-ax-mechanics-differences-from-gx.html
Drifting and grip physics are completely different. The "AX in GX" code only uses GX mechanics. You need to play the AX iso via Nintendont, or the actual arcade cabinet, to experience AX mechanics.

Unofficial AX leaderboards: docs.google.com/spreadsheets/d/14s4_ZCRkdsVrrSJB9nHHH9NXkRfVYfnbUhU5a9utcj4/edit?usp=sharing

Machine: Dark Grampus (Splash Whale, Dark Chaser, Euros-01)
Setting: 0% (all the way to the acceleration side)
Rules: T-Drift, no checkpoint skipping

Date: 2017.07.21 23:00 UTC-0700
Game configuration: Rev D AX iso run with Nintendont v4.441, played using a Gamecube controller. Running Homebrew on a Wii RVL-001 + Sony SF-32UY SD card.
Capture hardware: S-Video + Ezcap 116, Line In audio
Capture software: AmarecTV 3.00c with deinterlacing to 480p + x264vfw
Editing: AviUtl + x264guiEx, upscaled to 720p