Q&A Mechanics & Techniques Discussion

[Lavani]

Noticing that these all involve teleport moves so far, maybe you actually get to the ledge on Frame 1 if your character has no extra distance to travel when the snap happens?

Going through frame by frame, in both examples in the gfycat it looks like there's still one frame of ledgesnap with a frame 2 ledge grab. ( frame 1 frame 2). I'm not sure if it's just a coincidence, or if that means the frame before grabbing the ledge is invulnerable normally and reaching for the ledge from above skips said frame...or maybe my limited recording equipment happened to skip a frame twice in an unfortunate coincidence. It isn't a frame 1 grab in these examples at any rate, and Zelda does appear a bit above the ledge before snapping down to it. That all said, I wouldn't know if shorter snaps when closer to the ledge are a thing.

The gif linked in my post shows Ganon snapping at a downward diagonal through Luma, I've just been showing teleports as the training examples because I can't control two characters on 3DS and can't think of other recoveries that can both pass through a smart bomb blast and snap to the ledge while still invulnerable. This isn't limited to teleport moves or invincible recoveries at any rate.

 

[TheReflexWonder]

little mac's side B does with a super predictable path. does this help mac at all? lol.

It might if the edge meta revolves around guaranteed punishes on ledge snaps.

Lavani, it's worth noting that I tried going through the onstage Waft hitbox (lingering via hitting the Bike) with Ganon Up-B in the same manner (and other angles) and I didn't go through it once out of about a dozen tries. Of course, that could be human error on my part. I don't know if full Waft covers more space than Luma Rapid Jab, but it's worth noting that multihit moves with empty frames of no hitbox between hitboxes will not catch the one frame of vulnerability reliably; I tested with Wario D-Air and Up-B2, both of which longer but have a frame of no hitbox between each hitbox.

 

[Lavani]

Lavani, it's worth noting that I tried going through the onstage Waft hitbox (lingering via hitting the Bike) with Ganon Up-B in the same manner (and other angles) and I didn't go through it once out of about a dozen tries.

My assumption would be that Ganon's just being hit out of his upB rather than being hit out of ledgesnap, but I can't think of any good ways to test it solo.

 

[inconspikuous]

It might if the edge meta revolves around guaranteed punishes on ledge snaps.

so in mac dittos, dsmash on ledge vs. side B isn't necessarily guaranteed as long as the snap happens above the ledge? good to know.

 

[TheReflexWonder]

so in mac dittos, dsmash on ledge vs. side B isn't necessarily guaranteed as long as the snap happens above the ledge? good to know.

I don't know! We're trying to piece it together!

 

[MrTeddyBear]

Pikachu's Quick Attack can do this as well.

I think it would be useful later on if we got a list of all the moves that have this invincibility property, once we find out most of them.

 

[Natch]

http://www.reddit.com/r/smashbros/comments/3417l1/discoveries_on_living_longer_in_smash_4/

Here is a new thread recently posted on the Smashbros subreddit. I've no clue to it's validity, but I'd like it to be tested to see if it's true/false. If so, this may change the way we DI a lot,

 

[Shaya]

Well anecdotal, but for Zero Suit's Up-B, any form of holding left/right will kill you significantly earlier than holding down (at least from centre stage).

 

[deepseadiva]

I had another question about ledge trumping (is there a thread somewhere about it?):

@ TheReflexWonder mentioned once that ledge trumping shouldn't happen because you can always buffer a ledge action like getup or roll or something. It will happen on occasion regardless due to player error, and I was wondering about the mechanics of that? Are follow ups from a trump guaranteed? What kind of move would you need to combo. I'm thinking about most bairs of course, but maybe some characters have nairs, uairs, or side-b's that could work? I'm curious about the frame trap that's created on a trump, and what kind of window characters have to punish.

 

[Lavani]

Using 's numbers (I assume it's the same for most characters though I know tether recoveries are an exception), the minimum ledge hold time is 20 frames and being ejected from the ledge is 30 frames before being able to act. The sooner you grab the ledge to trump the sooner your own ledge hold timer ends, so we'd be looking at a +10~29 frame advantage; a late trump is only going to have a guaranteed followup with quick moves like 's 4f bair, but even Dedede's sluggish 17f bair is guaranteed off an early trump.

I haven't looked a whole lot into what moves other than bairs can be used off a ledge trump effectively, but I know Charizard can opt for upB against characters with a more vertical ejection angle like Fox.

 

[TheReflexWonder]

As is well-known, the ten-frame buffer window only allows you to buffer a single input at a time in Smash 4.

That said, there appears to be a set priority for inputs done within the normal ten-frame buffer window, regardless of the order buttons were pressed: Special > Shield > Attack > Jump. This was tested out of airdodge, knockback that doesn't put you in tumble, knockback that puts you in tumble, and tumble. At first glance, here's how I think this appears to affect the metagame:

Very weak attacks that put the opponent in the air after hitstun (things like some Jabs, Mac N-Air, etc.) tend to turn complex inputs (as in, involving more than one button) against the user. Reflexive shieldgrab inputs turn into an airdodge, giving you a pretty big window to punish. You can't buffer a jumping aerial or special, either, so in order to jump away from the situation, you have to be relatively passive, which may be useful for the attacker trying to follow up.

Also, buffering jump-cancel out-of-shield options is made slightly harder (though it was already super-hard due to being unable to buffer two inputs at once), due to Attack having priority over the Jump button. This is why if you press Attack while in shield hitlag and then press Jump, you'll get a shieldgrab; the game still wants to buffer the Attack input.

Some quirks: A+B Smash makes it so that pressing Special while also pressing Attack (either at the same time or by pressing Special while holding Attack in advance) turns the Special input into another Attack input instead. Also, you can airdodge out of the end of knockback that would otherwise put you in tumble, so input priority doesn't matter in that situation (since all other inputs do nothing in that instance).

 

[Pikabunz]

http://www.reddit.com/r/smashbros/comments/3417l1/discoveries_on_living_longer_in_smash_4/

Here is a new thread recently posted on the Smashbros subreddit. I've no clue to it's validity, but I'd like it to be tested to see if it's true/false. If so, this may change the way we DI a lot,

Yeah, you always want to be holding towards the stage for horizontal knock back. I know some people already do this and there was a video on it, but I can't seem to find it anymore. Though it seems no one really knows why DI works the way it does in this game. First of off, I don't think "vectoring" was really a thing in this game, not even before the patch. Holding up before knock back straight up increases your knock back by up to 20% depending on how far you're holding the control stick up, while holding down decreases it by up to 20%. The knock back reduction from holding down is separate from the knock back reduction you get from crouch canceling. Holding left or right have no effect on your knock back at all, but it still changes your trajectory. This why holding up to DI kills you earlier than no DI at all. Even though you're angling yourself up higher, you're also increasing your knock back by a lot which kills you. This is why you want to hold towards the stage. It's the only way to angle yourself up higher while not increases your knock back. So why doesn't holding down let you survive longer since it decreases knock back? It's because you're also angling yourself down which makes you travel almost completely horizontal that not even reduced knock back will help you survive, but it's still better than holding up or no DI at all.

Well anecdotal, but for Zero Suit's Up-B, any form of holding left/right will kill you significantly earlier than holding down (at least from centre stage).

This also explains why holding down is so effective against ZSS' upb. It's a horizontal knock back move that normally kills off the top with no DI. By holding down you're reducing your knock back by 20% and angling yourself down from the top blast zone.

 

[TheReflexWonder]

Can you DI moves that don't send you into tumble (due to there not being enough knockback)? I forget what special quirk the property of hitstun-that-doesn't-send-you-into-tumble has.

 

[Shaya]

I could be wrong for 6 years+ but yeah, no DI on pre-tumble knock back.
Tumble is a state after reaching a high enough knockback, which is 30 frames of hit stun and from Brawl's numbers around 6500 m/ph (Marth's tipper fsmash at 0% one of the few moves which achieve it IIRC).

 

[otter]

I had another question about ledge trumping (is there a thread somewhere about it?):

@ TheReflexWonder mentioned once that ledge trumping shouldn't happen because you can always buffer a ledge action like getup or roll or something. It will happen on occasion regardless due to player error, and I was wondering about the mechanics of that? Are follow ups from a trump guaranteed? What kind of move would you need to combo. I'm thinking about most bairs of course, but maybe some characters have nairs, uairs, or side-b's that could work? I'm curious about the frame trap that's created on a trump, and what kind of window characters have to punish.

Using 's numbers (I assume it's the same for most characters though I know tether recoveries are an exception), the minimum ledge hold time is 20 frames and being ejected from the ledge is 30 frames before being able to act. The sooner you grab the ledge to trump the sooner your own ledge hold timer ends, so we'd be looking at a +10~29 frame advantage; a late trump is only going to have a guaranteed followup with quick moves like 's 4f bair, but even Dedede's sluggish 17f bair is guaranteed off an early trump.

I haven't looked a whole lot into what moves other than bairs can be used off a ledge trump effectively, but I know Charizard can opt for upB against characters with a more vertical ejection angle like Fox.

You can ignore the minimum ledge hold time with some movement options such as rolling and jumping. You can easily react to being trumped and use these options while both chaaracters are on the ledge. It's why I only played the game for a month, honestly. They removed edgehogging and replaced it with nothing.

 

[TheReflexWonder]

You can't bypass the minimum ledge hold time; you can only buffer a command in advance so that it comes out on the first possible frames (which is like Frame 21 or something).

 

[TheReflexWonder]

I could be wrong for 6 years+ but yeah, no DI on pre-tumble knock back.
Tumble is a state after reaching a high enough knockback, which is 30 frames of hit stun and from Brawl's numbers around 6500 m/ph (Marth's tipper fsmash at 0% one of the few moves which achieve it IIRC).

That's what I thought. I imagine that's why things like the Link Jab infinite happened.

Well, in that case, people need to start trying to SDI out of stuff like Mario's U-Tilt, since holding a direction won't do anything until the point of tumble. This needs to get around.

 

[DungeonMaster]

You can't bypass the minimum ledge hold time; you can only buffer a command in advance so that it comes out on the first possible frames (which is like Frame 21 or something).

Tether recoveries are 7 frames hold time. A tether trump can be quite deadly.

 

[TheReflexWonder]

Public service announcement: If you're testing hitboxes and don't need a specific target, Mega Man is likely the best character to test it against. His standing animation and jumping animation both have no movement at all, and if a character has an item in their hand, it disables their idle animations (which is easily doable via Metal Blade). This prevents hitboxes from being affected by changes in hurtboxes caused by the character shifting around.

 

[Jaxas]

Public service announcement: If you're testing hitboxes and don't need a specific target, Mega Man is likely the best character to test it against. His standing animation and jumping animation both have no movement at all, and if a character has an item in their hand, it disables their idle animations (which is easily doable via Metal Blade). This prevents hitboxes from being affected by changes in hurtboxes caused by the character shifting around.

@ Dr. Tuen

 

[Dr. Tuen]

@ Dr. Tuen

Oh, this is awesome! A very useful reference indeed! I'll use it in my upcoming tests on zero suit samus.

I'm also really curious about this ledge snap vulnerability thing. The high recovery angle hypothesis is worth testing some more. I have a question regarding that testing process: Do smart bombs produce a hitbox on every frame? If not, how might we produce a lasting hitbox to test these ledge snaps? Makes me kind of want to make a TAS controller for it. Not that I have the skills.

Anyways, I know it's not a sterile testing environment, but I've managed what looks like a ledge-snap vulnerability punish. It agrees with the high-angle invincibility hypothesis, as this is a Dr. Mario recovering low. I hope this helps a bit! (the punish is about 20 seconds after the video starts.)

 

[Fox Is Openly Deceptive]

Tether recoveries are 7 frames hold time.

Actually I'm pretty sure that tethers don't have any ledge hold time at all (maybe one frame). Samus' ledge options just make it look like she's doing nothing for a little bit. Instead, see what happens when you hit away from the ledge to let go of it immediately, or try other tether characters. Tether Trumps are great.

 

[DungeonMaster]

Actually I'm pretty sure that tethers don't have any ledge hold time at all (maybe one frame). Samus' ledge options just make it look like she's doing nothing for a little bit. Instead, see what happens when you hit away from the ledge to let go of it immediately, or try other tether characters. Tether Trumps are great.

I think Izaw in Art of link 2 pinned it down to 7 frames for Link, that's where I got that number. 3:15 in this video: https://www.youtube.com/watch?v=PwpIxtg82I0
Shoutouts to @Afro Smash and @-_ellipsis_- for reverse charge shot trump videos: https://www.youtube.com/watch?v=NQBKOigBYlo and https://www.youtube.com/watch?v=WHxPEjV3mto
If you're trumped with tether it is in fact guaranteed, the b-air likewise. No video yet... soon...

Incidentally I just stumbled upon this thread today and I read the entire thing. You guys have been trying to figure out the rage and staleness formulas in detail, I empirically did exactly that with Samus, particularly her up-tilt and d-air.
These moves have perfect vertical knockback and several of Samus' combos only work within fairly tight ranges. The rage formula works out empirically to adjust the combo sweetspot by 1% for every 10% damage on Samus. Staleness increases it by 2% between the first move and second, with 1% increase thereafter. I mapped this out in great detail in this thread: http://smashboards.com/threads/the-...-string-list-and-knowledge-compendium.391853/
There is some non-linearity at the low and high end but for all practical purposes it's extremely accurate. I am in fact actually landing the 40 and 50 damage true combos in game now and I hope to have a video showing this shortly, this game can be played by %.

 

[Fox Is Openly Deceptive]

I think Izaw in Art of link 2 pinned it down to 7 frames for Link, that's where I got that number. 3:15 in this video: https://www.youtube.com/watch?v=PwpIxtg82I0
Shoutouts to @Afro Smash and @-_ellipsis_- for reverse charge shot trump videos: https://www.youtube.com/watch?v=NQBKOigBYlo and https://www.youtube.com/watch?v=WHxPEjV3mto
If you're trumped with tether it is in fact guaranteed, the b-air likewise. No video yet... soon...

Yes, well. To put it lightly, I certainly wouldn't trust Izaw with little facts like that. Because he's wrong. I know this because I was the one who found tether trumping in the first place (at least for the Links; I don't know if other characters found it earlier).

 

[TheReflexWonder]

Wasn't it some absurdly low number in Brawl, like three frames? I figure that's how people did the Z-Air ledge invincibility refresh for edgehogging.

 

[Pikabunz]

Has anyone looked into using aerials when trying to survive KO's? I've found that some make you survive a little longer and others killed you earlier. Like after getting back thrown by Ness at the edge, Pikachu went from surviving at 83% with DI alone to surviving at 88% with DI and using bair.

 

[Lavani]

Glad someone confirmed some aerials help you survive, I was going to ask about/look into that today because I read somewhere that it wasn't a thing anymore, and I had admittedly seen Rosalina start her uair animation but die anyway quite a bit recently (though doing some testing with Bob-ombs in training right now, it doesn't seem to be a helpful move).

Are there non-divekick A button moves that kill you earlier? I know some B moves will do some weird momentum shenanigans and kill you sooner if used out of hitstun.

 

[Shaya]

Really? Most instances I'm aware of it kills you earlier, to the point at which I taught myself not to have that brawl habit and often say to people when they look like they'll survive and suddenly "pop", "did you aerial?" "yes" "that's why!"

ZSS Up air! :<

 

[Pikabunz]

I've tested with Pikachu that fair and air dodging killed me earlier, but air dodge helped me survive 1% longer against vertical KO moves.

 

[TheReflexWonder]

Perhaps moves shift the center of your body in this game? I remember it didn't help in Brawl; doing something like Snake B-Air wouldn't keep you alive for longer vertically than other aerials when momentum canceling.

 

[Pikabunz]

Yeah, that's what I was thinking. I just went from surviving Sonic bthrow from 130% to 139% with Pikachu bair. That is a huge difference. I doubt any other move helps as much as Pikachu's bair since that move actually moves his body back a bit on start-up. Another random thing I noticed is that there's two knock back animations and the one that has Pikachu spinning helps him survivie 1% longer. So he can actually survive Sonic's bthrow up to 140% if he gets the spinning animation, but it doesn't appear to happen as often as his non spinning animation.

Survival percents with each move:
Side B - 123%
Fair - 128%
Air Dodge - 129%
No aerial - 130%
Uair - 131%
Dair - 132%
Nair - 132%
Bair - 139%

This is all with Pikachu's back facing the stage, which only happens with certain back throws.

 

[TheReflexWonder]

Most airdodges have people lean down/forward in their animations; perhaps that's why there appears to be a small difference with it. *shrugs*

 

[TheReflexWonder]

I don't read moonrunes, but there's frame data, apparently--

http://sixriver.web.fc2.com/ssb4/Character_data.htm

 

[Shaya]

It's not up to date on Marth at least

 

[Lavani]

There are some typos in there (Rosalina's dtilt/utilt IASAs are switched around, for one) but for the most part it seems pretty accurate from what I've looked at so far.

Something that jumped out at me - and I did confirm it - is Greninja's spotdodge being active frame 2, despite all the claims that spotdodges were universally nerfed to frame 4 startup like rolls before. That said, I'm also really glad to see dodge data in here considering its absence from MasterCore.

Rosalina's jab3 having a cancel point looping back to jab1 11 frames before the normal IASA isn't something I knew about either but explains a lot about how lagless that move felt when positioning Luma. That might also explain why jab1>jab2>jab3>jab1>jab2 tends to automatically buffer her rapid jab without a sixth A input.

Thank you very much for linking this.

 

[Ffamran]

I don't read moonrunes, but there's frame data, apparently--

http://sixriver.web.fc2.com/ssb4/Character_data.htm

Cool, it has end lag data too! Google translate works well enough and apparently Falco's Blaster does draw quickly if this is right at frame 11, but it stops at frame 58... That's one second of commitment. Aerial Blaster is frame 9 and ends at frame 49... Basically one second of commitment... There's also this (40) and (33) thing which could be how long the laser is out. Well, since I don't have a way to record and reliably count data, I'll just reference this for Falco's "missing data".

"Landing clearance" means landing lag right? I'm going to assume so since when Falco "lands" with Fire Bird, it's not that long.

 

[Lavani]

Since it probably isn't obvious for those unfamiliar with the terms, the 相殺 column is clanking.

◯ = Clanks
● = Clanks but continues uninterrupted
× = Transcendent (tangent: Little Mac's fsmash becomes transcendent angled down?!)
△ = Differing properties between hitboxes (ex. ROB nair has a transcendent hitbox and a clanking hitbox)

 

[Fox Is Openly Deceptive]

Little Mac's fsmash becomes transcendent angled down?!

Just confirming this. I tried to punch Toon's boomerang and got nothing while the straight forward f-smash made it disappear.

 

[TimeSmash]

Has anyone found any weird grab release stuff in general or on certain characters? Just curious, it's not exactly a technique, but I feel like the question is best suited here

 

[Lavani]

Has anyone found any weird grab release stuff in general or on certain characters? Just curious, it's not exactly a technique, but I feel like the question is best suited here

Well, there's this. Which really just boils down to "Marcina have a grab release combo on certain characters that would work on every character off fthrow anyway".

Rosalina can potentially get luma utilt off a grab release depending on his position, but that's both highly situational and less rewarding than going for a throw.

Sliding off ledges by pummeling after a dash grab with Falcon/Meta Knight is a thing, but I don't think you have frame advantage for it and depending on fall speeds you could just be putting yourself below your opponent offstage.

Lastly, if a character puts out a windbox while grabbed and forces their grabber off a ledge, it forces an aerial grab release on the grabber while the grabbee stays on the stage, ready to edgeguard.

Grab releases aren't really exploitable in this game, that's about all I've got.