Technical discussion archive

[Chesstiger2612]

Contains the conversation "PC Jump Drop, Buffer shield, potentially other questions"

Do you know how it works in PM? I can do it after Ike's ledgejump on WarioWare (for example) but not with Marth's full jump on PS2.
Also it would be interesting if there is always the one frame for this technique or if some characters jumps are too fast so that they just don't are in the area they need to be in on any frame.
Unfortunately, I haven't had enough success with my testing yet to find any general rule yet, maybe it is also stage + character + jump (full jump, short hop or ledgejump) dependent.

One other question I would like to ask you: Are there any good ways to shield with grab. If you buffer grab out of any lag animation you will shield. This might prove useful if you get shieldpoked and want to keep your techwindow open, similar to grab for L-cancel. I know you can also light press shield in PM but if lightshielding gets implemented or in Melee (although I don't know if grab buffers lightshielding there) it could be the only way to full shield without using your tech window.

I was away for a while, sorry.

What's pc jump drop? Never heard about that one. Regarding shielding with grab: I nowadays tech with grab, so I think that holding grab might also close the tech window.

Welcome back, mate

For the PC jump drop, I'll just quote the explanation I found:
From phanna's Platforms 101:

• PC Jump Drop - Essentially, if you midair below a platform at a point where you can land on it, you'll either go high enough to impact land on the platform (seen in the Normal Platform Drop), or go low enough that you'll just wind up on the platform without impacting (seen in the No-Impact Drop). If you happen to midair on the one frame that divides these two, you'll just pop up the platform immediately, and have no lag at all. Nifty, but mostly it's a flash mindgame - essentially an advanced version of the No-Impact Drop

Here is also a video example.

I just try to note down every time it happens to me (the Ike ledgejump on WarioWare was the first time I noticed it), but it is really hard to predict if it is possible with a certain character on a given stage, maybe its always possible but from my tests that doesn't seem likely.

Grab does indeed acticate tech window, you are right. So lightshielding might be the alternative. In the case of Melee, do you know if the border between lightshield/fullshield is a different one than the tech. If you would need to input more for a tech you might fullshield without activating your tech window.

Hmm, to me it seems that the pc jump drop is nothing more than no impact landing from as close to the platform as possible. The no impact land happens when you land during a jump or dj animation, so I'd think that it could be done with any character.

However I think the pm landing detection is still quite different from melee's, and I don't know how the ECB's generally work in pm, so it might not be possible with all chars (For example if the ECB of the char becomes too small and is oriented too high during the dj / jump). Also I don't know at which rate the ECB gets updated, so the previous actions you have done might impact your success. I seem to get no impact lands way more often in pm though, with marth especially.

I assumed it was the ECB stretching during the jump
("if you midair below a platform at a point where you can land on it")
and also stretching downward so you land immediately. It can't be just a no-impact-land because a perfect no-impact-land would make you land as your vertical momentum becomes negative in the slightest, while a PC Jump Drop can also happen while you do have upward momentum.

Also in te case I got it, I pressed jump and just instantly landed. Maybe some characters also don't need to jump to stretch their ECB enough but with jumping it can work, at least in my example.

It is a similar effect like the autoland aerial, which I always tried to explain with Brawl landing detection, meaning that some kind of hurtbox part (probably the lowest that triggers landing detection) moves down while you move through a platform causing you to land as soon as you pass through it. It made sense visually, because it fits with the animation. You could test it with many moves, for example Wolf's dair, MK's nair, Snake's dair&up-air.

I still assume the PC Jump drop works with ECB (because it was in Melee and there wasn't Brawl's landing detection), while autoland aerials probably work with that landing detection.

We need a debug mode for PM.

I've got a question regarding resets.
Like in Melee, the resets are happening if a character who failed to tech and is still in his respective "failed tech" animation gets hit and now lands during a "got hit in his failed tech animation" animation which seems to last 13 frames (like in Melee iirc).
In theory the same is true if you hit a character on the ground who isn't in his failed tech animation anymore if he doesn't buffer a roll, getup or getup attack anymore. The animation has the same length but in this case the resets never seem to work.
Do you receive less knockback when hit in the failed tech animation, is there some kind of automatic DI in the second case or is the landing detection different?
This also might be the reason why resets work different in PM.
As example: While Marth can jab reset a Marth with no DI (inner hitbox) until 35%, it doesn't even work with 0% if he isn't in the failed tech animation anymore.

About the Wario Ware ledgejump, it might have to do with that specific ledge because I could do it there with Link and Pit too.
Apparently it works with normal jumps too, I performed it with Peach on PS2 (without aerial so a DJC aerial cannot be the reason).

I played around a bit with resets, so i'll post some observations:

If you get hit during DownBound (the missed tech animation that lasts 20+ frames) or DownWait (The state after DownBound during which you can do gu options) by a move that doesn't knockdown, you'll be put into the 13(?) frame DownDamage. In melee you would be put to DownDamage if the move does less than 7.00% damage, so that's a bit different. So in pm you can reset with any spike or move that doesn't send you too high that doesn't cause knockdown. Other difference is that if you get hit during DownDamage, you'll be put into regular hitstun, while in melee the DownDamage simply starts over again. This means that resetting with more than one hit, for example hitting twice with foxes drill doesn't work anymore, and neither does double jab counter sdi out of reset, because the second jab won't put you in DownDamage.

This last part includes a bit of guessing: If you are grounded on the last frame of DownDamage, you'll be forced to do a gu action. If you are still in the hitstun while the DownDamage ends, you can choose the gu action which the happens after the hitstun ends. If the hitstun ends before DownDamage, you'll do a gu stand. I don't know whether the end of DownDamage while you are in the air cancels hitstun like in melee. Also I think kb works normally on char that's in DownWait or DownBound.

I tried that marth jab reset on marth, and I was able to get the jab reset at 34% on DownWait consistently. It was a bit hard to hit with the weaker hitbox because in the DownWait the marth lies so low that you'll have to be at a specific spacing. It's easier to hit with the weak hitbox during DownBound.

I don't know whether this is known, but I noticed that platform drop animations don't have aerial speed limits, and they don't slow you down at all, so you can go faster in the air with some characters by wd or run into platform drop. With luigi the effect of wd -> drop is exactly like in the vudu dash.

Aren't the first two contradicting each other? A spike lets no room for the DownDamage animation and the character goes into the hitstun on the ground animation if it is no knockdown. So the opponent will never be limited to one tech option.

My tests suggest that the opponent chooses a getup option without remaining hitstun applied, at least for spikes and meteors (animations on Wolf's and Falco's dair and jab reset seemed identical or at least not at a difference as big as the hitstun the dairs normally would have caused).

The end of DownDamage doesn't cancel hitstun, however the end of DownDamage doesn't even cause tumble if the hitstun is below the duration of the DownDamage animation.

Nice find on the platform drops!

I don't quite understand you. What is contradictory? If you spike with a move that doesn't knock down on player that's in DownBound or DownWait, he'll be put into DownDamage.

Forced get up stand still exists if hitstun is less than 13 frames, doesn't it?

Oh sry I forgot to remove that paragraph, misunderstood something something.

So as result, it is a good way to use a spike/meteor with less knockback than the environment collusion threshold because unless they input a getup option in hitlag, they will be forced to do a regular getup.

Or if the spike is weak enough they'll be forced to gu stand regardless.

I think I've got the data complete for your question in the statistics thread.
Fixed a few mistakes in it today

If you need any more information on the dashs and foxtrott I will search for the values.

Another question, fo you know how throws work? Pretty much any knockback rule contained the obligatory "except for throws", and there is this weight-dependent throw release point (?) ruining my current efforts to predict which possible followups are guaranteed and which not.

About shield pokes, if the same hitbox hits both the (uncovered) player and his shield, will it shieldpoke or not?
I saw many examples where players with bigger shields got poked and players with smaller shields didn't.
The other possible explanation would be that the graphical projection of the shields isn't representing its technical size accurately, or that the graphical projection of shield tilting doesn't represent its effect.
Or a hitbox that just hits the shield has priority

In melee the shield is prioritized over hurtboxes, meaning that if a hitbox hits both, it'll get blocked. In brawl it's the other way round, and I think so it's in pm, but I'm not sure.

In melee the shield is prioritized over hurtboxes, meaning that if a hitbox hits both, it'll get blocked. In brawl it's the other way round, and I think so it's in pm, but I'm not sure.

I had actually both cases where it should have hit both and either hit the player or the shield, so I assume it checks the hitbox with the most priority of those hitting (if multiple are hitting at the same frame).
Not sure yet what it makes if this hitbox also hits both, I might need to test with a move having one huge hitbox so it is secured both shield and player are hit.

You are right, the player gets hit (tested with Lucas up-smash and Zeldas up-air, they have pretty big hitboxes). It might be useful to know the hitbox priorities on hits often used for shield pressure.
Those moves with inner hitboxes having priority which are more likely to hit the outer player are better for shield poking.
Moves with many, small hitboxes have a disadvantage (opposite to Melee) because they are more likely to only hit a shield.

Another thing, it seems to be impossible to truly CC (no hitstun) any meteor/spike in PM. You will get less hitstun (probably 2/3 because thats how it works in Melee if an attack is too strong to truly get CCd). Exception is Bowser because of his light armor.
Tested with several spikes&meteors, for example Fox dair on DK, 0%, which he would normally be able to CC iirc.

Yeah hitboxes are always checked in some order and that can affect whether a poke happens or not.

The no hitstun cc is landing via asdi down after being launched into air, and landing animation replaces hitstun. Since forbidden A/S/Di applies to any move that sends at 0° or below angle, no spike can be land canceled in pm or melee when grounded.

I'm not exactly sure when landing out of hitstun happens: when you shine, jump and dair with falco, land canceling (almost) never happens. However if you jump into falco's dair that doesn't knock down, or if the falco utilt dairs, land canceling occurs always.

What do you mean by landcancelling, having no hitstun when landing?
I tested it a few times and which move I used before seems to make no difference to when the hitstun cancels, only the ECT of 80 knockback units makes a difference because from then on knockdown is caused.

Landing while in hitstun is what causes "no hitstun" in melee, so I call it landcancelling. That's not in common use though, I've heard people call that pseudo cc or whatever. What did you test with? In melee it's really hard to get anything to not cancel except falco's shine for some reason. It might be something kb or tumble related, because the landcancelling sometimes doesn't happen in team combos either, like samus dtilt into quick falco dair.

I tested with Falcos dair, with various moves comboing into it. Also tested shine -> other moves.
As far as I understood&tested it, knockbacks over the ECT cause tumbling, others don't, landing in tumble causes knockdown, landing without tumble causes landcancelling or hitstun while grounded if hitstun is left.

Didn't see any exceptions yet, but if there are some I have a possible explanation:
If you dair with Falco during the hitstun from the shine, the real knockback should just be reduced slightly because of knockback stacking and because the knockback of the shine is almost used up.
However, it could very well be that the game adds up these two knockbacks as vectors when determining the opposing player's position/momentum/etc. (so direction gets taken into account), while for the tumble criteria the knockback just gets added up and is enough to cause tumbling.
Just a theory...

I first also thought about adding up hitstun but then drill-type moves like Fox' dair and Pit's nair would let you follow up with anything.

Wait does uptilt dair actually work in pm?
It sure doesn't in melee.

Not as combo, just for testing I tested on "Control" CPUs in practice mode

So it land cancels?

It does cause tumble later than shine.
The real landcancel only works if hitstun ended, at least I have tested multiple setups and followups and this rule always holds still true.

Could it be that you are in knockdown when a spike/meteor is used against you in grounded hitstun?
Pretty difficult to test because moves are normally too slow to hit again before grounded hitstun ends

Powershields get less shield pushback in PM, while it was more in Melee.
Has good and bad sides, harder shield pressure escapes but powershield punishes get easier

less than regular shield or less than Melee

Less than both. Like this:
Shield pushback:
Melee ps > Melee/PM shield > PM ps

Less than both. Like this:
Shield pushback:
Melee ps > Melee/PM shield > PM ps

I think that's usually done completely by shield asdi, while the marths up b actually pushes you toward the stage.

Btw I've found the frame data on no-tech. Most characters' animation is around 25 frames.

In melee it's 26 frames for every char on floor. Here is more information: http://smashboards.com/threads/detailed-throws-techs-and-getups-frame-data.206469/#post-5896012

Also tech in place is also 26 frames, which means that same options should cover both TIP and DownBounce provided that they hit low enough and late enough.

Then it seems it is a bit shorter in PM, assuming it is the animation "DownBoundU" and "DownBoundD" (probably differing between back and stomach). I still wonder what animation name means what in P:M.

I wonder if some characters' air animations are different enough to allow things like:
Drop through a platform -> WD (similar how you can drop through platform->dair with falco and you will land on the platform if you do it frameperfect)
WD so you fall off a platform the stage -> WD back on

Yeah those work in pm with most chars I think.

Those making their landing detection higher when jumping can do this and those lowering it have an easier time doing automatic landing with jumping on the right frame.
Probably there are some characters inbetween who can do neither.

Standardtoaster sent me some data today answering some of the above questions:
The frames between consecutive dashes have to be at least 17, so CF, DK, Marth, Peach and Roy can do their foxtrots perfectly with if they cancel it on the last frame of his initial dash and then have one frame of slowing down before inputting the next dash.
Powershielding has 70% of normal shield pushback while it had (roughly estimated) 150% in Melee.

Do you know more about landing criteria in PM? Has it to do with bone landing detection or with ECBs?
There are so many possibilities still left in platforms and the weird landing habits of some characters.
I might want to work on a list for autolands, dropcancels etc...

No I don't unfortunately.

Drop cancels are very strange, there are more factors influencing it than one would assume. I got two or three randomly out of when I tried them.
Then I tried to get them consistently by TAS and didn't get a single one.

Maybe you should invite more knowledgeable people

Yeah but most of them are either PMBR (I wouldn't want to delay their work anyhow) or not as interested in PM. Do you have any suggestions?

Aerodrome, 1MachGo, Xeylode, Kink-link5, drinkingfood, Sartron, br8kage

8adge, BlackRainnfire, Gmaster171

Actually I'll just invite all of them now thx
EDIT: I found no user named br8kage. Also it turns out that you have some kind of invitation limit

umm idk off heart the answers,

idk when I'll look either because was at tournament this weekend and am at work now and I'm pretty sure I am simply going to DIE

We are in no hurry take your time. And thanks for participating.

Actually now that I think of it, I don't think pm uses bone landing detection, because quite often different parts of the body move quite far inside stage before the character lands. Also I think marth can ledgedash so that his feet literally go through the stage.

Sadly I can't do testing right now since my pm setup broke.

What are the questions, I have lost track over the evolution of this conversation

I think chess is currently most interested about this.

I would like to know exactly when landing after being hit happens: For example, if you shine dair with falco, the opponent almost always goes into grounded hitstun, but if you utilt - > dair or if the opponent just jumps into your dair, he will instead land and the landing animation cancels hitstun.

It might be that some bones count while some don't, because in some cases its fitting pretty well (many dairs where characters instantly land when doing the downward motion), but it could also be that the ECB is just

Yeah so I can do all frame by frame tests if you tell me what to do

For the reverse case (dropcancels) I got a few contradicting results.
I was testing on Battlefield with Falco's bair.
2 frames down are obligatory for the drop
If I input the aerial on the third frame, it will do a smash instantly (so it counts as grounded still)
If I input the aerial after that either nothing will happen (2 frames down, two frames neutral, aerial)
For one frame neutral position or three Falco will just drop through the platform in neutral. I don't really think anything changes for more than three frames.
Still, I got some drop cancels when randomly testing in real time.
Just realized I always got the dropcancel after hitting someone, why could that be?


Just as note, it started with me asking for the reason Ike can autoland after his ledgejump on Wario Ware's platforms by doublejumping. That brought us to the entire landing topic.

landing detection is done with knees,

As far as I know "land cancels" occur when you hit the ground in hitstun that would not have put you in tumble.

If you get hit by greater than or equal to 80 knockback, you will go into tumble after hitstun. If you hit the ground in tumble or in hitstun that would have sent you into tumble, then you will be put in a knockdown state.

One of the things that might be affecting this is that if you hit a character who is still in hitstun/knockback, and it's been greater than (or equal to?) 10 frames since the last hit, then the knockback vectors will combine. So for shine -> dair, if you hit them with the dair 10 or more frames after the shine, but before the knockback of the shine is done, the knockback vectors will combine (not "the remaining knockback" or something, the full initial launch vectors of both moves). In this case, this actually decreases the knockback of the move since the vertical components of the vectors are pointing in opposite directions.

One question I have on what I said above ^ if anyone knows is, is it greater than or equal to 10 frames, or just greater than 10 frames?

Also in this example, since the knockback is reduced when the vectors combine, does this also mean that the time in hitstun is reduced to the match the new knockback vector? I would assume so, but wondering if something weird happens like them taking full hitstun of what dair ignoring the knockback adjustment.


I had heard that it's when the shin's touch the ground, but knees might well be right. Regardless, it is when a particular bone on the character model comes in contact with a ground surface. Presumably it also requires you to have less than or equal to 0 vertical momentum to land when in addition to the contact.

I do know that this is one of the reasons it's so easy to do long wavelands out of Ganondorf's double jump, since he balls up in his roll of his double jump, and when you air dodge he shoots back to a vertical stance which shoots his legs down (making contact with the platform). As a result, the spacing for doing a completely left/right air dodge and having him land is more lenient when he's in that balled up double jump animation. The same is probably true of a number of characters.

I'm not really sure how any of that affects PC drops ect honestly. I'm guessing PC landings happen when you land with 0 (or close to 0?) vertical momentum, but that's just a guess. That would mean an animation where the legs lower near the peak of a jump height might make PC landings more lenient to space(?) And I would guess a lot of double jumps would fit this description (legs lowering around peak of jump).

I just asked about kb stacking in melee boards:



Also I'm 99% sure only the remaining kb is used for the 1st hit. Kb takes quite a bit of time to decay, and falco's shine -> shine would be ridiculous if the 1st hit kb refreshed when the second one hits. Also things like utilt chains would never work.

Hitstun/Hitlag will always be according to the last hit kb.

In case of shine -> dair, the dair doesn't put you in tumble but you still don't land. Also in melee I think I have had a scenario where shine -> dair had bowser land after the dair.

https://www.youtube.com/watch?v=z8qQ1jzw3ZE

This?

Oh yeah, I forgot about this video. And yeah, this.


Hmm yeah I think you're probably right. Not sure why I thought otherwise. Would be good to hear some confirmation on this if someone knows for sure, but the Falco shine -> shine example makes me think you're probably right.

Do you have any explanation why the dropcancels (drop through platform->aerial->land on the same platform) only worked when the aerial hit someone?

Do you have any explanation why the dropcancels (drop through platform->aerial->land on the same platform) only worked when the aerial hit someone?

Hmm, I just found out that a very staled shine -> dair so that the dair is hit within 10 frames of the shine doesn't cause bowser at 0% to land after he touches ground (melee).This makes me think that you have to have enough downward kb or just downward speed for the landing to occur, and that might be the only requirement. The reason why it's so easy to avoid opponent landing after shine is that he has so much upward kb left to work against the downward kb.

@ Chesstiger2612 That's exactly how they work in melee too, you need to hit with them. I don't know exactly why though. Maybe the landing detection system avoids you landing immediately when you drop by disabling landing for x frames after drop, and it doesn't account for hitlag.

The vectors are only added in the case that they are opposing vectors, such as up and down or left and right. Otherwise, the second hit overrides knockback even if they were still in hitstun, regardless of the ten frame window. When the vectors are added (meaning they oppose each other and subtract from each other's strength), hitstun is calculated first from the full knockback of the second hit, but the launch speed is calculated after the knockback addition (subtraction). In the case of shine->shine or utilt strings, the moves send in the same direction so their knockback is not added, the second one just replaces the first.


Basically ECBs don't refresh every frame, they change shape after a certain number of frames (i don't know what determines that though). When you drop through a platform the ECB changes shape so that is now below the platform as soon as you start dropping though so that you don't immediately land back on it again. So if you do an aerial, the ECB doesn't change to reflect the aerial's normal ECB right away, it takes a few frames. If it had changed as soon as the aerial came out, you'd land on the platform every time you dropped and nair'd, assuming the aerial has an ECB placed higher than the drop-through's ECB. Hitlag gives just enough time for the ECB to change to the aerial's respective ECB, so if the aerial is quick enough such that it can hit while the aerial's ECB would still be over the platform, the hitlag gives it the time to change to that.
But I'm not 100% sure that's how it works in PM, but aerodrome just saying that people land at the knees doesn't really account for all the other situations in which ECBs are needed to determine collisions. I figure brawl and thus PM have ECBs in the same manner too, and they are now being adjusted or have already been adjusted to match Melee.

Really good answer, thank you.
I also got the impression landing criteria can be different from situation to situation.

About dropcancel aerials, I've made a list of possible ones. For some of them you need to hit with the hitbox with the highest damage to get enough hitlag for the ECB to refresh.

Spoiler

Falco: nair, bair, dair
Fox: nair, bair
Ivysaur: dair (doesn't need to hit because of its upward push)
Link: nair
Luigi: up-air
Lucario: dair (doesn't need to hit because of its upward push)
Mario: up-air
Marth: fair, up-air
Samus: nair
Squirtle: up-air
Yoshi: nair, up-air
Zero Suit Samus: up-air

What exactly does opposing mean? Is the entire vector split into a vertical and horizontal one and it is checked if they are opposing, or does ist mean they differ by less than 90° ?
Lets say a character's hurtbox collides with a hitbubble with a 45° and one with a 135° angle, what would happen?

This is exactly how I thought it worked before asking Strong Bad, who gave different answer, so are you sure? And to be more presice, doesn't the game check both x-axis vectors and y-axis vectors, and if they are same, max is chosen and if they are opposite direction, they are added. Then the victim flies at x + y.

Yeah when I say opposing I mean after they have been broken down into Y and X vectors. The knockback angles have to be in different quadrants to have opposing vectors. 40° and 80° angle hits in the same direction wouldn't stack for instance. But 40° would stack with 40° in the opposite direction (which would just be 140° if you're assuming a static orientation). Even two downward hits could stack if their horizontal components were in opposite directions, but I can't really imagine a situation where this is likely to occur.
And yeah I forgot, if the first hit is stronger (and they are more than 10 frames apart) its knockback is used while the hitstun from the second is used.
So a generalized rule would be, when the hits are more than 10 frames apart, vectors are added if they oppose, otherwise the strongest is used, and hitstun from the second always replaces hitstun from the first.

The stacking must be done by components. Otherwise you would see shine di'ed to left -> shine di'ed to right do way more kb than two shines with no di, and it would be silly.

Yeah, only the opposing components are added. Didn't realize I didn't make that clear.

So I assume the effect of landing as pressing jump has to do with your ECB shape getting refreshed in the next frame after you pressed jump where it stretched, and you would need to be passing through the platform before too.
Is that reasonable?
And if yes, knowing the criteria for ECB shape refreshes would be the next step.

Aside from landing with jumps, landing with autocancelled aerials is also interesting, and at least some of them seem to work more reliably (on every stage)

I think I figured out some of the mechanics behind landing with jump.

Pressing jump always refreshes the ECB shape if you have a jump left and at least under some conditions if you have no jump left (if not always).
Another possibility is that it is just Brawl landing detection which only gets checked under certain conditions.

Now with the new animation, the character lands or doesn't land, depending on these factors:
- new relative position to the platform
- old relative position to the platform and old animation

I haven't found if that speaks for ECB or landing detection, because ECBs could have been shaped differently in the earlier animation and certain bones can move very abrupt between certain animations so it could also be landing detection.

But the previous animation has influence, for example ledgejumps make jumplanding much easier (probably because characters aren't standing straight), while only a few characters can jumpland with their DJs without ledgejump animation:

List of characters that can jumpland with their ledgejumps on WarioWare

Spoiler

Charizard
Ike
Link
Lucas
Marth
Peach
Pit
Wario
Yoshi
Zero Suit Samus

Lists of characters that can jumpland out of their standard air animation

Spoiler

Peach
Yoshi
(maybe their double jumps being slow helps here to trigger another ECB refresh, or they just move down slightly in the beginning)

Being in the jump animation still (especially impoetant for multijumpers) could also make differences but I haven't experienced any of them yet.

Aside from the missing parts, there is one thing I am not so sure about but it might be true...
Could pressing jump while not having a jump trigger an ECB refresh and potentially a landing because of that? It might be that it goes through the refresh loop before it checks if jumps are left. I think I could have seen such an "emptyjumpland" but it could have been a weird-looking no-impact land, too.

Has anyone more information?

What factors the boost grab window?
For example G&Ws boost grab that can bring out a hitbox before the grabbox, while his dash attack or grab isn't faster than other characters'.

That's a good ****ing question. I'm upset by the fact that ROB does not get a boost grab despite getting a DACUS. Wish I knew why so I could bitch at PMBR for such a dumb choice.

Dacus is affected by jump squat so it would be logical that boost grab is also

But apparently it doesn't, so I have no idea

That can't be all of it though because G&Ws dash attack is frame 5 (correct me if I'm wrong) and his jumpsquat is 4, so why does the hitbox come out?
Is there any kind of variable that determines if the hitbox gets kept even in the grab animation (sounds absurd)?
Different characters seem to have different windows (I wonder about the same thing with the gatling combo, jumpsquat also can't be the only reason for that)

G&W's dash attack comes out frame 3 (frame speed modifiers and stuff). Boost grab window are the first 3 frames of DA, so G&W can't get both a grab and a DA hitbox without hitting an opponent.
On the other hand, it seems like hitlag counts as the frame you hit on for the purpose of canceling your DA, i.e. if you hit on frame 3 with DA you can cancel it into both UpSmash and grab during hitlag (staying in hitlag, but performing the first frame of the corresponding animation during it). This is what allows G&W to cancel a DA into a grab/UpSmash after hitting.

If certain aerials can also lower your ECB, maybe you can use one during an empty hop to land a couple frames sooner and have it autocancel and get the grab out faster; for a faster tomahawk

Chesstiger, remove Xeylode and add this guy
http://smashboards.com/members/frakture.282608/

or just ask me

and be like

"aero our goddess pls add frakture"

I DID IT

Thank you for recommending @Frakture , he seems like a very knowledgeable poster.

About your question:

I am pretty sure I have got an old post covering moves that will make you land faster, however, most of these have their "accelerated landing phase" (when the knees that cause the landing detection are lower than in the noraml falling animation) after some amount of frames and not instantly. This normally leads to the move not being in its autocancel window anymore which negates the one or two frames you might have saved with that.
Testing is somehow hard because the differences are often subtle and might only be noticeable by framecounting.
To start with an example, I tested Marth's dair of which I know it instantly shifts landing detection down. It feels like a difference by one or two frames but it might be placebo.
With floaties it obviously maks a bigger difference and with Ivysaur's bair I'm sure that it makes SHFF faster, also its autocancel window is pretty long.

Notice though that in a SHFF you spend most time on the rise and the fastfall happens rather quick, so I think that it will mean some characters can probably speed their SHFFL by one frame up and some even two (maybe three), but not more and even then, it requires good inputs. The small things add up, so I consider this useful and will probably follow up with more research.

Ike can jump land on the lower platforms of Dreamland if you do a backwards fullhop and DJ on the 12th frame of airtime.

Also, on the topic of platforms, did you know that if you try to shield and shield drop on the same frame, you'll do a wavedash instead?

That only occurs with some characters, frakture

Really, what determines it?

Probably whether or not the bottom of their ECB moves back up through the platform when they airdodge
I haven't looked into it in detail in debug mode yet, I just remember not being able to perform it with more than a few characters last I checked

Probably whether or not the bottom of their ECB moves back up through the platform when they airdodge
I haven't looked into it in detail in debug mode yet, I just remember not being able to perform it with more than a few characters last I checked

That makes sense. It worked with all of the characters that I tried it with so I thought it worked for everyone. I'll look into it more.

I just tried it for literally every character and it worked for all of them, which makes sense because each of the characters' ECB start out on the platform, and because PM has a 1 frame physics delay, they can't possibly move underneath the platform before they airdodge.

The shielddrop wavedash thing seems hard but useful because you don't need the jumpsquat frames.

I tested it with buffering a shield drop direction out of powershield/shield but then it performs a normal shield drop.


Sonic has a similar thing too, where holding a sideways direction is enough to land on a platform while your jump would normally be higher. It is best noticeable on the standard transformation of Frigate Orpheon when you try to run off athe stage and DJ ASAP.


On another note, gatling combos don't work any more in 3.5.

The thing about shielddrop wavedashing is you can't get a full wavedash angle out of it because you're limited to the maximum shielddrop angle. The reason it doesn't work when buffering an input is airdodge only occurs on the frame you press it. If you buffer shielddrop out of any animation it will still perform the shielddrop; you have to press or re-press the buttons on the first active frame.

What do you mean gatling combos don't work? Like, do the hitboxes not come out, or does the hitbox not combo into upsmash? Because I'm pretty sure that G&W at least still has one.

Yes, but G&W is the only character that still has a gatling combo. It somehow relates to the attack frame of the dash attack, and the Gatling combo now has the same requirement as the dash attack cancelled grab which also only G&W can perform. In 3.02, Ivysaur, Falco, Wario and Jigglypuff could Gatling Combo, and in 2.6 even Sonic.

I calculated that you can get around (+- 3%) 3/4 of your WD length with the shield drop one. Still pretty good.

I didn't know that so many characters had gatling combos. What are the requirements for dash attack canceled grab? I know Ike can do a reverse dash attack canceled grab and still maintain momentum, but if he doesn't reverse it the grab has no extra momentum.

That's pretty impressive that you can get so much distance from a shield drop wavedash, considering how restricted the angle of the control stick is.

With the DACG I'm refering to the phenomenon where you get both the hitbox and the grabbox, and I don't know all requirements, but one is that the hit comes out at frame 3 or earlier. Probably the second requirement is a minimum of 4 on the jumpsquat frames (because Sonic could never DACG), but I'm not quite sure on that one. G&W is the only one that can do it afaik.

The reason is that you also go down with the control stick to reach the y-values for shielddropping and therefore you can get the same angle, and the same distance, as if you wavedashed in one of the lower diagonal directions. The specifics of the trigonometric functions also help you here, so to speak. If you want to be very sure that you don't roll, you will probably only get 70% though, but it will be more consistent.

With the DACG I'm refering to the phenomenon where you get both the hitbox and the grabbox, and I don't know all requirements, but one is that the hit comes out at frame 3 or earlier. Probably the second requirement is a minimum of 4 on the jumpsquat frames (because Sonic could never DACG), but I'm not quite sure on that one. G&W is the only one that can do it afaik.

The reason is that you also go down with the control stick to reach the y-values for shielddropping and therefore you can get the same angle, and the same distance, as if you wavedashed in one of the lower diagonal directions. The specifics of the trigonometric functions also help you here, so to speak. If you want to be very sure that you don't roll, you will probably only get 70% though, but it will be more consistent.

OK, I thought you meant DACG in general.

I know what you're saying about the wavedash angle, because you should be able to input a 45 degree downwards angle on the control stick which would give the horizontal velocity a magnitude of around 70% of the maximum, but because distance is proportional to the velocity to the second power, cutting the velocity to 70% cuts the distance to 70%^2 which equals around 50%.

You are right, my fault.

Sonic can do a Gatling Combo. In both 3.02 and 3.5 you can cancel your DA during hitlag into grab/UpSmash as if you were performing a DACUS/boost grab. In 3.02 the hitlag counteds as the frame you hit on with DA for those purposes, but in 3.5 it seems like it counts as the following frame for DACUS and the same frame for DACG (Sonic can Gatling Combo after hitting his DA on frame 2, but also grab after hitting on frame 3) the DACUS window was moved to be 1 frame earlier.

In 3.02 Sonic didn't have a DACUS due to his 3 frame jumpsquat, so he also couldn't Gatling Combo.

Rectification: It looks like DACUS-window is now frames 2&3 of DA instead of frames 3&4 like it was before (for 5-frame jumpers.) That explains pretty much everything.
(It also looks like you can't do the UpSmash on frame 2 with the C-Stick and like debug-mode doesn't recognize different attack inputs on consecutive frames as different, so you can't really test it without other means of frame advance.)

@8adge Would you be willing to Dolphin TAS record a secret technique Frakture discovered for me to use? I was thinking of including it in a larger video for PMDepot, unless you wanted to do something similar with it on your channel

@SpiderMad Why is it secret?

I counted the 1 frame sooner induced gravity from instant aerials/specials in Melee. It lowers your SH and FH only 1 frame from what it looks like; which I didn't think would be that apparent (good old constants?).

So if PM was given this back from Melee:
-an Aerial that Auto-cancels, reaches its IASA, or gets a 2nd aerial's hitbox out given you do it the first frame in PM would thus not work anymore.

-Your window for any Aerial to Action that requires the first be done on the 2nd frame or later is lowered when done on the first making as if you did it on the 2nd; and so on for 3rd..4th.. all being as if you did it on the 2nd.


I thought it was cool enough to shock you all a little bit, which it will when I complete the video now using 8adge's recordings (giving credit to him and Frakture of course).

And then after that; I want to make a little informing video on the physics delay impacting PM.
- Causes your first airborne frame to be the exact location as your last frame of jumpsquat vertically (your horizontal momentum is still carried); overall making you have more air time in PM by (1?) frame (and take longer to reach your jump's apex?). And I guess aesthetically would make all characters look slower jumping (for better or worse? Can't decide if it's more realistic looking or not; but probably worse in the sense of making the game look slower although it does add that extra frame for more Early/Double aerials to exist). Combined with the Melee mechanic not currently in PM mentioned at the top; this is giving PM characters doing 1st frame Aerials/Specials (2?) more frames of air time.

-ADs/DJs done the frame before landing don't take effect and you land, where in Melee they do. [Notably moderate detriment?] ADs/DJs done the 2nd frame before landing fall the next frame then activate on what would be a frame perfect one in Melee. [have to be done sooner]

-ADs onto platforms require being done 1 frame sooner both onto (you raise upward 1 frame more from your jump, then initiate the change) and falling onto (like mentioned above). This may be why I always felt platforms felt better to Waveland onto in PM; you can trigger it sooner. Unless there's other platform collision differences

-Perfect WDs have you go into this airborne animation but grounded frame; not sure if this effects if you get hit whether you're considered airborne if you got hit that very frame

-Oh and since DJs/WDs take 1 frame longer to move, things like DJ'ing away from an incoming attack are less effective

..it also affects/delays Walks/Dashes and Specials? Knockback?

Also 8adge mentioned this; although you can see it first glance with Debug Menu's collision points

Does somebody here know exactly what the 1 frame delay is? My guess is that since it's universal, it could be that on any given frame the physics engine first calculates new positions for all characters, and only after that updates their velocity vectors, whereas in melee it would be vice versa. That sounds easily fixable though, so maybe it's more complicated.

It's just that, but the issue runs really deep. PM updates your position, then from that determines the state your character is in (e.g. aerial/grounded), which then affects the actions you can perform, which in turn determine your velocity on the next frame.
In Melee the actions you can perform are determined from the last frame, then the actions you performed on that frame affect your velocity, which determines your position which then determines the state you are in.

E.g. say you're on the frame before you would no-impact land on a platform (i.e. 0 frames of lag) without any horizontal momentum and you input shield + diagonally downwards (near horizontal). In Melee you will, because you're airborn, airdodge diagonally downwards on the next frame, which in turn will make you waveland on the platform with the horizontal speed of the airdodge carrying over.
In PM you will land on the platform and thus be in a grounded state, making you perform a roll into the direction you would have wavelanded in Melee.
In both cases, without input, you would have been standing on the platform. With the same inputs in Melee you will be in special landing and moved to the side due to airdodge momentum, while in PM you will be standing on the platform, not having moved horizontally, in the first frame of the roll animation.
This also means that momentum delay makes no impact landings 1 frame faster than in Melee (you'll usually land one frame later, though, because the movement that got you there is also affected by momentum delay.)


Gravity in PM always affects your momentum on the first frame you move during a jump. I think in Melee it only does that, if you perform an aerial/special (this may also be a result of the mechanics around the momentum delay). If the latter is the case, you only get more air time in PM because character constants were changed to make short hop/full jump heights the same between both games, otherwise "momentum delay" (using the term liberally) would make you land sooner.

I'm pretty sure that jump constants were changed in PM because melee characters do get an extra frame of airtime. If their initial jump velocity was increased to combat the physics delay, they would follow the same flight path as in melee except they would have an extra frame of airtime in the beginning.

I don't think Melee characters get an extra frame of airtime, that's PM characters/

And making them follow Melee's exact flight path I don't think is possible until they fix the physics delay; the closest you would get is after that pseudo-airborne frame having them forcefully glitched upward a moderate amount to make up for it which would look weird.

Melee characters get an extra frame of of airtime in PM, compared to melee.

Oh sorry I read wrong, didn't read "characters" in my mind


I remember Strong Bad saying gravity occurs the 2nd airborne frame in PM over Skype; meaning PM just doesn't let a move induce it like Melee does during that 1st frame. How would momentum delay make you land sooner? That 1st airborne frame not having velocity yet from the delay makes you land later. Are you sure their constants are different/were changed to reach the same apex as their Melee counter-parts?

Also after reading this http://smashboards.com/threads/shortened-jumps-via-frame-perfect-aerials.322505/#post-14459857 I'm not sure if all characters lose 1 frame or 2 frames of air time.

Momentum delay by itself would increase airtime actually. Characters' grounded or airborne states are not tied to the amount of velocity they have. Due to the momentum delay, characters with with say, a 5 frame jumpsquat will enter the air on the sixth frame, but not until the 7th frame are they affected by either their jump velocitty of gravity. As such, that frame doesn't change your height, and just sits that as an addition to overall airtime, while through the rest of your jumping animation, your character gains height, comes to a stop, then starts falling, and lands (and goes into grounded state) only when their ECB collides with the stage. Effectively the momentum delay just stops their ECB from being moving for that frame, so it takes longer to progress through the jump arc by that one single frame.

Actually, it seems that characters in PM are affected by gravity in their first airborne frame. This is compensated for by increasing the initial jump velocity of melee characters. If they weren't changed, melee characters in PM would have lower jumps and shorter airtimes. Since they were changed, what you said is effectively what happens, making it impossible to notice frame 1 gravity through frame advance. The only way to notice it is to look directly at the game's code while it's running. @8adge knows more about that than me though.

I figure gravity wouldn't take effect until the second frame given that gravity is a kind of momentum change, meaning it should be subject to the same delay. Gravity isn't a constant thing, it doesn't take effect until you are airborne, so if it's delayed by one frame, it follows that it couldn't take effect the first airborne frame.

Changes in momentum can occur instantly, it's just they are not acted on until the following frame. I think that PM's order of operations is;

1. Apply positional changes based on the last frame's momentum
2. Act on any animations currently in progress (in this case it would be ending jumpsquat)
3. Accept and act upon any inputs
4. Apply any changes in momentum, including momentum from inputs or gravity. This change in momentum is stored and acted upon in the next frame.

I knew you can't use IASA to AD in Melee.. but can you not IASA specials as well? It won't let me do side-b or up-b with Marth's BAIR/FAIR IASA

That's a huge game difference

Everything Frakture said above follows my interpretation pretty much. My last usage of "momentum delay" may have been a bit too loose, but I was indeed referring to the differences in the physics engine between Melee and PM that have delayed momentum as a consequence and not that consequence in particular.


The second airborn frame is the first to be affected by gravity (it is applied to momentum on approx the first airborn frame (approx as in I didn't look at the code, but it is applied when Dolphin stops on the first airborn frame)), but as you don't move during the first airborn frame, that's also the first frame you move in from your jump velocity.


Take a look at these images, especially the values on the left (you probably all know how jumps look.)

View attachment 37385 View attachment 37386 View attachment 37387

Important to know is that Squirtle's gravity is 0.126 and his initial vertical hop velocity is 2

I edited the images to show the current height instead of that of next frame like it would have unedited. This makes it easier to look at, but height being calculated on the frame before is essentially what momentum delay is. There's a number of interesting facts to note in the pics above:

1. Gravity affects the first time moving from the jump. (the value increases by 2-0.126=1.874)
2. You already have velocity between the last frame of jumpsquat and the first airborn frame, but it has no affect on the next frame.
3. The jump velocity is already applied before the next frame, but apparently after the position for that frame was already calculated (or we wouldn't have momentum delay).
4. Positional values are calculated on the frame before. Note that I don't know the exact relative position we are in between the frames in the code. The correct statement would be that positional values for the next frame are calculated either very early in the transistion or even before the last transistion of frames.

The order of opperations appears to effectively be what Frakture detailed above. But if you take into account point 2. of the above, I think what actually happens in the game is closer to:

1. Set the character position to the values calculated last frame, check collisions, change character state etc.
2. Apply gravity, momentum decay etc.
3. Calculate character position for the next frame
4. Apply non-linear changes to the momentum (or momentum setting? Would have to test when knockback is applied when stacking, but that's also influenced by PM-code, so it may be hard to determine what Brawl's engine is doing)

@KuroganeHammer /Chesstiger plz add http://smashboards.com/members/schmooblidon.237646/

**** me this will take a while to catch up

My brain hurts.

Rewind a bit back to "PC Jump Drop." This is a Platform Warp, I made a thread about them a while ago, http://smashboards.com/threads/new-tech-platform-warp.385768/ I have only found NILs for Doc, Mario, Luigi, Marth and Ness, might be a few other low tiers.

The Platform drop > Aerial (need hitlag) > Land back on platform, is normally refered to as a Platform Cancel. Luigis use this the most because he has a larger window then most.

I'm glad other people are interested in smash's platform mechanics, there is a lot still to be explored even in melee.

There is a lot of PM talk here that I can only observe, but I'm gunna start labbing it up, see what I can find. The ECBs seemed more polished out, and more parallel across all characters. But they still have plenty of potential exploits and even manipulation (to what extent I do not know yet, but I know it's there).

The term "Platform warp" seems fine!
About the Platform Cancel (which I sometimes refer to as dropcancel, so if I accidentally do it now you know what I mean), I have a list on moves that can Platform cancel in PM:

Spoiler

Falco's back-air
Falco's down-air
Fox' back-air
Link's neutral-air
Luigi's back-air
Luigi's neutral
Luigi's up-air
Mario's up-air
Marth's forward-air
Marth's up-air
Samus' neutral-air
Sheik's neutral-air
Squirtle's up-air
Yoshi's neutral-air

Lucario's down air has the same effect, but for a different reason, and it doesn't cancel instantly, so it is up to you whether to count it or not.

In regards to possible platform warps I am very inexperienced.

Because of weird ECBs, Falco's nair doesn't need hitlag to land on the platform again, however, you land before the hitbox comes out, so it has probably no application. I didn't test all but some close candidates and I think it is the only move in the game with that property.

In PM, every character can platform drop->air dodge and land on the platform if executed frame-perfect. So platform drop->WD is also possible (for characters with long jumpsquat this could be useful).

Also as question at everyone, should I change the title of the conversation as the topics in the title are not the only topics in the conversation anymore. If yes, please suggest a title.

It doesn't seem like Platform Warps are possible in PM, as they do not have a set DJ ECB extension and thus no update. What you are performing with Falco is an Aerial Interrupt. This was the first thing I looked for in PM, because I knew Samus' upair AI was still there from melee. There are a huge amount of AIs in this game dude. I found 3 particularly good Ledgehop AIs, making it their fastest option back onto the stage.

Ivy Bair - http://smashboards.com/threads/ledgehop-bair-aerial-interrupt.388224/

Falcon Nair/Bair - http://smashboards.com/threads/ledgehop-bair-nair-aerial-interrupt.387740/

Diddy Nair - http://smashboards.com/threads/ledgehop-nair-ai.387706/

Those AIs can also be performed to land onto a platform too. There are many other AIs, I had a list on my PC but I don't have access to that right now. Most of the ledgehop AIs were outclassed by the character's other options, but are still useful for landing on platforms. There are also AIAs (Aerial Interrupt Attacks) in PM. An AI typically lands in early autocancel frames (but some you need to l-cancel), but an AIA makes you land at least 1 frame after the first hitbox frame, thus granting you a hitbox. Mario's Upair AIA makes a return, Metaknight has a Nair AIA, there may of been others I can't remember.

This Drop Wavedash is pretty cool, seems super hard though.
edit: scratch that, I was using 2 shoulders cos I'm an idiot. it's pretty easy with 1.
edit2: oh dude it works with shai drops thats sick!

What's it called when you jump from the ground and then DJ to land on a platform? I've been calling it jump landing but I don't know if there's an official name.

With characters like Yoshi, Peach, Mewtwo, Ness etc? I've known them as Double Jump Lands (DJL).

Ike actually. On Dreamland, if you backwards fullhop and DJ on frame 12, he will land with 0 landing lag.

I don't think there's any often used term. Kadano uses no-impact land term for those.

Lol PM. Falcon can do this too, probs others too. Some odd design choices there (although I guess you could say the same for melee's bone crushing concept). IIRC only yoshi and peach can do this in melee, because they literally dip downward, the ECB of doublejumps is directly affected by the previous state. They do update though, mostly on frame 10 (handfull of different values), that update is what causes the platform warp tech to work.

So PM traded the Melee DJ Frame 10 ECB update that causes platform warps, for a Frame 2 ECB update that causes DJLs.

tauKhan, although you aren't wrong, if it were me Id say that a NIL is the result of a DJL, like a NIL can be a result of a Platform Warp.

What exactly do you mean when you say ECB update?

It's when the ECB changes to a set size and position. Sometimes this update will look near identical to the previous frame, but if you manipulate the ECB, the update could be drastically different.

http://gfycat.com/ComplicatedSecondhandAmericanlobster
Frame 10 is the update, whatever you are doing before you doublejump. Frame 10 will always look like that. The gif shows how dropping from a platform shrinks the ECB, so that the update moves the ECB downward.

Perhaps I should not call the PM Frame 2 ECB an "update" considering that's just how it starts.

The fact that you can't use IASA for Specials (and AD) in Melee but can in PM is so big

I only tested Aerials though, someone see if Marth's d-tilt to down-b can IASA in PM but not Melee; though with any ground moves it's a little irrelevant since you can buffer walk

It is possible in PM, confirmed. Never seen it in Melee but I can't confirm it isn't possible there.

Isn't possible in Melee. Can't interrupt with shield in either game.

You can interrupt with buffered walk and that makes the buffered shield work

I forgot to test for certain if SH Fair Side-b/Up-b doesn't work with Marth in Melee

Not in IASA frames, but you can however DJ into them.

In PM If you wavedash off a platform or ledge while facing backwards and Nuetral-B you'll do it facing the other way (back to facing forward)

That's because the last direction you pressed was backwards when you wavedashed. Try pressing forwards while you're sliding on the ground and then neutral b.

How does that make sense? I thought it relies on your control stick position on the frame you did the Special/Neutral-B, with neutral not gonna be switchin your direction. And if you WD off and the SpecialFallLanding (10 frames) ends and your momentum is what carries you off (rather than ledge cancelling the SpecialFallLanding) then it doesn't seem to switch your direction with the Nuet-B // Same with a PC drop (Walk momentum to turn-around). So SpecialFallLanding seems to have a specific property for this.

If you input a direction on the control stick the frame you input a special, it will just do a directional special. The reason why you don't turn around after special fall wears off is because the input period to do a turnaround neutral b expires, and in order to do a PC drop, you have to press towards the stage, so the last direction you pressed it the direction you're facing.

Try falling off while still in special landing, except press backwards while still on stage. You shouldn't do a turnaround then.

Someone made a thread about this a while ago lol: http://smashboards.com/threads/new-mario-technique.384638

B-turnaround is a 20 frame window from when you hit a direction. In melee at least, likely the same in PM.

http://smashboards.com/threads/new-mario-technique.384638/#post-18292099

You're saying this would work in any instance then, as in jump in the air, press Opposite direction you're facing on control stick for a frame, then return it to neutral and advance 19 frames and input B and you're gonna cause it to be a B-turnaround'd Neutral B?

It might not be the same in PM
http://www.reddit.com/r/SSBPM/comments/2gmmg4/comprehension_wavebouncing_or_breverse/ckklda2

oh I coulda used this a video ago http://smashboards.com/threads/kadano’s-perfect-marth-class—advanced-frame-data-application.337035/page-26#post-17658858

Yeah, that's how things like reverse shdl are always done, you tap the direction you want to turn to before the b input, the window probably is 20 frames, I would trust @schmooblidon on this.

Yeah I checked and it's the same in PM.
Really it's 19 frames (same as melee), but that's because my definition has changed since I looked at this last.

So its like:
1 - Back
20 - Last frame to turnaround neutral-b

Also, B-reverse is a 5 frame window, and you will always reverse on the 5th frame.

For PM? That doesn't sound right at all

It's suppose to be like 3/4 frames, and you can get the reversal on the frame it occurs

I thought it was 4, are you sure?

It was 2/3 frame window in Brawl, PM increased the window by a frame to make it more lenient (and yet they don't really tell you this stuff)

b-reversal occurs when holding in the other direction on frame 4 of the special, so a 4 frame window to move the control stick there.

At least for the characters I tested, should be the same for all though as you need 4+ frames of startup to be able to b-reverse a special.

K let me try that again.

It's a 4 frame window, the reversal occurs on frame 5. If you press back on frame 1 you will Side-b of course. And yes you need to be pressing the direction on frame 5.

PMDepot's new slogan should be proudly misinformed; I need to make a new video on this information to correct some past ones.

I also think since the physics delay takes 1 frame before shines halt your momentum, approaching with Falco's SH FFall Shine to land on the ground and JC it (into pry grab on their shield) may work better than Melee.

I think you understood me wrong here. I meant you drop down from a platform, do Falco's nair and then land on the same platform again (the other aerial interrupts would have the effect when rising but not after a previous platform drop), while it doesn't require any hitlag. I think it is the only move in the game that does that (if you don't count "cheater" moves like Lucario's dair)

So like a platform cancel? What are the frames for this? I can't seem to reproduce it.

It happens a lot on moving platforms. Try using ROB's dair instantly while dropping through smashville's platform.

@ Chesstiger2612 add http://smashboards.com/members/auramaudegone.205470/ and http://smashboards.com/members/soft-serve.206736/ . Remove 1Mach if you still got that user per group limit; unsure why Aerodrome left.

 

[Frost | Odds]

For more like this, feel free to ask for an invite to the Project M R&D Skype Group, where all we do is talk about niche PM tech stuff.