Posts for Cronikeys

1 2
5 6
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
I’ve been following the progress of this on Twitter. Solid improvement for a great game. Gets a yes from me! Great work Dug
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
don't have time to look at this too technically atm (and let's be honest, it's a meme game), but nice job link. a bit more straightforward and not as "high octane thrill ride"-y as my WIP, but good enough quality for the vault I suppose. might try to improve one day since this game has always been a guilty pleasure of mine :) audio is also extremely underrated - hard not to feel nostalgic, especially listening to it on spike's temp encode
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
I have. No dice. You can get a similar visual glitch in dozens of other places Worth experimenting on w/ PAL though. This game is really prone to hardware differences affecting collision detection
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
CoolKirby wrote:
World 3 really does look tough! Is it possible to manipulate those stage hazards and enemies to appear at different times so they don't interrupt your movement?
I've redone it a few times.. the problem in most cases is that avoiding one causes you to hit another. For example, it's pretty fast to avoid hitting the two fish that pop out at ~2:35, but doing so offsets the rest of the level making it overall slower from what I could tell. Retaining health is obviously needed for the last level in that WIP I'll do some more testing before moving on. It'd probably require some sort of glitch involving the screen's alignment or pressing pause as an event occurs
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
[worked on it more so a lot of the original post was removed] edit: to avoid double posting - I got a lot further: Link to video bk2: http://puu.sh/gh8LP/6b78177c68.bk2 all feedback welcome. I'm anticipating getting 1 health in the next level (3), losing one in the one after (2), and damage boosting in the one after that (1) I should be fine on health for the rest of the game after that
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
Chiming in to say that the cause of the crash is hardware and not software related, so the distinction between N64/VC is a tad arbitrary. AFAIK the glitch works on emulators with recompiler just like it does on VC.. so IF a "vc" tas is done, it doesn't even have to be done on dolphin..... technically. That is to say there is no software difference causing this specific difference. I can't think of any precedent for time-saving, hardware-related glitches on this site, but an official ruling [?] would be cool since this issue has come up in other N64 games as well with hardware-dependent glitch differences (Paper Mario iirc has some vc exclusives). My 2 cents? It's a hardware glitch and shouldn't be in a tool-assisted run. The fact that it's possible on certain consoles due to the nature of emulation is irrelevant. Besides, it's in OOT 3D edit: this post is about gim, not the emulator bugs that skater posted
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
TAG wrote:
This has tickled my fancy, so what type of errors did you fix? And how? It would be great to get a very technical explanation of these fixes
the problem with saving was "fixed" by just updating something in bizhawk's game db (the thing that sets the config for each game) the physics issues will be fixed by attempting to make the interrupts that happen during the game's loop simulate hardware-conditions. it's a complicated fix that doesn't have that great of a solution (hence me aiming for 8 instead of 10). http://datasheets.chipdb.org/MIPS/R4300i_datasheet.pdf (see: frequency) i'm not going to provide any more explanations so as not to derail the thread. if you have any specific questions for me you can ask via pm or in the bizhawk/n64 emu boards.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
RingRush wrote:
...variety of emulator issues (most notably no version emulates lag properly and has reset recording). [...] In an unrelated note, I do have an emulator that runs the game reasonably well and has TAS tools (although lacks reset recording and thus cannot do a full game run). I am planning a few short, unoptimized demonstrations of some routes and tricks that a full TAS would use, so maybe expect those in the next few weeks.
I've been working on the emulator again for the BK crew (hyperresonance) and thought I'd kill two birds with one stone by addressing some of the DK64 issues. A couple hours ago I fixed saving in Bizhawk for DK64 (it was a minor fix), so full-route 'theory runs' are doable now. My next plan is to tackle physics issues. Mupen64plus has updated since it branched into Bizhawk (in lay terms, Bizhawk is using an old n64 emulator "core") so I'll either a) be updating Bizhawk to latest M64P (more complicated than it sounds), or 2) legitimizing the "Bizhawk" branch of M64P and fixing issues parallel to the m64p team Also, Isotarge (realtime runner) is working with me on the emulator. He has more experience with the game so he'll be a better judge of accuracy than me for specific tricks. - AFAIK Ringrush will be doing an Any% run for tasvideos without prior sram. That route is completable now, but I insist to anyone reading that you hold off on optimizing any TAS until I fix some of the physics issues (vines/trees teleporting you out of bounds, improper lag emulation, and intro/cutscene demos desyncing). The solutions to a few of these have already been found, so they just need to be rolled into Bizhawk. TLDR: The full route of DK64 is tasable now (now as in right now). On a scale from 1 - 10 the game's accuracy is at like a 6 for bizhawk (7 for my branch of mupen-rr) and my goal is to get that closer to an 8 before a TAS is started.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
AnotherUser wrote:
https://www.youtube.com/watch?v=w9S9lTTLp_I https://www.youtube.com/watch?v=8vFXcHTe3Dw Anyone want to verify these bee clips? I did it on NTSC pj64 1.6.
I can verify both on ntsc console related edit: you can leave that top room as bee if you enter it from oob even if the door isn't open
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
andypanther wrote:
It's not like you're not allowed to start from SRAM, there are several published TASes that do it.
There are other reasons for not using the SRAM route. My post on the matter lists several reasons. Regardless of whether it's allowed or not, I'd prefer to do the clean cart route and thoroughly explained why. I don't want to DCW us off topic so I'll keep this short: I know DKR didn't have a movie file of the dirty sram being created. The main reason for rejection was due to the fact that it did not complete the game on the file it started the game on. In BT's case, the actual file is completed despite using SRAM for the boss replay mode. So technically, yes, I'd be allowed to use dirty SRAM and stay within the rules of the site. I'm not going to. Edit: Side-note (and this is definitely off topic) your example for a run using SRAM is a developer intended NG+ mode. I believe the dirty sram rule was created with that in mind moreso than glitches involving warping to the end of the game. Further discussion of this should be done in a different forum, though.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
Patashu wrote:
Ok, this is a convincing FAQ to me. any% no dcw, let it be!
It will just be called Any% and it will be using all glitches.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
sack_bot wrote:
I'm going to say cheato% would be nice
a) tasvideos' policy on the use of in-game cheats is a grey area b) the majority of the run is cutscenes and it shows off subjectively the most boring parts of the game, performs a hard to spot glitch, and then beats the final boss c) http://www.twitch.tv/pjii/c/5936753 - the tas would be this only with more optimal movement.. so not much different
Patashu wrote:
IMO the first TAS for the site should be whichever category requires the least amount of effort, but I'm pragmatic like that ;)
In this case the one which would require the least amount of effort is cheato%, but as I said I don't think there's a serious (or warranted... see incoming rant) demand for that. Right now, realtime runners have 4 categories: "cheato%" which uses in-game cheat codes, "any%" which uses a cartridge with existing save data on it and performs the hard-to-spot glitch that warps to the final boss from level 2 using the boss replay mode cutscene of the final boss, "any% no dcw" which is any% without using the glitch that lets you warp anywhere in the game just by playing a cutscene from that area, and 100% cheato% = 27~ minutes any% = 42~ minutes (any% is a misleading name because it uses dirty sram) any% no dcw = 2 hours and 40 minutes~ (extremely rough estimate) 100% = 4 hours and 20 minutes~ (extremely rough estimate - this would use DCW and the best realtime BT runner has not done a run with that yet) In terms of "entertainment" which is kind of a nebulous concept, any% no dcw is the clear victor for realtime runs. Cheato% and Any% are more cutscenes than gameplay. Any% does have that cool detonator levitation glitch, though. Cheato% doesn't even have that. Three out of the four would require previous save data in order for a run to be considered optimal (exception being any% no dcw). The category I described is "Any%" but with the distinction that it is from a clean cartridge and without cheat codes. In length it would be somewhere between Any% "with save data" and Any% no dcw. It would be a 2 hour long long glitch fest (that unfortunately has tower of tragedy). In my humble (3 years + 1000s of hours spent on the game + the person who plans on tasing the game) opinion, this would be the most entertaining and least arbitrary of all possible category choices. Doing anything other than this would involve cheat codes or be a repeat of the "DKR Any%" fiasco with swordlesslink a few years ago: #3831: SwordlessLink's N64 Diddy Kong Racing "Glitched NewGame+" in 04:09.25 Again, to summarize: I want the TAS to beat the game as fast as possible without existing sram or cheat codes. That's the category/route I am proposing. Nobody runs it in realtime, so it will be a unique route with heavy glitch abuse. Pre-answering inevitable questions: Q: Why is the sram route boring? A: It's the intro cutscene, then the jiggywiggy cutscene, then the mayahem temple opening cutscene, then the Golden Goliath cutscene, then a reset, then the Glitter Gulch Mine opening cutscene, then.. an awesome cool glitch where you fly as detonator!, then the ufo cutscene, then the witchyworld opening cutscene, then the rest of the ufo cutscene, then the final boss cutscene, then the final boss cutscene again, then.. the final boss fight! 23 minutes of cutscenes in a 42 minute run. The glitch that's used to skip to the final boss fight is practically invisible in terms of execution. It involves playing 4 cutscenes. Q: Why are you against cheato? He's in the game, right? The codes used are not given in the game. They are developer codes, just like the ones in the sandcastle in Banjo-Kazooie. Allowing them in this game would mean you'd allow the world-opening and note-door-bypassing codes from Banjo-Kazooie as well. Q: Why should anyone care about a slower run? It will show off the many, many glitches in the later parts of the game that would be skipped if SRAM or cheats are used. Q: I am a realtime speedrunner. Why aren't you interested in my category called "any%"? The branding of "any%" by realtime speedrunners is not an acceptable label for TASvideos. At best it can be considered "any% using dirty sram." The existing 100% TAS of Banjo-Kazooie, for example, does not skip the opening cutscene. Speedrunners do. This is a difference between clean sram and dirty sram. Realtime runners are welcome to make "theory tases" of routes, but as for publication on TASvideos that version of any% does not belong. Q: Will the run include the DCW glitch? YES. It just won't be warping to HAG1 because that cutscene is not available on clean sram. It will be warping to TDL using the intro demo (which will be arguably cooler to watch).
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
andypanther wrote:
please don't get discouraged!
You have no idea how much I need words of encouragement these days for banjo-kazooie haha.. i've found so many things that have ended up being "emulator bugs" and not actual game glitches. fake ttc early is probably the biggest. i'm glad i thoroughly (jumped into that wall for like 15 hours) tested it on console, but RIP my hands and it's still hardly conclusive. Ugh. I've been primarily glitch hunting on console for the last few months for this very reason (so I don't find fake clips or fake jump possibilities.. which is probably a bigger concern for the tas routers)
andypanther wrote:
And I just randomly remembered something completely irrelevant: Another (apparently very unknown) difference between versions is, that some copies actually show more Stop 'n' Swop items in the secret ending! Specifically, I'm talking about the egg in MMM. Well, at least my old PAL cartridge does it... I love how many secrets those games still hold from us ^^
2 things: one, that's cool thanks for sharing. two, that reminded me: Blast Corps by Rare has a glitch where you can die on the credits screen. For the longest time nobody knew how it worked until this year Graviton (a speedrunner of the game) meticulously figured out it has to do with the direction you're facing when pressing start on the intro screen. I'm bringing this up because in Banjo-Kazooie, in-game demo desynchs, like in the SNS screens or any cutscene really, are extremely common. Occurrences I know about:
  • banjo missing the carrot at the end
  • dying in mad monster mansion intro demo causing back in time glitch (happened on ntsc console alledgedly without gameshark - and it's just a guess that this is what happened because the runner's story was spotty. see earlier tasvideos post on it)
  • dying in RBB during intro cutscene causing BIT glitch (might have happened on xbla - saw this on an old 2010 casual forum post)
  • falling off the ledge while getting the TTC egg in end SNS cutscene
The exact cause is unknown but it probably (definitely in my opinion) has something to do with how that Blast Corps glitch works. Link to video It's on my to-do list one day when I feel like staring at MHS (I use pj64, sue me) for hours, but thought I'd share it here in case any number-oriented people here are so inclined to look into them. I don't have a save state for any of the listed things, but there are videos floating about.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
Cool test run of "cheato%" or any% with in-game cheats, but it's far, far, far from optimized. Pjii's RTA WR is more impressive from a technical aspect. Anyway, I wanted to crosspost this novel I wrote in BK's thread: http://tasvideos.org/forum/viewtopic.php?p=400044#400044 As for this game, I still intend on TASing "Any%." I will not be using a pre-made save file, but I will be utilizing the DCW glitch to get clockworks early (the TDL chompa cutscene in the intro demo lets you warp to TDL without JRL). As I understand it, this is the category that would be most suitable for TASvideos submission? Clean cart any% w/o in-game cheats? Time would be around 2 hours. It's not a category that realtime runners do, but it is the fastest "new game" method w/o cheats. There's also this new glitch: Link to video If a discussion takes place where people would rather see a TAS that does use a pre-made save file, this would save a bit of time. There's also this: Link to video The TAS would skip the swamp completely using a glitch that still hasn't been entirely figured out, but this does provide an interesting out of bounds and an interesting look at how the game handles these flingy situations. Some weird stuff with them can happen in HFP, too.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
andypanther wrote:
I don't think we can use CEN64 for any serious testing yet. But interestingly enough, the emulator seems to work better with the classic "problem games" so far.
I'm going to try to not get too off-topic and I'm speaking in lay generalizations. Most of what I'm saying is relevant to the BK TAS. I'm moderately familiar with Mupen's code (I've read it all and know how most of it works) and can say most of its problems are due to the fact that it's emulating something beyond the scope of what 32 bit software can do. BK, BT, DK64, Blast Corps, Conker's Bad Fur Day, etc.all use a similarly programmed engine that, especially in BT/DK64's case, has a lot of 64 bit calculations occuring concurrently with 32 bit calculations. The juggling the software (mupen) has to do while switching between these values is what is causing a lot of the problems. With a single line change in the code, I can make some jumps harder, some more clips possible, and vice versa. The problem is at the bit level, so Mupen is basically hopeless in 100% accurate emulation. There is, however, a patch or bandaid available. It's something the 32 bit sibling of mupen, Project 64, has done for years: have an adjustable "counter factor" and Vi Refresh Rate. Neither of these have a hardware counterpart. They merely serve the purpose of changing the micro-calculations that occur at a machine code level. Changing the counter factor makes the jump at the opening more difficult or easier. Changing the counter factor makes the jump to Clanker's tooth more difficult or easier. Changing it can make NTSC perform "PAL exclusive" glitches, such as the RBA that uses a clip instead of the cauldron. A more extreme example is with Donkey Kong 64 where changing the counter factor can literally make you move at the speed of light and clip through literally everything in the game. Mupen64plus has recently added this bandaid (again, not lines of code that represent anything in the actual hardware). It hasn't been rolled into Bizhawk yet (I was working on it but I got sidetracked.. someone else might have taken over while I was gone for most of the Summer). BK is TASable, emulatable, somewhat accurate, etc. on mupen-rr and current release of Bizhawk, but with the next release adjusting the counter factor would make it more accurate to console but it enters a fuzzy area with movie recording because it would guaranteed not sync with anyone who had their emulator set to a different counter factor. Blah blah... all that is to say, CEN64 is being called an N64 "simulator" because it is attempting to handle the games as they were intended. As such, it can only run on beefy computers with 64 bit processors and will incorporate the hardware intensive conversions/translations whatever they're called of the calculations. The majority of N64 games were primarily 32 bit. Rare really pushed the system to its limits, so it will likely be a long while before CEN makes games like them cycle-accurate. So for this TAS, it should be started on Bizhawk with the latest mupen64plus rolled in (It's more complicated than just compiling the code of mupen64plus and dragging and dropping - about 20 files and hundreds of lines of code need to be changed. I already documented most of it, but I took a break over the summer that sort of pushed that aside.) Doing the TAS on this version of this emulator would give the best results for the foreseeable future (until the next-gen of N64 emulators become viable on consumer grade hardware). All that is to say: 1) The only way to guarantee a glitch is not a bug with the N64 emulator is to test it on console. This is true for most N64 Rare games currently. 2) A spec of dust landing inside your Nintendo 64 could potentially cause different glitches to occur in Banjo-Kazooie/Tooie and the other games I mentioned. So one person getting a glitch does not mean it is possible based on the software of the game. For an example, one person's cartridge desynched the intro and gave them 'back in time.' Another example, one person's cartridge let them clip through the floor in Mumbo's hut. Another example, one person's cartridge of Banjo-Tooie let them clip through the door into the weldar fight early. Another example, one person's cartridge let them fall out of bounds in the 810 Note Door room. 3) This is exactly why PAL glitch differences exist for BK. It really sucks. Best approach is reproducing known glitches on multiple (at least 2) consoles before incorporating them into the TAS route. I'd hate to see a TAS use something that isn't actually possible in the game, which right now is a lot of things (obviously nothing that diverges from the RTA route).
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
AnotherUser wrote:
Nevermind about this. Kirkq and Cronikeys tried to do this 2 years ago. -Idea for ttc early. There are a few bricks that are in the ground. The smaller ones that look like triangular prisms are solid rectangular prisms that jut into the ground. I'm thinking it might be possible to do the talon trot clip on these small pieces, but would require tas inputs to test. My idea is that because there is water that also exists slightly under, clipping will switch to swimming mode, and then you can swim to the loading zone for ttc. Though some water is not directly under the loading zone, so may need some special jumps to get to it.
This might help you or give you ideas related to TTC Link to video What's in that video isn't possible in the game and is just an emulator bug, though. At least as far as we know. I tried it for many hours, but couldn't get it on console.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
I showed hyper already, but you can shoot eggs at the flower hole instead of pooping them in. This should save a few more seconds with the new route. Also just for future reference here's an info dump on the CCW hole/gobi in general:
  • from nothing to jiggy requires 3 interactions with the area
  • 3 possible types of interactions are eggs in spring, gobi summer, and gobi fall
  • eggs in summer and fall will make a "you did something right" noise but don't count for some reason (this should be more thoroughly investigated)
  • an interaction can only be done once per time the area is loaded
  • the eggs are repeatable hence the glitch working
  • the "summer switch" reloads the area, so technically two interactions in the first visit to spring are possible without having to go into any unnecessary loading zones
  • gobi summer is skippable, gobi fall is skippable, or both are skippable
  • the three interactions means the flower has 4 phases (no flower, a flower, a bigger flower, and a bloomed flower)
    • summer gobi only appears when the flower is in 'a flower' state
    • fall gobi only appears when the flower is in a 'bigger flower' state
    • gobi disappears when you unload the area (e.g. using the fall or winter season switch)
  • the jiggy doesn't transcend seasons, so whatever season you perform the last interaction in is the season you will have to collect the jiggy in
  • if the last interaction is done in spring the game will softlock
  • bypassing the softlock with a loading zone by gaining control during the cutscene and heading to mumbo's (like in hyper's route) causes the state of the flower to save from 'bigger flower' to 'bloomed flower' but dying/save and quitting/resetting does not
    • gaining control is possible by getting hit when the cutscene is starting
    • while moving around with camera lock the hitboxes don't always work (this needs to be more thoroughly tested - specifically whether killing a wall bird's cutscene fixes camera lock.. if it's even possible to kill one while in that state)
    • getting hit by the bull seemingly doesn't work (hasn't been thoroughly tested. whoever tests this (i probably will) will need to check not only all of the bulls states but also his body vs head seem to have different properties)
    • the softlock occurs because the game is trying to return control to gobi, who isn't there in spring (ever)
  • if the flower is spawned in spring, getting it as bee is possible
shoutouts to ezazar, purplerupees, hyper, and whoever else knew about the egg interaction glitch. afaik it was never documented before ezazer (or maybe purplerupees). i wanted to info dump because it's an odd but extremely integral part of the ccw routing now -- side note: a speedrunner (kaptainkohl) was able to get what was thought to be a pal-exclusive oob in clanker's cavern on ntsc console. no one else has been able to replicate despite hours of testing. there's a video of him doing it though and it's easily possible on emulator (due to collision detection bugs in Mupen caused by 32-64 conversion) and PAL (again, due to teeny tiny differences at a machine code level.. i LOVE version differences) so whether or not it's "allowed" will probably need to be looked into. i'm not really even sure if it saves time. it lets you walk underwater. for now i'm inclined to say it's a hardware error for kohl (see: random back in time glitch happening for one guy's cartridge a few pages ago), but there's really no way to know for sure unless another runner on ntsc is able to get it.. so if you wanna help then jump into this wall on console!) Link to video ~it's a mystery~
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
goofydylan8 wrote:
Mitjitsu wrote:
That jumping noise is really irritating. Is that necessary to go faster like in Super Mario World?
From my tests of this game like 2 years ago jumping makes no difference, you move the same speed regardless of if you are jumping or not. On the run, I will say that it is a very boring, run right for justice run and that is why I stopped my attempt after barely doing research.
I got it for my third birthday when I was little and it holds a special place in my heart, so that's why I'm doing the run. There is some interesting stuff later on. It's not for everyone and you can always mute the lovely sound effects.
goofydylan8 wrote:
Unless I'm just dumb I can't find where you uploaded an actual WIP file so I can't compare the frames exactly but from the youtube video it appears that you have to stop on the toucan to clear the wall in level 1. If you time the ball throw properly you shouldn't have to halt your momentum. Like I said it is just the youtube video and not the actual file so I can't directly compare if it is slower or just a glitch of the video but something to keep in mind since there will be essentially 0 entertainment value so you will have to make sure it is completely technically sound to get it vaulted.
I'm using bizhawk so I couldn't upload to dehacked.. I can puush the WIP file if you'd like. It's in kind of an interim state right now (level 2 desynched a little bit due to the RNG of the ball enemies.. I think) I looked into the toucan thing. When I had tried that I was pushed back by some sort of invisible wall that only affects you from certain angles (ie, jump one frame later and you get hung on the wall and waste 10 frames doing that). Needless to say, I found a way to make it work with my health route. I beat your run right for justice demo of the toucan thing by 2 frames while still getting the useless perfume bottle, hitting the invincibility block, and ending with 6 instead of 3 health. So thanks for pointing that out! I'll post another WIP along w/ a hosted bk2 file next update with fixed level 2 and a v1 of world 2.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
I'm tasing this game on Bizhawk. If you've never heard of it before, it's an extremely difficult platform game that was well-received upon its release. Time sort of forgot about it. The movement speed is rather slow, so I am trying to make up for that with interesting actions while progressing. Here is a WIP of world 1 so far: Link to video The B orbs are health. I'm collecting lots of health in worlds 1 and 2 so I can use invincibility frames in world 3. If you've never seen world 3, it's a barrage of water hazards, flying fish, etc. Doing it without taking damage or collecting the extra health would probably add several minutes to the total length of the run, so I do plan on deviating from holding right and jumping to get those early on. The flashing "B" you see briefly that comes out of the block is like a star in Mario. It makes you invincible for a very, very brief period of time. It's not needed in 1-1, but in case people were wondering that's what it is. Glitches so far: nothing major. There's a corner clip at the beginning of 1-2 that I don't think was developer intended. Normally you have to ride the toucan down to get that health. This might be usable as a wall jump in stage 4-1 where palm trees (springs) are sometimes needed. Also, backwards jumping.. what NES game doesn't have that, though? If anyone would like to assist (mainly with finding tech or suggesting things as far as route/entertainment) let me know here. In the work in progress I didn't put a lot of effort into the boss fight and was just focusing on completing it as fast as possible. There's obviously room for play during that fight since the cat's speed is fixed. I'll do that before progressing.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
Great work guys :) Some small improvements possible, but overall a very entertaining and optimized looking run. Voting yes and thanks for the shoutouts in the comments
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
jlun2 wrote:
Try using Dolphin's Freelook feature to see what other stuff you can find.
The glitch to fall out of bounds that malleo posted was found about 6 months ago by Medibisu: http://www.youtube.com/watch?v=gFBF22Z2zn8 Him, nathanisbored (person who discovered fish glitch), and I all tested it on most screens very thoroughly. All of the game's NPCs and cutscene items, regardless of the progress you've made, are stored oob (including the star, machine, and boss of the third chapter). They aren't interactable as far as I can tell. There's something similar in a lot of Gamecube era games (like Twilight Princess has giant floating black boxes with characters from cutscenes in an area that you can even talk to but they have no dialogue). Also, on screens where you learn a move or curse, there's usually an invisible platform (or strip of water and boat pad in the case of Pirate's Grotto) below the ground where you can stand during the cutscene. The boat pad below the surface is interactable. I'd assume there's also a very thin railing you can ultra jump to below the room where you learn ultra jump. Just posting so someone doesn't have to retest all of that. None of that is useful ^ Things of note using that oob method include:
  • The key skip in Chapter 1
  • In the room in Rogueport with the fallen pillar blocking your path you can perform the glitch and then jump back so that you're partially in bounds on the edge. That allows you to inch your way back behind the buildings. You can't get back in bounds once you're back there, and even if you could being in West Rogueport early doesn't seem all that useful. You can't buy the blimp pass.
  • You can walk 'behind' other scenery, especially in Chapter 5. Most rooms in the game however have "boxes" around doorways that you can't get past even after performing the glitch.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
Sorry for double post, but I made a video of that key skip in Chapter 1: Link to video I'm not sure how much time it saves since this video I made is not optimized. It requires pixel perfect navigation while out of bounds. The door opens automatically after the curse chest so you don't have to worry about repeating this the second time you pass this area. Also, keys are door specific so you can't use the key you skip with this somewhere else.
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
Skip a Fish Glitch Link to video This saves time by skipping a fish glitch (~10 seconds maybe?). Basically, if you fall into the water you spawn back up top - meaning you only have to fish glitch by the warp pipes once. Chapter 1 Staircase Flip Skip Link to video This saves about 12 seconds by skipping the two cutscenes of the staircase flipping around. Early Boat Pad Outside Riddle Tower Link to video This saves about 30 seconds by skipping half of the background in this area. It requires a precise fish glitch to clip inside of this statue to reach the boat pad earlier. Pirate's Grotto Timesavers Compilation Link to video These tricks save about 48 seconds for TAS (31-36 realtime). They involve fish glitching to reach the Grotto key and skip two boat platforms as well as skipping two cannon fights which greatly affects inventory management. Additionally, there are two things I found which I didn't make a video of:
  • You can skip a key in Hooktail Castle by riding the purple block up, going into the next area, Goombella buffering past the loading zone of the doorway, and then jumping out of bounds (off screen) into the area below. It's extremely difficult. Since you have to use the door you're skipping twice normally, it is probably slower to do it this way. It might be possible to use the skipped key later in the dungeon in place of the key that you normally have to backtrack for after getting paper mode.
  • You can fish glitch out of bounds in the first area of Chapter 5 (the pirate chapter) and then walk out of bounds for a while and jump into the shortcut hole to the end of the chapter. It's extremely difficult and the game seems to be softlocked once you get there since the boss isn't there. You can backtrack a little once you're in that final area by falling into the water. This should be investigated more (there might be a way to get the toads to appear here).
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
Things tracked on status screens:
  • Superstar Status
  • Max Hammer/Boots
  • All Curses (4/4)
  • All Partners (with upgrades - meaning 14/14 Shine Sprites)
Items tracked in journal:
  • All Crystal Stars (7/7)
  • All Badges (85/85)
  • All Tattles (124/124)
  • All Recipes (57/57)
Things tracked on Pit of 100 trials sign:
  • All Star Pieces (100/100)
  • Pit of 100 Trials Completed
If the 100% definition included all of these, the run would be about 10 hours (4.5 hours longer than Any%). Speaking of 100%, this saves like 50+ seconds getting the Double Dip badge: Link to video
Experienced Forum User
Joined: 7/12/2011
Posts: 142
Location: Indianapolis
I found another skip that saves about a minute: Link to video It skips having to get a key, unlock a door, put in a light code, and then return back to this wheel. Edit: I also found this minor trick: Link to video
1 2
5 6