Data Kadano's perfect Marth class -- advanced frame data application

[Cort]

This thread is amazing.

 

[Kadano]

questions to kadano:

1. is it feasible to react to fox's ledgedash shine and be able to do anything? i was thinking of spacing myself right around fox's ledgedash -> usmash range, so i'm close enough to be a threat, but far enough away so that i can escape if i need to, but i don't know if i have the means to escape. e.g. shield, wd back, dash away

Whether spacing yourself just outside of his ledgedash → usmash range still poses a threat depends mostly on your opponents techskill. A Fox with perfect ledgedash → dash → upsmash execution could bring an intangible hitbox out as far as this:

If he doesn’t include the dash, he can get this far:

Because Fox can mix up usmash and grab, I think the only safe spacing against a ledgeborn Fox is outside the range in the second pic. Considering how close to perfection leffen and Hax have gotten with their ledgedashes, intangible usmashes like in the first pic are something Foxes might start doing, albeit not quite as far.
Shield is not safe because of the grab mix up you cannot react to (unless your reaction time is below ~100ms, which is highly unlikely if not humanly impossible). Wd and dash away are kind of safe to escape grab and usmash, but I don’t believe this is a situation where you can expect to punish Fox on reaction. He has many options from the ledgedash and doesn’t need to commit early at all.
If I know that my Fox opponent is proficient at ledgedashing, I simply accept that they can immediately threaten a large zone and retreat out of it. You still have center stage and can pressure from there. Being close to the ledge always poses the risk of getting grabbed, which destroys the positional advantage immediately.

2. do you know anything about how to DI against sheik and falcon? against these characters, sometimes i feel like they have pretty solid followups on every DI and that sometimes it's better to just DI weirdly to throw them off (e.g. partial DI, no DI, DI inwards, etc), but that's a little too janky for my liking, unless it's supported by numbers, in which case it ceases to be janky

I haven’t really looked into it yet, but it’s something I often thought about. Your suggestion should be enough to make me disregard my laziness, so expect some data on this within the next few days.

3. similar to question #2, but i kinda wanted to know about how to DI against marth's fthrow. i've believed for a while that using down-away DI is the best option (trajectory is roughly up-away, so down-away seems like it'd be the best), but i've heard of people using downward DI to great success.

additionally, i know/think that straight-away DI isn't too fantastic against f-throw, but based on my approximations and understanding of DI, straight-away DI and downward DI should have the same effect on your f-throw trajectory:

i assumed that f-throw trajectory was 45 degrees, so...
DI down: sin^2(135) = 0.5, 0.5*18 = 9 degrees change
DI away: sin^2(45) = 0.5, 0.5*18 = 9 degrees change
DI down-away: sin^2(90) = 1, 1*18 = 18 degrees change

my 45-degree assumption might have been inaccurate, but i don't imagine that it was so far inaccurate that it would explain the results

Fthrow has a 50° knockback angle. (How do I know? Toomai created a godly hitbox repository .xlsx you can find here (slightly modified by me to include Strong Bad’s knockback calculator and 0% and 100% pre-calculated knockback values))
Therefore, 320° DI would make you land the soonest. However, the difference between 315° DI (easier due to the southeast ridge) and 320° DI is negligible and already makes you escape both JC grab and fsmash even at 0%, so always using 315° DI should be a good and easy solution.

4. what exactly are the mechanics behind marth's first-hit forward B when it comes to recovery?

Because I can’t read internal Melee code, I can only tell you what I know from messing around with Dolphin. So while my findings should be very close to the truth, these mechanics probably work slightly different than I assume them to here.
The pseudo-jump flag is refreshed during the first frame of every normal landing animation. Landing with ↓B does not refresh the flag.
The pseudo-jump works similar to a normal second jump in that the introduced force is at work only for the first frame and then slowly diminishes.
Regardless of whether the flag was on prior to the start of the B, previous Δy is almost completely ignored and replaced for a small positive value if the flag was on or to just below zero if the flag was off.
During the entire duration of B, the control stick does not affect your position at all. Δx is kept just like it was in the previous action state.
Because of this, you should always wait for the knockback of the attack that sent you offstage to completely vanish and for your Δx towards the stage to reach its maximum (or maybe close to maximum) value before you start using B.
Here is a comparison between using B without the jump flag and merely falling:

Someone in my local community used to claim that using B without the jump flag didn’t help recovering except for the delay. Now I don’t think this is/was a widespread rumor, but in any case this picture proves it wrong.

5.

Thanks for telling him, haha. My main drive for creating the Jigglypuff grab conversion section was so that Pewpewu would get the much-needed early kills against Hbox. In his set at NCR, he landed very many grabs, but got next to nothing out of them, which made me really sad. At the end of the first match, he got a grab at 75%. Had he known all his killing options, I’m sure he would have won that game.

You are a god

A compliment I whole-heartedly return! I love watching you play Marth.
I hope this will help you, especially the Jigglypuff part.

 

[K.Louis]

Can/did you do a breakdown of how Marth's DSmash works with the hitboxes that blast them up and the hitboxes that blast them sidesways?

 

[Kadano]

Can/did you do a breakdown of how Marth's DSmash works with the hitboxes that blast them up and the hitboxes that blast them sidesways?

I started that! =)
Didn’t finish it back then, thanks for reminding me. Here you go:

 

[Bones0]

I started that! =)
Didn’t finish it back then, thanks for reminding me. Here you go:

That's Marth's GUA, Kadano... >.>

 

[Kadano]

That's Marth's GUA, Kadano... >.>

No.

 

[Bones0]

Oh, nvm. The other moves in there confused me. lol

 

[AceDudeyeah]

I find some of my opponents DI'ing my dancing blade so that by the 3rd strong hit, they are behind me.
Can they still do this with the same DI if I did the 2nd hit an upvariant rather than the normal neutral hit?

 

[Bones0]

I was messing with 2-up DB the other day and it doesn't seem to knockdown Falco until around 150%. They will probably just be able to shield.

 

[AceDudeyeah]

I use it as an anti-approach out of DD, so they're usually already in the air when the first DB hits em.

 

[Kadano]

I find some of my opponents DI'ing my dancing blade so that by the 3rd strong hit, they are behind me.
Can they still do this with the same DI if I did the 2nd hit an upvariant rather than the normal neutral hit?

Yes, they can, depending on damage, weight, falling speed and hitbox ID. I won’t go into details because it would be an insane amount of work to test all the factors.

 

[Bones0]

I use it as an anti-approach out of DD, so they're usually already in the air when the first DB hits em.

I feel like they'd reach the ground by the time you get DB 3 out, especially since a lot of people automatically DI down and buffer shield as soon as they start getting hit by DB.

 

[AceDudeyeah]

You are standing on stage facing the edge.
How much of a frame advantage do you get from dash dancing to wavedash backwards edgegrab vs simply turning around to wavedash backwards edgegrab?

Or what is the widest your dash dance can be to still have frame advantage?

 

[Kadano]

Depends on where exactly you are standing. If you are close to the ledge, for example where you end up after a ledgeclimb, turn-wavedash is faster.

 

[Swiper]

There should be a thread like this for every character, either way great thread.

 

[AustinRC]

We should make a sperate thread for the appeciation of this one. Just a thought.

 

[Bones0]

We should make a sperate thread for the appeciation of this one. Just a thought.

That would waste too many frames. Kadano would never approve.

 

[Kadano]

Edgeguarding spacies’ side-B

Very interested in your thoughts on this.

Been procrastinating on this way too long. Considering that there are thousands of different ways spacies can recover with side-B, I will only explain those where only a few of Marth’s options work in detail and treat the others in a more general way.
Every option is listed depending on the position Fox starts his illusion at, going from lowest to highest.
For Falco, everything starts 4 frames earlier.

1. Sweetspot to the ledge.

It takes Fox a total of 28 frames to reach the ledge from here; 26 if he shortens. If he shortens, he needs to use the second shortest length as the shortest one will make him miss the ledge.
Dtilt will whiff by far:

The only attack you can reliably hit him with in time is f-air. Dashing to the edge and fairing afterwards takes will take 16 frames until the hitbox is out where it can hit Fox. Thus, there are 10-12 leniency frames, depending on if – and when – Fox shortens. This isn’t true leniency, though – it’s the time during which you need to react. Thus, your reaction time needs to be below 167ms (10 frames) if you want to be able to hit Fox reliably (assuming frame perfection on dash, f-air and fastfall execution).

Edgehogging is also possible in time, but the frame window is really tight.
From the position Marth is standing, it takes 16 frames to do a pivot edgehog – just as much as the fair. Turn-wavedash is 18 frames, so if Fox shortens, he will grab it before you – assuming you are do it on reaction and have a <10 frames reaction time, which is asking pretty much.
If Fox ends up in this spot, it’s probably better to react to the position and not to the Illusion. Doing so should give you another 5 frames or so, which ensures that you grab the ledge in time. It’s important to make this decision very late so that Fox can’t react to it and decide to Illusion while he can still land on the stage.
If you time it that way, Fox has no options. He might opt for Firefox instead of Illusion, but covering that with b-air is really easy from the ledge.
2. Shorten sweetspot to the ledge from above.

The problem here is that Fox can still land on the stage and you can’t reach him where he starts up in time.
Many players choose to take guesses here. They fsmash or jab and hope that Fox will not shorten. But most Foxes will go for the shorten in the majority of cases, as hardly any Marth ever threatens the ledge in this situation.
It’s okay to guess. At the moment you need to start the dash / turn for the edgehog or the jab / ftilt, Fox already has committed to his shorten choice, unless his reaction time is below 5 frames for ftilt or 2 frames for jab, which is humanly impossible.
In other words, if you know the timing, you can play the guessing game with him and have up to 50% killing rate.

In case you don’t want to play the guessing game, you only have three options: double dtilt, jabs and Dancing Blade. Dtilt is most reliable on hit, but Dancing Blade is more reliable tohit. Jabs are the easiest to time, but also the easiest for Fox to counter.

Dtilt
The first dtilt will only catch him if he starts the Illusion not higher up than this:

The frame window for the first dtilt to hit him if he does not shorten is 13-15 of Fox’ Illusion.
To hit Fox if he shortened, you need to start the second slash during any of the IASA frames 19-21 if you chose frame 15 before. For every frame you did the first dtilt earlier, this frame window increases by one at the end, so if you dtilted on frame 13 of Fox’ Illusion, the window for the second dtilt is 12-23.

If Fox goes slightly higher:

dtilt will whiff if he does not shorten. You do have more than enough time to run up to him and punish with whatever you want. Also, because your hurtboxes are so low during dtilt, Fox’ illusion will never hit you.

Many probably think of jabs now, and indeed, for Illusions that start that high up, both jab 1 and 2 will hit full-length and shorten respectively. But the big problem with this option is that jab 1’s hitstun is so short that Fox can Illusion to the ledge after he got hit by it. This will transition to situation 1.

Dancing Blade also has a similar problem: if he doesn’t shorten and DIs the first slash away, the second slash will whiff and he can also Illusion to the ledge. Thus, it’s comparable to jabs, but has the advantage of an additional obstacle for the Fox player – for jabs, his DI doesn’t matter, but if he doesn’t DI or DIs in on the DB1, you can hit him with DB2
The frame window for the first slash to hit him if he does not shorten is 14-16 of Fox’ Illusion.
The frame window for the second slash to hit Fox if he does shorten is 7-10 of the first hit if you did the first slash on frame 16. It increases to 7-11 if you did the first slash on frame 15 and to 7-12 if you did it on frame 14.

It is possible for Fox to SDI the first hit in and land on the stage before the second hit comes out. Normal DI + ASDI, however, is not enough.

Both jabs and DB have the problem that even after the second hit, Fox can still Illusion to the ledge, and Marth is still in lag so he doesn’t have all the options of situation 1 available to him.
Thus, dtilt is the superior option, even though Fox can force his way onto the stage.

I feel these are the most important positions, if there are more you’d like to have analyzed, just post a screenshot of the illusion startup.

 

[AceDudeyeah]

Holy hell. This new technology.
Never even thought of using side-b to hit both illusion and the shortened.
Must incorporate.

 

[MT_]

Does side-B knockback scale with percent? I know the second hit does, but does the first hit? (I think it does very slightly). I feel that if they are holding in after the shorten and get hit by the second hit of side-B at a medium-high percent, they'll just DI it onto the stage. At low percents though it seems like a solid answer to shortens though. I'll have to start trying it out.

 

[Kadano]

Does side-B knockback scale with percent?

It does, but at a very low rate. All DB1 hitboxes have 25 KBG, so total knockback increases from 60 at 0% before hit to 70 at 100% before hit.

Anyway, at 80% before hit, full DI + ASDI in does make them fall down on the stage, but you can hit them with DB2→ if you time it perfectly (one frame window at 80%). If you manage to get this timing right, the third slash (DB3→) will also connect. It depends on spacing very much, the distance Fox travels between frames is so high that getting the correct hitboxes is hard / impossible to do intentionally.

 

[MT_]

DB3 will connect if you hit DB2? Even if they DI in?

All in all it seems like using DB at about 60 and lower might work well. I'm scared that if they DON'T shorten and I go ahead and hit both DB1 and DB2 that the second hit will allow them to DI onto the stage. But maybe I can get good at reacting to the shorten and only inputting DB2 if I see the shorten.

 

[Kadano]

DB3 will connect if you hit DB2? Even if they DI in?

Under the right circumstances, yes. The thing is, both DB1 and DB2 have four different hitboxes that all have different angles. Because Fox pops into your hitbox due to his ridiculous Illusion moving speed, you have hardly any control over the hitbox you get. Here are DB1 knockback angles:

And here DB2→:

All other values – damage, KBG, BKB etc. are the same for all four hitboxes.
And in case this isn’t obvious: If one hitbubble overlaps another, this illustrates that this bubble has priority. Priority chain: pink > salmon > cyan > sky blue.

 

[Bones0]

Can't you just dtilt once to cover full side-B, and then dtilt again if they shortened?

 

[Kadano]

Can't you just dtilt once to cover full side-B, and then dtilt again if they shortened?

Yes, you can. I didn’t consider it because I remembered that when I tested these situation half a year ago, I found that dtilt wasn’t reliable. However, that was for Illusions that start slightly higher, so at the position 2., dtilt is the better option, yes.
I rewrote the long post to incorporate the latest issues I discovered. Tl;dr: always dtilt.

Still, I don’t regret walking down the wrong path – the weird Dancing Blade angles are something I always wanted to research but I just haven’t had the incentive earlier.

 

[Bones0]

Can you double dtilt to catch shortens if they Phantasm at ledge height?

If you grab the ledge early, is it possible to react to a high Phantasm with a LH bair?

Does every legal stage have the same ledge structure in terms of how far down you can sweetspot?

 

[Kadano]

Wrote something in a different thread about pivot edgehogs:

Pivot edgehogs are not just hard, they also require an additional type of precision. Because the total sliding distance after the first turn frame is less than the maximum distance any character can move with a dash in one frame, you need to react to the exact distance to the edge with certain analog stick inputs to do the pivot edgehog successfully in every situation.
By holding forward, you ensure that the character approximates his maximum dashing speed. Any other inputs will decrease the dashing speed. By selecting the right option for every frame, you can always (haven’t confirmed that for every character, but I’m pretty sure it is) edgehog.

For example, if Captain Falcon climbs the ledge (to ensure an exact reference position) and then dashes back towards the ledge, holding full towards for 4 frames and then holding full backwards will make Captain Falcon spend one frame in his Turn animation and then fall off the ledge, successfully triggering the pivot edgehog.
If he instead dashes towards the ledge and holds toward for only one frame, he will fall off the stage after frame 5 (instead of 4 if he keeps holding towards the edge). If he again turns after frame 4 of his dash, he will not reach the ledge before his sliding motion has ended.
Now it might seem like he could simply wait one frame longer until he turns, right? Unfortunately, at that point he is so close to the ledge that he will immediately fall after frame 5 – to actually turn around, spending at least one frame in the Turn animation is required.

If you wonder what happens if forward is held for 3 or 2 frames: for 3 frames, everything is the same as for 4 frames. If forward is held for only 2 frames, Captain Falcon will slide off the ledge after Turn 3.

Another example: Marth. Again, we go to the ledge and climb up to ensure an exact reference position. Then we have Marth dash towards the ledge again. Here are the timings and their results:
• If Marth holds towards the edge the entire time during his dash, he will fall off after frame 6. If he presses backwards after frame 6, he will still slide off. [No pivot edgehog]
• If Marth holds towards the edge during the first 5 frames of his dash, then keeps his analog stick in a neutral position for one frame and then smashes it backwards, he will still slide off before he can turn. [No pivot edgehog]
• If Marth holds towards the edge during the first 4 frames of his dash, then keeps his analog stick in a neutral position for 2 frames and then smashes it backwards, he will enter his Turn animation and slide off after Turn 1. [Successful pivot edgehog]
• If Marth holds towards the edge during the first 3 frames of his dash, then keeps his analog stick in a neutral position for 3 frames and then smashes it backwards, he will enter his Turn animation and slide off after Turn 4. [Successful pivot edgehog]
• If Marth holds towards the edge during the first 2 frames of his dash, then keeps his analog stick in a neutral position for 4 frames and then smashes it backwards, he will enter his Turn animation, but he won’t have enough momentum left to slide off the ledge. [No pivot edgehog]

Can you double dtilt to catch shortens if they Phantasm at ledge height?

If what you understand as “ledge height” is what I illustrated with the screenshot in my previous post’s section 1, then no. Apart from the 4 frame timing difference, Illusion and Phantasm are pretty much exactly the same.

If you grab the ledge early, is it possible to react to a high Phantasm with a LH bair?

How high are you talking exactly?

Does every legal stage have the same ledge structure in terms of how far down you can sweetspot?

Yes and no, depending on your exact horizontal position and whether you shorten or not. I will assume you are only asking for Fox and Falco, as the ledge issue on YS only affects Marth’s Dolphin Slash and similar recoveries.

If Fox does not shorten on BF at the lowest position (perfect sweetspot, if you will), he will pass through the ledgegrab area before he even starts falling (which is required for ledgegrab):

By the time Fox starts falling, he is long below Battlefield in the spike area, pushed down by the ceiling.

If he does shorten, he reaches the ledge from the same position:

The ledges on YS, FoD and DL64 do not have this problem. Pokémon Stadium, on the other hand, does:

 

[Bones0]

If what you understand as “ledge height” is what I illustrated with the screenshot in my previous post’s section 1, then no. Apart from the 4 frame timing difference, Illusion and Phantasm are pretty much exactly the same.


How high are you talking exactly?


Yes and no, depending on your exact horizontal position and whether you shorten or not. I will assume you are only asking for Fox and Falco, as the ledge issue on YS only affects Marth’s Dolphin Slash and similar recoveries.

By ledge height, I basically meant low enough that a non-shortened side-B would hit the ledge and grab on. Your example had the side-Bs a little higher so that the spacie actually had to fall down to the ledge which I figured was what allowed you to get the second dtilt out in time. So if they go even with the ledge, do you still have time to double dtilt? I guess more importantly, do you even need to? I've always been kind of confused about there being two different timings for hitting shortens vs. full length side-Bs. it makes sense that if they are above the ledge they would hit me if I assumed they were doing a shorten and timed my dtilt accordingly, but if they side-B at the ledge, shouldn't I just time my dtilt for a shorten because that would also hit a full-length side-B going into the wall?

I didn't have any specific height in mind for the LH bair. I guess I'm just curious how fast you would have to react to be able to hit any height that is actually possible to hit. I'd be a lot less reluctant to grab the ledge before a spacie uses his DJ if I knew I could LH bair a DJ side-B on reaction.

My question concerning different stages, I was mostly wondering about the vertical limits to grabbing the ledge. Assuming you are sweetspotting horizontally, is it easier to go under Marth's dtilt on one stage than it is on another? I have always felt like I could sweetspot really low on FoD, but maybe that has to do with the graphic of the stage and the physical floor is in the same relation to the "edge" as all of the other stages.

 

[Kadano]

[…]if they side-B at the ledge, shouldn't I just time my dtilt for a shorten because that would also hit a full-length side-B going into the wall?

I don’t understand why you are still asking me these questions after I’ve already clarified that dtilt will not hit sweetspot, ie ledge height side-Bs:

Dtilt will whiff by far:

The only exception is on Yoshi’s Story, I will go into that later.

I didn't have any specific height in mind for the LH bair. I guess I'm just curious how fast you would have to react to be able to hit any height that is actually possible to hit. I'd be a lot less reluctant to grab the ledge before a spacie uses his DJ if I knew I could LH bair a DJ side-B on reaction.

“Any height that is actually possible to hit” – that would be a 1-frame reaction time, then:

For lower heights that an immediate ledgehop bair covers, it’s approximately 11 frames (if they are close to stage, you have 1-2 frames less because they burst through your bair zone earlier).
Realistic example with 10 frames reaction time:

However, Marth’s fastest edgehog from his standard edgeguarding position (pivot edgehog) takes 20 frames (turn → wavedash is 1 frame slower). Every opponent can easily react to that with an immediate side-B / jump → side-B and destroy your opportunity to react. Thus, I don’t think the edgehog is appropriate here as long as they either have their jump or can still side-B onto the stage.

My question concerning different stages, I was mostly wondering about the vertical limits to grabbing the ledge. Assuming you are sweetspotting horizontally, is it easier to go under Marth's dtilt on one stage than it is on another? I have always felt like I could sweetspot really low on FoD, but maybe that has to do with the graphic of the stage and the physical floor is in the same relation to the "edge" as all of the other stages.

Perfect sweetspots can’t be hit by Marth’s dtilt on any stage except YS. FoD should actually have a rather tight window for the side-B timing, as Marth can stand on the lower level part and still bring the dtilt out near the ledge.
All stages are the same in terms of dtilt reach vs. side-B ledgegrab window, except FoD and YS:

It’s hard to see in this image because of all the editing, but the ledge you can grab onto is nothing but a kind of red carpet that’s lying on the stage near the edge:

 

[MT_]

Perfect sweetspots can’t be hit by Marth’s dtilt on any stage except YS. FoD should actually have a rather tight window for the side-B timing, as Marth can stand on the lower level part and still bring the dtilt out near the ledge.
All stages are the same in terms of dtilt reach vs. side-B ledgegrab window, except FoD and YS:

Sorry to ask if you already made sure about this, but for Battlefield are you sure that Fox can sweetspot the ledge with side-B? I feel that if he goes low enough to not get hit by d-tilt that he'd actually go under the stage (if he didn't also horizontally sweetspot it). Maybe it's just a psychological thing that Foxes are scared that they will go under the stage, so they aim a little higher on the side-B sweetspots?

 

[Kadano]

Sorry to ask if you already made sure about this, but for Battlefield are you sure that Fox can sweetspot the ledge with side-B?

Yes, I am sure and also covered this just 3 posts before yours: http://smashboards.com/threads/kadano’s-perfect-marth-class—advanced-frame-data-application.337035/page-6#post-16035011
BF and PS do have an additional obstacle for the spacies player, but Bones was asking about maneuvering past dtilt and still being able to reach the ledge, and for that there is no difference between all neutral stages except YS and FoD.

 

[MT_]

Ooh sorry. Thanks!

 

[Kadano]

No problem. This thread is pretty cluttered, so I understand it’s easy to overlook something.

 

[Bones0]

I don’t understand why you are still asking me these questions after I’ve already clarified that dtilt will not hit sweetspot, ie ledge height side-Bs:

The only exception is on Yoshi’s Story, I will go into that later.

“Any height that is actually possible to hit” – that would be a 1-frame reaction time, then:

For lower heights that an immediate ledgehop bair covers, it’s approximately 11 frames (if they are close to stage, you have 1-2 frames less because they burst through your bair zone earlier).
Realistic example with 10 frames reaction time:

However, Marth’s fastest edgehog from his standard edgeguarding position (pivot edgehog) takes 20 frames (turn → wavedash is 1 frame slower). Every opponent can easily react to that with an immediate side-B / jump → side-B and destroy your opportunity to react. Thus, I don’t think the edgehog is appropriate here as long as they either have their jump or can still side-B onto the stage.


Perfect sweetspots can’t be hit by Marth’s dtilt on any stage except YS. FoD should actually have a rather tight window for the side-B timing, as Marth can stand on the lower level part and still bring the dtilt out near the ledge.
All stages are the same in terms of dtilt reach vs. side-B ledgegrab window, except FoD and YS:

1. I know if they sweetspot vertically that you can't get dtilted, but the vast majority of time spacies side-B, they don't go low enough to avoid dtilt. So assuming they are side-Bing into the ledge high enough for you to dtilt them, I'm wondering if I can use a single dtilt timing by assuming they will shorten and also still hit them if they don't shorten. (It sounds like this will be answered when you get to the specifics of YS dtilting anyway.)

2. I had thought about them jumping in reaction to a WD towards the ledge, and I was thinking I might be able to option select everything by either edgehogging, WDing off into an instant bair, WDing off into an instant DJ bair, and WDing onto the ledge and LH bairing delayed side-Bs.

3. That's crazy. I knew FoD had that weird, seemingly invisible, slant at the edge (weird stuff like Luigi's down-B turns diagonal on top of it), but I didn't think that affected edgeguards.

 

[1MachGO]

Hey Kadano, I have some questions regarding platforms and wavelands:

Question 1: If Marth is underneath or near a platform, what is the fastest way to waveland onto it? Is full hop>immediate DJ>waveland fast than full hop>waveland?

Question 2: When underneath/near a platform, is jumping up, wavelanding off the platform, and then immediately fast fall-fairing faster than a regular SHFFL'd fair? What are the speed differences of this tactic across different platform heights on tourney legal stages (excluding FoD)?

Thank you for your time, I really appreciate this.

 

[TKD+ITA+Mar=]

Kadano, what about using ftilt, fair, (or nair?) to catch fox out of his side-b when he does it higher than dtilt height? (Or does this require inhuman reaction time?)

 

[Kadano]

So assuming they are side-Bing into the ledge high enough for you to dtilt them, I'm wondering if I can use a single dtilt timing by assuming they will shorten and also still hit them if they don't shorten. (It sounds like this will be answered when you get to the specifics of YS dtilting anyway.)

Not for all horizontal Illusion starting points, except for YS.
I don’t plan on elaborating on YS specifics further than in the last two pictures.

Question 1: If Marth is underneath or near a platform, what is the fastest way to waveland onto it? Is full hop>immediate DJ>waveland fast than full hop>waveland?

Depends on platform height. For low platforms like the side platforms on YS, full jump → waveland is faster (10 vs. 12 airborne frames). For slightly higher platforms (DL64 side platforms, for example), both take 15 airborne frames.
Because you need to add 10 frames landing lag for these at the end, neither are ever faster than no-impact landing. Doing so, for example by immediately using your second jump under DL64’s side platforms, takes only 21 frames until you are actionable, counting from first airborne frame. Full hop → waveland takes 25 frames until you are actionable.

On DL64’s top platform, riselanding takes 33 frames while full jump → second jump → waveland takes 40 frames until you are actionable.

Question 2: When underneath/near a platform, is jumping up, wavelanding off the platform, and then immediately fast fall-fairing faster than a regular SHFFL'd fair? What are the speed differences of this tactic across different platform heights on tourney legal stages (excluding FoD)?

Interesting question! On YS’s side platforms, doing the waveland as close to the platform edge as possible makes the entire motion one frame faster than a shffl fair. In my opinion not worth the effort, as your positioning needs to be very spot on, and it’s less favorable on other stages, except FoD like you mentioned. On DL64’s side platforms, it’s 32 frames. On other side platforms, it scales with platform height.

Kadano, what about using ftilt, fair, (or nair?) to catch fox out of his side-b when he does it higher than dtilt height? (Or does this require inhuman reaction time?)

If you time and space them so that they cover a full length Illusion, Fox can shorten on reaction and escape all of these.

 

[TKD+ITA+Mar=]

If you time and space them so that they cover a full length Illusion, Fox can shorten on reaction and escape all of these.

Thanks a bunch Kadano! Just out of curiosity, how many frames would the fox need to shorten on reaction to ftilt? Because if for example it's something like 10 frames, I think looking at http://www.humanbenchmark.com/tests/reactiontime/stats.php that 90-95% of people wouldn't be able to react quick enough, which would be a margin of error I'd be willing to accept.

EDIT: In fact looking at http://www.humanbenchmark.com/tests/reactiontime/allresults.php and seeing that only the top 40 on the site can react that quickly (~166) to a much simpler stimulus, I'd be fine with a 10 frame reaction window.

EDIT2: This is a more realistic leaderboard of reaction times: http://www.humanbenchmark.com/tests/reactiontime/allresults.php?period=30&expert=on and it's even rarer for people to break 166 ms here.

 

[Kadano]

9 frames. So, very few Fox players will actually be able to react accordingly. They will probably guess from your positioning and go full length if you crouch or dash towards them and shorten if you stay on stage. That way, you need to guess their guess in order to hit them.

 

[Purpletuce]

Yeah I think 9 frames is a window where you could actually assume nobody will react to it. Also, I'm aware that this is 'advanced frame data application', but I think some of this should have some more reasonable inputs. For example, it is good to know that it is possible for you to dropzone fair on reaction to spacies going for a sweetspot, but with a window as small as 10 frames, I wrote that option off as inviable. Since that is the only option you mentioned, I'd appreciate if you maybe expanded on the options. Si