Frame Cancel

[Izaw]

Falco can L-cancel in Smash 4

https://www.youtube.com/watch?v=spHWz-9F7cQ&feature=youtu.be

Check out how it's done right here!

I posted it here because this possible with MORE characters than only Falco!

Bowser Junior is one of them. The rest you need to help me find !

You can Frame Cancel every aerial in the game.

The more hitstun it has, the more noticable the cancel will be I guess?

 

[ATH_]

Yessss, this. Just got done watching this, great find Izaw (or at least good video showing it off).
Falco is becoming such a threat, he's not terrible like people thought holy crap.

 

[Anomilus]

Already use Falco, so will practice for sure.

 

[LwYz.Ph.]

Great video, you should post this in the Falco and Bowser Jr. forums.

 

[Indigo Jeans]

So do you input the fastfall on the same frame the aerial connects, or can you input it during hitlag?

 

[ぱみゅ]

Amazing stuff. *hits the lab*

 

[Raijinken]

I'll see if I can figure something out for other characters. That said, I didn't realize that Bowser Jr function was noteworthy, I just thought it was a regular function of the attack.

 

[-RedX-]

Whoever can go through that pastebin text with most of the character data, it's probably a good idea to check Falco and Bowser Jr.'s Dair and figure out what's making this possible. That way, you can look at other characters and try to find something similar.
Of course, going through all that is a hassle so it's probably easier to just lab it up.

 

[san.]

Jr's has an addition for low lag dair on hit in the code. I dunno about Falco, is it cancelling the absurd hitlag with landing lag?

 

[ぱみゅ]

Whoever can go through that pastebin text with most of the character data, it's probably a good idea to check Falco and Bowser Jr.'s Dair and figure out what's making this possible. That way, you can look at other characters and try to find something similar.
Of course, going through all that is a hassle so it's probably easier to just lab it up.

http://i.imgur.com/A4yNjtA.png

 

[Lavani]

I'm guessing @ san. 's hitlag theory is what's happening, since nothing in the dump makes it look like Falco's dair should cancel like this otherwise.

If this works with the knee...

 

[dragontamer]

http://i.imgur.com/A4yNjtA.png

I agree. I was looking at that and was like, wtf? Auto-cancel window finishes on frame4.

I'm guessing this doesn't have anything to do with planned cancels. It has something to do with how hitlag interacts with landing frames in this game.

 

[Linkshot]

http://i.imgur.com/A4yNjtA.png

Thanks for using this!

So, I've done more research since I scrawled that, and my current theory is in this string of tweets: https://twitter.com/Chispshot/status/554845009779122176

Summary: If you fastfall dAir, on Frame 17 of the aerial attack, Falco will attempt to both calculate an air->ground transition and calculate a hitbox->hurtbox collision. The latter takes priority, and Falco clips through the floor, and the landing lag flag is unset. On the next frame, it notices Falco SHOULD have landed, sees no landing lag flag (which was previously set by 170 on Frame 5, and then unset on Frame 17), so it puts Falco into SoftLanding (which is generally 3 frames of lag).

 

[Indigo Jeans]

http://i.imgur.com/A4yNjtA.png

I agree. I was looking at that and was like, wtf? Auto-cancel window finishes on frame4.

I'm guessing this doesn't have anything to do with planned cancels. It has something to do with how hitlag interacts with landing frames in this game.

I think unk_170 sets the end for the early autocancel/beginning of landing lag, while unk_16F sets the beginning for the late autocancel

i.e. landing during F1-F3 of the aerial results in autocancel. Landing F4-F37 results in landing lag. F38-end results in autocancel. That's how I've interpreted it anyway

Still if you're supposed to hit with the first active frame of the attack that's well out of the range of the early autocancel, if that's what it is

Also kyokoro the move hits on F16 ;^)

edit: oh linkshot made that. unk_170 isn't unset when the hitboxes come out though. I don't think it is anyway...

 

[Linkshot]

Also kyokoro the move hits on F16 ;^)

In Brawl modding, I was always taught that the asynch timer is how many frames it -waits- since the beginning of the attack.
Ex: Anything before a timer is Frame 1, anything after asynch=1 is Frame 2

 

[Lavani]

If this works with the knee...

Hey

Hey guys

I'm subscribing to the hitlag landing cancel theory.

 

[Indigo Jeans]

In Brawl modding, I was always taught that the asynch timer is how many frames it -waits- since the beginning of the attack.
Ex: Anything before a timer is Frame 1, anything after asynch=1 is Frame 2

That was indeed the case in Brawl, but cross-referencing the data dump with 60 fps recordings of attacks shows that asynch timers give real values now. If an asynch timer says hitboxes are generated F3, then they are generated F3, as opposed to F4 in Brawl.

Hey

Hey guys

I'm subscribing to the hitlag landing cancel theory.

...did you get it to work with the knee?

Edit: oh the video didnt load

**** thats nice

 

[Kofu]

This is pretty interesting. Note that Bowser Jr.'s version is a little different; you actually have to connect with the last (landing?) hit on an opponent (not a shield) to have no lag on the attack.

Is this technique frame-perfect?

 

[Raijinken]

This is pretty interesting. Note that Bowser Jr.'s version is a little different; you actually have to connect with the last (landing?) hit on an opponent (not a shield) to have no lag on the attack.

Is this technique frame-perfect?

It's not a technique for Jr, it's just making the attack hit. Unlike most moves that have lower lag on miss than hit, his has lower on hit than miss.

But this is interesting. I'll be looking into it more.

And there went all of my hopes for Smash4 to not have nonsense like this to add a non-depth skill floor...

 

[Kofu]

It's not a technique for Jr, it's just making the attack hit. Unlike most moves that have lower lag on miss than hit, his has lower on hit than miss.

But this is interesting. I'll be looking into it more.

And there went all of my hopes for Smash4 to not have nonsense like this to add a non-depth skill floor...

Yeah, I know it's not a technique, it's just a property of the attack. This is pretty close to what he can do normally with his DAir, though. Can anyone test if this works if it hits a shield, too?

 

[Linkshot]

edit: oh linkshot made that. unk_170 isn't unset when the hitboxes come out though. I don't think it is anyway...

unk_170 is unset because the game tries to make Falco land but prioritizes the hitbox collision instead. The gameflow is usually:
air->ground transition -> unset landing lag variable, but the hitbox collision happened on the same frame, so it passed that during hitlag without actually going into landing lag.

"Is this technique frame-perfect?"

looks like it is, but you also have to be going fast enough to trick the game into thinking you hit the ground.

 

[Indigo Jeans]

Oh okay so you meant that this glitch occurs because unk_170 is unset "accidentally", right? If so that would make sense.

Yeah yeah unk_170 is normally unset upon landing, so if you land while your hitbox connects, it will unset the bit allowing you to land without lag. I get you

 

[Raijinken]

Hey

Hey guys

I'm subscribing to the hitlag landing cancel theory.

Something to observe: During the Knee's hitlag on Diddy, in the canceled vid, Falcon's legs hit the ground, much like in his landing animation. In the uncanceled, he's still hovering for a moment.

In fact, it looks to me like the difference in landing lag is the same as the time it takes him to land in the uncanceled animation. So rather than necessarily canceling part of the landing lag, it looks to me like that lag is simply starting sooner.

 

[Kofu]

So far it's only been proven with high hitlag moves, right? Additionally, this wouldn't happen to have any chance of working with multi-hit aerials, would it?

 

[ぱみゅ]

Maybe it actually cancels the attacker's hitstun and begins the landing?

I need to watch the video again.


EDIT: Kofu, I was wondering just the same, maybe it applies to all hitboxes

 

[-RedX-]

Can't seem to get this with Ganon's Dair. :/

 

[Lavani]

Something to observe: During the Knee's hitlag on Diddy, in the canceled vid, Falcon's legs hit the ground, much like in his landing animation. In the uncanceled, he's still hovering for a moment.

In fact, it looks to me like the difference in landing lag is the same as the time it takes him to land in the uncanceled animation. So rather than necessarily canceling part of the landing lag, it looks to me like that lag is simply starting sooner.

I was just about to mention this.

All this seems to be is landing during hitlag and starting the usual landing lag earlier. Both Falcons shield 30 frames after the dust graphic.

I also feel like I've been getting different landing points so maybe this isn't actually frame perfect? Feels like I've been able to move at variable points after the knee when I manage to do this successfully, but I haven't been recording these followup attempts to verify.

 

[Raijinken]

I was just about to mention this.

All this seems to be is landing during hitlag and starting the usual landing lag earlier. Both Falcons shield 30 frames after the dust graphic.

I also feel like I've been getting different landing points so maybe this isn't actually frame perfect? Feels like I've been able to move at variable points after the knee when I manage to do this successfully, but I haven't been recording these followup attempts to verify.

Seems likely that it would work at a range of times during hitlag, and thus would give you variable success with an "optimal" timing being the earliest valid frame of the hitlag in which the fast fall can help. That would explain why it's easy enough to notice on Falco and the Knee - they've both got massive freeze-frame counts. Zelda's aerials, and maybe Ness's, Gdorf's, etc, would probably make good test subjects.

 

[Linkshot]

I think I just got it working with Dedede's bAir, so this is pretty universal.

 

[Lavani]

Got it to work with Ganon's dair. Even combo'd a usmash after it vs 20% Fox.

 

[Ffamran]

So far it's only been proven with high hitlag moves, right? Additionally, this wouldn't happen to have any chance of working with multi-hit aerials, would it?

Falco's Fair wouldn't because of it's landing hit, so that's out of the question.

I think, since my understanding of frame cancel is poor, @Pacman9 said that Pac-Man's Bair can frame cancel, but I have no evidence aside from Pacman9's word. From the Character Competitive thread:

Yeah. Pac-Man also has those unique frames of no landing lag in his bair. The difference is, the move cancels before the hitbox comes out. However, it does give us a way to catch z dropped fruit without any form of landing lag and gives us a lag less option to land in general like mega man's nair.

 

[Fox Is Openly Deceptive]

Something to observe: During the Knee's hitlag on Diddy, in the canceled vid, Falcon's legs hit the ground, much like in his landing animation. In the uncanceled, he's still hovering for a moment.

In fact, it looks to me like the difference in landing lag is the same as the time it takes him to land in the uncanceled animation. So rather than necessarily canceling part of the landing lag, it looks to me like that lag is simply starting sooner.

You have a point, but it's not entirely true. Step it forward one frame at a time from the peak of his jump, and you'll see that they desynch, with the one on the right simply falling faster, and because of this he does in fact connect with the knee closer to the ground and there are a couple of frames in which the one on the left has to continue falling before he lands, but the difference between how much time it takes to land and how much landing lag is cancelled is by no means the same. Keep stepping forward one frame at a time and you'll see.
So no, it's not just a matter of starting the landing lag earlier (in this sense). There's something else going on.

It might be that the landing lag starts as soon as the move connects (as you 'land' at literally the same time), and the immense amount of hitlag simply means that a lot of the landing lag frames are spent in hitlag so that by the time you're out of hitlag you only have like half or less of your landing lag to go through, or in some cases perhaps none.
What this would mean is that it is a universal technique, but the more hitlag you have, the more landing lag frames you 'save'.

 

[Linkshot]

Got it to work with Ganon's dair. Even combo'd a usmash after it vs 20% Fox.

This is the best sentence I will hear all day.

 

[Kofu]

Falco's Fair wouldn't because of it's landing hit, so that's out of the question.

I think, since my understanding of frame cancel is poor, @Pacman9 said that Pac-Man's Bair can frame cancel, but I have no evidence aside from Pacman9's word. From the Character Competitive thread:

That's actually different from what I know. It's comparable to IASA frames before the hitbox comes out. It's fairly easy to do with Bowser Jr. and Game & Watch's BAirs and is somewhat common.

 

[Trifroze]

Not sure if said already, but the way this works from what I frame-by-framed is it cancels the user's smash stun by having you land during it. Moves with more smash stun like the knee benefit the most from this.

 

[Indigo Jeans]

If the mechanics behind this really are landing while simultaneously connecting a hitbox, then there's no reason for this to not work with nearly every move in the game.

Not sure if said already, but the way this works from what I frame-by-framed is it cancels the user's smash stun by having you land during it. Moves with more smash stun like the knee benefit the most from this.

"Smash stun" (hitlag?) and landing lag are two different and independent things though. Even if you were to land while bypassing the hitlag, you should in theory still experience the normal landing lag

Falco's Fair wouldn't because of it's landing hit, so that's out of the question.

I think, since my understanding of frame cancel is poor, @Pacman9 said that Pac-Man's Bair can frame cancel, but I have no evidence aside from Pacman9's word. From the Character Competitive thread:

If it cancels before the hitboxes come out then it's just an early autocancel. As kofu said its a normal thing found on a lot of characters (Villager's DAir and UAir for example). This is different because you're still connecting your hitboxes

 

[Ffamran]

That's actually different from what I know. It's comparable to IASA frames before the hitbox comes out. It's fairly easy to do with Bowser Jr. and Game & Watch's BAirs and is somewhat common.

Interruptible as soon as possible frames? Yes, I Googled it. In a way, it's like frame canceling, right? Interrupting the move cancels out the rest of its frames which leads to what Bowser Jr., Pac-Man, and Mr. Game & Watch can do with their Bairs or having Falco's Fair transitioning to the landing hit.

... I've done that so many times; I interrupted moves on accident and get severely punished. It's what I get for fast-falling (Dr.) Mario's Fairs. Don't ask, it's like a thing I like to do for some reason. It's not even for stylish purposes, I just like to throw fast-fall Fairs for some reason. Landing them is still awesome, though, like fast-falling to avoid a Falco ditto's Bair and throwing a Bair to counter. Anyway, I digress.

 

[Shaya]

So the theory is...

Both players experience hit lag on hit.
On moves with extra hit lag, if the attacking character would land in the next frame of the move (or it came out at the same time or some other similar scenario), the difference (theory) between the standard hit lag and the extended hit lag would be subject to landing lag frames.

A few questions/assumptions that need to be proven:
Is it the hit lag difference or is it any hit lag being neglected (even 1.0 modifiers), what happens on moves with less hit lag? (divide by zero)
Falcon's knee with some 10 odd base hit lag going to 15, is he getting 5 frames less lag than he normally would, or is it 15? 10? what?

And does the move have to hit on the same frame it comes out, or is the only condition landing a frame after you hit an opponent.

 

[Raijinken]

If the mechanics behind this really are landing while simultaneously connecting a hitbox, then there's no reason for this to not work with nearly every move in the game.

Has anyone started testing it on special moves? Probably harder to test due to timing and most specials having less landing lag than normal aerials, and since most specials have their own animations, I'm not sure it'd even be noticeable. Maybe on something like an aerial Falcon/Wizard's Kick onto a grounded opponent.

 

[Ulevo]

This is quite exciting. I'll see about testing it with a few other characters.