Some Yoshi Information

[DD151]

I tried to determine which WBK attacks exceed the 120 KB reduction on Yoshi's DJ armor, but I was getting some funny numbers that weren't corresponding to empirical results. For example, I calculated DK's side-B KB on Yoshi to be 151.27 (for all hitboxes, grounded or aerial), but Yoshi never flinched out of his double jump when he got hit by the attack. The same anomaly occurred with Sheik's F-smash (first hit), which I calculated to inflict 124.34 KB.

I think it might be worth asking someone like Kadano to look into this? There are some possible explanations for my results; Kadano's formula for WBKB might be incorrect, I might have made a computational mistake, or Yoshi's DJ armor can't be broken by WBK moves.

 

[Sashimi]

I'm not sure which formula you used for WBK moves, but as far as I know, it is this:

((((p/10+p*d/20)*(200/(w+100))*1.4)+18)*(s/100))+b

Where
- p is 10 (for normal moves it is the damage of the character after being hit)
- d is the WBK value (for normal moves it is the damage of the attack)
- w is the weight of the character
- s is the knockback growth (always 100 for WBK moves)
- b is the base knockback (seems to be 0 always for WBK moves)

So Sheik's Fsmash 1st hit would deal 59.7307692308 KB by my calculation. If I'm wrong, please correct me, but this seems more consistent with Yoshi's armour.
n any case, it is easy to test for whether or not the WBK moves will knock Yoshi out of his armour. For the finnicky moves like Pikachu's Uair the formula is useful but for most moves I was just going to get two controllers and hit Yoshi while he's double jumping (or crouching since 120*2/3 is equal to 80). It wouldn't take long since there's no need to test for any % threshold.

This formula also gives me 79.3903743316 KB for Fox's Shine on Marth, so although I don't know the exact KB value Marth is supposed to receive from that move, it seems at least fairly accurate.

 

[DD151]

I'm using the KB formula provided in Kadano's Melee mechanics video: https://www.youtube.com/watch?v=c93Fwem-ngw

k = (s/100) * ((14*(p+d)*(d+2) / (w+100)) + 18) + b

So the first problem here is that the formula you posted doesn't simplify to Kadano's formula. Simplified, your posted formula is:

k = (s/100) * ((14*p*(d+2) / (w+100)) + 18) + b

But that's okay for the WBK attacks, because in Kadano's video, he says that for WBK moves, (p+d) = 10 and d in the (d+2) term equals the WBK value. Since in your formula, p = 10 for WBK moves, they end up being the same thing.

For Sheik's F-smash (first hit), the values that I have are:

d = 156, w = 108, s = 100, b = 0
k = (100/100) * ((14*10*(156+2) / (108 + 100)) + 18)
k = 124.35

So do I just not have the right value for d? Because unless I'm using the wrong equation or simplifying the terms improperly, that's the only possible explanation for the error. Using w = 87 (Marth), Fox's down-B (d = 80) does indeed have 79.39 KB.

(EDIT: your calculation is consistent with d = 60 for Sheik's F-smash (1st hit).)

Now as for the normal non-WBK attacks, for which the discrepancy in the equations is significant, I also tested Sheik's N-air (clean), F-air, and U-air (clean) in training mode and confirmed that the %ages calculated using Kadano's KB equation are accurate.

For the finnicky moves like Pikachu's Uair the formula is useful but for most moves I was just going to get two controllers and hit Yoshi while he's double jumping (or crouching since 120*2/3 is equal to 80).

According to the spreadsheet, Pikachu's U-air does not have WBK on any of its hitboxes.

 

[Sashimi]

I'm using the KB formula provided in Kadano's Melee mechanics video: https://www.youtube.com/watch?v=c93Fwem-ngw

k = (s/100) * ((14*(p+d)*(d+2) / (w+100)) + 18) + b

I got my formula from smashwiki so I guess I shouldn't be surprised that it's wrong. I'll use this one from now on, thanks!

For Sheik's F-smash (first hit), the values that I have are:

d = 156, w = 108, s = 100, b = 0
k = (100/100) * ((14*10*(156+2) / (108 + 100)) + 18)
k = 124.35

So do I just not have the right value for d? Because unless I'm using the wrong equation or simplifying the terms improperly, that's the only possible explanation for the error. Using w = 87 (Marth), Fox's down-B (d = 80) does indeed have 79.39 KB.

(EDIT: your calculation is consistent with d = 60 for Sheik's F-smash (1st hit).)

Yeah, Sheik's WBK value is 60 according to Master Hand. I don't know if the spreadsheet is right but 156 is a very high number for a WBK value.

Now as for the normal non-WBK attacks, for which the discrepancy in the equations is significant, I also tested Sheik's N-air (clean), F-air, and U-air (clean) in training mode and confirmed that the %ages calculated using Kadano's KB equation are accurate.

This is good to know, thanks! Again, I appreciate you posting here.

According to the spreadsheet, Pikachu's U-air does not have WBK on any of its hitboxes.

Huh, you're right. It always seemed like it was set knockback to me. Ignore that then, haha.

 

[DD151]

I'll check out Master Hand and correct the WBK values, then, and then I'll see if I can post some findings.

Huh, you're right. It always seemed like it was set knockback to me. Ignore that then, haha.

It's probably because Pikachu's U-air has low damage and low KBG on all of its hitboxes. The trajectory for the tailspike is actually not even negative, it just looks negative because a 0 degree KB angle means that the character who gets hit starts falling instantly.

Random thought, it might be worth figuring out at what %age Yoshi's D-smash has greater KB than his D-tilt vs. various characters. D-tilt has a slightly lower trajectory (28 degrees vs. 30 degrees) but the difference is almost negligible.

 

[Sashimi]

Yeah, Dtilt vs Dsmash is on my to-do list now that you've shared the KB formula with me. Quite a while ago I tested for it originally with the SSBwiki one but obviously that was wrong and I ended up with Dsmash being stronger at 0%, which is clearly not true because Dtilt causes tumble at 0% and Dsmash doesn't. I didn't bother looking for a new formula at the time, but I will definitely get back to that (as well as recalculating the % thresholds for optionless tap resets).

 

[DD151]

Sorry to steal your thunder, but since I had the time, I went and did the D-tilt vs. D-smash comparison. It's the second sheet on the spreadsheet (same Dropbox link as before). I only eyeballed the attacks vs. Bowser in-game, so the numbers mostly haven't been verified. For each character, I have the D-tilt KB, % at which the near D-smash hitboxes out-KB D-tilt, and % at which the tipper D-smash hitbox out-KBs D-tilt. D-smash is assumed to be unstale and I didn't include the second hit because that seemed irrelevant.

Also, when you have the time, I have a couple of questions that I'd like to investigate, but I don't have the tools.

1. Can you check until what %s U-air combos into D-smash (or D-tilt) on fastfallers? I've seen U-air -> D-smash used as a DI mix-up near the side of a stage at low %s and I want to know exactly how useful this is. Obviously a target can get hit at various initial heights with the U-air, but assume for now that the target is grounded and Yoshi is doing the earliest possible DJC U-air (off, say, a jab reset).

2. Is an aerial tech-chase possible on reaction? What I mean is, suppose that Yoshi hits the opponent with an attack that forces him to tech upon hitting the ground. Is it possible for Yoshi to jump in the air before the opponent techs and cover tech in place + missed tech, techroll in, and techroll away? I've hypothesized that from the apex of a short hop, Yoshi can fast-fall U-air to cover tech in place + missed tech and then DJC F-air to cover techrolls, but I can't confirm if this would work in practice.



Minor correction on a description in your opening post: for Egg Toss angles, there are 29 possible angles within each shaded range, so there are a total of either 59 or 61 possible angles. I'm not sure if 29 angles include the max forward and max backward angle, but schmooblidon on reddit says that there are 61 total possible angles.

EDIT: Oh! I have a third question that maybe you can investigate if it's not too much trouble. Actually, now that I think about it, it's quite complex, so you can approach it however you like.

3. At what %s can DJC N-air lead into U-tilt, F-tilt, or D-tilt? Obviously there are a lot of variables here, but I'm curious if it's possible to follow up a stray N-air on an enemy at low % with something that can either set up a combo or improve stage position that is not a dash grab.

 

[Sashimi]

Sorry to steal your thunder, but since I had the time, I went and did the D-tilt vs. D-smash comparison. It's the second sheet on the spreadsheet (same Dropbox link as before). I only eyeballed the attacks vs. Bowser in-game, so the numbers mostly haven't been verified. For each character, I have the D-tilt KB, % at which the near D-smash hitboxes out-KB D-tilt, and % at which the tipper D-smash hitbox out-KBs D-tilt. D-smash is assumed to be unstale and I didn't include the second hit because that seemed irrelevant.

No need to apologize, it's very helpful. I don't do this for any kind of recognition, I just want this information to be available! Also, thanks to Magus's new code, I will be able to double check those %s ingame with more accuracy than just eyeballing it so don't worry about that.

Also, when you have the time, I have a couple of questions that I'd like to investigate, but I don't have the tools.

1. Can you check until what %s U-air combos into D-smash (or D-tilt) on fastfallers? I've seen U-air -> D-smash used as a DI mix-up near the side of a stage at low %s and I want to know exactly how useful this is. Obviously a target can get hit at various initial heights with the U-air, but assume for now that the target is grounded and Yoshi is doing the earliest possible DJC U-air (off, say, a jab reset).

Yep, this won't take too long. I will add this to the OP under "combo options."

Edit: Done. The combo will work at slightly higher % if they DI left or right, but for now I am not including DI in my combo %s because there are so many variables.

2. Is an aerial tech-chase possible on reaction? What I mean is, suppose that Yoshi hits the opponent with an attack that forces him to tech upon hitting the ground. Is it possible for Yoshi to jump in the air before the opponent techs and cover tech in place + missed tech, techroll in, and techroll away? I've hypothesized that from the apex of a short hop, Yoshi can fast-fall U-air to cover tech in place + missed tech and then DJC F-air to cover techrolls, but I can't confirm if this would work in practice.

This is interesting, and it will take a bit longer but I will let you know! Off the top of my head, the Uair part wouldn't be too hard but DJC Fair would have a small frame window. Given 16 frames to react you'd have at most 5 frames to hit the techroll with Fair (frames 36-40 of the techroll). Otherwise it's just a matter of being able to cover the distance of certain backwards techrolls that go quite far like Falco's. I'll have a complete explanation later.

Minor correction on a description in your opening post: for Egg Toss angles, there are 29 possible angles within each shaded range, so there are a total of either 59 or 61 possible angles. I'm not sure if 29 angles include the max forward and max backward angle, but schmooblidon on reddit says that there are 61 total possible angles.

Good to know, thanks!

EDIT: Oh! I have a third question that maybe you can investigate if it's not too much trouble. Actually, now that I think about it, it's quite complex, so you can approach it however you like.

3. At what %s can DJC N-air lead into U-tilt, F-tilt, or D-tilt? Obviously there are a lot of variables here, but I'm curious if it's possible to follow up a stray N-air on an enemy at low % with something that can either set up a combo or improve stage position that is not a dash grab.

DJC Nair > Utilt, Ftilt, Dtilt or Dsmash are all reasonable (though Utilt can be harder to space) but at low % you will have to watch out for ASDI down. If you can catch your opponent in the air or otherwise not holding down when you hit them with Nair, the follow-up should be pretty free. Once you reach mid % (around 30+) you can techchase off of the Nair against fastfallers and wavedash into a tilt against everyone else (Ftilt being the easiest, Utilt being the most rewarding). At that point you can also dash forward into another DJC Nair against non-fastfallers. At high % all of this works with weak Nair.

I'm still figuring out the exact %s for these combos but I will edit them into the OP as I go along.

 

[DD151]

Edit: Done. The combo will work at slightly higher % if they DI left or right, but for now I am not including DI in my combo %s because there are so many variables.

Thanks, that's fine. The effect of DI on viable combo % is probably negligible. Also, because there's variation in DJC U-air timing (since we're human and not 100% frame-perfect), I think the takeaway from the data is that DJC U-air from the ground or near the ground is a viable DI mix-up on fast-fallers until around 60%.



I was hypothesizing a little bit while looking at the D-throw techchase on Falco frame data and wondering if dash DJC wavedash is a more guaranteed follow-up than dash wavedash or if you could just do a longer dash wavedash. Quick comparison by the numbers:

- The current dash wavedash (dash 5 frames, perfect wavedash) paradigm requires 20 frames before Yoshi reaches Falco and can act. Yoshi is actionable on the 3rd frame after Falco hits the ground.

- A dash DJC wavedash (dash 1 frame, longest DJC wavedash) paradigm requires 26 frames before Yoshi reaches Falco and can act. Yoshi is actionable on the 9th frame after Falco hits the ground.

Given that we're assuming a 15 frame reaction time, Falco's techroll animations are only distinguishable from frame 2 onwards, and we're only considering options to cover tech in place no slower than U-tilt, Yoshi doesn't need to be actionable until around frame 17 or so. I'm wondering if it's possible to place Yoshi closer to Falco's contact point with the ground so that he can get more reliable coverage on techroll away with dashgrab. Like, Yoshi can dash 7 frames into a perfect wavedash, or something.

(I really need to get 20XX for my Wii so I can test things out on my own instead of dumping them on you, lol.)

EDIT: in fact, it seems like a slightly longer dash should yield better dashgrab punish windows on both Falco and CF because currently Yoshi doesn't seem to get as close as possible to their contact point, unless Yoshi's dash loses horizontal momentum after frame 5. Also, since Yoshi's D-throw trajectory is 80 degrees, I think a DI angle of -17 degrees can put the opponent slightly farther away than a DI angle of 0 degrees.

 

[Sashimi]

I was hypothesizing a little bit while looking at the D-throw techchase on Falco frame data and wondering if dash DJC wavedash is a more guaranteed follow-up than dash wavedash or if you could just do a longer dash wavedash. Quick comparison by the numbers:

- The current dash wavedash (dash 5 frames, perfect wavedash) paradigm requires 20 frames before Yoshi reaches Falco and can act. Yoshi is actionable on the 3rd frame after Falco hits the ground.

- A dash DJC wavedash (dash 1 frame, longest DJC wavedash) paradigm requires 26 frames before Yoshi reaches Falco and can act. Yoshi is actionable on the 9th frame after Falco hits the ground.

Given that we're assuming a 15 frame reaction time, Falco's techroll animations are only distinguishable from frame 2 onwards, and we're only considering options to cover tech in place no slower than U-tilt, Yoshi doesn't need to be actionable until around frame 17 or so. I'm wondering if it's possible to place Yoshi closer to Falco's contact point with the ground so that he can get more reliable coverage on techroll away with dashgrab. Like, Yoshi can dash 7 frames into a perfect wavedash, or something.

You are right. I should have written "about 5 frames" because it was meant more as a general idea rather than the exact number of frames you should dash. However, with 5 frames, Yoshi ends up here before he needs to act:

This is pretty close to where Falco lands. But I think I will redo the frame windows for a longer dash to see if it makes a difference. The window for regrabbing the techroll behind is very large as it is right now, so it wouldn't hurt to move a bit further forward before reacting to the tech option.

(I really need to get 20XX for my Wii so I can test things out on my own instead of dumping them on you, lol.)

EDIT: in fact, it seems like a slightly longer dash should yield better dashgrab punish windows on both Falco and CF because currently Yoshi doesn't seem to get as close as possible to their contact point, unless Yoshi's dash loses horizontal momentum after frame 5. Also, since Yoshi's D-throw trajectory is 80 degrees, I think a DI angle of -17 degrees can put the opponent slightly farther away than a DI angle of 0 degrees.

This is true, but the difference is very small. The ideal DI angle (-10°) is rounded to 0° by the game so like you said, -17° is the best you can get. I used Dolphin TAS input here to test for the best possible angle:

Stick coordinates (255,128):

Stick coordinates (255,89):

At most I would think it makes a 1-frame difference for following up. Falco does land 1 frame sooner as well with about -17° DI, but like you said, Yoshi doesn't need to be there right when the opponent lands, so this doesn't make much of a difference.

 

[kofinater]

Am I correct in saying that the frame disadvantage on parry is the same as on shield since yoshi doesnt take stun? And that shield drop up air comes out on frame 6? From the frame advantage chart it seems like if you lose on shield drop upair then the upair was late right? Granted you weren't going to whiff. Same with sffl nair? Playing other characters my powershield -> jab loses a lot but that chart really highlights how powerful yoshi's parry/shield drop is. I'd like to work on parry -> shuffle nair more, it seems liable to miss but super fast.

 

[DD151]

I don't think that's 100% correct. Yoshi's shield doesn't take shieldstun, but it should still take hitlag, whereas invincible hurtboxes don't take hitlag. So the difference between frame advantage on parry and frame advantage on shield drop should be the number of frames of hitlag.

PerhapsMan, I figured out what's up with the KB equation discrepancy that I mentioned in a previous post. It's actually so simple that I can't believe I missed it. In your equation, p is the target's % after damage is applied, whereas in Kadano's equation, p is the target's % before damage is applied. That's why Kadano's equation uses p+d, because that equals the target's % after damage is applied.

In a similar vein, you should note in the OP whether the %s listed for combos, tumble, etc. are for before any of the hits are applied or after the hits are applied.

 

[Sashimi]

@ K kofinater Like DD151 said, frame advantage after a parry is always higher than shielding because of the opponent's hitlag.

@ D DD151 If I am reading correctly, the equation I am using still provides the right knockback value?

Also I thought I had written somewhere that the tumble % and combo ranges were for the target's % before being hit. I will fix that now, thanks!

 

[DD151]

@ D DD151 If I am reading correctly, the equation I am using still provides the right knockback value?

Yeah, as long as you're using the % after hit for your equation.

@ Sashimi , how do you figure out your combo %s? I'm trying to use 20XX in develop mode so I can get frame-perfect execution and hitstun indicator, but Fox keeps on randomly DI'ing in training mode.

 

[Sashimi]

Yeah, as long as you're using the % after hit for your equation.

@ Sashimi , how do you figure out your combo %s? I'm trying to use 20XX in develop mode so I can get frame-perfect execution and hitstun indicator, but Fox keeps on randomly DI'ing in training mode.

I use vanilla training mode with the combo counter. It's very basic but that's why I don't factor in DI or anything like that (so many variables anyway - eventually I might do that). To make sure it's accurate, I TAS anything low % that requires fast inputs and I check the upper value a ridiculous number of times by trying the combo until I can make it work or until it is clear that it won't combo.

 

[DD151]

Okay, I asked this question in the other thread:

This actually prompts two Egg Lay related questions: does swallowing a target cause the move to have different frame data, and what's the formula for an opponent breaking out of the egg?

After investigating this in develop mode, I have the following findings:

Yoshi always executes the first 19 frames of the Egg Lay animation. Keep in mind that Egg Lay has a grabbox frames 17-21. After the first 19 frames of the animation, or when the grabbox makes contact with a target, whichever is later, Yoshi then has a 30 frame animation where he swallows the target and poops out the egg. This animation is independent of the target's weight - I tested this with Bowser and Pichu. This animation can't be interrupted by a jump, and I don't think it can be interrupted at all.

The target is ejected as an egg on frame 26 of the pooping animation. I think the target can begin inputs to escape as early as 2 frames before he becomes an egg; not sure on this. I didn't test this extensively.

A target that doesn't attempt to mash out at all will be trapped in the egg for 199 frames (YoshiEgg) and will break out into the Fall animation on frame 200 if they haven't. If the target is inputting A/B/X/Y/up on the frame before he breaks out of the YoshiEgg animation, then he will break out directly into a double jump or an attack depending on the input. The YoshiEgg animation is an 8-frame loop that goes from 0 to 7. The first frame of the YoshiEgg animation when the target is pooped out is frame 1 of the animation, not frame 0 - it goes from 1 to 7, then it loops repeatedly from 0 to 7.

The YoshiEgg duration seems to be independent of the target's weight and %. I tested Bowser @ 0%, Pichu @ 0%, and Pichu @ 108%. I don't have enough time to test how mashing reduces the YoshiEgg duration such that I can come up with a formula (see EDIT), but I do have some useful results.

If the target mashes 1 of A/B/X/Y every frame, he will be trapped in YoshiEgg for 13 frames and escape on the 14th.
If the target inputs a direction on the control stick every frame, he will be trapped in YoshiEgg for 13 frames and escape on the 14th.
If the target mashes 1 of A/B/X/Y every frame and inputs a direction on the control stick every frame, he will be trapped in YoshiEgg for 6 frames and escape on the 7th.
The L/R buttons, C-stick, and D-pad have no effect on YoshiEgg duration. The target cannot input 2 of A/B/X/Y on a particular frame. I'm pretty sure the Z button does nothing as well, but Z frame advances in develop mode, so I don't know how to test.

Finally, Nana. Nana actually does seem to mash a tiny bit right when she enters YoshiEgg. Nana is always trapped in YoshiEgg for 185 frames and will break out into the Fall animation on frame 186. Popo's inputs have zero effect on how long Nana is trapped in YoshiEgg.



EDIT: So based on this information. I have the following hypotheses. Each input reduces the duration that a target will be trapped in YoshiEgg by 14 frames. This is based on the fact that Nana escapes from YoshiEgg 14 frames faster than a target that attempts no inputs at all (correlating assumption is that Nana attempts exactly one input, presumably so she can fall towards Popo). Additionally, inputs take effect starting the frame before YoshiEgg comes out. This is based on the fact that assuming 1 input = 14 frames, there needs to be 14 inputs for a frame 14 escape and 7 inputs for a frame 7 escape, which is consistent with the data. I don't have the time right now for further testing to confirm this.

 

[Sashimi]

Fantastic. A few questions:

- You mentioned that the target starts on frame 1 (second frame) of YoshiEgg. I am guessing that you tested with a target who was a higher port # than Yoshi. If Yoshi was P2 and the target was P1 (for example), I would expect the target to start on frame 0 of YoshiEgg. When I get a chance, I will confirm/deconfirm this.

- If Popo is in the egg and Nana is still close, is it possible for the player to control Nana's actions? If you don't have time for this, I will check it as well.

It is a bit scary to think that the target can break out so early, but I suppose mashing this well is not only very difficult but also basically guarantees that the target will commit to an attack or a jump after breaking out. Otherwise this is all very cool! It is interesting that L/R aren't useful for mashing. I would guess that Z actually would be useful given how it is basically R+A, but I will check this.

 

[DD151]

- You mentioned that the target starts on frame 1 (second frame) of YoshiEgg. I am guessing that you tested with a target who was a higher port # than Yoshi. If Yoshi was P2 and the target was P1 (for example), I would expect the target to start on frame 0 of YoshiEgg. When I get a chance, I will confirm/deconfirm this.

You are correct. A target with a lower port # than Yoshi will start on frame 0 of YoshiEgg.

When I tested this on a target with lower port #, the fastest breakout with only A/B/X/Y inputs was on frame 15 (14 frames in YoshiEgg). Contrast this with the result that I obtained on a target with higher port #, where the fastest breakout with only A/B/X/Y inputs was on frame 14 (13 frames in YoshiEgg).

- If Popo is in the egg and Nana is still close, is it possible for the player to control Nana's actions? If you don't have time for this, I will check it as well.

Yes, I tested this quickly with D-smash, F-tilt, down-B, and neutral-B. Nana executes all of these moves with the 6 frame input delay. I think she is not under player control unless she is in Wait (Nana will run towards Popo after Popo gets pooped), and she will not respond to C-stick inputs.

I have another observation which agrees with my hypothesis that Nana's CPU is inputting control stick directions to try to follow Popo. If Nana gets pooped out as an egg and Popo dash dances around Nana, Nana will break out of the egg faster because she's trying to follow Popo by inputting left and right on the control stick. I didn't count the number of frames that Nana was trapped in the egg this way, but what happens here is that when a player in the egg registers an input, the YoshiEgg frame counter increments by 2 several times, and then it increments by 1 afterwards if the player hasn't registered another input. When Nana was stuck in the egg, her YoshiEgg frame counter would temporarily start incrementing by 2 every time Popo changed the side on which he was standing, and then it would return to incrementing by 1 after the expected interval.

I will edit in more findings as they come along.

It is a bit scary to think that the target can break out so early, but I suppose mashing this well is not only very difficult but also basically guarantees that the target will commit to an attack or a jump after breaking out. Otherwise this is all very cool! It is interesting that L/R aren't useful for mashing. I would guess that Z actually would be useful given how it is basically R+A, but I will check this.

Mashing out is either dangerous or slow for someone who doesn't know how to mash out. However, I think the optimal way to mash out is to do the following concurrently:

- Wiggle the control stick left and right (to prevent accidental jumping)
- Sweep across A/B/X/Y once, then sweep back

The A/B/X/Y sweeps amount to 8 inputs, and the player should finish those inputs before breaking out so that the last input prior to breaking out is left or right on the control stick. As Ice Climbers, I can consistently mash out from a grounded Egg Lay before the egg hits the ground. This means that against any player who knows how to mash out, a lot of follow-ups are unsafe or not guaranteed. I've seen someone - I think it was Kofi - follow up Egg Lay with Egg Toss, which definitely can be escaped assuming good mashing.

 

[DD151]

I was watching some match videos vs. C. Falcon and I noticed that Yoshi's D-throw doesn't force him to tech on the ground. I don't think that D-throw tech chase on C. Falcon is a guaranteed setup unless he can't jump out before hitting the ground if he DIs fully away or something.

I also noticed that when aMSa knocks C. Falcon on a platform, he likes to read the tech with U-air below the platform. I think this is sub-optimal. What I'd like to know is, on all tournament-legal stages, can Yoshi cover techs on platforms with the following flowchart:

- F-air for tech in place, missed tech, and techroll one direction, L-cancel
- D-smash or F-tilt for techroll other direction

Based on frame data, if Yoshi times his F-air to come out on frame 20 of all animations where the opponent hits the ground, he can cover 3 tech options. If Yoshi lands on frame 21, L-canceled landing lag renders him inactionable until frame 32. Turnaround F-tilt and D-smash come out on frame 39. So theoretically this should work with perfect timing. I would like to know roughly where I should aim the F-air on each platform to cover the most area while also covering one edge of the platform.

 

[Sashimi]

You're right, technically Falcon does not have to tech after Yoshi's Dthrow. Likewise, Yoshi doesn't have to tech after Falcon's Dthrow. I chose to post the techchase information anyway since many Falcons will tech after the Dthrow even if they don't have to, and the techchase flowchart applies to any other situation where Falcon does have to tech such as after Nair, Ftilt, Utilt, Uair, Fair, etc. As long as you can get to the same position by the time you need to cover tech in place, your options are identical.

I haven't started any frame data for platform techchases, but I usually try to do exactly what you described when I am playing. Fair > Dsmash certainly seems like it can cover every tech option on a platform (except edge-cancelled tech, of course). I think Dsmash will also work on its own if the opponent starts the tech on one side of the platform. Ftilt might not quite have the range that is needed for longer platforms, but it might work so I will test it.

I have also noticed that Leffen used to Uair from below platforms as well, but I agree with you that it isn't optimal.

 

[DD151]

You're right, technically Falcon does not have to tech after Yoshi's Dthrow. Likewise, Yoshi doesn't have to tech after Falcon's Dthrow.

Doesn't C. Falcon's D-throw frame trap Yoshi? Like, I think at most %s, Yoshi is still getting kneed out of D-throw regardless if he tries to jump out because his DJ doesn't go upwards.

I think U-air below platform can be a tech chase on reaction (DJC forward or backward to U-air if opponent techrolls away from Yoshi), but Yoshi's position after the U-air isn't conducive to follow-ups. An alternative to the F-air / D-smash coverage is U-air / D-smash coverage, which should work just as well because U-air's hitboxes are actually slightly larger than F-air's hitbox. Yet another possibility is the fastfall U-air / DJC F-air reactive tech chase that I described before, though I don't know if that works on all platforms.

Something else that I've thought about is using a full jump B-air to cover the entire platform - it has active hitboxes over a span of 21 frames, which can cover the entire vulnerability period of a tech roll.

 

[kofinater]

Could some one figure out the inputs for fade away djc nair? When you dj to start moving back and cancel momentum but don't turn around

 

[Sashimi]

Could some one figure out the inputs for fade away djc nair? When you dj to start moving back and cancel momentum but don't turn around

From the OP:

Latest Retreating DJC
1 Double Jump backwards
5 Aerial

Earliest Reverse DJC
1 Double Jump backwards
6 Aerial

So the inputs are:
1 (holding backwards) double jump
2-5 (no longer holding backwards) A

 

[Icyo]

How does Vectorman do this egg lay bounce? https://youtu.be/RxNNZxE_-r4?t=5m34s

Also, what kind of effect does Sheik's needles have on Yoshi's hitboxes?

 

[Sashimi]

(I think you mean Egg Roll) And I am not sure exactly how it is done, but I know an easy set of inputs that can reproduce it: Jump and quickly (but not too quickly) use Egg Roll.

If I had to guess, I think it has something to do with landing at or close to the apex of the little hop you get from starting the Egg Roll. There isn't much practical use for it.

Can you clarify what you mean by "what kind of effect does Sheik's needles have on Yoshi's hitboxes? Do you mean whether or not Yoshi's attacks will clank/win/lose when meeting the needles? I was going to look into this at some point and I will post here soon when I have the answer.

If you mean something else, please let me know.

 

[DD151]

I was thinking about Yoshi's options to cover get-up options after a missed tech and the first thing that I decided to investigate was crouch canceling get-up attack. I glanced at the knockback spreadsheet and then did some testing in training mode. Here are my findings:

The vast majority of get-up attacks in this game have 80 BK, 50 KG and deal 6% unstaled. The two notable exceptions are Peach and Zelda, whose get-up attacks have 70 BK instead. Yoshi cannot crouch cancel any get-up attack at 0% without sustaining some knockback. At or near 0%, if Yoshi crouches adjacent to the opponent, he can still punish some get-up attacks with D-smash. However, this does not have coverage on get-up roll behind. In general this is just a very poor option.

So supposing that the opponent misses a tech and Yoshi cannot jab reset in time, what are the best option coverages for get-up?

 

[Sashimi]

Given about 15 or so frames to react, it is possible to react to all of the getup options as long as you are close enough when you need to react (same as techchasing). Unless I am mistaken, the standard for getup options is 30 frames for neutral getup, 49 frames for getup attack and 35 frames for getup rolls. Some getup attacks might be too fast for a reaction, I am not sure.

Generally, I try to react to the getup option. I don't think there are any option selects here, unfortunately.
- You can punish neutral getup exactly the same way as you would punish tech in place, except that you have more time to react.
- You can dodge the getup attack on a read with charged Fsmash or Yoshi Bomb. I like to parry the getup attack into Fair at lower %s (but watch out for the second hit if you parried the first hit) or at high % if I want to react instead of read.
- You can dashgrab the getup rolls, but you have less time to react. I don't think something like run cancel Dsmash is fast enough for getup rolls.

This is mostly from playing since I don't have a lot of frame data for this stuff, but I will put it on my to-do list and get back to you with more information.

 

[DD151]

I think parry -> DJC N-air is an option select for neutral get-up and get-up attack, but it only works if you parry the earlier hit and space properly to not get hit by the later hit.

One idea that I have is to crouch and then do the following on reaction:
- D-smash to cover neutral get-up and roll behind
- CC getup attack and then dash grab or dash attack to punish

Not sure how that would work in practice. If Yoshi gets hit by the first hit of get-up attack (usually frame 19), he has 29 frames to do something after being hit. 11-16 of those would be getting the grab hitbox out, some of those would be hitstun (not sure how many), and the rest would be dashing.

Aiming for a 2/4 or 3/4 coverage is pretty good because that covers the scenarios where a target is knocked down near the edge or on a platform. Yoshi can set up the former somewhat often with a N-air or a D-tilt in neutral. Also, I think this is a scenario where punishing roll away with dash attack can actually be useful because roll away often backs up a target to the edge.



EDIT: I have more suggestions for notable combos to add to the front page. I might be able to test these if I have the time, but I have a gauntlet of exams coming up and I'd rather use any time I spend on Melee to practice tech skill.

- Ground F-air -> ground down-B
- Ground F-air -> aerial down-B
- Ground F-air -> earliest waveland on side platform (Battlefield, DL64, YS, PS) -> ground down-B
- Ground F-air -> earliest waveland on side platform (Battlefield, DL64, YS, PS) -> F-smash or D-smash (I suspect this would only work against fast-fallers)
- U-tilt -> ground down-B
- Ground U-air -> ground down-B

 

[Sashimi]

Using Falco's tech and getup information as an example, Falco's getup roll away moves him less than his techroll away (and takes longer to start moving). Despite it being only 35 frames long instead of 40 (techroll), it seems like you would be able to cover it on reaction with dashgrab or dash attack.

If you spaced yourself very close to the first hit of the getup attack, it would probably knock you out of range of the second hit even at 0%, and the rest of the options are easily covered with Dsmash or turnaround Dsmash, so I think this would work.

CC dashgrab against the getup attack would go like this:
Frame 1: Falco starts the getup attack. Yoshi crouches.
Frames 17: Falco's getup attack hits.
Frames 18-19: Both characters are in hitlag.
Frames 20-21: Falco is in hitlag, Yoshi is in landing lag (from ASDI down).
Frames 22-23: Yoshi is in landing lag.
Frame 24: Yoshi can dash forward.
Frame 25: Yoshi can start dashgrab.
Frame 35: Yoshi's earliest possible dashgrab hitbox (full reach on frame 40).
Frame 50: Falco is actionable again.

This seems lenient enough to allow for considerable knockback from the getup attack. As well, Yoshi's dashgrab moves him forward quite a bit so I think it would be very reasonable to CC dashgrab. The first hit of Falco's other getup attack is on frame 19 and that seems fine too.

 

[DD151]

Excellent work. It seems like Yoshi can CC and punish the first hit of getup attack with other options such as wavedash D-smash and wavedash F-tilt with some leniency. It takes 22 frames for both attacks to come out after frame perfect dash wavedash. However, these options can't punish the second hit of most getup attacks, whereas the dash grab and dash attack can punish the second hit (usually frame ~25) with some leniency.

Unfortunately I don't know how easy it is to tell whether Yoshi is on the first hit side or the second hit side. I know that the getup attacks are unique based on whether a character is on his front or back, but I can imagine that it would be difficult to evaluate character position in addition to memorizing the proper getup hitboxes for each commonly used character.

 

[DD151]

Tested grounded F-air to grounded down-B. All %s are inclusive and recorded before the first hit. KO %s are on FD with no DI.

F-air to grounded down-B (33% fresh):
C. Falcon - 53% to 117%, 89% KO
Falco - 54% to 110%, 80% KO
Fox - 48% to 102%, 75% KO
Ganondorf - 36% to 85%, 71% KO
Jigglypuff - 17% to 39%, 35% KO
Luigi - 26% to 60%, 58% KO
Marth - 30% to 74%, 61% KO
Peach - 25% to 56%, 51% KO
Pikachu - 28% to 66%, 55% KO
Samus - 29% to 58%, 58% KO
Sheik - 33% to 81%, 64% KO

That should cover most of the relevant characters. Obviously there is human error involved and I can't be bothered to refine the %s further (especially at the low range where timing needs to be tight).

 

[Sashimi]

Thanks. I will check the numbers and add them to the OP. I knew that Jigglypuff could be KOd off the top with Fair > Yoshi Bomb, but I expected the % range to be smaller!

 

[DD151]

This testing takes a surprisingly large amount of time. I would've also tested F-air to aerial down-B, but testing the min %s is sooo tedious. I think it's more useful to make sure that F-air to aerial down-B works at the max % of F-air to grounded down-B and then work up from there, because if possible, F-air to grounded down-B is always better.

 

[Sashimi]

Yeah, it does take a long time. That's why there are so few entries in the "combo options" section. I have quite a few started but not yet finished.

There is some overlap between the KO % of Fair > Aerial Yoshi Bomb and the combo range of Fair > Grounded Yoshi Bomb, at least for some characters. In any case, thanks for the help again!

 

[DD151]

Tell me what you have and I might look at them from time to time. I will probably ignore low and bottom tier characters, though.

 

[Kibbles_n_Bits]

Hey PerhapsMan, for the matchup specific section, could you add the Peach whiff grab when yoshi lands in front of her? It's a tight timing but because of the way Yoshi lands he can avoid her grab sometimes. It's pretty dope.

 

[Sashimi]

I don't think it is limited to Peach, but here you go:

Basically, Yoshi's big head moves into the z-axis, and the rest of his body is very low during his landing animation, so he can't get grabbed here.

It lasts for about half of his regular landing animation, all of his jumpsquat, and frames 2-4 (and kind of 5, he sticks his arm up on this frame but it isn't near his front) of his L-cancelled Nair landing. Yoshi's Bair also has a very low landing animation, but this doesn't move along the z-axis so his head can still be grabbed if you are facing Peach.

Timing Nair on-shield to abuse this is very difficult and can be messed up if Peach grabs at the wrong time, but autocancels, tomahawks and jumping can all be used to avoid her shieldgrab quite easily. It is not really spacing dependent since the part of Yoshi that doesn't move along the z-axis is very low to the ground for all animations listed.

I might not put it in the "matchup-specific" section since there I think are a few characters whose grabs are affected by this, but I will put it somewhere.

 

[DD151]

Man why can't Yoshi crouch like his jumpsquat, that is almost Jigglypuff tier grab evasion...

EDIT: For what it's worth, all of Yoshi's hurtbubbles can crouch under Ganondorf's standing grab except for his head, which can sometimes get grabbed if Ganondorf is spaced properly and his grab hitbubble comes out when Yoshi is bobbing slightly upward in his crouching animation. If Yoshi's back is turned towards Ganondorf, he can't perform a standing grab on Yoshi at all.

 

[Sashimi]

Yep, Yoshi can crouch under Zelda and Ganon's grabs, as long as his head is spaced properly. You can also crouch under moves like Ganon Jab/Ftilt.

 

[flieskiller]

It reminds me when I found Yoshi can't dash grab a standing Ganondorf, because the tongue goes behind the legs in the Z-axis... I was surprised