Wednesday, August 16, 2017

Advanced FD Uthrow Combos on Fox and Corrections

By now most of you have seen the "simplified tables" for combos on Fox and Falco in my previous posts, which were meant to make things more memorizable and easy to execute without losing much effectiveness.  Some simplifications, however, hid some important complexity, so now that my readers have a good bird's eye view of how the combos are supposed to go, it's time to delve into some details.

The biggest simplifications were that I treated both toward DI and side DI to be the same and each DI category had the same break points for their followups.  Break points in the percent ranges were hardly even a real thing, given the non-binary nature of the game, only suggestions for ease and simplicity. I felt that this was needed in order for players to be able to handle all the information, but after a lot of practice, I was able to recall more than I thought possible.

When you think about it, it doesn't make too much sense that the percent ranges listed under "no DI" would be the same as those under "side DI" when the followups were often different.  It just happened to fit pretty well together that you could start doing soft uair->regrab on no DI at around the same percent where you should start doing tipper uair->Fsmash on side DIs, for example.  For those starting out with doing percent-based combos, it made it really easy to break the combo into sections.
The same is true for the conflation of toward DI and away DI.  Uthrow does not have a perfectly straight-up trajectory, so the full side DIs will have slightly different trajectories that slightly affect the followups.  For the percent ranges given, everything still works regardless of the inaccuracies resulting from the simplifications, which only leave out things that do work instead of listing anything that doesn't work, but the difficulty of each cell varies, and the goal of ease of execution was sometimes sacrificed for memorizability.

Memorizing can be practiced as well as execution, so what does a truly optimal uthrow combo strategy look like?

Transitioning from the chaingrab into the juggle

The transition from repeated regrabs into doing uairs or utilts is one of the most important parts of the combo and where many make mistakes even if they are chaingrab masters, including top players.  The percent limits of the chaingrabs on Fox and Falco are affected by the differences between the different DIs.  Instead of putting a hard limit on when to stop chaingrabbing for any DI, you should end the chaingrab once they approach the limit for the particular DI they've done.  Against Fox, this means going for uair->regrab on away DI at about 50% but regrabbing if they do toward DI here, waiting until about 55% if they continue doing toward DI.  The higher the percent at which you decide to uair, (up to a much later point, of course), the easier it will be to hit the uair, react to the DI and get your regrab.  The trajectory for toward DI is slightly lower and further away at any given time, so you want to do your uairs at higher percents than you would for away DI.  Again, the old table still works, but this is a slight optimization.  

That's if you still plan on taking a purist no-utilting path that I strongly suggested in the past.  Even if you avoid the difficulty of taking chaingrabs right to their limit, the uair->regrab transition from no-DI or slight toward DI is fairly difficult to pull off, requiring a pivot uair, as a SH back uair is weak to DI away.  This occurs at 47%.  The chaingrab limit on no DI is at 46% (away: 52, toward, 61 (58 is practical)).  Taking this path then requires both a frame-tight regrab or two and a difficult pivot uair if the opponent is doing no DI or slight toward DI.

Here's how I would map out the difficulty of the relevant percent ranges:

21-30: easy
31-43: very easy
44-52: very hard
53-59: hard
60-64: medium
65-75: easy

Utilts at around this percent hit too hard if the opponent DIs in the direction of Marth's back, so the alternative given was a mixup between walk utilt (or middle-hit usmash) and turnaround utilt.  However, this gives your opponent a chance to escape.

But, it's possible to skip this difficult part of the chaingrab if you utilt earlier, where you can still follow up on any DI.  The percents where you can uthrow-(no DI)->utilt-(any DI)->regrab, 29-34, are a bit too low to make it past the hard range.  37-44 seems to be a good range to front-hit tipper utilt easily on side DIs (walk forward for away DI) while having a turnaround back hit utilt for no DI and slight DIs that won't give them a chance to escape.  These utilts also hit hard enough on away DI and no DI that you'll be able to follow with an uair->regrab.  Utilt->uair->regrab will put you right into medium or easy percents.  Less ideal is if they DI toward on the utilt and you're forced to dash regrab, however "hard" or "medium" is better than "very hard."

The only stipulation is that these utilts may be susceptible to SDI.  If you find that they are too far away to regrab, you can still mix up between a dash attack or dash SH tipper fair, both leading to a regrab at those percents.  Dash attack loses to toward DI, while fair beats it.  If you can't even get those, a tech chase should still be possible. You need to be able to react and adapt on the fly to strange DI scenarios, which brings me to the next section.

Note: this strategy doesn't work as well on Falco, since his utilt followup range on side DI is about the same, while his very hard range occurs later (53-59).  This means that unless you can get an utilt->uair->regrab combo instead of just an utilt regrab, you won't skip the very hard range at all.

Uair regrabbing unreliable?

Sometimes they will go too high on your uairs.  Maybe you did them too early, maybe your opponent ASDI'd up.  What I have neglected to mention was that there is a middle height between uair->regrab and uair->tipper-Fsmash.  It is usually possible to avoid this middle height by timing your uairs carefully, but sometimes things don't go as planned.  It's very difficult to get the right timing to get uair->regrab at 60-64 as I suggested in the table, anyway.  The solution of course is to do uair->uair (both tipper).  At low uthrow->uair percents you might even need to do three tipper uairs.

You will just have to judge when you need to go for a second uair and when you can regrab in the percent ranges between early uair->regrab range and uair->Fsmash range.  Remembering the percents from the table can give you a good gauge, though. 

Soft Uair unreliable?

While we're talking about uair regrabs, the soft uair regrab loop was a major feature of the flowchart.
There are actually two soft uairs, a middle sword hit and weaker body hit.  The percent range I gave for soft uair->regrab was 65-79, however, you can do it with the middle hit as early as 55.  Body hit only causes Fox to enter tumble at 67.  The thing is, it's really hard to hit the middle hit, and it's hard to tell that you've done it sometimes.  The slight difference in knockback can mess you up at later percents where you have very little time to react to DI and so little time to differentiate between trajectories as well.  So, you should honestly only be going for this past 67, up to about 75, since if you hit the middle hit there, they'll be hit too high.

The other issue with the move is its weakness to toward DI.  The natural trajectory sends Fox slightly behind Marth, so toward DI sends them far away.  The 80-94 range for soft uair->Fsmash is quite unreliable because you can't tell if you need to pivot or not.  SDI makes this even more difficult.  Half the time I settle for the soft Fsmash into an edgeguard, which isn't bad but it isn't ideal.

An alternate up to about 75 is to do a SH back tipper uair->Fsmash, but that has a weakness to DI away.  Pivot uair into pivot Fsmash solves that issue, though it is difficult. 60-88 you can pivot Fsmash to either direction, and beyond that, up-B.

Just taking the Fsmash or up-B is especially advisable when you are not in center stage.  An extra 10-13% damage tacked on before your Fsmash is not an improvement if it forces you to send them in a worse direction.  That segues nicely into a discussion about stage position.

Hold your ground

Using ikneedata, I came to an important conclusion: launch position and direction can be more important than launch percent.  The distance they end up off stage is very dependent on where you hit them from.  I devoted a good portion of the Fox table post to the topic of using fair juggles in cases where you want to hold stage position, but there is more that you should know about how it affects your optimal followups.

If it's a good idea to use fair you have stage position, it's a good idea to use uair when you don't have it, since there's a chance they will DI toward and the position will reverse in your favor.

If you're off to a side of the stage, taking an Fsmash or an up-B is a strong option, especially when you don't have the option to fair, as is often the case when they do no-DI.  I did recently learn that you can pivot fair, which results in a decent mixup with pivot Fsmash, but that will still only work up to 72 or so.

It's not always necessary to go for a mixup with fair.  Instead you can take the edgeguard mixup from a low percentage Fsmash or a high percentage up-B.  However, sometimes it is really obvious when your opponent is DIing toward you, you just have the read.  Many players, even top players, can't resist holding toward, especially because DIing away against soft fair is also dangerous for them by the edge at high percents.  The correct answer would be to do no DI, but players tend not to do that in fear of strong hits.

Fsmash and Up-B

So now you might be wondering when you can land these strong hits, really.  In my table I said up to 95% is when you can dash pivot Fsmash.  However, this is ridiculously difficult on away DI, and still very difficult on toward DI, and in fact you need to do a long dash-wavedash.  Pivots for earlier percents still require practice.  Being realistic, I say you should start trying it at 85% and when you're comfortable with that, up to 90%.  It's also possible to land soft uair on side DIs, with it being more practical on away DI, at around these percents.

Charging Fsmash has a noticeable effect.  For every three frames you charge, you get the equivalent of a 2% higher pre-hit percent Fsmash.  You can pivot charge Fsmash against no DI if you're able, or you can do standing charge Fsmashes on side DIs (up to 69% on away DI, 72% on toward DI).  This tends to make a bigger difference on smaller stages or when you're off to a side of the stage.  On the other hand, charging telegraphs your Fsmash, letting the opponent execute optimal DI and SDI.  If you decide to charge Fsmash, make sure it will kill outright or lead to a guaranteed edgeguard.

Some players say they are able to react to Fsmash no matter what, but that doesn't seem to always be the case.  To exploit this, Fsmash can be used as part of a mixup with fair.  Players tend to have some DI pattern, watch for it.  Even trickier is to mix up what direction you Fsmash them in, with the use of dash dances and pivots when you uthrow them with no DI.  

Up-B percent listings were just incorrect.  I couldn't seem to land reverse up-B past the percents given.  Using non-reverse up-B (yes, with the strong hit), I was able to get the following, including in real matches:

Away: up to 130
Neutral: up to 125
Toward: up to 120

Again, start lower and practice it until you can get these consistently, it's worth it.  You can mostly disregard what other options I had at 95-120.  Up-B is criminally underrated, just don't use it when they won't be sent far away enough off stage for you to fall to the ground and edgeguard.

It's also possible to up-B in either direction on uthrow no DI via pivot jumping in order to face the other direction before double jumping back.  This sacrifices some of the percent range of possibility, depending on how fast you are, but it is very much worth it to be able to reverse positions with a strong hit if you're cornered.

Juggling out of uthrow (side DI)

So, later in the Fox table post I got into juggle options, but they were mostly not accurate, talking only about tipper fair, and the ranges tested were for no-DI, assuming that everything would still work the same way for toward DI and survival DI.  In fact, these numbers still assume toward DI. Survival DI, the DI that's meant to survive Fsmash, will affect the numbers and followups a bit.  Up-B is a good fallback when things go wrong, or you can use numbers further in these listings to continue with long juggles.  You want to keep your juggles short, this is not a stylistic choice.  The less mixups you have to do, the better.  And if dair is possible and will kill, just do it, I won't always list it.

Here's how I now think the juggles should be done (t means tipper, s means soft):

57-76 SHtFair->Fsmash
A quick fair and a tight FF is required on the low end of this range, and a big delay on the fair is required on the later end.  Alternatively, you can do a SHsFair or SH Uair to continue the juggle.

75-95 SHsFair->Fsmash
Like in soft uair combos, you don't have much time to react to properly align a tipper Fsmash.  Soft Fsmash is still pretty good at this percent.  I've found that charging Fsmash after landing and then releasing when it's time helps with reacting for the tipper.  Alternatively, you can do a SHtFair or SH Uair to continue the juggle.

96-105 SHsFair->dair/fair only
You can dair most of the time in the other ranges, but that's all you can do at this point.  If dair won't kill, use a fair to end it.

106-120 just Up-B

up to 100 or so
FH tipper fairs are okay, but SH fairs tend to give you more options.  They do have the advantage over soft fairs for giving you more time to react, but soft fairs carry further.  The way you hit the tipper fair on the way up can be varied greatly, being possible to hit with the bottom part of the fair.

As I later discovered, on uthrow side DI you can do soft uairs, which can be done at about 90-105 for away DI and 85-95 for toward DI.  Reacting to DI a little bit easier at this high percent, and you can follow with a tipper Fsmash at 90-100 if they don't do toward DI.

SH tipper uairs in high percentage juggles (up to 105) are aiming for away DI into a dair.  It's usually a dair because away-DI'd uairs will carry them across the stage.  This makes SH tipper uair as good an option to beat away DI than up-B when you're off to one side of stage.  They're equally effective at beating away DI, but toward DIs will reverse position in your favor if you uair, and up-B is slightly better against no DI here.  Fairs have slightly less DI coverage in these situations.

Choose your aerials depending on how much room you have on stage left.  Tipper uairs and soft fairs carry further, but you'll some room on stage to dash in order to follow up on them.  Tipper fairs and soft uairs can be used when you're lacking room, or you can end it there with a strong hit or a dair that would only work if it isn't DI'd in as your option coverage for away DI.

At many points in the juggle, you can end it with up-B.  This can come into play especially when they do no DI on your SHtUair at a fairly high percent, above uair-(no DI)>Fsmash range (80+) or fair-(no DI)->Fsmash range (75+).  Or, when they just happen to go too high for Fsmash due to the opponent's DI or your mistake in timing the aerial and you need a strong hit or don't have a good juggle followup.

There are several other ways to continue juggles, such as with side-Bs or reverse fairs and bairs, but they rarely find usage right out of uthrow and are often not necessary or worth the risk.  However, there are some situations at high percent, as part of extended juggles, where these options are viable, along with neutral B, generally when they are just out of reach above you or above their effective percent range for the usual options.

Further research needs to be done on juggles from uthrow no-DI.  The juggles that I listed for 95-110 no DI are not that great, often if you're close enough to the edge where reverse aerial juggles are useful, you're also close enough for up-B into edgeguard to be a more consistent option.  The side-B->uair->up-B combo is useful in theory at center stage, but I have yet to land it in a match.

That's enough for now

I probably could have broken this up into multiple posts, but the sections flowed so well together! Next time I'll go over any loose ends regarding uthrow combos on Falco on FD.

Saturday, August 12, 2017

About Me and This Blog

Let me tell you a little about myself.  I started playing Melee late 2009, a couple months after I started playing Soul Calibur 4 competitively.  I focused much more on SC4 and SC5, and achieved some pretty good tournament results.

However, I remained bad at Melee, finally seeing steady improvement after I quit Soul Calibur (dead game, waiting for SC6) and started focusing entirely on Melee.  

Melee is quite different from Soul Calibur, being much faster, more chaotic, and more mechanically complex which greatly affects the mindset of the players.  In Soul Calibur and many other traditional fighters, learning and practicing your combos is one of the early steps you take before being a serious competitor, and the necessary information is readily available on their community sites.  However, it's difficult to find any good information on combos (or really anything else) in Melee.  Especially when I started playing, there were absolute seas of misinformation that endlessly frustrated me in their incompleteness or inaccuracy.  Combo videos were not demonstrations of basic combos, but montages of rarely applicable combos.  The differences in player mindsets certainly don't end there, but this is a problem that I feel capable of solving as a currently unranked (barely) San Diego player.

Take this old chaingrab guide for example:

  1. 1. 0 - 16%: Regular regrabs
  2. 2. 17 - 32%:
  3. 2a. No DI or slight behind DI: Pivot regrabs.
  4. 2b. Any other DI: Regular regrabs.
  5. 3.0. 33%:
  6. 3.0a. No DI: Turnaround uptilt -> regrab
  7. 3.0b. Slight behind DI: Uptilt -> regrab
  8. 3.0c. Any other DI: Regular regrab -> pummel
  9. 3.1. 34% (the reason it's different from 33% is that uptilt starts becoming unreliable against no DI because of its weird "not-upwards" hitboxes. If you actually get the correct hitbox, it'll still connect into a regrab, so if you feel confident that you can get it every time then just follow the chart for 33%):
  10. 3.1a. No DI: Rising SHFFL uair --> regrab
  11. 3.1b. Slight behind DI: Uptilt --> regrab
  12. 3.1c. Any other DI: Regular regrab
  13. 3.2. 35 - ~59%:
  14. 4a. No DI or slight behind DI: SHFFL uair -> regrab
  15. 4b. Any other DI: Regular regrab
  16. 5.0. ~60 - ~65%: Pummel at least once before throwing -> delayed SHFFL uair -> regrab (they should now have at least 80%, but less than 90%)
  17. 5.1. ~69 - ~75%: Pummel at least once before throwing -> delayed SHFFL uair -> tipper (post-tipper damage should be around 105%)
  18. 6.0. ~80 - ~84%: Pummel once or twice before throwing -> delayed SHFFL uair -> tipper (post-tipper damage should be around 110%)
  19. 6.1. ~85 - ~87%: Delayed SHFFL uair -> tipper (post-tipper damage should be around 115%)
  20. 7. 90+%: Either upthrow -> tipper or upthrow -> weak uair -> tipper

The problems with this guide are endless.  There's directions for 33% only, in a game where staling is going to affect slight differences? It isn't even true that utilt becomes unreliable due to hitboxes, it only arguably becomes unreliable due to higher knockback at later percents.  If you want the right hitbox, you want to do turnaround utilt.  30-33% is a fine place to do utilts that aren't too susceptible to escapes by SDI.  35-59 chaingrab on all DIs? That ain't Falco, and even the chaingrab on him ends at 52 on no DI.  No direction on how to uair against slight DI toward in that range.  66-68 and 76-79 are just skipped.  90+% you can't "uthrow -> tipper" and there's no details as to what do against side DIs.  No mention of uthrow->up-B?

This and guides like it are all people had to work with for years.  SSBM tutorials put out a chaingrab guide in 2016, but its directions for transitioning out of chaingrabbing into the juggle or finishers are vague, suboptimal and misleading.  And nobody gives any special attention to comboing Falco.  Kadano has put out great combo tables, but his Sheik data seems to contain mistakes.  His Falcon table was excellent, but overly detailed and lacking any guidance as to what followup one should choose.  What I mean by that is, nobody should be interested in learning every possible followup in every situation.  What's important is knowing the best followups for the situation -- not just what can be done, but what you should do.

Even my own guides on this blog have had their problems that have either fixed or saved for a later "advanced guide" to deal with advanced anti-combo techniques, which I have neglected to mention because I don't want non-Marth players to know about them until I've developed a good way of beating it.  

Just play by feel, they said...

Most of the time when I ask players about combo options, they reply "I just do it by feel."  This makes a lot of sense.  Most characters don't need to memorize combo percents, and can just look at the situation and see what they can continue the combo with.  The problem is, this doesn't work.  Players are not nearly maximizing their punishes unless they have done some kind of labbing.  Falcon players and Sheik players have quietly done their own homework on comboing Marth and other characters to their fullest, and Marth players need to do the same to catch up in the meta.

And it just isn't enough that Marth players know their stuff when it comes to combos on FD.  The only combo guide that takes into account platforms at all has been Kadano's uair platform tech chase guide, which, as it turns out, doesn't have as much applicability as many players seem to think.  Doing the setup right out of uthrow, which is to drift or FF uair on reaction to tech options, only covers tech in place with a tipper uair past 43% or so, and 50% for Falco, and something really high for Falcon.  It also won't cover all options if you're to the side of the platform and side DI'd to the platform.  And people have no idea what do out of those uairs, should they hit, they just go for more uairs, racking damage until they escape, with no good plan for a strong combo finisher.

Every other character often seems to have a well-developed metagame for their punishes, which are often much simpler than Marth's.  They often can get away by feel because they don't benefit much by thinking past the next hit.  Marth can actually be punished if he messes up some of his combos or if attempts to continue a combo when he should instead apply pressure or a tech chase mixup.  Marth seems to be the only character whose top players often seem to be flailing around in their combos without a plan.  You need a plan with Marth or else you won't finish your combo well.  I'm not as good as the top Marths, but knowing what I know now, they very often make obvious mistakes.

Playing by feel to some extent is unavoidable, given the amount of possibilities, but the more you know, the better at playing by feel you will seem to be.  If you know what you're doing, the gaps in knowledge that you encounter and need to close with intuition will be much smaller.

This blog is meant to give you the knowledge necessary to comfortably navigate the complex punish game of Marth.

What's to come

After months of labbing and testing in real matches I have a large portion of the uthrow combo game on platform stages mapped out, so hit that follow button.  I'll also tie up any loose ends in the FD combo guides.  My level of play isn't sufficient to take my suggestions only by my word, but I will be backing everything up with hard data and game theory.

Wednesday, August 2, 2017

The Art of Juggling

The Falling Juggle

In the *video game* Super Smash Bros. Melee, in *Esports* competitive play (Adsense's crawler has no idea what this blog is about, just making it clear)...

Marth wrecks a lot of characters just by being below them, but somehow, they still make it down. Why? What's the best way making sure you never lose the exchange? What are they doing correctly when they manage to get down? I'm going to try and answer these questions. Marth's Uair beats everything head to head, so it seems like it should always win, but there are a few cases where it doesn't:

1. They hit you before it comes out
2. They trade favorably
3. You do it too early
4. They move too far to the side 

Sound familiar? The juggle situation is it's own kind of vertical footsies or "neutral game," except it's in your favor: they don't have the option to retreat or stay still except with an expendable double jump, and your hitboxes are more disjointed than theirs. So if you let them down, depending on the character, it's because you timed and positioned the uair wrong.

Here's how I suggest you make it easier:  Time and position the jump so that you encounter them with as little upward momentum as possible. This slows it down just in case they try and surprise you by closing the distance when you jump. If both players move toward each other, they close the distance twice as fast as normal, making it difficult to react and time the attack properly.

You also want to be as close to right below them as possible. If they are already drifting to the side, you should dash just past their current position. Their drift is aiming to avoid directly challenging Marth's uair from above, getting a better angle where they can pit their fair or bair against the side hit of your uair or your fair instead of trying to beat your uair with their dair.  Or they are trying to avoid challenging your aerial at all, by getting out of range, or enticing you toward them and suddenly changing their horizontal direction with a double jump or air dodge.  By the time you react to their change in direction, they may be too low to the ground and far away and you won't be able to continue your juggle.  This is especially the case if you jump toward them with momentum.  

In order to account for both possibilities, them drifting away and them changing momentum, I recommend shield stopping before jumping to stop horizontal momentum and staying close to the ground. To stay close to the ground, you either want to short hop or full hop to a platform. Instant double jump for that middle height jump might be useful too. If you jump too early, they have time to wait out your jump or move to the side. If you jump too late, you'll be rising right into them and your timing will have to be stricter and you're in worse position if you miss. The key is to get a sense of the right height the enemy character needs to be before you choose to jump, based on how fast they fall. Other than their double jump, they only have their fast fall timing, their aerial drift, or some character specific ability to vary where they will be in the near future behind the veil of reaction time.

Purple Rectangle has little choice but to fall into Blue Rectangle's attack.

Staying under them without too much horizontal momentum allows you to follow their drift with your own drift, and staying close to the ground allows you to land and disengage or throw out another attack, if need be.  You don't always have to time to attain this perfect positioning, especially after disengaging, but you want to get closest to this ideal as possible.  And as long as you're not completely overextending with high vertical or horizontal momentum, you can punish any air dodge attempts, as well.

You do want to kill them eventually, though.  Endless uairs doesn't do you much good. You want to lead them toward the edge of the stage, eventually offstage, without a double jump at high percent if possible, to secure an easy edgeguard. This can be done by staying slightly toward the side of center stage instead of right under them. Mixing in fairs and bairs will get the job done at high percent.  There are also mid percent ranges for each character where soft fair Ken combos become worth going for.  Look forward to a post about that in future.

Character specific stipulations

If you choose your jumping position, height and fastfall timing properly, there's little that many characters can do, but some have ways of handling it. Double jumps, fast fall speed, good downward hitboxes, and horizontal momentum can make it nearly impossible to consistently juggle a character. Against the fast fallers, they fall too fast so they have to be extremely high up for you to have time to get below them and jump early enough to not be on the way up by the time they get to you. Their FF is too fast to react to, so you have to guess their timing. If they double jump at the right timing, you can't continue the juggle easily. If you stay close to the ground, as with this strategy, you won't be flat out punished, though.  But you should mix in standard dash dancing more often than not.

On the other side of the extreme, Puff has enough horizontal momentum to sort aerial dash dance to throw off your below positioning, and can simply move to the side and beat you with a bair. You can still limit options if you execute this strategy well and mix it up, but you have to take great pains to maintain an offensive stance against so many double jumps.  The same is true against Kirby, except he has a more threatening dair and the down-B that punishes any over-extended aerial chase.

Middle weights are where you'll see the most difference using this strategy. Even if they double jump, you can simply fast fall to the ground and follow them with another dash shieldstop SH. Floaties will want to LRA+START and switch characters. 

Sheik and Pikachu (and Pichu) have very little to help them come down, though Sheik benefits greatly from getting to the side of you, and Pikachu's double jump is quick and low enough that any imperfection at executing this strategy can leave you with not enough time to consistently stop Pikachu's FF aerial after he does it.

In the Marth ditto (and against Roy), side-B and sideways aerial drift is used, but proper execution of this method makes it matter very little.  Only a very precise, non-boosted side-B which alters Marth's fall speed just enough and not too much can save him, and if you simply wait it out, it is punishable on reaction.  Mixing up the heights and timings of your aerials makes it pretty much impossible for side-B to ever work, nor will a falling aerial save him.

Peach will try to float just outside of your SH uair range, but you don't use rising aerials like it recommend, it won't matter if you whiff, you can just land and utilt or jump again before Peach can get to you.  You can just follow her drift and wait out her float, mixing between dash shield stop SHing and delaying the aerial and doing one right away.  The Peach player needs to hard read your aerial timing and drift within a limited period of time in order to get down.

Samus can bomb to avoid an attack, but if you are executing the falling juggle, this gains Samus nothing but a brief period of sideways drift which can throw you off.  You can repeatedly dash shield stop SH aerial against a bombing Samus until platforms force you to do something else.  Her only hope is use the bomb drift (she briefly gains the ability to nearly instantly change directions in the air) in those situations.  Just stay safe, you don't want to risk taking any damage against this character, as you need to hit her a lot more than she hits you, with the exception being when you can go for mid-percentage Ken combos.

The Marios tend to use their dairs or air dodges to get down, but this loses to the falling juggle.  Capes are about as ineffective as Marth's side-B.

Luigi and ICs can use their down-B and side-B respectively and mash just enough to get over your aerial and then fall into you.  In this case, the falling juggle still works, but you need to mix it up a bit.  Completely disengage, baiting it out, and you get a free punish.  Mix up your aerial timing and they'll be hard-pressed to get their hard read.

Falling uairs can still work against Ganon, DK and the Links, but the timing is not easy.  You definitely want to mix in baiting them out with dash dance and punishing their end lag to skew the risk/reward further in your favor.

Not sure what to do against Yoshi.  He can easily time his double jump armor on reaction to a falling aerial, and a rising aerial is really dangerous to do.  Maybe SH, fall, and shield or dash away? He must DJ and do an aerial if he wants to come down and be safe against an aerial.  There's also side-B to worry about as well, that pops him up a bit.  That's if you're going to jump at him at all, it might be better to just stay grounded, dash dancing and shield grabbing his aerials when he reaches that predictable spacing.

Let's see, who's left? Zelda's never getting down.  Mewtwo is almost always getting down with up-B, you have to hard read where he will go.  Ness is not getting down except if he gets to the side of you and tricks you with a DJ aerial that can suddenly come at you diagonally (just bait it out).  Bowser will be trying to up-B away, I think, just follow him.

Stage specific stipulations

Three-platform stages complicate things. You have to see what platform they are trying to land on and aim to land there with the double jump. If they make it ambiguous by being positioned between the platforms, wavelanding off or platform dropping off the top platform to a low platform can work well. 

Watch for air dodges into high platforms, this will beat this strategy.  The counter-play is to save your DJ, bait it out, and DJ with an aerial or a NIL grab, or to do an aerial that barely lands on the platform.

It can be quite easy to juggle opponents on FD and PS because they have no top platform for refuge, with the exception of PS transformations where they have very safe places to escape to.


So that's about it.  The TLDR is that you should shield stop under them, jump earlier than you might think and hit them with falling uairs as they fall toward you. Stay close to the ground or platforms when you're in the air to stay safe. Lead them off stage.  You can't always do this if they're not high enough above you.  Recognize when you can and can't do it, and acknowledge the weaknesses and understand the mixup when you can't do a falling juggle.  This really important part of Marth's game, as it replaces a combo-based punish game against characters that don't have elaborate tables and flowcharts of true combos to work with.