Q&A Mechanics & Techniques Discussion

[Vipermoon]

https://m.reddit.com/r/smashbros/comments/4lltzk/ally_wavelands_onto_dreamland/

So about this...ally basically 'wavelanded' onto the stage. It should have to do with the same mechanic that allows the lagless ledge getup, but Id expect him to have airdodge landing lag...instead he seemed to have been flashing with i-frames until midway into the backthrow?

Ally took air dodge landing lag. I didn't notice invincibility after he landed.

We've seen this before. Usually it's called the lagless get-up. Ledge cancelling a double jump.

 

[Darklink401]

Ally took air dodge landing lag. I didn't notice invincibility after he landed.

We've seen this before. Usually it's called the lagless get-up. Ledge cancelling a double jump.

As I said, I know it follows the same principle as the lagless get-up, I just hadn't seen it with airdodges (with jumps, and moves yeah)

I guess rather than invincibility, he was just flashing because...water? After rewatching it and airdodging myself into the ground, I guess it just seemed like it had no landing lag because by the time he slid towards M2k, he already could act xD

Still kinda interesting that he slid with that tho xP

 

[EpicSonicLatios]

Hello

I'm not sure if this is the right place to post this, but I have a question regarding safety on shield data. When the data says -X number, does that mean that any move that takes X or fewer frames to hit (assuming the move has enough range) will punish the opponent before they can shield? I am pretty sure that this is the case, but I just want to make sure that their isn't any other factors in play. I ask this because I am trying to create a punish on block guide for my main, Fox.

 

[Funen1]

That's essentially how it works. In a general sense, those numbers mean when you can act again after using a move vs. when your opponent can act again after shieldstun - moves that are minus on block could be punished by any move that comes out in that number of frames or faster, assuming you're frame perfect and all that.

 

[meticulousboy]

Out of curiosity, do counterattacks that scale with damage received get stronger with rage?

 

[Kofu]

Out of curiosity, do counterattacks that scale with damage received get stronger with rage?

Rage is a flat boost to all base knockback (IIRC), so yes.

 

[Vipermoon]

Rage is a flat boost to all base knockback (IIRC), so yes.

Not base knockback, total knockback.

 

[A10theHero]

Reposting this question in case someone can provide some answers:

How would I account for gravity when calculating vertical knockback (also for knockback that has both a horizontal and vertical component)?
Based KuroganeHammer's website doesn't have any formulas mentioning gravity. But I did find a formula on Smash Wiki that says to just add (g - 0.075) * 5 to the vertical component of knockback, though I'm kinda skeptical about it.

Does anyone with access to the game's coding see anything about calculating gravity there? If the exact formula isn't known, what do we know about gravity in Smash 4?
I have some procedure in mind that I could do to try and figure out a formula, but it definitely wouldn't be perfect. It likely could only be exact to the first decimal place. And it'd be a lot of work. ;_;

 

[samosa]

Reposting this question in case someone can provide some answers:

Does anyone with access to the game's coding see anything about calculating gravity there? If the exact formula isn't known, what do we know about gravity in Smash 4?
I have some procedure in mind that I could do to try and figure out a formula, but it definitely wouldn't be perfect. It likely could only be exact to the first decimal place. And it'd be a lot of work. ;_;

This was posted some pages back.

Gravity Knockback is calculated with the equation ((G*21.6) - 1.84) * 5, where G is the character's gravity attribute (ranges from 0.19 for Fox to 0.05309 to Jigglypuff - gravity values lower than and including Donkey Kong's 0.08505 will actually result in vertical KB reduction)

Plugging fox's gravity into the equation gives us the value 11.32, which is added to the vertical component of KB calculated by either the standard or training mode KB formula (both found here, but the pre-gravity KB is what you use the SpaceJam calculator for). Horizontal/vertical KB components are found with trigonometry:

horizontal KB component = (standard/training KB) * cos(launch angle of the move)
vertical KB component = (standard/training KB) * sin(launch angle of the move)

add the gravity component to the vertical KB component to get the gravity-adjusted vertical KB component, then plug the old horizontal KB component and the new vertical KB component into the pythagorean theorem:

Gravity-Adjusted KB = sqrt( (old horizontal KB component)^2 + (new vertical KB component)^2 )
Gravity-Adjusted Hitstun = floor(Gravity-Adjusted KB * 0.4) - 1




Rage is only part of the standard VS. mode KB formula (it does not factor in at all to training mode KB, which can be useful). It's a number between 1.0 and 1.15 that the standard KB formula uses as a multiplier in the very last operation, but before gravity is added. To the best of our knowledge, the general order of standard KB/hitstun calculation goes like this:

1. Raw KB is calculated from attack damage/staling, target damage, target weight, attack KB growth, and attack base KB (not necessarily in that order)
2. Raw KB from step 1 is multiplied by factors for rage, crouch cancel, grounded meteor, and target charging smash (last 3 are constants). This is the Pre-Gravity KB.
3. Pre-Gravity KB has the target's Gravity KB component added to its vertical KB component using the process described above. This is the Gravity-Adjusted KB.
4. Hitstun is calculated using Gravity-Adjusted KB (the result of steps 1-3). The number of hitstun frames calculated here is not modified by any subsequent step.

Any changes to launch angle/trajectory/KB received by the target that are the result of the target's DI take place after step 4.

 

[Lavani]

I'm a bit skeptical about the vertical KB reduction. Playing around with edited low gravity values, the knockback distance doesn't seem much different, and with the reduced falling acceleration characters actually start dying earlier off the top.

 

[Funen1]

This brings up a question I have about the gravity formula now. The wiki page currently says the gravity formula is (g-0.075)*5 instead of what Clamhat had posted. I'm aware that some things on the wiki haven't been updated, and this may well be one of them, so instead of asking which one is supposed to be right, I'll ask where Clamhat got his/her gravity formula from, whether from the game's coding or something else, and whether anyone else like KuroganeHammer could back it up.

Lavani Out of curiosity, which characters were you testing this with? Using the wiki's formula, there are still seven characters that would get a slight knockback reduction because their gravity is less than 0.075, but this is obviously less than what you'd get with Clamhat's formula.

 

[KuroganeHammer]

Clamhat got it from me over twitter direct messages I think

 

[Lavani]

I was just doing various gravity value edits on King Dedede and Mewtwo ranging from 50 to 0.001 a month or two ago.

An absurdly high gravity value like 50 makes characters just die instantly to any move that induces tumble knockback unless the angle is 0°/180°. Gravity isn't factored into knockback pre-tumble, so <80 KB hits still behave normally.

I didn't mess around a whole lot with the lower gravity values but I recall Dedede being able to die off the top at like 60~80% to whatever move I was using (probably RosaLuma usmash) with some of the lower value edits.

I'll do some more thorough testing editing Diddy's gravity and using Mewtwo's uthrow tonight I think.

 

[KuroganeHammer]

I think because units of knockback is a speed per frame rather than a length per frame that it creates wonky interactions with low gravity and falling speed.

 

[Funen1]

Then I suppose the (g-0.075)*5 formula on the wiki page was more a piece of leftover info from Brawl before anyone dug deep to find out what it was in Smash 4? If it's been confirmed to have changed, I'd be willing to edit in the new formula.

 

[Horseketchup]

Does anyone happen to know what the standard air friction and air deceleration values happen to be that are applied to the entire cast? Like on KuroganeHammer it lists the values, but apparently those are just the values that certain characters have on top of that base value. Just trying to get a sense of how big the value differences actually are across the cast when factored in with the base values.

 

[Lavani]

Alright, ran into several hiccups that delayed this, but I have some numbers for gravity now.

Tests were done with Mewtwo uthrow vs DK on Dream Land Ω with no DI, measuring kill percent (NOT red lightning %) for each gravity value given.

0.09505: 152%
0.08505 (default): 154%
0.07505: 155%
0.06505: 156%
0.06000: 156%
0.05505: 156%
0.05309: 156%
0.05000: 155%
0.04505: 154%
0.03505: 149%

It appears gravity values below 0.08505 do reduce vertical knockback, but only to a point. Kill percents plateau around 0.065~0.055, then begin to drop again at some point afterward. Since Jigglypuff is the lowest gravity character at 0.05309 and her gravity value still resulted in a 156% KO, this doesn't mean much in terms of real-game, but if there is a peak rather than a plateau it's possible Rosalina might get the most KB reduction from gravity instead of Jiggs.

I might look at this more later with a weaker move like Palutena jab to see if I can find the peak.

 

[Big O]

Alright, ran into several hiccups that delayed this, but I have some numbers for gravity now.

Tests were done with Mewtwo uthrow vs DK on Dream Land Ω with no DI, measuring kill percent (NOT red lightning %) for each gravity value given.

0.09505: 152%
0.08505 (default): 154%
0.07505: 155%
0.06505: 156%
0.06000: 156%
0.05505: 156%
0.05309: 156%
0.05000: 155%
0.04505: 154%
0.03505: 149%

It appears gravity values below 0.08505 do reduce vertical knockback, but only to a point. Kill percents plateau around 0.065~0.055, then begin to drop again at some point afterward. Since Jigglypuff is the lowest gravity character at 0.05309 and her gravity value still resulted in a 156% KO, this doesn't mean much in terms of real-game, but if there is a peak rather than a plateau it's possible Rosalina might get the most KB reduction from gravity instead of Jiggs.

I might look at this more later with a weaker move like Palutena jab to see if I can find the peak.

Logically I think it works this way because vertical KO resistance is based on weight and your current fall speed. With more gravity you reach your max fall speed more rapidly, but the initial velocity increases due to the gravity offset. With less gravity the opposite is true and the offset become negative instead of positive.

Just based on the math, there will be a point where the gravity is so high that you instantly die due to a ridiculously high gravity offset KB. Likewise, there will be a point where the gravity is so low that your reduced fall speed during the launch makes you die earlier despite the reduced vertical KB you receive from having low gravity. The most beneficial gravity value will depend on how high your fall speed is. The higher your fall speed, the more gravity in general benefits your vertical KO resistance since your maximum fall speed kicks in sooner and slows you down more rapidly I also suspect that the optimal gravity value will be slightly different based on ceiling height as well.

 

[Funen1]

Something else about hitstun calculations is throwing me for a bit of a loop. The formula for hitstun includes subtracting 1 at the end, but one move I've been looking into, Ness' D-throw, is being a bit inconsistent with this. For reference, Ness' D-throw deals its last hit on frame 50, and I'm assuming that that frame counts as one frame of hitlag (technically indistinguishable from any other frame, but it's still there, unless I'm mistaken). Against someone like Bowser, the full hitstun formula, including the gravity component, says he should be dealt 34 frames of hitstun from D-throw after that one hitlag frame I mentioned, and tests in Training confirm this. However, with characters like Mario, Luigi, or Peach, the formula says they should be dealt 33 frames of hitstun in Training, but their buffered double jumps don't come out until frame 35 after the lone hitlag frame, meaning they're also experiencing 34 frames of hitstun. I don't have any vids or anything right now, but it should be simple enough to reproduce. But unless my formula was mistaken (which I'm pretty sure I got right, after checking it multiple times), I'm not sure where this discrepancy is coming from.

I also have a question about the Sakurai angle. I know that the opponent is sent in different directions based on their percent and whether they were grounded or airborne, but just to make sure nothing else is at play, are these launch angles substituted into the knockback formula directly in place of the 361 value?

 

[Literal Bacon]

So I noticed, if you time an attack right after you fall off of a ledge correctly, you'll do a full hop instead, while also keeping your double jump, is that something new?
https://www.youtube.com/watch?v=G5pSgK4qaYg

I find it's easier to do if you dash toward a ledge and input a fair, if you set the control scheme to have A+B to smash on, you can press B to do aerials instantly too if your c stick is set to tilts.

 

[samosa]

So I noticed, if you time an attack right after you fall off of a ledge correctly, you'll do a full hop instead, while also keeping your double jump, is that something new?
https://www.youtube.com/watch?v=G5pSgK4qaYg

I find it's easier to do if you dash toward a ledge and input a fair, if you set the control scheme to have A+B to smash on, you can press B to do aerials instantly too if your c stick is set to tilts.

It seems like what is explained in this thread
http://smashboards.com/threads/weird-ledge-mechanic-jsc-jump-squat-canceling.437213/

 

[Literal Bacon]

It seems like what is explained in this thread
http://smashboards.com/threads/weird-ledge-mechanic-jsc-jump-squat-canceling.437213/

Ahh, it is, thanks.

 

[Literal Bacon]

It seems like what is explained in this thread
http://smashboards.com/threads/weird-ledge-mechanic-jsc-jump-squat-canceling.437213/

Man, should've posted this here way sooner, found it a long time ago.
How do I delete threads I created?

 

[Sonicninja115]

Man, should've posted this here way sooner, found it a long time ago.
How do I delete threads I created?

Contact the forum moderator and ask them for help.

 

[Literal Bacon]

Contact the forum moderator and ask them for help.

Who would I contact?

 

[ぱみゅ]

Use the "Report" button and fill up the "reason" field, any available mod can work it afterwards (most likely I'll do it if it is issued shortly), though it's unlikely for it to get deleted, the thread might only get closed.

Also, please avoid posting multiple times in a row like you did above.

 

[Ssbm_Jag]

Simple yet important question. I cannot find an answer on the wiki. In Smash 4, how early in terms of frames can someone drop ledge once they've started grabbing the ledge. I know in melee it's 7, but testing it out in smash 4, it feels no where close to 7 frames. How many is it?

 

[Vipermoon]

Simple yet important question. I cannot find an answer on the wiki. In Smash 4, how early in terms of frames can someone drop ledge once they've started grabbing the ledge. I know in melee it's 7, but testing it out in smash 4, it feels no where close to 7 frames. How many is it?

The answer is 20 frames. It can't be buffered though, so it's difficult to do it in perfectly 20 frames. The other options are available after the same amount of time, but they (other than normal get-up) can be buffered.

 

[Funen1]

KuroganeHammer I have a quick question about how Shulk's "knockback received" multipliers are applied. Going back to Clamhat's quoted post above, where exactly does the knockback multiplier for Shield and Smash fit into the formula? Is it when calculating each individual component (i.e., before the components are squared) or after doing the full Pythagorean theorem for the "gravity-adjusted knockback"?

Also my previous post still stands if you or anyone else could help out.

 

[KuroganeHammer]

It's a multiplier on the result of the formula.

 

[Rango the Mercenary]

Can we go over all of the characters that can chain Fairs offstage, possibly for an early percent kill or gimping?

Sonic
Shulk (Jump)
Pikachu

These are all the ones I know can use Fairs repeatedly offstage for a chain. Some of them I don't believe are escapable with airdodge until they run out of jumps. Can anyone add to the list of characters (if I missed any) and how to DI out of them properly?

 

[Sonicninja115]

Can we go over all of the characters that can chain Fairs offstage, possibly for an early percent kill or gimping?

Sonic
Shulk (Jump)
Pikachu

These are all the ones I know can use Fairs repeatedly offstage for a chain. Some of them I don't believe are escapable with airdodge until they run out of jumps. Can anyone add to the list of characters (if I missed any) and how to DI out of them properly?

Did you see 9B's Mewtwo combo? Most of that was true. Mewtwo can chain 4 or 5 fairs in a row, the only escape option is by DI'ing away and this option is only open for the medium heavies, not the upper heavies or supers. I have video proof.

 

[Vipermoon]

It is dificult, but possible, to escape from Cloud and Marth chaining Nairs and Fairs, respectively.

 

[thehard]

Let's talk about foostooling off of a shielding opponent, a mechanic I was aware of but haven't thought of in practical terms until now

I can see it being a nifty mixup option when your opponent is looking for a telegraphed landing aerial to shield grab. If you input the full jump you can escape retaliation. Inputting the short jump seems riskier but if your opponent isn't prepared you could dair them or something (Megaman seems like a good pick for this). What about burning your double jump and surprising your opponent by bounding off their shield anyway?

You can also cancel an opponent's grab animation with a footstool, and they activate on F1.

 

[I speak Spanish too]

Is it possible to osicilate your foxtrot cancels as frequently as Bidou without bidou?

 

[Sonicninja115]

Do we have Roll distance rankings? And does roll distance directly influence RCG (Roll Cancelled Grab)? I am trying to find out the mechanics behind Boost grab and RCG. I am pretty sure RCG distance is decided by roll length and Boost grab is decided by speed and/or traction.

 

[Yikarur]

Is it possible to osicilate your foxtrot cancels as frequently as Bidou without bidou?

Yes, most Falcon players for example are doing it all the time. Just needs a bit of practice.

 

[moofpi]

Do we have Roll distance rankings? And does roll distance directly influence RCG (Roll Cancelled Grab)? I am trying to find out the mechanics behind Boost grab and RCG. I am pretty sure RCG distance is decided by roll length and Boost grab is decided by speed and/or traction.

Roll Comparisons
https://youtu.be/kkWdfYfBDDY?list=PLp-1erL0_ll25oeqkp82K9KXUNA_lJA4c
DA Cancel Grab and Roll Cancel Grab
https://youtu.be/Nt8lNOpXt6U?list=PLp-1erL0_ll25oeqkp82K9KXUNA_lJA4c

If you have any more questions about them, I'm sure @TheReflexWonder would have the answers you seek.

 

[Jams.]

So, this Twitter post about Luma's uair recently came out, and there's a gigantic difference in kill percent between horizontal DI and diagonally downwards DI on the 108° hitbox. I was wondering if this is due to this phenomenon described by KuroganeHammer , where diagonal downwards DI moves the launch angle into the vectoring threshold. In that case, the difference in kill percent can be explained by the fact that diagonal downwards DI results in both a more favourable launch angle, and a raw reduction in the move's knockback due to vectoring.

If that is the case, it seems like proper DI can result in gigantic increases in survivability for vertical kill moves within a certain range of launch angles.

 

[Lavani]

65-115° is the angle where "VI" doesn't work. DI would put it at 118°, which means "VI" would also take effect and reduce knockback.

And yes, this is why moves with a ~70° angle like Charizard's uthrow get gutted so badly by down-diagonal DI, since it's both optimal DI and reduces knockback on top of that.