Smash Ultimate Frame Data (E3 Demo)

Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#1
Hi everybody, I was willing to bet money they'd announce a global test smash demo at E3 and the dataminers would get to the juicy frame data sooner than anybody else. That didn't happen, but it appears the E3 demo is making an appearance at more events, such as the RAGE tournament in Japan, and CEO 2018. With such an abundance of footage, I decided I ought to work on counting all the frame data of the playable characters so that people are more informed about how specifically game speed has been altered. It's also a game very similar to Smash 4, so I can understand what I'm looking at. It's not extensive, if a move isn't used in a particular way, then I can't get accurate data, and I refuse to make educated guesses despite having done this sort of work in the past. I also left sources for all of my counts if anybody wants to proofread me

It should be noted that any total frames or landing lag values discovered could potentially be lower, since I can't technically know when a move ends unless my hands were on the controller buffering something or holding shield. Also please understand that this is a demo of the game and nothing we've seen and counted is necessarily permanent. I'd rather not have people in here freaking out about buffs, nerfs, or the lack of either. There are character subforums for that. What I am willing to do is answer questions.

  • How did you count frames? On youtube videos you can advance frame by frame using the comma and period keys. Not all footage is 60FPS though. Check the settings to make sure that there is the option for 720p60 or 1080p60 and select it. If there's no 60, the video is not in 60 FPS, and any counts would be inaccurate.
    [*]Where can I find Smash 4's frame data to compare? You can use the charts I helped write, or Kurogane Hammer.
    [*]What's the difference between Total Frames and FAF? Total Frames is the total duration of an attack. FAF (First Actionable Frame) is the first frame you can act after that attack. So in short: Total Frames = FAF - 1.
    [*]Does EVERY aerial have less landing lag? Pretty much. The two exceptions I can confirm are Bayonetta's N-air and B-air landing lag being unchanged. But I think everything else is reduced that I've discovered.
    [*]What's the landing lag for air dodges? 10 frames for the Brawl air dodge and 15 for the directional air dodge. Much less than Smash 4, but I guess they gotta keep up with so many attacks being this fast.
    [*]I see this attack all the time, why don't you have its startup? I need a hit on the move's first active hit frame. Otherwise any data I put up would just be misinformation about a move that isn't any slower. Like Cloud's Up-Air for instance. I know its startup is unchanged, but I'm still looking for that first hit frame that nobody will ever land unless they mispace the move hilariously.
    [*]I see this Smash Attack used all the time, why don't you have its startup? Smash attacks often have very early charge windows and I don't like having to guess around them for the proper startup. Even professional players charge their smashes unintentionally for a frame or two when the charge window is as early as frame 3. So for smash attacks, I only counted the ones that weren't charged at all.
    [*]Why doesn't Samus, Pac Man, Mega Man, etc have more data? Because nobody is playing them. I did the best I could hunting through free for all matches with the camera zoomed out. Plus some moves are just never used, or never used properly.
    [*]Can I help add data to these charts? Sure! I added notes on the first sheet about this game's odd visual quirks that interfere with accurate counts. If you count something I don't have, be sure to give me a timestamp. I will not add anything to my charts unless I can personally confirm for myself. Also, for returning characters, here's a resource for what their moves used to look like. It helps figure out what the first frame of an animation such as dashing.
    [*]Will you do this again when the game releases? All 68+characters? Boy that sounds like a fun vacation, but yeah probably. I won't have to if somebody manages to datamine the game's code, but I'll certainly be curious enough to do this once I have the game in my hands. Plus doing my own recordings and counting them is several times faster then hunting for moves I haven't seen yet across hours of footage.
Hopefully I'm not wrong about stuff. I'd hate to be the source of any misinformation so if you have a correction or want a recount on something, I'd be willing to do that.
 

Ffamran

The Ace Tank Driver?
Moderator
Joined
Aug 25, 2014
Messages
14,333
#2
For Zelda, I got different numbers for her landing lag. Dair's the closest at 13 frames while Nair, Fair, and Bair were 2 frames slower. That said, counting frames at times is tedious and sometimes you can't tell exactly if the person did their next action right after. In other words, I am probably wrong with the numbers I got for her and other characters I looked at. We still get an idea of what the game's and some of the characters are like, though. To fill in some data for Zelda, Nair is frame 6 like in Smash 4 and you can see this in the Smash Ultimate Direct video where she's shown doing a U-throw to Nair on Meta Knight and Uair's landing lag was 14 by my count.

My observations for Zelda are a bit of a mess since they're scattered in different posts on their general discussion and social thread. I started with this post that checked her standing grab, Nair startup, and jab startup: https://smashboards.com/threads/zelda-social-dins-fire-is-free-fall-free.453509/page-6#post-22120100.

Jab, Dtilt, dash attack, Fair, and Bair startup along with unsure checks for Side Smash, dash grab, pivot grab, Nayru's Love, and Farore's Wind startup was this next post: https://smashboards.com/threads/zelda-social-dins-fire-is-free-fall-free.453509/page-6#post-22120380. Video link is in the quoted posted.

Fair sharing Bair's frame 6 startup was reconfirmed here: https://smashboards.com/threads/zel...s-free-fall-free.453509/page-11#post-22134425, and here: https://smashboards.com/threads/zel...s-free-fall-free.453509/page-11#post-22134523. The second post also goes into the start of landing lag checks for Zelda's aerials, Nair, Uair, Fair, and Bair, and an unsure check and guess of Ftilt's startup. I was unsure of Fair and Bair's landing lag in this post that will later be checked again in a later post.

This talk about relative landing lag reductions has landing lag data of Fox, Ike, Bowser, and PAC-MAN, pulled from threads on Smashboards: https://smashboards.com/threads/zel...s-free-fall-free.453509/page-16#post-22168157. It was edited to reflect the finalized landing lag checks I did for Zelda in the linked post below.

Landing lag was finalized by this post: https://smashboards.com/threads/zel...s-free-fall-free.453509/page-16#post-22168998.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#3
Jab, Dtilt, dash attack, Fair, and Bair startup along with unsure checks for Side Smash, dash grab, pivot grab, Nayru's Love, and Farore's Wind startup was this next post: https://smashboards.com/threads/zelda-social-dins-fire-is-free-fall-free.453509/page-6#post-22120380. Video link is in the quoted posted.
Regarding Pivot Grab, those are tricky to count. When I was Smash 4 testing, my theory was that there were a few frames of forced turnaround animation not accounted for in the game's code as part of the move's startup. It's unfortunate that we lack hitbox visualizations for her in Smash 4 as far as I can tell. And I dumped my footage of her from testing as well, so I can't provide an immediate answer for this. A lot of grab startups have been shifted about in the game so we can't count out the possibility that pivot grabs have been altered just as often.

Good links! I'll search through them for the counts other people have done.
 

Ffamran

The Ace Tank Driver?
Moderator
Joined
Aug 25, 2014
Messages
14,333
#4
Regarding Pivot Grab, those are tricky to count. When I was Smash 4 testing, my theory was that there were a few frames of forced turnaround animation not accounted for in the game's code as part of the move's startup. It's unfortunate that we lack hitbox visualizations for her in Smash 4 as far as I can tell. And I dumped my footage of her from testing as well, so I can't provide an immediate answer for this. A lot of grab startups have been shifted about in the game so we can't count out the possibility that pivot grabs have been altered just as often.

Good links! I'll search through them for the counts other people have done.
Yeah, that's something I'm unsure about as well and made some of the moves like the time I counted her Side Smash in that post was when she did it from a turnaround. We did have a hitbox visualization thread for her in Smash 4: https://smashboards.com/threads/zelda-hitbox-visualization.434820/, and everyone else for that matter, but I think the gifs don't work anymore and I'm not sure why.

The SSB wiki does have pages showing hitboxes, though. Here's Smash 4 Zelda's grabs: https://www.ssbwiki.com/Zelda_(SSB4)/Grab. Her pivot grab looks like this.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#5
Yeah, that's something I'm unsure about as well and made some of the moves like the time I counted her Side Smash in that post was when she did it from a turnaround. We did have a hitbox visualization thread for her in Smash 4: https://smashboards.com/threads/zelda-hitbox-visualization.434820/, and everyone else for that matter, but I think the gifs don't work anymore and I'm not sure why.

The SSB wiki does have pages showing hitboxes, though. Here's Smash 4 Zelda's grabs: https://www.ssbwiki.com/Zelda_(SSB4)/Grab. Her pivot grab looks like this.
Well the pivot grab in the video looks correct, so I added it. By the way, the N-air startup showed me something I missed. The first few minutes of the Smash Presentation video had gameplay recorded in 30FPS, despite the video itself being 60FPS. When I saw that I assumed the whole thing would be untestable. However, the gameplay starts being 60FPS once they start showcasing individual fighters. Giving me some more video to work with.

CEO has a bunch of juicy footage so far too. Unfortunately almost nobody appears to have a 60FPS recording of these matches. Bummer.
 

Ffamran

The Ace Tank Driver?
Moderator
Joined
Aug 25, 2014
Messages
14,333
#6
Well the pivot grab in the video looks correct, so I added it. By the way, the N-air startup showed me something I missed. The first few minutes of the Smash Presentation video had gameplay recorded in 30FPS, despite the video itself being 60FPS. When I saw that I assumed the whole thing would be untestable. However, the gameplay starts being 60FPS once they start showcasing individual fighters. Giving me some more video to work with.

CEO has a bunch of juicy footage so far too. Unfortunately almost nobody appears to have a 60FPS recording of these matches. Bummer.
Really? I didn't notice that since I only went to the sections showcasing the individual characters and onward. That said, I do remember going back yesterday when they showed everyone and the part with Fox and Falco looked like Falco's jab seemed like it skipped frames or the frames didn't move. I thought it was my laptop acting up or the recording having a hiccup there.

For CEO, if Nairo recorded anything, then he'd be a source for off-screen 60 FPS videos. MVG's videos are weird. The videos are on-screen and 60 FPS, but for some reason the footage is really messed up. I remember trying to check Zelda's Dair and I got numbers ranging from 18 to 23 which makes no sense. 18 frames is the same as Smash 4's and close to her landing lag for Dair in Melee and Brawl, but 23 doesn't work when the developers said they're trying to lower landing lag and speed up characters. I had 18 as a placeholder since it's close to what she had and it could have been unchanged like Falco's Dair landing lag is unchanged (for some dumb reason) until I looked at a different video and got 13 frames for it.

I looked at your section for Fox and I did check him a while back. It would fill in ground Blaster startup, rate of fire, and recovery, Fire Fox's launch, and Reflector's recovery frames. I referenced Smash 4 Fox's frame data for Blaster and Fire Fox and the startup looks similar. I reasoned Blaster's startup as when Fox experiences recoil. There's an explanation on it in my post: https://smashboards.com/threads/wil...itch-speculation-thread.453400/#post-22124771.

This post has landing lag from Skarfelt Skarfelt , FAF for Fox's Utilt, Dtilt, and Reflector: https://smashboards.com/threads/wil...peculation-thread.453400/page-2#post-22131471.

Fox's rapid jab finisher recovery guess: https://smashboards.com/threads/wil...peculation-thread.453400/page-2#post-22133387.

I think Fox's Up Smash and Side Smash were found by D DT Raw : https://smashboards.com/threads/wil...peculation-thread.453400/page-2#post-22133995.

I also did some checks on Snake and one thing that is missing is Snake's C4 plant on people and the ground. You have the same frame count for Snake's C4 plant on people, but don't have his C4 plant on the ground which I got 22 frames for. Snake's C4 detonation is weird in that I don't know if it's him pressing the button which I counted as frame 20 is the startup or if it's the first active frame for C4 explosion, frame 25. I would assume him pressing the button means the C4 is going to blow up regardless since it's activated by that time, but that might not be the case. My total frames for him were higher as well at around 47 to 51 frames for planting it on people and 49 to 50 when planting it on the ground. Other things would be that dash attack is at least active to frame 10, has around 55 total frames, and I got 42 startup frames for Side Smash instead of 41. Post link: https://smashboards.com/threads/snakes-virtual-reality-training-center.455533/#post-22123585.

It should be noted that these were from when I started doing frame counts, so I probably messed up somewhere. Jump frames being the biggest one since I thought the first airborne frame counted, but it doesn't and it's everything before that which is how every Smash game, KoF XIII, KoF XIV, and Street Fighter V did it. Anyway, Zelda's finalized landing lag would be the most recent and Falco's which I didn't post here since he's not playable and Zelda's initial posts, were the first ones I did.

Edit: The Ike boards' E3 demo build frame data thread: https://smashboards.com/threads/e3-build-information-changes-and-damage-values-ike.455509/. It would add in information for Down Smash, Dair, Aether, and Quick Draw.

Marth boards' E3 demo build frame data thread: https://smashboards.com/threads/pre-release-frame-data-–-video-analysis.455434/. Up Smash, Side Smash, Fair and Counter. Different numbers for dash attack, Dolphin Slash: https://smashboards.com/threads/pre-release-frame-data-–-video-analysis.455434/#post-22125444.

Edit 2: Edit Harder: With CEO, Abadango uploaded two videos of him using Inkling and Zaki using Ridley that is 60 FPS. Problem is I don't know what their moves are exactly, but the videos are available at 60 FPS. First video: https://www.youtube.com/watch?v=QrBfeyTDpFo. Second video: https://www.youtube.com/watch?v=hB5K3JhdmCc.
 
Last edited:
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#7
For CEO, if Nairo recorded anything, then he'd be a source for off-screen 60 FPS videos.
Boy I hope his camera made it to that event. He had a lot of footage from E3. I'll take any good source. Even if it's a free for all between casual players. See casual players tend to button mash, and that means buffered inputs. They also tend to use underutilized moves which fill out a lot of my blank spaces. Also thanks again for the links. I've been adding anything I can find a source for.

So I haven't done frame data on any throws thus far, because I was intimidated by the new hitlag. In Smash 4 it's easy to tell since damage displays on the first frame of hitlag. Not so in Smash Ultimate. However, I just remembered that the final hit of any throw has (gasp!) no hitlag. So damage does indeed display the very frame when the victim is launched. That should allow me to get endlag counts on throws. The only problem for comparison are weight based throws. We don't know weight values in this new game or whether the throws that were weight based still are or if the reverse is true.

Edit: Okay, it doesn't look like damage occurs on the frame they're launched, so we still have a problem for comparing to Smash 4. Still going to have something written down for these moves though.
 
Last edited:

Thinkaman

Moderator
Moderator
Joined
Aug 26, 2007
Messages
5,904
Location
Madison, WI
3DS FC
1504-5749-3616
NNID
Thinkaman
#8
First, great job.

Much of this is likely to change but we're all adults who know that. Still a great quantitative look into the current direction.

As far as broad stroke implications, I actually worry about this low landing lag. I dislike any simple aerial attack being safe-on-shield regardless of spacing, so I generally preferred Smash 4's slightly higher values to Melee. (Though I both liked Smash 4's broad AC windows and would have still have reduced quite a few specific landing lag values by a good chunk, like Ganon.) That said, Smash 4 and Melee's (l-canceled) values are really close for most characters, and it's unwise to fixate on landing lag vs. other more significant differences in the games.

The cost of too-low landing lag was evident if you ever messed with Smooth Landing, which cut all landing lag by a third and put it well below Melee. It introduced a lot of dumb and repetitive true combos, and made people like Sheik and Ryu very carefree on hitting shields. It wasn't as silly as Smash 64, but it was still a can of worms.

From what we've seen and confirmed in this data, Smash Ulti's landing values are universally less than Melee's by a nontrivial amount, and close to Smooth Lander values. Uh oh!

But 2 things put my mind at ease:

First, looking at the data, several key moves (especially certain landing dairs, and Ryu's moves) have a much smaller landing lag reduction than most. Given that these moves were the problem with Smooth Landing, I think it's safe to say that we shouldn't expect those sort of issues.

Second, the big new game system that interacts with shield safety (besides 3f jumpsquats making OoS aerials more viable) is the new shield parry. This new system is harder to do by accident, but unlike old powershielding can be accessed even if your shield is already up. This, in theory, is a more reliable and safe tool against braindead shield pressure aerials. If this is correct, then it matters less is a Sheik fair or Ryu nair is otherwise safe on timing alone.

Spot dodging (and rolling) will be less viable against aerials regardless, but defensive pivots should remain just as good. (And easier to execute) The real unknown remains the long-term impact of this new shield.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#9
But 2 things put my mind at ease:

First, looking at the data, several key moves (especially certain landing dairs, and Ryu's moves) have a much smaller landing lag reduction than most. Given that these moves were the problem with Smooth Landing, I think it's safe to say that we shouldn't expect those sort of issues.

Second, the big new game system that interacts with shield safety (besides 3f jumpsquats making OoS aerials more viable) is the new shield parry. This new system is harder to do by accident, but unlike old powershielding can be accessed even if your shield is already up. This, in theory, is a more reliable and safe tool against braindead shield pressure aerials. If this is correct, then it matters less is a Sheik fair or Ryu nair is otherwise safe on timing alone.

Spot dodging (and rolling) will be less viable against aerials regardless, but defensive pivots should remain just as good. (And easier to execute) The real unknown remains the long-term impact of this new shield.
I also feel that the perfect shield mechanics have to pull a lot of weight when it comes to this game's prospective meta. If they're too easy to perform and basically make any multihit move unviable then I think we'll be back in a dull meta where people just shield more often than they attack. But if they're not easy enough to perform then characters with the best frame data and move speed will continue to dominate. Putting your shield down for a punish is a lot more risk than putting it up in the first place so I feel like this mechanic shows a lot of promise. Give heavyweights just a few more frames advantage on their punish and they'll put out as much damage in one tilt as a short Sheik combo.

I did find a good example of the Perfect Shield in action. Sheik perfect shields a Kirby Fsmash. Kirby is stuck in shieldlag longer than Sheik and his move resumes for just 4 frames before Sheik can act for a punish. Kirby's Fsmash frame data is unchanged so Sheik has earned 31 frames of advantage on Kirby with no 7 frame penalty for dropping shield first. I have to assume that the attacker will suffer more or less shield lag depending on how much damage their attack does. That's probably what nintendo Treehouse meant when they said parrying a heavier move gives you even more time for a punish. Kirby's Fsmash has been staled enough to do a little over 12% in that example.
 
Last edited:

Thinkaman

Moderator
Moderator
Joined
Aug 26, 2007
Messages
5,904
Location
Madison, WI
3DS FC
1504-5749-3616
NNID
Thinkaman
#10
If they're too easy to perform and basically make any multihit move unviable then I think we'll be back in a dull meta where people just shield more often than they attack.
Informed reports suggest that you cannot parry in shield stun; Sol reports a test finding that Mac's f-tilt2 could not be parried after blocking f-tilt1. (At best, you could parry only f-tilt 1 and try to punish)

Multihit moves appear to be natural anti-shield options, as they should be.
 
Last edited:

san.

1/Sympathy = Divide By Zero
Moderator
Joined
Oct 26, 2008
Messages
5,621
Location
Rochester, NY
3DS FC
4957-2846-2924
NNID
Sansoldier
#11
Nice finding in Ike's jab2 being 2 frames instead of 3. (I was tricked by the damage increase being 1 frame later)

Side B, same timestamp. I counted the FAF as 35, down from 45. It was tough to count so I can be off (another source lists it as 34).

I think his nair landing lag is 8 instead of 7. Is there a case where it looks like it is 7?
https://www.youtube.com/watch?v=fe4Nb9nGLJE&feature=youtu.be&t=1m45s


Also, any data on shield stun at all?
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#12
Nice finding in Ike's jab2 being 2 frames instead of 3. (I was tricked by the damage increase being 1 frame later)
I really hope they fix this in the final build. I've pretty much given up on trying to count anything past the first hit of a multihit move. And it just makes sense for damage to update on the first hitlag frame.

Side B, same timestamp. I counted the FAF as 35, down from 45. It was tough to count so I can be off (another source lists it as 34).
Good find, I missed that one.

I think his nair landing lag is 8 instead of 7. Is there a case where it looks like it is 7?
https://www.youtube.com/watch?v=fe4Nb9nGLJE&feature=youtu.be&t=1m45s
My timestamp has 8 too. I must have miscounted.
Also, any data on shield stun at all?
Can't help you at the moment on this. I've just been focusing on the moves themselves. Good thing I haven't tried though as I only found out yesterday that the presence of items shuts off the 1v1 damage modifier. I'll try to keep my eye on how shields behave and jot down some numbers. If you're taking a crack at it, remember that shields have a visual bug in the demo where they are still there on frame 1 of shield drop, shield jump, shield grab, etc.

Informed reports suggest that you cannot parry in shield stun; Sol reports a test finding that Mac's f-tilt2 could not be parried after blocking f-tilt1. (At best, you could parry only f-tilt 1 and try to punish)

Multihit moves appear to be natural anti-shield options, as they should be.
Well I hope that's sufficient. Odd example though if true. Mac's F-tilt punches are 8 frames apart. You could drop shield in between. If I were designing this perfect shield mechanic, I would make it impossible to parry additional hits from the same move after you normal blocked one of them.
 
Last edited:

Thinkaman

Moderator
Moderator
Joined
Aug 26, 2007
Messages
5,904
Location
Madison, WI
3DS FC
1504-5749-3616
NNID
Thinkaman
#13
Good thing I haven't tried though as I only found out yesterday that the presence of items shuts off the 1v1 damage modifier.
Key find btw. I missed this too.

Well I hope that's sufficient. Odd example though if true. Mac's F-tilt punches are 8 frames apart. You could drop shield in between. If I were designing this perfect shield mechanic, I would make it impossible to parry additional hits from the same move after you normal blocked one of them.
There are general reports of shield stun being longer, though I haven't heard anyone nail anything down yet. Just lots of non-quantitative anecdotes.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#14
Added data for the non playable characters in a separate sheet. I know this is gratuitous since the 38 non-playable characters are probably the least finished. However, a lot of what they show clearly has a focus on what has changed and I doubt we'll get more footage of them in the near future.

Informed reports suggest that you cannot parry in shield stun; Sol reports a test finding that Mac's f-tilt2 could not be parried after blocking f-tilt1. (At best, you could parry only f-tilt 1 and try to punish)

Multihit moves appear to be natural anti-shield options, as they should be.
Well I just seen a perfect shield on the fifth hit of Fox D-air. That move is not a true blockstring in Smash 4 but it's still a problematic example of how you could perfect shield whichever hit you want on a multihit move.
 
Last edited:

Thinkaman

Moderator
Moderator
Joined
Aug 26, 2007
Messages
5,904
Location
Madison, WI
3DS FC
1504-5749-3616
NNID
Thinkaman
#15
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#16
So I've been looking at blocked moves in order to pinpoint a shieldstun formula. But I fear such work may be useless because I don't see shields operating with any sort of consistency like in Smash 4. Here are my prevailing theories at the moment. One of these may serve as an answer to why I've been getting so many different results in the same match.

  1. - This build has no shieldstun.
  2. - Alternatively, the build may have shieldstun, but it fails to stop the player from performing OoS options such as a shield jump during shieldstun
  3. - The Attacker and Defender suffer disparate amounts of shieldlag. This would be the third game in a row to have this issue, but the first time it could be deliberate. With the attacker suffering more shieldlag arbitrarily in order to make moves more punishable. Perhaps only when it comes to aerials?

My most absurd examples that appear to support all three answers as potentially right:
- Ike's Fresh N-Air, 9.4% damage. 11 shieldlag, 1 shieldstun. This is my support for Theory #1, since if there's just 1 shieldstun then I'm willing to bet the concept of shieldstun no longer exists. and decades of blocking moves has left our players simply waiting in their shields because they expect it to be there. Note that Zelda shield jumps rather than shield drops, which supports theory #2 as well. This one is interesting. As an expert on shields the last game I can tell you there were two ways to tell when shieldlag ends and shieldstun begins. The first is when the attacker's move continues with its animation. The second is when the defender starts suffering shield push. Shield push is almost non existant in this build, but it appears Zelda starts getting pushed at frame 8 of shieldlag. If we reinterpret it this way, then we have 7 shieldlag and 5 shieldstun. Which while both numbers are less than what they'd be in Smash 4, at least sounds like a reasonable block state and supports Theory #3
- Zelda's Fresh F-Air 21.4% damage 17 shieldlag, 6 shieldstun. (I seen a fresh B-Air do that amount and I assume it's still the same as F-Air). Theory #3 is the only way I can explain this game's shieldstun in a way that isn't "It's broken and doesn't work". And it may potentially hold up despite this example. Here's the thought process. We don't see Ike get pushed by the attack until the frame shieldlag ends for Zelda. So shieldlag is the same between targets yes? Consider it this way. Zelda's F-Air has a hitlag modifier, and a lot of damage besides. What if you could hit a shield hard enough that the attacker overcomes their shieldlag penalty? The only thing left that's weird is the idea that Zelda's massive damage attack only dealt 6 shieldstun. Again, note that Ike doesn't drop shield but shield jumps, so that's why this supports Theory #2 as well.



Edit: shoot I was far from done with this post and hit the enter key while trying to format.
 
Last edited:

Thinkaman

Moderator
Moderator
Joined
Aug 26, 2007
Messages
5,904
Location
Madison, WI
3DS FC
1504-5749-3616
NNID
Thinkaman
#17
So I've been looking at blocked moves in order to pinpoint a shieldstun formula. But I fear such work may be useless because I don't see shields operating with any sort of consistency like in Smash 4. Here are my prevailing theories at the moment. One of these may serve as an answer to why I've been getting so many different results in the same match.

  1. - This build has no shieldstun.
  2. - Alternatively, the build may have shieldstun, but it fails to stop the player from performing OoS options such as a shield jump during shieldstun
  3. - The Attacker and Defender suffer disparate amounts of shieldlag. This would be the third game in a row to have this issue, but the first time it could be deliberate. With the attacker suffering more shieldlag arbitrarily in order to make moves more punishable. Perhaps only when it comes to aerials
#1 simply does not square with reports of increased shield stun, like M2K complaining he "couldn't do anything" following Zero's Ivy bairs. It's too big a pill to swallow that every top player who has played the game has simply obeyed placebo shield stun, particularly in a game with significant buffering.

#2 seems the most likely way of explaining the two linked clips. "You can jump out of shield stun." This does strike me as rather polarizing towards OoS aerial punishes, especially low-hitting SH-fairs.

#3 is worth considering. This would be, at least in the second and especially first linked examples, an extreme case though.

Option #4 is "Attacks have individual specified shield stun values or modifiers." However, would Ike's nair really be set to 1 frame, and Zelda fair to 6? Very dubious.

Option #5 is a bug or otherwise somewhat-incomplete implementation, causing some degree of inconsistent behavior.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#18
#1 simply does not square with reports of increased shield stun, like M2K complaining he "couldn't do anything" following Zero's Ivy bairs. It's too big a pill to swallow that every top player who has played the game has simply obeyed placebo shield stun, particularly in a game with significant buffering.
No offense to M2K but Johns are not evidence and I've never known top players to be bastions of knowledge regarding game mechanics even when the game isn't new. People see what they want to see and parrot everything they hear. Furthermore, optimally punishing a move you blocked isn't an easy task. When a move hits your shield and has you sit there for a total of ten frames, that's barely enough time for your eyes to confirm you blocked a move, much less the 9 or so frames of input delay added on to your response. People aren't robots, I seriously wouldn't expect any shield activity to be frame perfect when people are playing a genuine match where they impulsively guess with their shields. And matches are all the data I have to work with. Finding that N-Air example was straight up luck, not the norm.

Now if I had this game in my hands with my equipment, it would be a different story. Ten minutes of actual tests would be all I need. What I can say is that, assuming the shields in this demo work as intended, which I'm not convinced of, shieldstun has definitely not increased nor matched the levels of Smash 4. Not even close. You'd think it would be if the formula hadn't changed and everybody was dealing more damage in 1v1s. However In the same match, here's Zelda entering shield drop after 5 frames of stun from Ike's fresh B-Air (17.64%). Would have been 12 shieldstun in the previous game's formula, and there's no shield jump this time. Not a single move I've looked at matches Smash 4's level of shieldstun while looking like the defender was responding appropriately. I thought I found one with this Ike B-AIR doing 12 shieldstun, but that previous link a few sentences ago refutes it.

Maybe this game's shieldstun formula has more factors. Like, say, giving less advantage to aerials because those would be so great for pressure. But tentatively I'm putting it in the same lot as footstools and jab locks. Mechanics that ought to be working but just don't yet. And if they don't fix the visual bugs concerning hitlag and shields by the time the game releases I'm going to scream.
 

Ffamran

The Ace Tank Driver?
Moderator
Joined
Aug 25, 2014
Messages
14,333
#19
Zapp Branniglenn Zapp Branniglenn , just noticed you started adding in the nonplayable characters. Here's my numbers for Falco and Wii Fit Trainer for comparison.

I started with Falco on this post: https://smashboards.com/threads/ultimate-falco-gameplay-general-discussion.455401/#post-22208126. The references were the Ultimate Direct video and Falco's trailer both of which are linked in the post. You could possibly get his Down Smash data as well, but I'm not sure how to factor in charging frames which someone in the Fox boards apparently was able to figure out.

The most recent one would be this post: https://smashboards.com/threads/ultimate-falco-gameplay-general-discussion.455401/post-22178839. It's primarily focused on a Reddit thread that stated you can see the hitboxes for Falco's Blaster and Falco Phantasm in his trailer when he's fighting Cloud. That narrowed down his Blaster and Falco Phantasm. For you section on him, it should add in jab 1's startup, Fair's startup, Dair's startup and landing lag, Blaster's estimated startup, and Falco Phantasm's startup and active frames.

The one difference I have is that I counted Falco's dash attack as being frame 7 instead of 8. Also, I think the developers may have added in a penalty for Falco if he continues to fire Blaster similar to Bayonetta's increased landing lag if she combos with Witch Twist and/or Afterburner Kick and correct me of I'm wrong, lands with Bullet Arts extended aerials. When Falco uses Blaster in the Ultimate Direct, he only fires once and I estimated him as having 31 recovery frames since he returns to idle and doesn't do anything for a few frames. In his trailer, Falco fires twice and immediately uses Falco Phantasm and it takes him 33 recovery frames. Also, you can see jab 1's startup. It's frame 2 in every game including Melee where his jab 1 was Fox's. The video just cuts to the first startup frame and Wolf gets hit on frame 2.

For WFT, here's the post: https://smashboards.com/threads/sur...it-trainer-thread.453425/page-2#post-22176024. It would only add in jab 1's startup, Ftilt's startup and active frames, Nair's first hit startup, Uair's startup, dash attack's startup, Sun Salutation, and an educated guess on her standing grab having the same frame 6 startup in Smash 4. I did try to guess the total frames on WFT's moves that hit someone or something in the trailer in a weird way that is explained in the post, so ignore those numbers if you think it was an improper way of getting those numbers.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#20
Zapp Branniglenn Zapp Branniglenn

The most recent one would be this post: https://smashboards.com/threads/ultimate-falco-gameplay-general-discussion.455401/post-22178839. It's primarily focused on a Reddit thread that stated you can see the hitboxes for Falco's Blaster and Falco Phantasm in his trailer when he's fighting Cloud. That narrowed down his Blaster and Falco Phantasm. For you section on him, it should add in jab 1's startup, Fair's startup, Dair's startup and landing lag, Blaster's estimated startup, and Falco Phantasm's startup and active frames.
Whoa! rare footage of a debug menu. That's so cool. Wish developers still left these things in their games. And yeah if we assume those hitbubbles are accurate then that would mean Falco's Blaster and Phantasm may be hitting as early as Frame 9 and 18 respectively. It's not the sort of definitive evidence I like, but I'll make a note of it at least. As for Jab 1, the animation has been touched up compared to his Smash 4 version, and I can't be certain frame 1 of that jab is the precise frame 1 we're cutting to in his trailer. I don't doubt that it's probably still frame 2 but I only count definitive evidence for my charts. For Dash Attack, that's definitely an 8. The animation is unchanged from Smash 4 where you'll notice both of Falco's feet are off the ground on Frame 4. And Dsmash, F-Air, and D-Air give me no data to work with based on how they're used.

As for WFT, the Jab 1 is a frame 5 hit in the video, but I suspect it's a late hit. WFT has famously terrible hitboxes and I suspect Villager's Fsmash had only then placed him close enough to it. I think proper testing will find it's still frame 4 since villager is a short target. Similar situation with N-Air, that's a frame 11 hit in that video when it could still be starting on frame 9. F-Tilt is another late hit but you did remind me to make a note that it has more active hit frames which is the only change we can definitively gleam from that video. U-Air is another late hit on frame 7. Dash attack also appears to be a frame 8 late hit in that shot. Her Grab is in progress so I can't count it based on the same criteria of other moves that we cut to. And Sun Salutation is again a scenario where I need a clear first active hit frame.
 

Ffamran

The Ace Tank Driver?
Moderator
Joined
Aug 25, 2014
Messages
14,333
#21
Whoa! rare footage of a debug menu. That's so cool. Wish developers still left these things in their games.
Some people have optimistically speculated that the release build will feature the ability to turn on hitboxes and kind of frame data in training mode; there was the coincidence of Falco Phantasm doing 7% on the ground and on grounded targets and the hitbox having a white 7 and angle marker. The main reason, according to self-claimed developers, was that hitboxes usually don't look that pretty and that developers wouldn't spend time making something look that clean and detailed without a reason since they're trying to get the game going and fast. While that is great and all, part of me is skeptical the developers would do this and I'd rather have a record feature in training mode first over the ability to see hitboxes. Part of that comes from people can datamine games and get that information while a record feature is something we can't put in easily if at all. Anyway, I digress.

As for Jab 1, the animation has been touched up compared to his Smash 4 version, and I can't be certain frame 1 of that jab is the precise frame 1 we're cutting to in his trailer. I don't doubt that it's probably still frame 2 but I only count definitive evidence for my charts.
Would a comparison with the Everyone's Here section of the Ultimate Direct work as backing evidence: https://youtu.be/L93H7YC-83o?t=88? When Falco appears, he does a jab 1 and jab 2. The footage runs at 30 FPS for some reason as noted earlier and it's not a proper way of counting, but when checking Falco's jab 1, it only takes one frame advance to see him hit with it where it has the same blue attack effect when he hits Wolf in his trailer and the frame before that in Falco's trailer has no attack effect.

For Dash Attack, that's definitely an 8. The animation is unchanged from Smash 4 where you'll notice both of Falco's feet are off the ground on Frame 4.
I'm still seeing frame 7. The starting frame is when Falco shifts from his running animation to having his feet point towards the screen. From there, both his legs are off the ground on frame 3 and Fox gets hit on frame 7. Frame 6 is when the blue attack effect starts to appear behind him and frame 7 is when it completely appears as an attack effect along with the red hit effect from Falco hitting Fox. I don't think there is a hiccup in the video and it's a minor change, really.

...D-Air give me no data to work with based on how they're used.
Understood with the startup needing definitive evidence, but Dair's landing lag shouldn't be a problem with this, right?

As for WFT, the Jab 1 is a frame 5 hit in the video, but I suspect it's a late hit. WFT has famously terrible hitboxes and I suspect Villager's Fsmash had only then placed him close enough to it. I think proper testing will find it's still frame 4 since villager is a short target. Similar situation with N-Air, that's a frame 11 hit in that video when it could still be starting on frame 9. F-Tilt is another late hit but you did remind me to make a note that it has more active hit frames which is the only change we can definitively gleam from that video. U-Air is another late hit on frame 7. Dash attack also appears to be a frame 8 late hit in that shot.
Regarding those moves for WFT, I did make a note of them in my post and compared them to in Smash 4. In particular, jab 1 is probably frame 4 like in Smash 4 and like in Smash 4, it's frame 5 when it can hit lower or has a large, low hitbox appear allowing it to and as noted, Villager is a short target.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#22
Japan Expo, SD Comic Con, Jump Victory Carnival, all giving us lots of Smash Ultimate footage, but nothing in 60 FPS. I got really enthusiastic with the quality of footage from E3 leading to these charts, but the recent hunt for more data has been fruitless. EVO is coming up and I don't want my heart broken again.

Twitch is supposed to have a way to advance frame by frame on clipped footage but it doesn't work for me.

Understood with the startup needing definitive evidence, but Dair's landing lag shouldn't be a problem with this, right?
I assume we're talking about this particular D-Air? Falco does nothing after landing. Giving me no hard indication of which point the landing lag has concluded. In Smash Bros, the point at which you can act out of a move very seldom matches the point at which the animation finishes and the character resumes their idle stance. If I compared side to side with Falco's old D-air landing lag animation, I'd say the animation is sped up by 9 frames. Used to resume his idle stance on frame 32, now resumes his idle stance on frame 23. But it's just not the hard evidence I'm looking for. Bottom line is, expect less landing lag on his D-Air, but don't get too attached to the particular number of 14 landing lag since we don't know for sure.

As for his jab, again I don't dispute it's probably still frame 2, but I have a methodoly of hard evidence in my charts and would rather not stray from it just to confirm something is probably the same as it was in the last game. If I did, then it begs the question of why I don't go to this trouble making educated guesses on every move we've seen and I hate to increase the potential for mistakes. Like a Sith, I only deal in absolutes.
 

Ffamran

The Ace Tank Driver?
Moderator
Joined
Aug 25, 2014
Messages
14,333
#23
I assume we're talking about this particular D-Air? Falco does nothing after landing. Giving me no hard indication of which point the landing lag has concluded. In Smash Bros, the point at which you can act out of a move very seldom matches the point at which the animation finishes and the character resumes their idle stance. If I compared side to side with Falco's old D-air landing lag animation, I'd say the animation is sped up by 9 frames. Used to resume his idle stance on frame 32, now resumes his idle stance on frame 23. But it's just not the hard evidence I'm looking for. Bottom line is, expect less landing lag on his D-Air, but don't get too attached to the particular number of 14 landing lag since we don't know for sure.

As for his jab, again I don't dispute it's probably still frame 2, but I have a methodoly of hard evidence in my charts and would rather not stray from it just to confirm something is probably the same as it was in the last game. If I did, then it begs the question of why I don't go to this trouble making educated guesses on every move we've seen and I hate to increase the potential for mistakes. Like a Sith, I only deal in absolutes.
I did not know about that about landing lag. I knew you could act before the total animation of moves in general for games, but not, in particular for Smash 4 Falco's Dair, that there was a 9 frame difference from the total landing animation and when he can act out of it. That said, 23 landing lag would be a safe number for me. Other Dairs did receive lowered landing lag, but as far as I know, no other Dair received a drastic startup reduction, what looks like 6 less frames, from frame 16 to 10, that Falco did. So, part of me feels like they could be justifying his Dair keeping its 23 landing lag because of the startup change they decided on. The other thing is that most frame data for characters in Smash and fighting games in general are carried over, so it would also not surprise me if they just carried over its landing lag at least in the initial phases of the game's development.

Anyway, I think, and this is just me talking to myself at this point, a safe bet if it (probably) did have its landing lag reduced is around 18 frames. That was its landing lag in Melee and while it is higher than some others like Cloud's whose Dair would have similar startup, but 2 less landing frames, they could do something like bring back its ability to auto-cancel from his hop. 14 would put it close to Brawl's 12 which could work considering it doesn't have its frame 5 startup or the ability to auto-cancel from a hop.

For jab, I was just wondering it the other video could work for a comparison and backing evidence.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#24
I did not know about that about landing lag. I knew you could act before the total animation of moves in general for games, but not, in particular for Smash 4 Falco's Dair, that there was a 9 frame difference from the total landing animation and when he can act out of it. That said, 23 landing lag would be a safe number for me. Other Dairs did receive lowered landing lag, but as far as I know, no other Dair received a drastic startup reduction, what looks like 6 less frames, from frame 16 to 10, that Falco did.
Oh, I assumed you wanted me to look at his Dair since we see him land with it. I didn't even notice the startup difference.Well that I can at least jot a note down of even when I don't have numbers. A reminder to get numbers for it when the game releases. Not sure how to interpret its animation difference from Smash 4. I'll be honest, comparing the animation to Smash 4 that looks like frame 5 or 6 startup to me. Because a fast fall occurs eight frames before the kick while he's still in his full hop animation. And you can see his fall speed looks quite normal during the move - since doing an aerial immediately voids a fast fall until you re-input it. I'm not sure I believe that they gave him a frame 5 spike, but that's what it looks like to me. The fastest spikes in Smash 4 were Olimar's and Ryu's D-Air at 8. As well as Villager's if the RNG gave him three turnips. Hey if Falco has his Melee D-Air let's hope it's NTSC and not PAL.

As for jab, yeesh are we still on this? Look, when I have the game in my hands, Falco's frame 2 Jab 1 will be the first thing I confirm. Sound good? I'm totally on the same page with it still being frame 2 since the evidence we've seen makes mathematical sense. I'm just not okay with using totally different methodology of evidence just to confirm one thing.
 

Thinkaman

Moderator
Moderator
Joined
Aug 26, 2007
Messages
5,904
Location
Madison, WI
3DS FC
1504-5749-3616
NNID
Thinkaman
#25
In case you haven't seen, there is a parallel Japanese effort that is focused on grabs, dodges, and landing lag:
https://docs.google.com/spreadsheets/d/1k3APyiQk0fu1KCqq-t9Fk2zmmSvcEH8rBy6GVAHEtQk

Translation of the first page:
  • Grab
    • Standing
      • Before
      • After
      • Change
    • Dash
      • Before
      • After
      • Change
  • Dodge
    • Standing
      • Before
      • After
      • Change
    • Forward
      • Before
      • After
      • Change
    • Back
      • Before
      • After
      • Change
  • Air Dodge
    • Nondirectional Landing Lag
    • Directional Landing Lag
Obviously the characters are in labeled numerical order, which is a very nice reference anchor. Snake/Squirtle/Ivy are greyed to indicate that they are comparing to Brawl data, and ICs only has some Melee values listed.

The character pages list various attacks with ダメージ (damage), 全体フレーム (total frames), or 着地隙 (landing lag), always listed with both before and after.

The big takeaway here is better tether grabs, worse non-tether grabs, better spotdodges (before staling!), roughly similar forward rolls, and worse back rolls.

Edit:
Zapp Branniglenn Zapp Branniglenn After comparing both documents, almost all values recorded in both documents match exactly. The onyl exception I saw was PAC-MAN grab total frames, which you have at 64 vs. 58.
 
Last edited:

J0eyboi

Smash Ace
Joined
Dec 28, 2017
Messages
534
#26
I think it's pretty safe to assume shield stun isn't properly functional yet, but I'd like to point out that these numbers are very close to what they would be in Brawl. Brawl's shieldstun formula is [damage] * .375, rounded down, and because all of these examples were used in a short hop, they would deal 85% of the listed damage. 17.64*.85*.375 rounded down is 5, and 21.4*.85*.375 rounded down is 6. The Nair one is definitely odd, though. Even by Brawl standards it should have 2 frames shieldstun, not 1. (Edit: I checked the timestamp and it looks to me like there were 2 frames of shieldstun(Edit: nope, nevermind))
 
Last edited:

Thinkaman

Moderator
Moderator
Joined
Aug 26, 2007
Messages
5,904
Location
Madison, WI
3DS FC
1504-5749-3616
NNID
Thinkaman
#27
I produced a quick and dirty informal graph for a friend, using this data as one of my 3 sources for Ultimate:



I wish I had auto-cancel data for each game as well, as that is really the other half of the story. (For example, the second outlier on the Smash 4 data? That's Sonic, who has very generous AC windows.) Unfortunately, that data doesn't exist for Melee or Brawl in an accessible format, and would be asinine to try and pull from Ultimate vids.
 
Last edited:

SSBBDaisy

Smash Lord
Joined
Nov 24, 2007
Messages
1,374
Location
Florida
#28
OMGoodness yes I really want to know the auto cancel windows in smash bros ultimate in comparison to smash 4. I hope they are drastically improved/adjusted to where they actually matter.
 

J0eyboi

Smash Ace
Joined
Dec 28, 2017
Messages
534
#29
OMGoodness yes I really want to know the auto cancel windows in smash bros ultimate in comparison to smash 4. I hope they are drastically improved/adjusted to where they actually matter.
You main Rosa, your autocancel windows definitely matter.


Anyway, I figured I should also mention in this thread that the shield drop animation in the demo is FAF 12, but can be jump cancelled earlier (Ike jumping out of shield drop animation before the parry frames even end)
 

SSBBDaisy

Smash Lord
Joined
Nov 24, 2007
Messages
1,374
Location
Florida
#30
You main Rosa, your autocancel windows definitely matter.


Anyway, I figured I should also mention in this thread that the shield drop animation in the demo is FAF 12, but can be jump cancelled earlier (Ike jumping out of shield drop animation before the parry frames even end)
They would matter if it didn't mean the auto cancel windows showing up so late after I hit a shield to where I can get punished. Unlike in brawl where the auto cancel windows were early enough to keep pressuring safely.

Example

Brawl Peach fair in comparison to Smash 4 Peach fair. The auto cancel window on brawl is early enough to pressure safely with her fair.


EDIT: also in regards to Rosa. I can tell you 100% confidently. There is proof. Rosa's Fair now auto cancels from a short hop in ultimate but not in smash 4.
 
Last edited:

Amazing Ampharos

Balanced Brawl Designer
Joined
Jan 31, 2008
Messages
4,565
Location
Kansas City, MO
#31
They would matter if it didn't mean the auto cancel windows showing up so late after I hit a shield to where I can get punished. Unlike in brawl where the auto cancel windows were early enough to keep pressuring safely.

Example

Brawl Peach fair in comparison to Smash 4 Peach fair. The auto cancel window on brawl is early enough to pressure safely with her fair.


EDIT: also in regards to Rosa. I can tell you 100% confidently. There is proof. Rosa's Fair now auto cancels from a short hop in ultimate but not in smash 4.
I'm interested in this proof of Rosalina's fair auto-canceling out of a short hop in Ultimate. That would be extraordinarily helpful; her fair has by far the best hitboxes of any of her aerials and also is a really good combo starter if you connect but the high landing lag and unfriendly auto-cancel window always made it a lot less applicable than other options just due to how risky it is. This is a very large buff for Rosalina if true and if the move retains most of its 4 dynamics otherwise.
 

SSBBDaisy

Smash Lord
Joined
Nov 24, 2007
Messages
1,374
Location
Florida
#32
I'm interested in this proof of Rosalina's fair auto-canceling out of a short hop in Ultimate. That would be extraordinarily helpful; her fair has by far the best hitboxes of any of her aerials and also is a really good combo starter if you connect but the high landing lag and unfriendly auto-cancel window always made it a lot less applicable than other options just due to how risky it is. This is a very large buff for Rosalina if true and if the move retains most of its 4 dynamics otherwise.
It's in her character trailer. Try getting a short hop fair to auto cancel with her in smash 4 and it won't happen. She can't auto cancel fair in tr4sh I mean sm4sh from a short hop but she can in ultimate. The point I'm getting at is they are changing the auto cancels I just was wondering if they had actually made them effective this time when it comes to approaching cause neither of my characters could make any good use from their auto cancels cause the window happens literally at the end of the aerials. U can't safely pressure with that.
 
Last edited:
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#33
In case you haven't seen, there is a parallel Japanese effort that is focused on grabs, dodges, and landing lag:
https://docs.google.com/spreadsheets/d/1k3APyiQk0fu1KCqq-t9Fk2zmmSvcEH8rBy6GVAHEtQk

The big takeaway here is better tether grabs, worse non-tether grabs, better spotdodges (before staling!), roughly similar forward rolls, and worse back rolls.

Edit:
Zapp Branniglenn Zapp Branniglenn After comparing both documents, almost all values recorded in both documents match exactly. The onyl exception I saw was PAC-MAN grab total frames, which you have at 64 vs. 58.
Thanks for running a comparison. I only add data I have evidence for but I looked over some old pac man footage and managed to confirm the 58 number.

Anyway, I figured I should also mention in this thread that the shield drop animation in the demo is FAF 12, but can be jump cancelled earlier (Ike jumping out of shield drop animation before the parry frames even end)
More Shield jump cancelling. I think I'm gonna be sick but thanks for bringing this to our attention.

I think it's pretty safe to assume shield stun isn't properly functional yet, but I'd like to point out that these numbers are very close to what they would be in Brawl. Brawl's shieldstun formula is [damage] * .375, rounded down, and because all of these examples were used in a short hop, they would deal 85% of the listed damage. 17.64*.85*.375 rounded down is 5, and 21.4*.85*.375 rounded down is 6. The Nair one is definitely odd, though. Even by Brawl standards it should have 2 frames shieldstun, not 1. (Edit: I checked the timestamp and it looks to me like there were 2 frames of shieldstun(Edit: nope, nevermind))
Good catch on the shorthop damage modifier, as I was not privy to that information at the time.

(For example, the second outlier on the Smash 4 data? That's Sonic, who has very generous AC windows.)
Real talk I thought Sonic's U-Air had no landing lag state in this demo when I witnessed him land with it exactly one frame after the second kick. But then I looked up the AC window from Smash 4 and it suddenly made sense.

Tangentially related but Donkey Kong's Down B has no landing lag state in this game. I'm assuming for now this is a bug since it had 30 landing lag in the last game.
 

Nobie

Smash Champion
Joined
Sep 27, 2002
Messages
2,034
3DS FC
2809-8958-8223
NNID
SDShamshel
#34
I'm thrilled and terrified by Mewtwo's 8 frames of landing lag on fair. Yeah, moves staling on shield and the short hop modifier mitigate it somewhat, but you know who loves it when opponents shield too much? One big psychic angry cat.
 
Last edited:
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#35
Starting to hunt for data not related to attacks such as (fresh) dodges, techs, and other actions. Here's a source for Smash 4's data for comparison. I'm also curious how much dodges will "stale" with overuse, but that sort of data will be tricky to get since I have no idea how long it takes for dodges to refresh and players typically don't spam more than two dodges at a time. Nothing notable to report other than I can confirm back rolls are worse than forward rolls. Tech actions appear to still be uniformly the same. It looks like you are left prone for 24 or 25 frames on a missed tech instead of 26 in the last game which is an indirect nerf to Ganon's Side B and Snake's DThrow. As well as a nerf to jab locks if those ever get implemented.

I think there are some other patterns with attacks that are worth pointing out

Multihit moves finish quicker: Not all of them, but many moves have shortened gaps between hits. Falcon's N-Air, Charizard's Usmash, Corrin's Dash attack. I have some theories about why these sorts of moves got this adjustment. (1) programming multihit attacks to link together properly even at high rage is hard. Just have the attacks come out quicker so the victim doesn't escape. (2) We want this move to end quicker but shaving off its endlag may make it arbitrarily difficult to punish. (3) we want this move to not be perfect shielded when the victim lets go of shield too early. In the last game, multihit moves slow enough that they weren't a block string were generally a good thing since the defender could mistakenly put down shield too early for their punish. But in this game, putting down your shield on a multihit attack will be a free perfect shield. And depending on the advantage state, may lead to a free punish. I know I wouldn't want to use a move that was just as unsafe on block as it was when they dodge it.

Multijabs are better: Sort of related to multihits but this feels like a much more wide trend for this particular move. As far as I've seen, these moves might be doing less damage per hit, but you'll land more hits since the gaps between have been shortened to one or two frames on many of them. I think the ultimate result will be multijabs being easier to escape with SDI but harder to escape with normal DI. Unless SDI remains this weak in the final build in which case multijabs will be harder to escape period. And if multijabs link better because of them hitting quicker then you can generally call that a buff since the user is in full control of when these moves end.

Buffed Dash attacks: It makes sense for these moves to be made more relevant now that they compete with every other grounded move as an option from a run or dash. :ultbayonetta::ultcorrin::ultdk::ultfox::ultike::ultpalutena::ultpikachu::ultsheik::ultsonic::ultvillager::ultzelda: All have some confirmed buff to their dash attack's frame data. :ultpit:also appears to have some kill power on his. :ultkirby: may be a tad slower than the previous one, but multihit dash attacks have a harder time crossing up shields while this new one should have a lower launch angle giving Kirby vitally rare opportunities to edgeguard so it's a buff in my eyes. :ultwario: also has a new dash attack entirely, but it's hard to say whether it's a buff since the old one did have some gimmicky applications and I know what Wario needs most is a move with decent kill power that isn't a slow Smash attack or waft. One thing I do find egregious however is them missing the worst dash attack in the game :ultmarth:
 
Last edited:

Thinkaman

Moderator
Moderator
Joined
Aug 26, 2007
Messages
5,904
Location
Madison, WI
3DS FC
1504-5749-3616
NNID
Thinkaman
#36
I concur with all of those observations and reasoning, including all 3 motivations for faster multihits as likely factors.
Though dash attack or not, the way things look currently I am not remotely concerned for :ultmarth:'s viability, haha.
 
Joined
Jan 28, 2014
Messages
7,809
Location
Buried under 990+ weapons
3DS FC
1504-5709-4054
#38
Dash Attacks are overall buffed... I remember hearing the opposite previously.

Ah well, like half of those ranged from “literally just Fair/Nair/Bair instead” to mediocre anyway.

Pit’s is practically a straight buff yeah. It has a little less horizontal range but slightly better frame data (like literally a frame right?), same damage, seemingly good vertical range, (Dtilt > DA maybe?) and a surprising amount of kill power for Pit. Though... Most people are surprised about literally any good thing Pit has lol so YMMV.
 
Joined
Apr 13, 2014
Messages
1,636
Location
Santa Ana, CA
#39
Pit’s is practically a straight buff yeah. It has a little less horizontal range but slightly better frame data (like literally a frame right?), .
That source I initially found was one frame less, yes, and I still have it there in my charts But I've been looking at his dash attack and have since found several more sources where the move's frame data is unchanged from the previous game, so that's what I believe and have noted in my charts weeks ago. My only explanation is there being a frame drop in the footage. It was from a four player free for all, and I assume the E3 build isn't perfectly optimized to never have a frame drop ever. I can confirm something new for Pit however. His F-Smash has five less endlag.

Speaking of frame drops wow has that been a problem lately with recent footage from the London and Arcadia tournaments. I don't know if they're using shoddy equipment or if it's a video encoding failure but that footage is really unstable. Although it being 60FPS at all basically makes it the only viable footage to have appeared in like a month so beggars can't be choosers. These sorts of frame drops can be counted around depending on when they occur.
 
Joined
Jan 28, 2014
Messages
7,809
Location
Buried under 990+ weapons
3DS FC
1504-5709-4054
#40
I think I’m speaking for all of us when I say that you shouldn’t sweat it so much... You’re doing great work man.

The Fsmash buff... The Sakurai patented “let’s buff this character but not really lol” buff. Probably doesn’t mean anything on shield because of the (lack of) shieldstun, it should still be punishable by nearly everything. Barely matters on whiff because that’s still a solid half a second+ endlag. Dunno. Kinda hope they buff something really stupid in there like Dsmash with the rest of it lol.

Hey, have other Fsmashes been buffed? If others (perhaps Link’s, that’s also two hits) that could be evidence that the shieldstun changes are intended and are being designed around.
 
Last edited:
Top