ClaudeFFT

2023 Claude Speedfast Review

Jan 8th, 2024
282
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 17.13 KB | None | 0 0
  1. 2023 Speedrun Review (also I finally made a pastebin account lmao)
  2.  
  3. To preface, I already talked about all of this in a pretty elaborate video. Check that out here:
  4.  
  5. https://www.twitch.tv/videos/2026449771
  6.  
  7. I'll try to keep this pastebin kinda brief (at least compared to past ones), and will let that video provide more details if you want them.
  8.  
  9. To begin with, I had 5 main goals I set for 2023. Those were:
  10.  
  11. 1. FFT NA Any%: 3:54:XX
  12. 2. Mystical Ninja Starring Goemon (100%): 1:54:XX
  13. 3. Sword Master NA Any%: 0:05:4X
  14. 4. Kirby Super Star 100%: 1:14:XX
  15. 5. Super Marisa World (100%): A run with 7 disasters or less
  16.  
  17. I succeeded in the first 3, am very close to the 4th, and learned all the strats for the 5th (but never did a run or got consistent at any of it). So overall, I consider this a pretty successful year!
  18.  
  19. My 2022 review pastebin, for reference:
  20. https://pastebin.com/ixZB2rWq
  21.  
  22. Onto the main review:
  23.  
  24. ===
  25.  
  26. Donkey Kong Country 3 (Any%)
  27. 47:29 -> 47:27 (Jan 12th)
  28. https://www.twitch.tv/videos/1705939123
  29.  
  30. Carryover from 2022. I wanted to play this category again in honor of Shaaf's GDQ run, with the idea being to implement a lot of the new minor movement stuff + the 2-1 zip + one skillful 2-1 jump on KAOS to hopefully make PB easy. Instead, it took forever like it always does because Bleak is impossible.
  31.  
  32. While I will say it was fun to learn all the new stuff, actually grinding for a PB is just way too annoying. The run I ended up getting had a huge fall in Tracker Barrel Trek that cost 10+ seconds, but because I got first try Bleak Skip it didn't really matter. It's really the dynamic of the run that I find too annoying - the early game is boring / doesn't feel like it matters until you get past Bleak, and then if you do get past Bleak on a good run it becomes really easy to mess up the actual hard stuff in W6 / W7. If I do play this again, it will be for 46 - but I'm not sure when that'll ever happen. Should just stick to 103%...
  33.  
  34. Jungle Book
  35. 17:52 -> 16:18 (February 3rd)
  36. https://www.twitch.tv/videos/1729888417
  37.  
  38. The capstone to Jungle January. The goal for this game was to get a run with both vine warps, which was tech known for years but considered too difficult for RTA. This game has a lot of precise tricks other than vine warps, as well as some difficult platforming stages like Waterfall. It also features a 5+ minute autoscroller, which concludes with fighting Louie 2 (the hardest video game boss of all time).
  39.  
  40. I succeeded in my goal of getting a run with both vine warps (which save a billion years), however the first one was very slow (I even died once attempting it). The second was instant though, so that was pretty nice. The run itself contained a few other mistakes, so sub 16 is definitely possible (especially if I go for one more prs-boost), but this was a good step forward in Jungle Book history. I hope one day Newpants redoes the TAS...
  41.  
  42. Donald Land
  43. 17:58 -> 11:53 (March 22nd)
  44. https://www.twitch.tv/videos/1773287636
  45.  
  46. D o n a l d L a n d. I always said I would learn this game someday, and while I technically learned it for the 12 Hour Challenge last year, this was my attempt at playing it "for real". The initial goal was to see if the Oasis World wrong warp was possible RTA, but after about a half hour it became incredibly clear it was strictly TAS only, so I gave up on that pretty quickly.
  47.  
  48. At first, this game was very difficult - Donald is pretty unwieldly, and the game is very punishing if you're not consistent at Burgerless (the act of skipping all hamburger pickups to avoid the forced visit to the shop, which wastes ~10 seconds each time). I also had a lot of difficulties with the random movements of the generic enemies, as well as keeping up Burgerverse (breaking the layout of the level by scrolling the screen really fast). However, eventually everything clicked, and I got pretty consistent at everything past the early game (first 3 levels impossible, though).
  49.  
  50. I made a theory TAS for this game early on, and funnily enough it was slower than what my final time ended up being. Just goes to show how much I improved over my weeks of running it. My run even got a Croc Skip, which is (apparently) sub-pixel perfect. This game's a lot of fun, and I'll probably try to improve it someday. Put Donald in Big 20, you cowards ! !
  51.  
  52. Sword Master NA
  53. 05:57 -> 05:43 (April 21st)
  54. https://www.twitch.tv/videos/1800776657
  55. https://www.youtube.com/watch?v=kp2kj2fE2GQ
  56.  
  57. More NES fun. I've been running Sword Master since 2007, so no surprise I felt like running it again last year. Despite being only 6 minutes long, this game is pretty rough - not only do you need to be consistent at 1-frame taps for double & triple jumps, but the luck requirements have only gotten more brutal as time's gone on. I estimate only 10% of runs will obtain the luck needed to make it past the first ~2 minutes of the run, and even that is being kinda generous.
  58.  
  59. For this grind, two things changed. One, I switched to using a SNES controller, which made the 1-frame taps much, much easier. This allowed me to do a few more strategies that were previously too difficult for me (and even then, they were still kind of hard). Second, I drastically improved the EXP route, which involved farming the rocks in the second-to-last level. Both of these improvements contributed heavily to being able to improve my PB as much as I did, although it still took a very long time (400+ attempts). Still, pretty happy with the final product, and will definitely go for sub 5:40 at some point (with an even harder route).
  60.  
  61. Sword Master J
  62. 06:15 -> 06:08 (April 28th)
  63. https://www.twitch.tv/videos/1808492819
  64.  
  65. Sword Master, but the Japanese version. This version is a bit more jank due to some sound issues and the fact that it can pause randomly for no reason, but it's a bit less stressful due to how much easier the multiple jumps are. The fact that you can't skip level outros by pausing make it slower overall, though.
  66.  
  67. I usually attempt to improve my J time after a big NA grind, so I kept tradition here. I didn't realize it when I got it, but this run was actually extremely good. In terms of both luck & execution, this is probably the best speedrun I did all year - the sole blemish is having an unlucky fight against the Flail Guy, costing a few seconds. So, I'm not sure if I'll ever try to actually improve this. It's possible I'll find enough strat optimizations to make coming back worth it, but even then, the luck requirement seems pretty daunting. Will just focus on the NA version moving forward most likely.
  68.  
  69. Final Fantasy Tactics (NA Any%)
  70. 3:56:13 -> 3:54:43 (June 2nd)
  71. https://www.twitch.tv/videos/1836468755
  72.  
  73. F F T. My NA Any% time was the worst of the 4 categories that I knew, so guess it was time to get back to that.
  74.  
  75. I've been running this category for over 10 years at this point, so I know it pretty well. As always, I tried to find some optimizations - but because you don't really do anything in this category, it's pretty hard. However, I did manage to succeed a little bit - I found a conditional strategy that uses Rad if he can join with very high starting Wizard JP. This allows you to unlock Time Mage on him via spillover, to potentially one-round Bariaus Valley, which will save ~40 seconds in a perfect situation (assuming very high starting Wizard JP). Anyways, I didn't actually get that in the run - but it was still cool regardless :)
  76.  
  77. The run I got was alright I suppose. Watching it back, it didn't seem very good at all - but it had a good start and didn't completely blow up, so I suppose that'll do it. 5 Encounters also helped, but eh - that feels kind of expected at this point. No Scrolled Orlandu also lost a lot more than I remember in this category... oh well.
  78.  
  79. Had I kept grinding for a few more weeks, 3:53 seemed pretty realistic... maybe even 3:52. But, I'm experiencing diminishing returns on the category, and sub 3:55 has historically been considered the "good" time - so I decided to call it there. Now that I have decent times in all 4 categories that I know, I might try and investigate other ones - probably New Game+ (that category's complicated, will explain when we get there) and potentially Glitchless stuff.
  80.  
  81. Street Fighter II - The World Warrior (Hardest)
  82. 10:38 -> 10:22 (June 21st)
  83. https://www.twitch.tv/videos/1852783011
  84.  
  85. A chill 1s PB game to take it easy while preparing for Goemon. I first ran this game back in 2017, and it's an okay run - you mostly spend the entire run trying to throw-loop the opponent (SF6 foreshadowing?). I use Chun-Li because I think she's the fastest (mostly for having Lightning Legs for the bonus games), but obviously I did no real testing or science.
  86.  
  87. I did a small amount of prep for this game - I watched a few other runs, and improved my strats for Zangief, Claw, and Dictator. I ended up getting this PB extremely fast, with the last 4 fighters being extremely good. If the first 8 fighters had been just as cooked, it would have been close to sub 10... maybe. Anyways, it was fun to play this again, but only because it lasted for one day. I don't think I will return any time soon, but you never know.
  88.  
  89. Aladdin
  90. 17:55 -> 17:45 (June 22nd)
  91. https://www.twitch.tv/videos/1853616643
  92.  
  93. Another easy 1s PB game. I first ran this in 2016 (solely to cut the Twig), and have been an infamous Aladdin hater ever since. However, the game is actually not that bad - or maybe I've just learned how to appreciate games more over time? Either way, running this again was kind of fun. You mostly just hold right and mash the cape thing.
  94.  
  95. The biggest improvement for these attempts was that I changed my control setting so that the cape was on X (instead of a shoulder button), since I learned from Marisa that mashing L or R is kinda not good. My old PB had a death in it, and this one didn't - so that's p much where all the time save came from. It's kind of crazy how much time you can save in this game if you optimize all of it - but I'll leave that to the professionals. May return to this if I'm ever super desperate for content.
  96.  
  97.  
  98. Mystical Ninja Starring Goemon (Any% w/ Thaisamba Skip)
  99. 1:11:57* -> 1:08:42 (July 5th)
  100. https://www.twitch.tv/videos/1864624570
  101.  
  102. The first of the 3 Goe runs. While making practice saves for 100%, I accidentally found a trick that saves 4 and a half minutes - WHOOPS! Unfortunately, it's 1-frame (just like everything else in the game), and you only have one chance at it. That trick is called Thaisamba Skip, and you can find more info on it here:
  103.  
  104. https://www.youtube.com/watch?v=IOg3CyojWBI
  105.  
  106. Well once that trick got found, I didn't know what to do. So, I did some attempts to get a run equivalent to a 1:12 (a time that would be one minute off my old PB without the trick), and succeeded in just two days. Afterwards, I went back to 100% and let other people enjoy grinding the funny frame.
  107.  
  108. ZAR recently got a sub 1:07:30 in this category - huge congratulations to him! As for me, I'll probably end up going for a 1:07 at some point, but I'm pretty Goe'd out at the moment. I'll talk about why in these next two runs.
  109.  
  110. Mystical Ninja Starring Goemon (100% No Thaisamba Skip)
  111. 1:55:54 -> 1:54:42 (September 22nd)
  112. https://www.twitch.tv/videos/1932850098
  113. https://www.youtube.com/watch?v=letOj-KS7Gk
  114.  
  115. The grind that took the longest this year by a significant margin. To prepare for this, I found a lot of smaller room optimizations (maybe adding up to 30 seconds or so?), and combined with all the faster strats that ZAR found a few years back, I was hoping this wouldn't be that bad. But... it was :)
  116.  
  117. Before running this category again, I considered this to be the most difficult speedrun I knew. And, after grinding it for as long as I did, it does in fact remain that way. It's just extremely brutal for so many reasons - the length, the third dimension, the fact that pretty much every important trick is one frame... by the end of September, it really started to be too much for me. I was also experiencing some issues with my N64 capture, which certainly didn't help matters much. Although I was getting some pretty consistent paces by the end of the grind, I felt an enormous sense of relief when I finally got 1:54.
  118.  
  119. While 1:53 seems doable (I had a bad early game, missed two mokkoris while also getting one 2nd try, and had a pretty slop Festival Temple Castle), I don't think I'll ever run this specific category again. I got the time I wanted with all the new optimizations, and am pretty content moving on. Plus, most of the stuff I missed in this run, I ended up getting in this next category...
  120.  
  121. Mystical Ninja Starring Goemon (100% w/ Thaisamba Skip)
  122. N/A -> 1:51:15 (October 6th)
  123. https://www.twitch.tv/videos/1944720247
  124. https://www.youtube.com/watch?v=nhzKK4HY5Mk
  125.  
  126. Hey, you know how relieved I was to be done playing Goemon 100%? Let's do it all over again, with a do-or-die trick! At least I lowered my standards for what type of time I wanted.
  127.  
  128. Basically, it's the same as normal 100%, except you skip Thaisamba. I thought this would be very straightforward, but in the process of routing it, I figured out more about how Thaisamba Skip actually worked, and realized the route I had in mind wouldn't be possible without hitting ANOTHER critical 1-frame trick. So, I had to do a whole bunch of rerouting to make sure the deathwarp halfway through was still possible, which tbh I didn't really feel like doing at the time. But, I stuck it out, and managed to get a completion with the proper health route - even though the run itself was kind of bad (it missed a ton of mokkoris).
  129.  
  130. When (if?) I return to 100%, it'll be to try and improve this variant. 1:50 is extremely doable, but the real goal would be 1:49. Will I be strong enough to go for it in 2025? Who knows. I'm also interested in routing out True 100% someday, where you get all the maps, compasses, dungeon keys, warp points, Level 2 weapons, as well as some other stuff - but that'll have to wait.
  131.  
  132. Castlevania: Aria of Sorrow (Soma Glitchless)
  133. 29:51 -> 28:44 (October 15th)
  134. https://www.twitch.tv/videos/1952347486
  135.  
  136. Aria October returns. The complete opposite of Goemon, this run has very few precise tricks, and is instead a comfy movement game (at least going for 28). This round of attempts was probably the easiest so far, since I didn't have to actually learn a lot of new stuff - instead, all I did was optimize what I did in my 29 from last year. I didn't have to do too many attempts, and I was able to make plenty of mistakes and still achieve this time. Pretty convenient too, since there were a lot of games I wanted to play casually in October - so I'm happy I didn't have to dedicate too much time to this.
  137.  
  138. Looking towards 27 this October, I'm not sure if I should just have a very clean run with strats I already know (plus cutting out some safeties), or just skip the shopping for a big time save and learn Julius + Chaos properly. Will see when the time comes, I suppose.
  139.  
  140. Donkey Kong Country (All Stages w/ O-Gang)
  141. IDK lmao -> 36:02 (November 2nd)
  142. https://www.twitch.tv/videos/1968552789
  143.  
  144. Funny game. This marks the third time I've played this category, with the previous two being to prepare for events (first for a DKC Festival, and then for trilogy commentary at AGDQ). I had a 00:32 IGT before, and the goal was to get a 00:31 IGT. I played this for exactly one day, and improved by a billion - mostly by cheating more (Perils warp, Antics Blaster, etc.).
  145.  
  146. Speaking of Antics Blaster, that was the main reason I decided to do this again. I came up with a pretty consistent setup using Diddy (the slower method), so I'm happy I was able to show that off. As far as jumprolls go, I only attempted TuT, and didn't get it a single time. I don't plan on grinding this again anytime soon, but if I do, I guess I'll go for 34:13 and then truly quit forever.
  147.  
  148. Kirby Super Star (100% Without Resets)
  149. 1:19:29 -> 1:15:43 (January 4th, 2024)
  150. https://www.twitch.tv/videos/2022490732
  151.  
  152. It's Kibbi! I first learned this run in 2019, and it was one of the most fun speedruns I ever learned. I then proceeded to not play it for 4 years for some reason.
  153.  
  154. This is the game I'm actively running, and it definitely seems like there's still a ton to improve & optimize. Because this game is so movement heavy, I've just been gradually improving my strats in every room, since I've already implemented most of the big pieces of tech (other than Jet Spring Breeze). The plan is still to play until the end of February, and hopefully get a 1:13? I love Kirby Super Star for the 1-Chip Super Famicom.
  155.  
  156. ===
  157.  
  158. Alright that should cover it. Again, I didn't go into a lot of detail because I talked about all of this in video format, so check that if you want more details. As far as future plans go, I'd like to keep things kind of open. But, for the sake of the 2024 Review, I'll state these as my 5 main goals (with the goal of accomplishing 3 or more):
  159.  
  160. 1. Kirby Super Star 100% for 1:13:XX
  161. 2. Final Fantasy Tactics New Game+ for a good run with 3 or less encounters
  162. 3. Mystical Ninja Starring Goemon Low% for any completion with all 3 critical frames (Flute Skip, Thaisamba Skip, Festival Temple Castle Warp)
  163. 4. Aria of Sorrow for 27:XX
  164. 5. American Gladiators for 00:04:51
  165.  
  166. The rest isn't too set in stone, so we'll just see what happens. No plans to stop speedrunning anytime soon either - too many good games out there :3
  167.  
  168. - Claude
Advertisement
Add Comment
Please, Sign In to add comment