Update: The New First Active Hitbox frames For Pikachu on 3ds and WiiU

[CarbuncleHero]

Here is the first frame pikachu's moves hit on in Wii U's ver. 1.02 and 3ds's ver. 1.05:
Jab: 1
Dash Attack:3
Tilts
F:3
U (facing backwards):4
D:4

Smashes
F (15%):8
U:5
D:4

Aerials
N:2
F:5
B:2
U:2
D:7

Grab:3
Dash Grab:4
Pivot Grab:5

*I tested both the 3ds and the Wii U; everything was the same between them. Something to keep in mind is although pikachu looks to have gotten quicker between the versions, it seems everyone is still using previous 3ds frame data so how much pikachu improved compared to the others is yet to be seen. I'm not really sure if there were no changes to the other characters. I honestly can't believe no one else did this first :O. If I should test any else, let me know and I'll figure it out!

This is wrong see @ dettadeus 's post below.

 

[CarbuncleHero]

​

Interesting things:

• Frame 2 aerials everywhere! Holy cow! FFAC fair into Nair is a true combo btw.
• Frame 3 grab!
• Even though jab is now frame one the end lag is still the same...
  
• Up Smash comes out on frame 5, good gosh. I have something else to show everyone soon that will show you how ridiculous f-smash's sweetspot(18%) coming out on frame 9 is. edit:i don't really feel like showing the rage kill percentages list but sweespot fsmash is better than upsmash at killing with and without rage
  
• Dash attack does 10%, goes through shields, is a kill move, and now I know it comes out on frame 3. It also clanks with basically every tilt and smash in the game, leaving an even amount of time for both players to perform a move, barring a very few amount of strong smash attacks. I don't think many characters can compete with pika's speed in a clank situation. It is a very, VERY underutilized move.
• Heavy Skull Bashes HB comes out on frame 6 if you don't charge. That move is like my favorite move, it's so silly. There are set ups into this move, and it kills really early! Frame. Six.
  QA's HB is not within human reaction time lol. And Thunder Wave/Jolt comes out on frame 10 (this is important to me because I think that thunderwave is best used in close quarters)
• Pika's jumpsquat is 2 frames now instead of 3. This means SH nair and dair essentially get pseudo -1 frames, since these are likely to be used right in your opponent's face. 4 frames for nair to hit off of the ground.

 

[Glicnak]

This is actually amazing, and of course I encourage you or anyone (who isn't in midterms right now... darn midterms) to continue this for the other characters, although I understand how much work it would be.

That being said, this is insane! they basically doubled the speed on all of pikachu's moves. I can't imagine that they actually would do this to only one character

 

[CarbuncleHero]

This is actually amazing, and of course I encourage you or anyone (who isn't in midterms right now... darn midterms) to continue this for the other characters, although I understand how much work it would be.

That being said, this is insane! they basically doubled the speed on all of pikachu's moves. I can't imagine that they actually would do this to only one character

I would test maybe a few other characters myself but I'm testing some things out for pikachu right now. A lot of his abilities are underused or misunderstood so I'm spending my time exploring his options. It's a bit time-consuming, but if you want me to check a particular character really quick to see if they changed, I'd be glad to. The data-mining for 1.05 should come out soon so there won't be a long wait until other changes are found.

 

[Glicnak]

Oh its okay thanks for the offer though and I hope you find some interesting stuff!!

 

[Bergbuur]

Never heard ffac, fast fall a? c?
What does it mean?

 

[A10theHero]

Never heard ffac, fast fall a? c?
What does it mean?

Fast fall auto-cancel forward aerial.
It's when you do a short hop with Pikachu and then a fair, followed by a fast fall. The auto-cancel part refers to the fact that there will be no ending lag at the end of the attack.

 

[JayWon]

Holy crap these frame data is insane! Frame data for Thunder and quick attack please!

 

[CarbuncleHero]

Holy crap these frame data is insane! Frame data for Thunder and quick attack please!

Quick Attack's hitbox comes out on frame 8.
Quick Attack's second hitbox hits on frame 15.
Thunder's bolt comes out frame 1 but the hitbox starts frame 7.
The Hitbox for the base starts on frame 17 from the ground. Interestingly, the base connects later for Short hop into thunder, the hitbox starting at 19 frames, and earlier for Full hop into Thunder, at 16 frames.
Thunder's invincibility starts when the bolt connects i believe (i'm not sure how to test that) the invincibility ends on frame 27, lasting only 9 frames (frames 17-26).

 

[dettadeus]

there's no way some of this can be right, like... frame 3 standing grab? what? 6 frames is the usual startup for standing non-tether grabs, I don't think any character had a faster standing grab in all of smash. everything else seems 1-3 frames off as well.

how did you test/find all this?

 

[CarbuncleHero]

I tested this in training mode on both versions and got the same results for each. I used 1/4 hold L frame and went frame by frame to get the results. This is easily testable, I've gotten very consistent results. If this is wrong then let me know. The reason it seems 1-3 frames off is because version 1.03's data is different than 1.05 and that's where the previous data for pikachu came from. I also tested landing lag against previous versions against 1.05 and found it to be different as well.

 

[dettadeus]

that explains why they all seem to be off, because training mode isn't really accurate lol. i could believe a few frame 2 aerials, maybe a frame 1 jab, but i really doubt stuff like the grab/smash frame data

 

[CarbuncleHero]

Okay, i see your point but there has definitely been a change. I'll test the grab with a video right now and edit this with the results. But so far it seems accurate. I'll use smash mode too so there's no worry there.

 

[dettadeus]

http://smashboards.com/threads/pikachu-moveset-data-wip.381156/#post-18115682
might want to take a look at this thread, lol

 

[CarbuncleHero]

Okay. I just tested grabs twice with a video of both smash mode and replay of smash mode. I got 4 frames for grab and 5 frames for dash grab. The reason it is one frame off is because in training mode i took the frame when mario switches animation, but my phone has poor quality so it's hard to see that frame on video so I simply didn't count it. It similarly is hard to tell when pivot grab begins (since i can't tell by the press of the button) so I don't have frames for that one.
However, 3 or 4 frames for standing grab is earlier than 6 frames, so by both accounts, the data has changed.

http://smashboards.com/threads/pikachu-moveset-data-wip.381156/#post-18115682
might want to take a look at this thread, lol

I already did, which is the point of this thread :/.

 

[dettadeus]

I really doubt 1-4 startup frames were shaved off nearly every one of pikachu's attacks though, and it definitely doesn't feel that way. A frame 3 grab is literally unheard of in any smash bros game.

 

[CarbuncleHero]

Well, i don't know about other smash games but I got these from the 3ds and Wii U. If you don't believe this you don't have to. Unless it's been verified it's not set in stone, and this hasn't by anyone but me. If you think it's wrong you are free to disprove it. I got the same results consistently, but I don't really know where to get another person to test this out. Sorry, for the trouble though, I don't know what else to say to prove this other than I got it directly from the game and that it doesn't match the previous version.

 

[dettadeus]

"got it directly from the game"? i still don't know how a low-quality cell phone recording is more solid, directly-from-the-game proof of frame data than the exact numbers pulled from the game, even if it is the previous patch. i personally can't disprove it because i'm not dantarion but it doesn't entirely seem like it needs to be disproved to begin with.

i'll try to see what i can do on my own console later today i guess.

 

[CarbuncleHero]

I got most of the data from the game itself. Frame by Frame.
The video served its purpose. "A patch changed the data here is what I think that data has changed to" is what this thread is. I honestly don't see how it can get anymore direct than this. Unless you can no longer get 1 frame using training mode, then I don't really know. I can't really believe that two separate games would give the same results, a video, and testing against a previous patch wouldn't show that at least something in pikachu's frame data is different. Not just that but pikachu's landing lag is different as well, I attempted multiple times.
If you find that everything is still the same then that's fine. Please let me know and thanks for helping out.

 

[dettadeus]

have you tried your method with other characters? the 1.0.4 frame data is available for all characters, you could see how your method holds up. if it's consistently 1-4 frames off the 1.0.4 numbers for more than one character i would highly advise taking a closer look at your method.

 

[CarbuncleHero]

Good point.
I already noticed something different with mario so I'll get the smash and tilts for him or anyone else you want. But I'm not sure how this proves the method wrong. I think that the 1.04 data doesn't match the wii u's 1.02 data so wouldn't 1-4 frames off mean i'm correct?

 

[CarbuncleHero]

Double Post
Here are mario's for the 3ds, done pretty quickly.
jab 1
grab 3
ftilt, dtilt, utilt are all 3
fsmash 8
upsmash 5
dsmash 3

There's probably a mistake in there but I want to get this out quickly so I can compare the result to the wii with this one without any bias. I can get 1.0's data for the 3ds if you need that too. keep in mind I don't know mario's data (only his ftilt because i just looked it up)

jab 1, grab 3, ftilt 3, dtilt 3, utilt 3, fsmash 8, dsmash 3, upsmash 5,
It's all the same for the wii u.

 

[dettadeus]

well the dantarion dump says:
jab 2
grab 6
all tilts 5
fsmash 15
usmash 9
dsmash 5

so you seem to be off by quite a bit on some of them, but average between 1-4 frames off again. and mario definitely did not get 7 frames of startup shaved off his fsmash lmao. i think you might want to either rethink how you're going about this or wait for people to pull the actual data.

 

[CarbuncleHero]

:/
Uhh... i feel like you aren't reading my posts at all. But if you say it's right, you're right.

Why wait? the data is in the game. The method is reliable if not valid. I got what I wanted, I shared it, and left it up to whoever reads this to decide whether or not they'd like to believe this or prove it wrong. I read the data, it could say pikachu is made of cheese for all I care. I think that dantarion's patch data is correct for 1.04 but not for the 1.05 or 1.02 for wii u. I really don't understand why you keep bringing that up, I know already (hence "update" and "new"). To be absolutely clear: I am saying these two sets of data are different.

edit: I just got an idea. Dantarion's data says that mario's jab won't hit on frame 1, correct? Humor me a little, try to have it not hit on frame 1 in training mode. If you really want to humor me revert back to 1.0 on the 3ds and try it again.

 

[TL?]

Think about it. Every move startup for every character you've tested is consistently half of what other and more reliable sources have reported. You're telling us that startup has halved across the board for two characters and yet no other source confirms that and not a single player has noticed any difference in speed just from playing. You're basically telling use that the game has changed DRASTICALLY and you're the only one to notice. While it seems a lot more likely that simply your testing methods are incorrect.

 

[dettadeus]

i just booted up my wiiu and went into training mode with the 1/4 speed thing, and i can see why you might still think that stuff is so fast. you probably can't hit the l button for literally a single frame so that only one frame advances in the game. try setting the cpu to control, have it grab a timer, and then try again. you'll notice that you can get pikachu to be fully extended with the jab pose but the other player does not get hit until a little bit after.

i hope that helps a bit, because i'm done arguing this with you lol. i would be surprised if any frame data got changed, because as far as anyone knows it was just some bug fixes and 8-player-mode stage updates.

 

[CarbuncleHero]

Think about it. Every move startup for every character you've tested is consistently half of what other and more reliable sources have reported. You're telling us that startup has halved across the board for two characters and yet no other source confirms that and not a single player has noticed any difference in speed just from playing. You're basically telling use that the game has changed DRASTICALLY and you're the only one to notice. While it seems a lot more likely that simply your testing methods are incorrect.

Fair enough. but if I am wrong then shouldn't this be disproven fairly quickly. I am saying that 1.04 and 1.05 are different. The same method was used for both. Just saying it wrong won't solve anything? Saying 1.04 has this for pikachu's frames therefore this means that 1.05 is the same isn't credible. Nor have I claimed to be. I tested this and relayed my results. This is what I found and no one has refuted it (without just saying it's flat-out wrong that is).

i just booted up my wiiu and went into training mode with the 1/4 speed thing, and i can see why you might still think that stuff is so fast. you probably can't hit the l button for literally a single frame so that only one frame advances in the game. try setting the cpu to control, have it grab a timer, and then try again. you'll notice that you can get pikachu to be fully extended with the jab pose but the other player does not get hit until a little bit after.

i hope that helps a bit, because i'm done arguing this with you lol. i would be surprised if any frame data got changed, because as far as anyone knows it was just some bug fixes and 8-player-mode stage updates.

You used the method, said "this is wrong, here's a better way" and changed it, then came back to tell me I was wrong? This is a little frustrating, how am I supposed to take that seriously? Unless something has changed, using 1/4 hold L mode is a known way to get 1 frame. What makes it that the so the timer gives precisely one frame? I don't care to argue this any further myself. I don't know what would even convince you anyway.

 

[dettadeus]

ok so since you are very obviously not getting this, i made some videos for you

pika:

mario:

their jabs are not frame 1. your method is inaccurate and your data is wrong lol. end of story, just wait for dantarion to get the updated 1.0.5 frame data and chances are it will be identical to the 1.0.4 frame data.

 

[CarbuncleHero]

ok so since you are very obviously not getting this, i made some videos for you

pika:

mario:

their jabs are not frame 1. your method is inaccurate and your data is wrong lol. end of story, just wait for dantarion to get the updated 1.0.5 frame data and chances are it will be identical to the 1.0.4 frame data.

A-Are you pausing and unpausing the game to get the frame data? How do I even begin to start? 1 frame is 1/60th of a second, you know that right? Try that again with Fmash.
I'm sorry but I don't even know what to say about this? I am legitimately worried that that last post was a crude joke and I missed the punchline.

 

[Hoenn]

A-Are you pausing and unpausing the game to get the frame data? How do I even begin to start? 1 frame is 1/60th of a second, you know that right? Try that again with Fmash.
I'm sorry but I don't even know what to say about this? I am legitimately worried that that last post was a crude joke and I missed the punchline.

If a move comes out on frame 1, then it comes out on the first frame of animation
If you are able to pause before a move comes out (Like Detta did), then it is not frame 1

He even unpaused the game to prove that the move hit, What Detta posted was absolutely legitimate

 

[CarbuncleHero]

If a move comes out on frame 1, then it comes out on the first frame of animation
If you are able to pause before a move comes out (Like Detta did), then it is not frame 1

He even unpaused the game to prove that the move hit, What Detta posted was absolutely legitimate

Ahh true. I see. Well, this is in fact proof.
I'll go see what I went wrong, but i'll edit the post i the meantime. However, I got the same results each time I tried, and this method was being used by other people. ??? I don't know what the problem is.

 

[dettadeus]

nobody is using that method, lol. people have the frame data ripped straight from the game. it doesn't get any more accurate than that. as TL? said above, there is literally no way that multiple characters could have had half their startup lag shaved off and nobody notice it, ESPECIALLY something like Mario's fsmash - a 7 frame reduction (just over 1/10 of a second) is a LOT more noticeable than a 1 frame reduction.

the most likely reason your method would be incorrect is that you aren't pressing and releasing the l button fast enough to get only one frame of progression in the game; you would have to press L and then release it within four real frames - 1/15th of a second. it probably takes more time than that for the trigger to reset to neutral because of the spring, so it's entirely possible that you can't get single frame progression in 1/4 speed training mode without further slowdown from something like a timer.

 

[Angiance]

Even though this frame data is extremely inaccurate, your efforts are still admirable. If you can find a way to -accurately- gather frame data, it'll be beautiful

 

[Hoenn]

The problem is that it is VERY VERY hard to consistently release the L button for a specific amount of frames (1 frame in your case). You can easily mess that up and skip frames. I personally use pause buffering the way Mew2King does it when I want to test frame data.

Like, if I want to check if a move is frame 3, I use pause buffering to see if the move is slower than a 2 frame move. Then I take a 4 frame move and see if it is faster than the 4 frame move.

To pause buffer you need 3 controllers though
here are the steps.
1. Pause the game with a designated pausing controller.
2. Using the other 2 controllers, hold the inputs for the moves that you want to come out at the same time. (If you want to jab hold A)
3. While holding the inputs, unpause the game, both moves will come out on the first frame.
4. Which ever move hits is faster

 

[CarbuncleHero]

nobody is using that method, lol. people have the frame data ripped straight from the game. it doesn't get any more accurate than that. as TL? said above, there is literally no way that multiple characters could have had half their startup lag shaved off and nobody notice it, ESPECIALLY something like Mario's fsmash - a 7 frame reduction (just over 1/10 of a second) is a LOT more noticeable than a 1 frame reduction.

the most likely reason your method would be incorrect is that you aren't pressing and releasing the l button fast enough to get only one frame of progression in the game; you would have to press L and then release it within four real frames - 1/15th of a second. it probably takes more time than that for the trigger to reset to neutral because of the spring, so it's entirely possible that you can't get single frame progression in 1/4 speed training mode without further slowdown from something like a timer.

I see that now.
and I hate to argue further but where else would I have learned to do this? Other people were using it and I noticed it worked so I used it myself. I just reverted back to 1.0 and everything is fine and I can get the data frame by frame with no problem, as I could in 1.04. (i don't even use the gamecube controller) For some reason it doesn't work in 1.05. I had no reason to believe otherwise, and you are the only person that gave any proof against it. But you are right, thanks for clearing this up.