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

[Wafflesaurus]

does anyone understand the mechanic behind getting a full/strong backwards dash after executing a throw? so dashing backwards after a u-throw when chain grabbing for example. I get so many turn arounds into nothing or my character will dash much slower than it normally would when dashing forwards

ever since I got a new controller getting a clean dash back after an action is almost impossible (I imagine the analog stick's increased stiffness is just not working in favor of the mechanic)

There's a link to the wiki on post #1
http://www.ssbwiki.com/Turn

Read section Smash Turn. Basically you need to enter the area that says smash turn in 1 frame (see image), otherwise you get the slow turn.

 

[Bones0]

Tech-chasing with Marth
General approach: forward throw → wavedash forward (45° below straight forward = SE / SW notch), check for their tech option during wavedash lag
If they don’t tech or tech in place, you just wait and react accordingly (grab for tech in place, dash JC grab for get-up roll, shieldgrab for get-up attack).
If they tech-roll, you do dash JC-grab.
In theory, it would be better to do the longest dwd (18° below straight forward) on DI away and a shorter wavedash on DI in, but you’d need to react within 11 frames. The sources I know quote 384 ms as the expected value for visual choice reaction time, which is 23 frames. So reacting to DI away or in within 11 frames is most likely not humanly possible. Thus, I think it’s best to do an intermediate length wavedash, so you can cover both DI+techroll away and DI+techroll in, although these two options will require being very close to frame perfect with your reaction and follow-up.

Fox

#|DI|Fox’s option|Your reaction|Necessary reaction time|Input frame window
1| |No tech|Wait until Fox goes for 2/3/4/5| – | –
2| |Get-up attack|Shield grab|

(This is a work in progress right now. I still published this terrible post as-is as a placeholder so I don’t forget about completing this. There is a lot of stuff to consider that I haven’t mentioned here, so it will take many hours to even find out how to go about this. I hope I can complete this in a useful way eventually.)

Are you sure the type of reactions you are talking about should be considered visual choice reaction time? Recognition reaction seems more like a more appropriate description of what is actually happening during reaction tech chases. When I think about how I react to tech options, I don't weigh each possibility equally. I sort of assume I will be doing one action and then if I have a positive reaction to a single stimulus, I alter my default behavior for a new action.

So for instance, the hardest get-up options to discriminate between are GUA and stand. As soon as I identify the get-up as an in-place option, I limit my actions to either shield (GUA) or grab (stand). I ASSUME I will need to grab, but if I recognize the GUA startup, I shield instead. Idk what frames each option becomes distinguishable, but Falco's GUA hits on 17 so I think it's safe to assume that is a reasonable reaction time for players since many people can shield GUA on reaction while also alternatively grabbing stand on reaction (lasts 30 frames so you have to input it by 23). When distinguishable frames come into play, you're looking at reaction times that are definitely sub-20. I'm not even saying the study you read was wrong, but at the very least, it doesn't account for players who are both experienced and practiced at identifying and reacting to certain stimuli.

 

[Kadano]

Are you sure the type of reactions you are talking about should be considered visual choice reaction time? Recognition reaction seems more like a more appropriate description of what is actually happening during reaction tech chases. When I think about how I react to tech options, I don't weigh each possibility equally. I sort of assume I will be doing one action and then if I have a positive reaction to a single stimulus, I alter my default behavior for a new action.

So for instance, the hardest get-up options to discriminate between are GUA and stand. As soon as I identify the get-up as an in-place option, I limit my actions to either shield (GUA) or grab (stand). I ASSUME I will need to grab, but if I recognize the GUA startup, I shield instead. Idk what frames each option becomes distinguishable, but Falco's GUA hits on 17 so I think it's safe to assume that is a reasonable reaction time for players since many people can shield GUA on reaction while also alternatively grabbing stand on reaction (lasts 30 frames so you have to input it by 23). When distinguishable frames come into play, you're looking at reaction times that are definitely sub-20. I'm not even saying the study you read was wrong, but at the very least, it doesn't account for players who are both experienced and practiced at identifying and reacting to certain stimuli.

I agree with you. I also always thought that tech-chase reaction is something between simple and choice. Or, rather, multiple recognition reactions that happen quickly after each other, so I guess that we are slightly slower than pure recognition since we have to keep what comes afterwards in mind too. But the recognition value is the one that comes closest, I guess.
I can’t find any values for recognition reaction averages, do you know any?

 

[Bones0]

I agree with you. I also always thought that tech-chase reaction is something between simple and choice. Or, rather, multiple recognition reactions that happen quickly after each other, so I guess that we are slightly slower than pure recognition since we have to keep what comes afterwards in mind too. But the recognition value is the one that comes closest, I guess.
I can’t find any values for recognition reaction averages, do you know any?

I don't, but I tried this online reaction test and on my third set of 12 trials (going for 100% accuracy), I got 326.14 ms, which is ~19.6 frames. The lousiness of this so-called experiment doesn't elude me, but at least it gives some indication that sub-20 frame reaction times are possible for the type of reacting done during most game situations. I'm sure with an improvement in my diet, exercise, and sleeping (I'm awake at 5 in the morning...) I could significantly improve my score. I'm curious how I would do if they flashed 3 go/no-go images in quick succession because I think that is the closest way of simulating reaction tech chasing. I think getting a group of Smashers to take an online test like this one is the best way to figure out what is reasonable reaction time. Even if there are issues with computer performance influencing the average, it's better than nothing.

 

[dreamhouse]

reaction time stuff...

I'm planning on building an arduino based reaction time tester and bringing it to a major at some point so we can have a more reliable, relevant, data set to work with. I'm personally most interested in finding out what the fastest reaction times are in a low latency environment audially and visually, but if I can figure out a way to test other things like choosing the correct reaction to a more complex stimulus, or how length of focus effects reaction time I will.

Unfortunately, I don't have an ETA, but it's something I 100% will do eventually.

 

[DeadPigeon]

im sure you know how from 0-17(ish), fox and falco can DI slightly behind marth to force him to do turnaround grab instead of standing grab, but since it is slight DI, it can throw off marth's judgment.

would you have any way of showing the SLIGHTEST possible DI (i.e. the closest to no/neutral DI) where marth is forced to turnaround grab (gif or still image is fine)? i usually try to judge the CG based on fox/falco's position relative to marth's hand (from the uthrow), and i feel like if they're barely behind marth's hand, you can still turnaround grab, but i guess i want to know the visual cues with 100% certainty

i mean, i get it would change with percent

but idk, maybe there's something you know that i don't

Upward control stick smash inputs keep their smash input properties for 4 frames. If the game checks for strong upward during these 4 frames and you keep the control stick held upward, the corresponding action will be triggered. This is what you call a hold-buffer.

This kind of behavior is universal, so it works for shine and landing lag (indeed starting on frame 1), shield stun, IASA and many other things, except SDI.

The thing is, the DI values that make the difference are so slim that it’s hardly possible to notice them if the Fox player knows how he needs to hold the control stick.

For example, at 0% before throw, DI with x=163 will keep Fox in Marth’s no-turn grab range, but DI with x=164 will make Fox evade it, so Marth needs to turn. Here is a gif that shows the difference between 163 and 164 for the first 5 frames of hitstun:

If you can tell these differences in a match, you deserve an award of some kind.

To clarify, does turn around grab miss at 163? If not, then do you know the values of behind DI where you can first turn around grab?

Choosing an arbitrary number, lets say that turn around grab first works at 150. Then you can really only need to be able to tell the difference between 150 and 156 DI, and 156 and 163 DI. In other words, if you try to turn at 156, then you could be up to 7 degrees off on reading your opponent's DI and still get the grab. Is this correct?

 

[Kadano]

To clarify, does turn around grab miss at 163? If not, then do you know the values of behind DI where you can first turn around grab?

Choosing an arbitrary number, lets say that turn around grab first works at 150. Then you can really only need to be able to tell the difference between 150 and 156 DI, and 156 and 163 DI. In other words, if you try to turn at 156, then you could be up to 7 degrees off on reading your opponent's DI and still get the grab. Is this correct?

Ha, if only it were that easy.
I’ve long planned a video on this topic (Marth’s chain grab). I currently have PAL Fox from 0-7% damage before uthrow mapped out for fully stale and unstale uthrow (that took something between 10 and 20 hours). I’m still not sure how I can properly explain this in a video, and I certainly won’t manage to do that here. I can give you a quick explanation, but I don’t want to go into details since I’ll have to do this again for the video anyway. Frankly it takes way too much work to research and document this properly than I’d want for a post that will get lost on page 39 of this thread.

This is a map for PAL Fox’s fully-stale uthrow DI options. Control stick x on Fox’ controller is tested, y is always at neutral position.
The grey bar symbolizes the horizontal dead zone of the controller, the white dot in the middle is 128 (controller value) / 0 (Melee value). These have been mapped from GCC values, which means that I didn’t use Magus’ input display, so the ranges’ offset to the white dot can vary up to ±3 depending on your stickbox alignment / calibration. The ranges itself are unchanged by this, so to compare DI options they are still perfectly valid.

Green lines are standing regrab, purple is turn regrab. Background colors denote frame window for Marth’s regrab (black=0 red=1 orange=2 etc.).

The overlap between stand and turn regrab is very small, 2-3 values depending on percent and staleness. I‘ve taken save files at the positions just outside of that area (which means that only one of the two options will connect there), and the visual difference between a Fox who DI’s just enough to escape stand regrab and one who DI’s just enough to escape turn regrab is so slim that it’s hardly possible to tell when looking at the still frames. I’m absolutely certain that I will not be able to tell in a match whether my opponent DI’d barely left or right of the overlap area. Maybe I would with months of practicing this, but even then I’d have to memorize it for almost every single percent, and keep track in my head of my current uthrow staling.
I try to avoid making absolute statements when there is possibility for a contrary outcome, but I think the chances that a Marth player can always correctly recognize the amount of DI are almost zero.
For the aforementioned reasons, I won’t take screenshots to show this here. You’ll have to wait for my video or test it yourself / have someone else do it.

I also (most probably) won’t answer follow-up questions to this, due to the same reasons. I write this mainly to explain why I’m taking so long. (And it’s not like I’ll be finished with the chain-grab video soon, I expect the total workload to be about 800 hours.)

 

[Bones0]

Ha, if only it were that easy.
I’ve long planned a video on this topic (Marth’s chain grab). I currently have PAL Fox from 0-7% damage before uthrow mapped out for fully stale and unstale uthrow (that took something between 10 and 20 hours). I’m still not sure how I can properly explain this in a video, and I certainly won’t manage to do that here. I can give you a quick explanation, but I don’t want to go into details since I’ll have to do this again for the video anyway. Frankly it takes way too much work to research and document this properly than I’d want for a post that will get lost on page 39 of this thread.

This is a map for PAL Fox’s fully-stale uthrow DI options. Control stick x on Fox’ controller is tested, y is always at neutral position.
The grey bar symbolizes the horizontal dead zone of the controller, the white dot in the middle is 128 (controller value) / 0 (Melee value). These have been mapped from GCC values, which means that I didn’t use Magus’ input display, so the ranges’ offset to the white dot can vary up to ±3 depending on your stickbox alignment / calibration. The ranges itself are unchanged by this, so to compare DI options they are still perfectly valid.

Green lines are standing regrab, purple is turn regrab. Background colors denote frame window for Marth’s regrab (black=0 red=1 orange=2 etc.).

The overlap between stand and turn regrab is very small, 2-3 values depending on percent and staleness. I‘ve taken save files at the positions just outside of that area (which means that only one of the two options will connect there), and the visual difference between a Fox who DI’s just enough to escape stand regrab and one who DI’s just enough to escape turn regrab is so slim that it’s hardly possible to tell when looking at the still frames. I’m absolutely certain that I will not be able to tell in a match whether my opponent DI’d barely left or right of the overlap area. Maybe I would with months of practicing this, but even then I’d have to memorize it for almost every single percent, and keep track in my head of my current uthrow staling.
I try to avoid making absolute statements when there is possibility for a contrary outcome, but I think the chances that a Marth player can always correctly recognize the amount of DI are almost zero.
For the aforementioned reasons, I won’t take screenshots to show this here. You’ll have to wait for my video or test it yourself / have someone else do it.

I also (most probably) won’t answer follow-up questions to this, due to the same reasons. I write this mainly to explain why I’m taking so long. (And it’s not like I’ll be finished with the chain-grab video soon, I expect the total workload to be about 800 hours.)

I know you said you probably won't answer followups, but I just have to clarify: Are you saying if a spacie gets the proper DI at low percents, it's basically impossible for Marth to regrab without (essentially) guessing?

This also relates to a question I've had for a while. I know there are moves like Fox's usmash which are best DIed horizontally even though they don't hit vertically because you can't DI at all angles. Can you compensate for these impossible angles by slight DIing? it would seem that by tilting the stick left or right for Marth's uthrow, it is equivalent to DIing at an angle that is supposed to be impossible (e.g. 91 degrees). Or am I mistaken and the deadzone at the center of the stick prevents the same angles as the ones around the edge? I don't have a very good grasp on how slight DIs are affecting trajectories, I guess.

 

[Meru]

At what percents can you uthrow -> dair fox/falco? assuming you have at least enough time to reach your full dash speed

 

[Littlebear]

Do you have any diagrams like this: https://pbs.twimg.com/media/CEQqMT-WIAAKFGG.jpg
for the back air? I see people getting revere hits with the bair into down air, and it looks really cool, but i'm not sure what part of the sword to hit em with.

 

[Bones0]

Do you have any diagrams like this: https://pbs.twimg.com/media/CEQqMT-WIAAKFGG.jpg
for the back air? I see people getting revere hits with the bair into down air, and it looks really cool, but i'm not sure what part of the sword to hit em with.

https://www.reddit.com/r/SSBPM/comments/2yhhot/how_do_reverse_hitboxes_work/

 

[Nicholas1024]

When trying to sneak out a SHFair against Falco lasers, is it faster to shield the laser and then SHFair out of shield, or just get hit by the laser and jump immediately after the hitstun? Also, how big of a frame difference is it?

 

[Lawn Chair]

You get hit by the laser and just fair. There is no reason to jump

Shielding the laser means you take the shieldstun and have to leave your jumpsquat too

 

[Stride]

You get hit by the laser and just fair. There is no reason to jump

Shielding the laser means you take the shieldstun and have to leave your jumpsquat too

How are you supposed to fair without jumping or already being in the air?

When trying to sneak out a SHFair against Falco lasers, is it faster to shield the laser and then SHFair out of shield, or just get hit by the laser and jump immediately after the hitstun? Also, how big of a frame difference is it?

Lasers do 3 frames of shieldstun, but 8 frames of hitstun (5 if crouch-cancelling); so you can jump sooner after shielding a laser than you can after being hit by it.

 

[SpiderMad]

How are you supposed to fair without jumping or already being in the air?


Lasers do 3 frames of shieldstun, but 10 frames of hitstun (5 if crouch-cancelling); so you can jump sooner after shielding a laser than you can after being hit by it.

They don't do 10, are you crazy?

 

[1MachGO]

They don't do 10, are you crazy?

According to the knockback calculator, they do 9 frames of hitstun.

However, since its a projectile, Falco won't experience hitlag even though the opponent will. This means the opponent is stunned for 13 frames.

This then applies to shields as well. The opponent experiences shieldstun AND hitlag for a total of 7 frames. Theoretically, Falco's laser is +3 on block which is pretty insane.

 

[Stride]

They don't do 10, are you crazy?

Checking it again now, I realise I was stupid and forgot that the animation could be cancelled (and also didn't notice the discrepancy between the crouch-cancelled hitstun and the normal hitstun which would have indicated an error), so it's not 10 frames of hitstun. However, it is still 8 frames of hitsun (with 4 frames of hitlag, for a total duration of 12 frames): 4 frames of hitlag followed by frames 2-11 of the DamageNI animation, which is interruptible on frame number 10 of the animation (with the input on frame number 9). My number for when crouch-cancelling was accurate.

According to the knockback calculator, they do 9 frames of hitstun.

However, since its a projectile, Falco won't experience hitlag even though the opponent will. This means the opponent is stunned for 13 frames.

This then applies to shields as well. The opponent experiences shieldstun AND hitlag for a total of 7 frames. Theoretically, Falco's laser is +3 on block which is pretty insane.

I got my values by testing manually with frame advance.

Am I correct in interpreting every frame on which frame 1 of the damage animation is played to be hitlag, and every frame after that (starting from frame 2 of the animation) to be hitstun; or should the last repeated frame be considered hitstun? Regardless, my value for hitlag + hitstun is 1 frame less than yours, so either you input the wrong values into the knockback calculator or I've miscounted every time I've rechecked my values without understanding/realising where I'm going wrong.

 

[Lawn Chair]

How are you supposed to fair without jumping or already being in the air?

be in the air already

 

[dreamhouse]

Kadano, is there a spacing where shdf beats fox's out of shield jump forward Nair, but still hits their shield if they don't do an oos option? Assuming no shield di.

 

[Stride]

You get hit by the laser and just fair. There is no reason to jump

Shielding the laser means you take the shieldstun and have to leave your jumpsquat too

be in the air already

You can't be in the air all the time, nor should you be. It doesn't make sense to say that there's no reason to jump, unless you're talking about running of an edge after being hit instead (which isn't always possible); you have to get into the air somehow in order to fair.

 

[1MachGO]

Checking it again now, I realised I was stupid and forgot that the animation could be cancelled (and also didn't notice the discrepancy between the crouch-cancelled hitstun and the normal hitstun), so it's not 10 frames of hitstun. However, it is still 8 frames of hitsun (with 4 frames of hitlag, for a total duration of 12 frames): 4 frames of hitlag followed frames 2-11 of the DamageNI animation, which is interruptible on frame number 10 of the animation (with the input on frame number 9). My number for when crouch-cancelling was accurate.


I got my values by testing manually with frame advance.

Am I correct in interpreting every frame on which frame 1 of the damage animation is played to be hitlag, and every frame after that (starting from frame 2 of the animation) to be hitstun; or should the last repeated frame be considered hitstun? Regardless, my value for hitlag + hitstun is 1 frame less than yours, so either you input the wrong values into the knockback calculator or I've miscounted every time I've rechecked my values without understanding/realising where I'm going wrong.

I am not entirely sure. I believe hitlag is freezeframes whereas hitstun has an animation?

And I actually think you are right in counting it as 12. I want to say that there is a strong bad post somewhere where he talks about laser hitstun and 12 sounds familiar. Don't know for sure, though, I'll double check the spreadsheet.

 

[Nicholas1024]

Is there a good way to practice acting on the first possible frame out of a down tilt? I'd like to improve my follow-ups and movement after placing a D-tilt, but it's difficult to tell if I'm hitting the IASA window as quickly as possible, or if I'm missing a few frames.

 

[Wafflesaurus]

Is there a good way to practice acting on the first possible frame out of a down tilt? I'd like to improve my follow-ups and movement after placing a D-tilt, but it's difficult to tell if I'm hitting the IASA window as quickly as possible, or if I'm missing a few frames.

best way is to record yourself and analyze your videos frame by frame. Kinda hard without that. You could play around with 20xx and make cpu spam actions out of shield (roll, shine, grab, nair etc) but still wouldn't be able to tell if you are being frame perfect but it would help you get a feel for it and the correct follow ups for each action.

 

[dreamhouse]

Is there a good way to practice acting on the first possible frame out of a down tilt? I'd like to improve my follow-ups and movement after placing a D-tilt, but it's difficult to tell if I'm hitting the IASA window as quickly as possible, or if I'm missing a few frames.

Achilles made a code to show a yellow overlay during your iasa frames for d tilt.

http://smashboards.com/threads/color-overlays-for-iasa-frames.401474/#post-19120928

 

[A_Reverie]

Achilles made a code to show a yellow overlay during your iasa frames for d tilt.

http://smashboards.com/threads/color-overlays-for-iasa-frames.401474/#post-19120928

Color overlays are the bomb. Very helpful for this and for visualizing shield stun.

 

[dude it's raining]

https://www.reddit.com/r/SSBM/comments/2oozfc/trajectory_tables/

In that excel doc, under the gravity section, I just want to clarify some things.
Is the first matrix of numbers velocity after a given number frames, and the next matrix degrees of change in trajectory due to gravity?

If not, how can I figure out how gravity affects trajectory? I'm trying to calculate the maximum percent an ASDI CC can work. I think I can figure out the trig of it if I can get some standards (ASDI distance, etc).

 

[Kadano]

https://www.reddit.com/r/SSBM/comments/2oozfc/trajectory_tables/

In that excel doc, under the gravity section, I just want to clarify some things.
Is the first matrix of numbers velocity after a given number frames, and the next matrix degrees of change in trajectory due to gravity?

If not, how can I figure out how gravity affects trajectory? I'm trying to calculate the maximum percent an ASDI CC can work. I think I can figure out the trig of it if I can get some standards (ASDI distance, etc).

The second one is just gravity change to quickly see when max gravity is reached (a 0 can be recognized visually faster than repeated values).

You should be able to figure out all of that with the Magus iso.

 

[FlamingForce]

Whats a good way of going about ledgedashes and figuring out the possible intangibility from one? I'm trying to figure out the maximum amount of intangible frames Y.Link can get from a perfect ledgedash, but I have no experience using TAS.

EDIT Nvm found your list for this, thx Kadano.

 

[dude it's raining]

The second one is just gravity change to quickly see when max gravity is reached (a 0 can be recognized visually faster than repeated values).

You should be able to figure out all of that with the Magus iso.

I'm still troubleshooting, but it keeps crashing dolphin when I hit versus mode.

I've figured everything out except for quantifying the ASDI distance in terms of degrees of trajectory influenced (or speed in units per frame). I also consolidated the knockback formula into 1 excel string (I have a different one for WKB): =<CC>*(<BKB>+0.01*<KBG>*(18+1.4*(<Hitbox dmg>*(<Hitbox dmg>*(1-<staling>)+<Victim dmg>)*0.05+(<Hitbox dmg>*(1-<staling>)+<Victim dmg>)*0.1)*200/(100+<Weight>))) I'm also consolidated my knockback angle formula, which takes into account gravity and TDI, including the dead zones and the octagonal shape of the control stick's max range.

Once I have the ASDI distance, I can calculate the outputs and finish my CC % list.

EDIT: All I need now is to quantify the ASDI distance in Knockback units. What I did for gravity is added is as a downwards KB vector to the KB vector at an angle affected by TDI. I was able to use trig to calculate the new angle given the KB (i.e. Damage). All I need to do is add the ASDI to the gravity in KB units.

 

[AustinRC]

@ Kadano

If ice climbers jab marth and pop him up does he have time to dair them?

So lets say marth fthrow's pikachu and the pika tech's in place at a decent percent (not sure which ones work best.) But could he land a dair after the tech? I know pika's tech in place is freaking amazing and I'll likely have to play against Axe again soon so I wanna get at least something figured out.

I'll likely be in here more often asking tons of questions. :D

 

[DeadPigeon]

@ Kadano
So lets say marth fthrow's pikachu and the pika tech's in place at a decent percent (not sure which ones work best.) But could he land a dair after the tech? I know pika's tech in place is freaking amazing and I'll likely have to play against Axe again soon so I wanna get at least something figured out.

Pika needs to be over 20% for dair to knock him up. At this percent, you can dair on prediction. At 100% I could only get a phantom to hit.

 

[Link24a]

Hey, I just wanted to know how long youre invincible for after you drop from the ledge. i saw http://smashboards.com/posts/16761069 but im not sure if thats what im looking for/how to read it

 

[Kadano]

Hey, I just wanted to know how long youre invincible for after you drop from the ledge. i saw http://smashboards.com/posts/16761069 but im not sure if thats what im looking for/how to read it

29 frames. I’m not sure if I ever wrote a post on all the details, but they should be easier to understand by watching this animation frame by frame: http://gfycat.com/TidyScaryAzurevase#?frameNum=102

30 frames intang are granted upon completing CliffCatch and entering CliffWait. However, the first frame of CliffWait can’t be skipped if you want to drop from the ledge, so you’re left with 29 after drop.

 

[Nicholas1024]

In the Marth ditto, are there any legitimate mixups out of F-throw/D-throw for people that are always holding down? In particular, what can hit on down+away vs straight down vs down+in DI?

 

[Lawn Chair]

In the Marth ditto, are there any legitimate mixups out of F-throw/D-throw for people that are always holding down? In particular, what can hit on down+away vs straight down vs down+in DI?

Pretty sure you can pivot fsmash one of those DI's

 

[DeadPigeon]

@ Kadano , I couldn't find a formula on how DI affects KB angle. I know that full perpendicular DI gives +/- 18 degrees, but does it scale linearly? In other words, is the formula something like this:
((DI angle - KB angle)/90)*18 + KB angle
where KB angle is the move KB angle and DI angle is the value of your control stick (assuming that DI angle used is the one within 90 degrees of KB)?

 

[dude it's raining]

@ Kadano , I couldn't find a formula on how DI affects KB angle. I know that full perpendicular DI gives +/- 18 degrees, but does it scale linearly? In other words, is the formula something like this:
((DI angle - KB angle)/90)*18 + KB angle
where KB angle is the move KB angle and DI angle is the value of your control stick (assuming that DI angle used is the one within 90 degrees of KB)?

I posted the formula here:
http://smashboards.com/threads/how-to-di-almost-every-move-in-melee.384442/#post-20113132

Also, see here:
http://smashboards.com/guides/that-doesnt-work-and-i-should-know-it.610/

I'm assuming the angle scales linearly with the 18 degrees without control stick displacement taken into account.

 

[1MachGO]

@ Kadano

So dair isn't exactly the safest aerial to use offstage and I was wondering how to maximize its range.

What potentially hits the farthest (offstage) whilst still allowing Marth to recover?

dash 1 frame>short hop>dair?

or backwards short hop>backwards dair?

 

[dude it's raining]

I had an idea that you could crouch cancel shine by TDI'ing up on the control stick and ASDI'ing down with the c-stick.
The idea seemed too good to be true, and I haven't gotten it to work.
It was based on the suspicion that TDI is calculated before ASDI, but since they're both received on the last frame of hitlag (I think), I guess they happen simultaneously, so you can't DI off the ground and then ASDI down to CC. And I guess the same thing happens with gravity on the 1st frame (which I'm assuming is only ignored on the 1st frame when you jump and don't aerial on the first airborne frame).
It'd be really cool to find a way to CC shine and other 0 degree semispikes, but I haven't succeeded.

 

[Kadano]

@ Kadano

If ice climbers jab marth and pop him up does he have time to dair them?

If you SDI up from ~15% on, yes, but their jab 2 will connect before your dair.

So lets say marth fthrow's pikachu and the pika tech's in place at a decent percent (not sure which ones work best.) But could he land a dair after the tech? I know pika's tech in place is freaking amazing and I'll likely have to play against Axe again soon so I wanna get at least something figured out.

I'll likely be in here more often asking tons of questions. :D

Yes, that’s possible. It should be possible to techchase him with wavedash inbetween to close up to his teching spot.