Even newer possible deku castle warp (ocarina ng+) any%

[gamestabled]

Assuming sockfolder and robdogs trick works on oot3d, this game is dead.

New route is along the lines of:

Ocarina ng+
Deku sticks/ nuts
get sword dont equip
WWT mido skip? if not possible get lullaby
beat gohma
cutscene set? whatever it ends up being.
warp to castle (frame perfect bullshit)
equip sword and beat ganon as child.

Let's make this happen.

Sock just got deku entrance cutscene to work if you die after killing gohma. But we cant die and keep ocarina, but there will be other possible cutscenes we can use!

[Nook]

If we can't keep Ocarina, how will this work?

WWT Mido skip doesn't work, just was testing it yesterday. Still have to go see Zelda :}

[gamestabled]

You dont have to use that cutscene hopefully, and we could use another Im assuming to get the warp. Or just leave deku and come back would work too i think. Or even would playing suns song warp you to the beginning of deku if we got that?

And actually you wont need to get kokiri sword, you could kill ganon with deku nuts. Once this trick has a more decent setup, Im definately destroying this game

[Nook]

I will admit I'm pretty confused on what to do once we reach Gohma, what happens once we reach her?

Edit: Getting sun's song means getting Deku Stick(s) for Keese in the Royal Family's Tomb. You can do a broken Deku Stick on them, but it's a bit harder (just tried it)

[CloudMax]

I will try not to get involved in this at all since I'm working on BiT in skyward sword. I do not want to forget how BiT works again...
Anyway, an easy way for you people to find out if this trick even works at all on the 3ds is to simply use farore's wind out of the boss room after defeating gohma, right?
I mean, this wouldn't be useful in a speedrun, but it is a way to test if you even can warp to ganon's castle on the 3ds version without the game crashing and stuff.

[Nook]

Great idea!

[gamestabled]

I have tried warping (gohma room->231 room, deku warp) with farores wind, and every time so far it has crashed. I tried with the farores in other places and it worked fine, so it does seem as if ganons castle as child crashes. I wasnt oob, my system did crash. So i guess for now this route isnt happening, unless you guys want to try to make it work.

Unless, does the previous cutscene watched affect crashing? That could still make all the difference

[Nook]

I got crashes too.

[mzxrules]

If you're you're using Farore's Wind wrong warping as a test, you have to climb to the top of the deku tree and set your point outside the compass room (top floor). Setting FW while in B2 causes a crash because tower collapse lacks a map with the same number.

[gamestabled]

Just reading that made me so   ... Cant wait to test this now!

it makes sense too because deku boss->deku tree DC warp also crashes

EDIT: I've tried Compass room->top floor in Gohma warp a couple times now and Im still getting crashes. I wish I understood this better so I could have an idea of what else to try... is just trying every room in deku tree worthwhile, just to see if getting to tower as child is even possible here?

[mzxrules]

I took a further look into this with a ram watcher, and discovered something amazingly funny...

If you perform the set-up I mention, the deku tree intro is completely wiped out of memory, meaning that literally appears to work on 64 only by sheer dumb luck.

If you have set up the FWWW properly (by exiting ghoma's boss room, head up to compass room/leave, place FW, immediately ww from Gohma) and the game has crashed, this most likely means that you will have to try setting other cutscenes as your last cutscene to see if you can somehow get the ww to work.

Otherwise... no dice.

[Nook]

Okay, how about Saphire or Ruby cutscenes?

[gamestabled]

Im looking now at pages to try and understand this trick better.

if the cutscene value falls outside the normal range of values used for that scene, the game loads garbage data, resulting in broken cutscenes or crashes.

The cutscene values listed on this site are for n64 1.0, but hopefully they are related in the same way on the 3DS version.

Tower Collapse=8037BAC0
Does somebody know what digits need to match up for a successful warp? (if thats what even determines it)
Console runs use Deku Intro which is 803825B0.
All I understand of that code is that Deku Intro is after Tower Collapse, and the final digit I believe is the number of the cutscene that plays in that scene.

More things: Goron Ruby=80369FA0
Sapphire=80375FF0
ZL=80368410
Drain Well=803708C0 *This one is pretty similar, and we need ZL anyways, worth a try? Idk how to get in well early in a run though.

Anyways, I probably dont understand this as well as I think I do, but I'm disappointed this hasnt worked yet and Im a little desperate.

[mzxrules]

You're quite a bit off, so I created this image to show you exactly what I've been looking at in ram



Alright, so when you wrong warp, you are setting that current entrance (or last entrance) index to a different value. This is what is manipulated by FW.

Stage number (2 bytes) determines the cutscene number to play when reloading a scene. This is the "cutscene value" the ZSR page is referring to. This value is only used to tell the game the number of the cutscene that should be played next.

The cutscene pointer holds a memory address (as do all pointer values). This points to the part of ram where the cutscene is loaded in ram. This value changes depending on what cutscene needs to be played.

Now, three things can happen when you wrong warp.

• You can have a "perfect" cutscene warp where the entrance index + stage number matches up so that the calculated "Real" entrance index lines up with the proper entrance index table. The cutscene pointer is set properly, and the cutscene data is loaded properly
• You can have an "imperfect" cutscene warp where the entrance index + stage number does not match up, but you still fall within a scene that has a certain numbered cutscene. The cutscene pointer is set properly, but the cutscene data appears to not reload properly resulting in odd glitches with the cutscene
• You can have a "corruptable" cutscene warp where the entrance index + stage number causes you to load a scene with no cutscenes. The cutscene pointer is not changed by the warp itself, and the cutscene data can become "corrupt" in the sense that the last cutscene is no longer in memory.

In this situation, we are hopefully dealing with wrong warp type #3 and not something like the lost woods stuff. Now the theory in this case is that the wrong warp doesn't work because the last cutscene pointer ends up pointing to gibberish instead of a cutscene. So the idea is to change the cutscene pointer so that we've exausted all options, or the pointer points to some garbage data that causes the cutscene playing to crash, but allow us to progress.

You're wrong to look into v1.0 cutscene pointers for patterns (which is what you're doing) because we have no way of knowing what the memory actually looks like.

[Norkix]

I'm quite sure Grezzo built from 1.2. You should do your research with that as well.