Some Yoshi Information

[DD151]

Lightshield trigger tricking

I mentioned in a response above to dvatch that one possible way to get a consistent lightshield is to trigger trick L or R such that pressing the trigger until just before the click yields a lightshield of ideal "lightness" (approximately close to a Z shield).

Terminology (for clarity):

A digital trigger press is pressing L or R past the click.
An analog trigger press is pressing L or R, but before the click.
Pressing "all the way down" will refer to an analog press to just before the click.
Trigger tricking is manipulating the 0 value of the L or R trigger button by holding the button down before plugging in the controller or recalibrating the controller using X + Y + start.

Determining depth of trigger trick

Plug your controller into port 1. In 20XX, go to the Debug Menu > Training Codes and set Debug Level to Develop. Start a match in Vs. Mode or go into Training Mode.

Here we have to use Magus's physics input display. Press X + Dpad right to turn it on. Next, press L + Dpad right to highlight the first set of floats and then press L + Dpad up/down until it displays Item -> System Analog L/R.

In this display, the left value is the analog depth of the L trigger and the right value is the analog depth of the R trigger. Make sure that both read 0.00000 or close to it. If they don't, unplug/replug your controller or recalibrate it by holding down X + Y + start.

A brief digression on mechanics: if you recall from above, there are 141 discrete analog inputs ranging from 0 to 1, inclusive. Inputs that are 0.30000 or less do not register for commands such as lightshielding or L-canceling. You must input at least a 0.30714 analog press for the lightest possible lightshield. Additionally, you may notice that even when you press a trigger "all the way down," there is still some variation in the analog depth based how much pressure you're exerting on the trigger. Therefore, even when trigger tricking using this method, there will still be some inconsistency in analog depth, so it's humanly impossible to consistently get the lightest possible lightshield.

Press a trigger until the display reads a depth of approximately 0.85000. Hold the trigger at that depth, then press and hold X + Y+ start to recalibrate the controller. You will know when the controller is recalibrated when both analog L/R input displays reset to 0. Next, press the trigger "all the way down" several times and observe what analog depths you get. They should be fairly close to 0.35, which is the strength of a Z lightshield, with some variation.

Play around with this and try to lightshield out of different actions to see how consistently you can get desirable lightshields. If you find yourself frequently failing to bring up lightshield, then you need to either press the trigger "all the way down" a little harder or you need to recalibrate the trigger to a shallower depth (such as 0.80000). If you find yourself frequently shielding a little too hard, then you need to recalibrate the trigger to a deeper depth.

Testing

After calibrating the L trigger to the desired depth (0.85000), I semi-randomly attempted to lightshield in the middle of my movement practice routine by pressing L "all the way down" and logging the input from Magus's physics input display. I did this 40 times.

Mean analog L input: 0.35625
Mode analog L input: 0.34286
Lightest analog L input: 0.31429
Heaviest analog L input: 0.4

This is pretty good. In 40 attempts, I lightshielded all 40 times, with the heaviest input still being acceptable (in my opinion). It took a little bit of getting used to to lightshield consistently out of some actions than others. It's probably more desirable to get heavier lightshields than to occasionally miss a lightshield because the input is too shallow.

Real-life applicability

Okay, so now you have a good idea of how to trigger trick your controller to get consistent lightshields. The problem, of course, is that in a real-life setting such as a tournament, you're not going to be playing on a 20XX setup where you can meticulously calibrate your controller with Magus's assistance. Somehow you'll have to do this without any precise feedback from the game at all. Unfortunately, I haven't thought too hard about this, but I have a couple of ideas.

1. Mark or etch your trigger with a line that corresponds to the target calibration depth. This can lead to some precision issues because you have to eyeball how far to press each time you plug in your controller.
2. Make an object that can consistently press the trigger to the target calibration depth. This can be as simple as a hard object that somewhat fits the contours of the controller. The drawback is that you have to carry around this object and take care to not lose or damage it.

Controller scheme problems

Since all Yoshi players, in my opinion, should already be trigger tricking one trigger for guaranteed parries, using a lightshield trigger trick would require both L and R to be offset from their standard calibrations. If you light press the non-parry trigger for L-cancels, this can be a problem unless you consistently L-cancel by pressing "all the way down" (since lighter presses no longer register). The obvious solution is to L-cancel with Z.

 

[dvatch]

Thank you for such a detailed reply @DD150. I didn't know that the shoulders could be trigger tricked in the way you describe.

I tried the method for .85000 depth and the lightshield works great! As you warned, I did have trouble L canceling. I will probably press the trigger down to around .60000? unless I figure out how to L cancel with Z.

With a pencil, I marked the back of my R button to test how well i could reproduce the analogue value. It seemed that the most variation came from tilting the trigger inward/outward while pushing the trigger to the marker. Tilting the trigger inward produces a value about .20000 higher than if the trigger is tilted outward. The marker seemed to line up fairly well with both variations. Therefore, I would suggest tilting the trigger inwards before marking so that the new analogue value won't be at an unusable number.

Ill try out this controller config against real people today.

edit: Im switching to L canceling with z. Lightshielding easily is very nice.

 

[Sashimi]

Sashimi you should come on the Discord more often so that we can come up with ideas about what to lab.

I have been very busy since GOML but I will try to do that from now on.

 

[Klaww]

I'm not too active here so I'm not sure if anyone else has mentioned this before or not, but in the RECE section of the main post, it says that Yoshi's double jump has a diagonal variation that allows you to drift backwards to grab the ledge.

I think what's actually going on is that the distance of your double jump depends on the analog position of the stick when the jump button is pressed. If you just slightly hold backwards on the stick while you double jump, you will barely move at all, pretty much allowing you to turn around in place.

I've started trying to pull off RECEs by tilting the analog stick backwards a tiny bit to drop off the ledge (about as far as you would need to tilt it in order to slow-walk) while tapping the Y button, then throwing an egg. I'm not sure if that's actually an easier way to do it, or if it's just me but I figured I'd bring it up just in case. You don't need to drift back to grab the ledge though, you'll immediately grab it as long as you didn't tilt the analog stick too far away from the ledge initially.

 

[Ten of Nine]

I'm not too active here so I'm not sure if anyone else has mentioned this before or not, but in the RECE section of the main post, it says that Yoshi's double jump has a diagonal variation that allows you to drift backwards to grab the ledge.

I think what's actually going on is that the distance of your double jump depends on the analog position of the stick when the jump button is pressed. If you just slightly hold backwards on the stick while you double jump, you will barely move at all, pretty much allowing you to turn around in place.

I've started trying to pull off RECEs by tilting the analog stick backwards a tiny bit to drop off the ledge (about as far as you would need to tilt it in order to slow-walk) while tapping the Y button, then throwing an egg. I'm not sure if that's actually an easier way to do it, or if it's just me but I figured I'd bring it up just in case. You don't need to drift back to grab the ledge though, you'll immediately grab it as long as you didn't tilt the analog stick too far away from the ledge initially.

This is the way that I used to initially do it as well, but it's difficult to pull off back to back with optimal I-frames. Same with the suggested method in the OP.

The easiest way to do it, and the way that I can consistently get it back to back is with a fast Street fighter type motion on the analog. In the midst of rolling the analog stick you Egg when your analog reaches the up-left to up position -

<>

So assuming you are on left edge your basically rolling from left to up right with the analog and anywhere between the Up-left to up spots press B while rolling the analog to egg. This method is like 95% consistent for me now and it times out perfectly for optimal regrab.

Side-note- I've noticed that starting a little up left instead full left can help with timing and consistency

 

[CAUP]

Just was watching some amsa and noticed that a laser passes through him around 10:47 of this video while dashing. I know there are some weird things that can happen with Yoshi on the z axis and was just wondering if this specific situation is known.

http://youtu.be/icyYH2_lDZ8

 

[Ten of Nine]

Yoshi's Jump squat and landing animation can go right under Falco and Fox's Standing height laser. By extension his WD, WL, and andy AC can as well.

So unless they are low to really low with the Laser cancels Yoshi can go right under them (it might also have to do slightly with the 3D aspect <- IT DOES).

edit- I just labbed it a bit Yoshi's head actually skews to the side a bit as he lands or goes to jump. This is different to his crouching stance. So his top part of his head (eyes and eye brows) sort of side step a bit into the Z axis in jump squat and landing leaving only his body and tail in the linear plane.

 

[tauKhan]

When you look closely, you can see that aMSa starts jump right as the laser comes at him and the jumpsquat makes yoshi slip under the laser.

 

[CAUP]

Yeah I noticed that after posting. I think I have heard about Yoshi's jump squat avoiding projectiles before.

 

[Ten of Nine]

Here are some sweet action shots, the last 2 are crouch to show the big difference and Yoshi's Roll

All the pics are a dash WD which is optimal and easiest to pull off consistently
You can just WD forwards and backwards as well (against the laser direction) but it's tighter of course
Empty land > immediate jump is lenient, while moving forwards it's even easier (kinda tough to time though, it's what Amsa accidently did in the VoD)
You can also dash jump squat under but it's almost frame perfect

 

[deeseejay]

Any advice on Yoshi's shield pressure?

I use fair on shield alot into either jab pressure or ftilt and I seemed to get grabbed out of the fair A LOT.

What am I doing wrong? Not fast enough? Hitting too far into shield? I'm definitely comfortable with the l cancelling.

 

[Ten of Nine]

You are probably doing what I see a lot of Yoshi mains do including AmSa.

You probably aren't mixing up timing/pressure or aren't utilizing more technical pressure options, plus you should be conditioning them if they are reading you.

Fair is already negative on shield even when fully unstaled; on a practical level it's probably -3 to -4 on shield on a TAS impossible level it's -1. Double Jab is negative on shield to a react-able time. Fair is also highly telegraphed because of it's animation, plus it's slow so it's easy to prep for and they just have to react or read what you follow with.

Single Jab into something else would be better. First Jab is not really react-able but readable if you do the same timing EVERY time.

AmSa gets punished attacking shields a lot be cause he does Fair or Nair into double Jab/tilt almost every time same timing or he DJC Nairs into immediate FH or SH Nair and floats for a whole 20-35 more frames just waiting to get easily punished (that is a really bad option, not worth risk).

Here are better choices, and they need to be rotated through or mixed up -

Fair or Nair > crossup DJC Nair(s)

Fair or Nair > Jump away then reverse DJC Uair or Nair with different timings

Fair or Nair > spaced SH Fair (pull back avoids retaliation and the late timing can punish spot dodge)

Fair or Nair > Jump away RDJC Egg Lay

Fair or Nair > a barrage of optimal DJC Nairs into something else
---- For the above they will most likely buffer roll or do something OoS this is what you want to predict and punish
-- You may get lucky and they get nervous, it only takes 4 Nair hits as they hold shield for it to break. Punish that optimally with a Fair into big combo or KO.

Fair or Nair > jump straight up and wait to do a move, this can catch an OoS option they try

Fair or Nair > Jab > Dash away and if they attack OoS run in and punish or parry the aerial then punish

Fair or Nair > Single Jab > Dtilt > whatever (Dash way DD Dash grab can work if they like to hold shield a lot)
Ftilt can be used but it is the most punishable on shield, far more than spaced Dtilt.

A really interesting experimental option is -

Fair or Nair > SH or FH Dair (drift behind-on top them or down onto them)

Similar to how Falcon and Ganon's sometimes do. If they hold shield it will most like be broken unless they shield DI or escape somehow. It has a ton of large hitboxes, so it can also shield stab their head and feet as it reduces the shield.

If they get hit or shield poked they will freak and try and start SDIing if they react in time. Most can't punish it on reaction even for how slow it is....drifting behind them as you dair down can avoid getting grabbed or other fast ground move. This is to be rarely used since it's readable and reactable once they look out for it. But the start up is kind of ambiguous in the animation (looks like DJ) so it seems to work as a mix up more than it should.

 

[tauKhan]

Physical powershield storing (Previously known as rapid powershield, I believe)

Before you get your hopes up, I haven't found a way to store a projectile reflectbox, the following will only apply to powershielding normal, non-projectile attacks.


Terminology

GuardReflect: One version of Yoshi's action-state during pre-shield. GuardReflect starts when doing a full digital L/R press while able to shield immediately. GuardReflect lasts 6 frames and starts powershield window. It also makes Yoshi invincible and he can jump out of it, which enables parrying. GuardOn can be converted into GuardReflect if digital shield is input for the 2nd frame of GuardOn. Yoshi becomes vulnerable when doing so though.

GuardOn: The other possible action-state for pre-shield. Happens if shield is initiated with analog input (lightshield) or buffered digital shield input. Lasts 6 frames like GuardReflect. Yoshi can't jump out of GuardOn.

Powershield count: A number that's initialized to 4 when GuardReflect starts. The number normally ticks down by 1 near the end of every frame of GuardReflect until it's 0. If the Powershield count is greater than 0, powershieldable hitbox gets powershielded when it connects with shieldbubble. This can normally happen only during frames 3-4 of Yoshi's GuardReflect, since Yoshi's shieldbubble isn't up on frames 1-2. Note: this definition was invented to describe what seems to happen. I don't know how this actually works on code level.


Storing

Any time Yoshi's GuardReflect gets interrupted by anything while his powershield count is greater than 0, the powershield count gets locked up and stored to the value it was on at the time. Anything that intercepts the GuardReflect will work including spot-dodge, roll, jump, grab and shielddrop. For example, if you start a spot-dodge on the 2nd frame of GuardReflect, the powershield count gets locked at 3 because it was initialized to 4 at the beginning of GuardReflect, and ticked down once at the end of the first frame of GuardReflect. Likewise, if you spotdodge on the 3rd frame of GuardReflect, the powershield count is left at 2.

You can store up a powershield count of 4 by sliding off of an edge on the exact same frame you start GuardReflect. With other setups the storing is limited to 3, since it isn't possible to start GuardReflect while performing other actions. The wildest setup to store up powershield count of 3 I found is probably transitioning from GuardOn to GuardReflect, which makes Yoshi vulnerable and getting hit immediately. It's also possible to powershield a move on GuardReflect frame 3 and slide off of an edge immediately as a result of shield knockback (and/or shield [A]SDI). That way it's possible to store up powershield count of 1.


Powershielding with stored powershield count

When you have successfully stored a powershield count greater than 0, a number of next powershieldable hitboxes you block with shield get automatically powershielded. Usually the powershield count ticks down by 1 every time you shield a move, which means often you will powershield a number of moves equal to the powershield count you initially stored. Shielding a hitbox that cannot be (physical) powershielded and thus won't be, like a projectile, will still tick down the powershield count by 1. However shielding a move that causes 0 hitlag(puff's sing, <1% dmg hitboxes in NTSC 1.0) doesn't cause the powershield count to tick down.

Note that starting GuardReflect initializes your powershield count to 4 and it then starts ticking down normally, so if you perform GuardReflect while having a stored powershield count, the storage will be lost. Therefore after storing powershields, you have to get your shield up by lightshielding or by buffering digital shield input out of lag so that you start with GuardOn instead of GuardReflect.


Benefits

As many have noted here, physical powershielding unfortunately does very little for Yoshi as it doesn't allow him to cancel GuardOff into moves and neither does it increase shield knockback. As far as I know, the only benefit is not taking shield damage when hit, so one would perhaps be able to use light shield a bit more liberally when he has ps stored up, which can be done easily enough.


Stored ps depletion

Stored powershield count doesn't seem to wear off over time at all. The main ways of decreasing or losing the count is blocking attacks, losing stock or activating another GuardReflect. The most bizarre detail about the ps storing I discovered is that if Yoshi does a turn (Dash dancing counts too) or walks or performs side-b or successfully grabs an opponent (whiffing doesn't count), that seems to set the powershield count down to 1 if it was >= 1 beforehand. If you don't have a ps stored up, none of those options raises the powershield count though. There might be more actions that affect the ps count that I haven't found out yet, I'd be delighted if anyone can point them out.


Shield start-up deviation

If Yoshi has stored ps count, shielding a hitbox on frames 1-3 of GuardOn seems to use up extra powershield count: If Yoshi blocks on frame 1 he loses all 4 ps, on frame 2 the count decreases by 3 and on frame 3 it decreases by 2. I suspect this might be a counter measure to multi-powershielding with Yoshi, since on other characters the powershield count seems to tick down during shieldstun which Yoshi doesn't receive. However if Yoshi gets pushed off of an edge immediately after shielding an attack, the powershield count seems to always tick down only by 1 like normally.


Special thanks to Ponkapa for helping me confirm most of the details.

 

[Sycorax]

if Yoshi does a turn or walks or performs side-b or successfully grabs an opponent, that seems to set the powershield count down to 1 if it was >= 1 beforehand.

F*ck this game. Seriously.

Edit: Nice work though, TauKhan.

 

[AXE 09]

Holy crap
Thank you so much for finding out how that works. Finally I have an answer after all these years lol.

I also always wondered why Marth's counter will randomly do a powershield animation. It looks like it has similar properties.

EDIT:
Wow, moves that do less than 1% (like Ylink's staled up-b) make Marth's counter do a powershield animation, but the counter doesn't activate lol. Interesting

 

[tauKhan]

Wow, moves that do less than 1% (like Ylink's staled up-b) make Marth's counter do a powershield animation, but the counter doesn't activate lol. Interesting

In NTSC 1.0 that is?

I didn't realize the connection with Marth/Roy counter, thanks! It makes sense. I was thinking earlier that the difference between Yoshi and everyone else in this aspect might be Yoshi's changed GuardOn not initializing the ps count properly. If only there was a way to shield directly without GuardOn/GuardReflect first, that way other chars might be able to used stored ps.

 

[AXE 09]

In NTSC 1.0 that is?

I didn't realize the connection with Marth/Roy counter, thanks! It makes sense. I was thinking earlier that the difference between Yoshi and everyone else in this aspect might be Yoshi's changed GuardOn not initializing the ps count properly. If only there was a way to shield directly without GuardOn/GuardReflect first, that way other chars might be able to used stored ps.

Yes in 1.0. Sorry I didn't clarify lol

 

[V3ctorMan]

What were using the boards...I feel old (er)

 

[Dralro]

What were using the boards...I feel old (er)

It only took pages of a giant info and tech to finally get V3ctorman to respond, I don't even play Yoshi but that is still funny and Vector is still awesome.

 

[Kibbles_n_Bits]

Hey, does anyone know if people have frame data on Yoshi's drop-zone aerials?

I just got drop-zone n-air sorta down and was wondering if people have been looking into potential applications in edge-guarding n such.

 

[Stax_Machine]

Drop zone nair is really frame tight, its much safer to drop zone fair and its easier to recover from but if people are good they can meteor cancel it. If you want to drop zone nair with it being a bit more practical if the stage has platforms waveland off of one and nair, the extra height makes it really easy to recover and since the move is started sooner you can even fast fall it a little bit to mess up the opponents timing. Another edge guard thats not talked about is moonwalk into fair, it works better than you would expect and it's a swag overload. Other swag overload option are while recovering or to punish people for getting in your face for edge cancel eggs it to invincible back-air onto stage from ledge, it combos into everything you could ever want but it's really hard to execute and maintain invincibility (it is possible though, grinded it for a bit with hitboxes on).

I've been really trying to optimize my shield pressure but yoshi sucks vs shield, everything is fraudulent, djc nair is good behind shield, jab into bs can be effective with mix ups, spaced d-tilt is useful, and put it all on uptilt/up angle f-tilt shield poke can work if the other persons di is bad. Only real thing I've found is if the opponent is higher percent nair into d-smash seems to almost always shield stab.

But we still have one thing to optimize that I haven't really seen anyone talk about (I also haven't looked the hardest) but remember when I said we have jab into bs, yeah that bs needs some revising, I know I've done this (did it once a really long time ago and couldn't replicate it, did it yesterday during a game so I know I wasn't imagining it) but jab into short hop back air (hit all 4 hits and while the back air is out we drift past their shield so those pesky shield grabs can't nab us *cough except for falcon* and then before you land we have to buffer an up-air right before we land. If we can do this we'll deal a ton of shield damage and be behind a shielding opponent all for about the same risk as a jab into any other option. Does anyone know the frame data on this? and how many djc nairs I would need after this sequence to break the shield? I don't really understand all the frame data for shield stuff.

 

[moofinmoofin]

Thanks for all this info everyone. Study time!

 

[deeseejay]

At what percents do upair juggles on spacies become possible? If it's zero I'm not doing it anywhere near fast enough.

If it's not zero what's the best follow ups and 0% percent from dthrow/ftilt etc.

I also struggle sometimes to follow up with the upair if they hard DI away, again is this an issue of me not doing fast enough?

Cheers!

 

[Stax_Machine]

up air chain: the up air chain works at 50% for fast fallers, you get 2 (guaranteed) on spacies and 3 on falcon because he is more heavy but they will most likely DI to platforms, you should mix up the ending of your up air chains with nair/upair to DI trap, use bair/fair with platforms to get d-smash (bair > dsmash works almost 100% of the time on spacies but falcon can get out of this, I've yet to find a consistent finisher out of up air chains for falcon)

0% grab: on fast fallers get the grab, d-throw, djc nair, they can not tech this and are in an awkward landing lag, immediately dash grab them (they will be at around 30 now) and you can d-throw nair them again/uprilt frame trap them for jab resets till you get them to 50% for the up air chain. the grab after the nair at 0% is not guaranteed but people often miss the roll and once they do start rolling you can read and punish it

0% in general: you should aim for grab or a low fair, if you get the low fair up angled f-tilt is your best option but uptilt will work as well as they get popped up for more tilts but they can SDI this down and you can react with a jab reset (up tilt has more lag after so the jab reset is harder to get off of the SDI, if you can catch them landing from a jump or when they arn't cc'ing landing back air is a good way to build %, you get 20% if you hit all 4 and if you l-cancel it you can pivot ftilt to make it almost 100%safe, but if they cc the bair you will be in trouble, another good way to get percent at low percent (its not safe) but its hard to punish even with cc and that's to get a djc nair into rising nair as you hop away from them so they cant punish two nairs puts them over 20%

getting the up air chain more consistently: to get the up air chain consistently you will need to have really crazy reaction times but it is possible, make sure to use the c-stick for the up air so you can get the full horizontal DI to help keep you underneath them. After the l-cancel landing you will use the time from landing the up-air and landing lag to react to the DI, and try to immediately get the dash and then jump into the next djc up air.

Extra Challenge: I'll tell you a really good way to practice up air chains as you'll need to be almost perfect to get this to work but you'll never miss hard DI away follow ups again if you can do this. Go into home run contest, pick yoshi (obviously), and what you will do is dash at the sandbag, hit weak djc nair, l-cancel, dash, rising strong nair (this is the same tech I mentioned to get damage on low percent foes; at this point you and the sand bag will be falling off the stage but as you land you will be able to land a falling up air before you and the sand bag land, from here just continue to chase and up-air the sand bag but in order to continue getting the up airs you will need to execute them almost perfectly (getting the dash after the l-cancel is the hardest part but it is also the most important) you should be able to continue this juggle till time runs out. It's also fun to do this and kinda like a yoshi alternative to the falcon nair homerun challenge, you can even practice getting double jump landing cancel to smash attacks for the finisher to send the bag even further as time runs out

 

[Stax_Machine]

shield pressure update: I mentioned in a post earlier that short hop bair to up air on shield is a good option... and it's really good, I can implement it in game now granted i'm still not the most consistent at it but I assumed djc nairs would be the best option at this point but I've discovered that you can land both hits of d-shash after the upair which is really good on shield and the second hit has a really really good chance to shield stab. You can get grabbed between the up-air and d-smash just like between the bair to up air but they will only have like a 2 frame window to get the grab so it's almost safe, especially since they will most likely miss the timing the first time they try and eat the up air, we can spread more yoshi jank tricks and make people think we're fox and have shield pressure, and I've found a way to get nair instead of up air after the bair, you have to time the djc with the x,y buttons instead of buffering it with the tap jump like you would for the up air as you can't return the stick to neutral in time for the nair, so if nair isnt stale, bair to nair (can be shield grabbed between hits) but the nair to d-smash should be safe on shield as they can not grab you during nair > dsmash (in order to not get grabbed this way the nair can't be stale and must be hit very low to the ground) and since nair comes out quicker than upair it should be easier to execute consistently but lacks the coverage to jump out of shield options as nairs hit box is below us