Quick Question, Quick Answer Thread

[thexsunrosered]

Airdodge cancel throw.

Airdodge to catch and then flick the C-stick to throw it. (You can press A as well.. just stating the easiest way)

:leek:

oh alright, I wasn't doing it fast enough then. Thanks susa

 

[SuSa]

Is there any way to calculate when a move would KO given the stages blast radius, the moves knockback, the moves hitstun, and the moves trajectory (factoring perfect DI, but no SDI)?

I ask because of the following:

Spoiler

Snakes Utilt KO %'s with perfect DI + momentum cancel in TRAINING MODE

Mario: 119%
Luigi: 117%
Peach: 110%
Bowser: 134%
DK: 141%
DiddY: 119%
Yoshi: 126%
Wario: 126%
Link: 139%
Zelda: 106%
Sheik: 110%
Ganon: 134%
TL: 117%
Samus: 122%
Zamus: 110%
Pit: 121%
Ice Climbers: 114%
ROB: 124%
Kirby: 109%
Meta Knight: 105%
King D3: 152% O_O
Olimar: 111%
Fox: 115%
Falco: 119%

Snake's KO %'s with NO DI, on FD, IN TRAINING MODE:
http://www.smashboards.com/showthread.php?t=236143

Notice how MOST (not all..) fall within 13-17% of the no DI data?

However these %'s were gathered in training mode (which is an issue), and I'm unable to test the in versus mode (requires 3 controllers + a working copy of BAWrl)

So I thought to myself.... "Is there any way to calculate this data?"

If we can find a way to calculate basic vertical KO moves (DDD utilt, Snake utilt, Fox Up Smash, etc.) then I see that as a step forward.

Things that I think would make a difference:

Size of stage [From Platform to Top Blast Zone] -- We have this data now.
Trajectory of move --- We can acquire this, but factoring DI is tricky.....
How fastfalling effects trajectory (complex formula time? :x)

:leek:

 

[rPSIvysaur]

There are too many variables that we have not been able to make a conclusive formula.

 

[Indigo Jeans]

There are too many variables that we have not been able to make a conclusive formula.

QFT. Joe's way ahead of you. Or was. Check his doc.

 

[Yikarur]

I have one problem since years
Yoshis double jump has weird actings.

Yoshis double jump cancel can be cancelled by an aerial or an airdodge.
even the turnarround animation is instantly cancelled.
but if I do the aerial/airdodge close to the ground the double jump doesn't get cancelled, the animation does fully end and it then buffers the aerial

on battlefield, stay under a side platform.
00:00 Input Jump and Hold Jump
07:88 Leaves ground
20:66 Jump + Back
21:65 Airdodge/Aerial

see what happends ._.

 

[A2ZOMG]

Most importantly nobody even knows exactly how knockback works. We don't exactly have a precise formula that determines knockback distance as far as I know.

 

[SuSa]

Most importantly nobody even knows exactly how knockback works. We don't exactly have a precise formula that determines knockback distance as far as I know.

I wish I had a working copy of BAWrl so badly right now.

DI allows for ~20-25° variance from the default without accounting for SDI and reversibility. So to formulate the effect of DI you need to look at hitbox flags (esp special collision), SDI modifier, and angle. Also knockback and damage values in relation to character since pre-tumble hits are not DIable... and whether the move has a discrete multihit... and... Its not possible, don't bother trying.

Sorry, even simple accounting for kbg, bkb, ∂kb, damage, character % and weight vs angle (just change the angle of a move ±20° to test) is extremely daunting as a task

• Is there any way to determine an exact variance for "default" moves? If we can find that variance that is one variable down. (The maximum that a move can be altered, although I see this as being visually debatable unless someone can measure it in some way...)
• Another variable would be hitstun (I'm assuming hitstun is actually what I'm assuming it to be... which is the time before you can input an action such as an aerial after bein ghit) |||| Do we have any information on this? (I know we have it on hitlag), if so we have a point at which we see a further change in the trajectory due to the ability to fastfall.
• Character % would be what is being solved as "x".
• BKB, KBG would be variables we gather from the move itself (EG: Snake's utilt)
• ∂kb (what the **** is this?)
• Stale moves should not be factored. Everything should be assumed fresh, because while I know how stale moves effect damage I still don't know how it effects other things. If we can find it's relation to other things than we can change the formula to factor in different rates of "stale" (wow that sounded weird)

What I know we have to find for sure:

• Relation of knockback and weight, and how weight lowers the knockback of a move.
• How fastfalling counteracts vertical knockback.
• The rate at which knockback weakens over distance (would need to be tested on a large stage)
• Ceiling boundary heights. (We have this, but we need to find the relation of everything else before it's appliable)

Of course figuring all of this out seems impossible.

Anyone have any theories for a way to test this information? I have a few but I can't really... put them into words... not without a copy of BAWrl to ensure I'm not imagining a few things (it involves the data at the end of the match)

:leek:

 

[Luxor]

Finding the exact angle that DI affects you at would be pretty neat.
Hitstun formula... nope. Don't got it.
BKB, KBG, ∂KB are all built into the move, and all closely related so w/e.
Stale moves could be ignored.

In a controlled environment, with moves dealing 0% and controlled knockback via PSA, we could perhaps come up with a workable model. Not even a good model, like this, but one like this. And it would still take a lot of work.

 

[SuSa]

Testing that exact angle (I'd think) wouldn't be that hard given a way to test it accurately and using controlled angles via PSA.

My idea:
70 and 90 degrees. See if you can DI from 70 to 90 perfectly and vice versa? Test this with a few other angles as well..

Hitstun formula is.... still a problem. =\ I don't really have any ideas for this. Given controlled BKB, KBG, ∂KB, and controlled %'s - using frame-by-frame to figure out the amount of frames one is in hitstun for and trying to work out a formula that works at every % may work?

As for the amount of work, I think the most work would be calculating a hitstun formula and the other relations (blast zones, stage, etc.) a workable model could be made for a move.... figuring out the rest of the moves would be a metter of changing variables and number crunching (which I can find someone to create a program to do so.... or go to the drawing board, suck it up, and actually learn Ruby/C++)

:leek:

 

[Luxor]

I think it would be easiest to ignore BKB totally at first and just get KO %s for a control move on FD. Then we do that for a bunch of sample KBG's (and angles?) on Lucario from the middle of FD. After that, we introduce a few multipliers for position on stage (easy maths + my dataz), weight multiplier (got them), and subtract a correction factor in accordance with BKB. That should get a *workable* model, theoretically, with a minimum of work (although still a lot).
I guess you don't really *need* to test angles, since ignoring gravity, the same KB at different angles will send you the same distance, and you can math that together with the sine or cosine of the angle to see which blastzone it will pass first and at what % that will happen... this part was mostly just so I could say "sine" or "cosine" outside of math class, so testing angles would probably be better. Still.

And then a little applet would be nice.

 

[Indigo Jeans]

Iirc, BKB = base knockback, no? So uh, how do we ignore that? I say get a sample move and test BKB on a stage with the BDU texture, then check out trajectory and launch distance in some terms of some working of BDU and hitstun/post-hitstun frames. Imo a good test move would be something with a trajectory of 70, no electric multiplier, maybe 8 frames of hitlag, and obviously average BKB. KBG and weight KB would be irrelevant as you're testing on a Lucario/Ivysaur @0%, and your attack is at the 1.0 damage multiplier(duh). Kill per cents don't seem necessary when you're in the beginning stages of constructing a working model.

And theoretically, if you want to test for KO per cents without DI/FFing/w/e in as neutral an environment as possible, you'd have to do so in a perfectly circular, large stage, with the test subject suspended in the center, so all trajectories are tested in a non-arbitrary way, thus only testing for knockback-based per cents... but that's never going to happen.Irrelevant now that I read it over, but I don't wanna delete it lol. Anyhow, current test methods aren't well constructed, as you're just getting favorable per cents, not static per cents. People test moves with largely horizontal KB at stages' horizontal extremeties, and other moves on stage centers. This is flawed imo.

Tl;dr start testing BKB trajectories, then BKB trajectories with survival inputs, and compare launch distances/distances away from favorable blastlines in BDU. THEN get to the heavy stuff.

Please excuse the terrible wording and lack of construction... I'm tired but I still wanted to give my thoughts.

 

[KayLo!]

the BDU texture,

Link?

I have a grid texture for the futuristic custom stage builder, but I'm not sure if it's in BDU.

 

[Luxor]

Iirc, BKB = base knockback, no? So uh, how do we ignore that?

I figured you'd test a move with X KBG and try several different BKB values to see how much the kill % varies by, and correct for that in the end approximation. I honestly don't think BDU and counting frames of hitstun, etc., will really add anything until we have a hitstun formula, and it seems like a lot lot of work.

 

[A2ZOMG]

If it's worth knowing, Falco's B-air seems to be the one move in Brawl that is purely knockback growth based. It has zero bkb and weight kb.

 

[saviorslegacy]

http://www.youtube.com/watch?v=syUZnbX9nCo
Any idea on how it works?

I think it might have to deal with the infinite second jump thingy....

 

[SuSa]

Looks exactly like an SJR thing to me.

:leek:

 

[rPSIvysaur]

I looked at it in VirtualDub frame by frame. It's not a double jump but it literally has no jumpstart lag.

 

[Pierce7d]

Perhaps it needs it's own thread then

 

[thexsunrosered]

how exactly does knockback growth work? I'm pretty seasoned as far as staleness goes, but knockback growth as always eluded me in terms of understanding. I know that moves like ness' bair have really large amounts of knockback growth, and that using them when the opponent is at a high percent will kill them mad quick, but that just seems like any move to me, use it at a high enough percent and it'll kill lol

 

[KayLo!]

I think it's how much knockback grows for each percent of damage.

 

[Pierce7d]

It's simple.

Base knockback is a default amount of knockback, that a move deals regardless of percent. D3's Down Throw will send you a small distance even at 0. Warlock Punch will send an opponent a significant distance, even at 0.

Knock Back Growth sends opponents farther, the more damage they have. Marth's Dtilt has very little knockback growth, so it doesn't kill until around 400%. On the other hand, DK's Fsmash will K.O. you at around 80%, because the gain in knockback the move gets from the opponent having higher percent is great.

 

[Yikarur]

I have one problem since years
Yoshis double jump has weird actings.

Yoshis double jump cancel can be cancelled by an aerial or an airdodge.
even the turnarround animation is instantly cancelled.
but if I do the aerial/airdodge close to the ground the double jump doesn't get cancelled, the animation does fully end and it then buffers the aerial

on battlefield, stay under a side platform.
00:00 Input Jump and Hold Jump
07:88 Leaves ground
20:66 Jump + Back
21:65 Airdodge/Aerial

see what happends ._.

woooosh D:

 

[saviorslegacy]

Apparently the Olis have known about this for a while and I am getting that "old" spam. What I am curious about it why it works.

Is there anything else out there that is similar to this?

 

[Yikarur]

I would look into it BUT I can't see the video at all ^__^ it's Gaga.

 

[saviorslegacy]

I would look into it BUT I can't see the video at all ^__^ it's Gaga.

What does that have to do with anything?
Don't tell me, it is copyrighted in Germany?

 

[Yikarur]

seems so..
"this video contains content of UMG and is not available in your country"

something like that ^__^...

 

[saviorslegacy]

seems so..
"this video contains content of UMG and is not available in your country"

something like that ^__^...

Now that is gay. I'll have to change the video and upload it with a different song for ya.


Here's the video for your Yikarur:
http://www.youtube.com/watch?v=At9U0WKWByU
If that music gets blocked that is dumb, since it is German. <,<

 

[KayLo!]

Does anybody know where to find collision info for projectiles?

I can't find the hitbubble info for Din's Fire, just stuff about Zelda's animation while she's doing it.

 

[Luxor]

Did you check the articles and subactions? I'd bet it's under articles.

 

[A2ZOMG]

Yeah it's under articles (use OSA2 to find the offset for the hitbox of Din's Fire and other projectiles). Din's Fire is a weird projectile in particular because it is listed as having extremely low damage because there is a damage multiplier that takes effect as the move stays out. A small change to the damage or knockback parameters of that move makes a HUGE difference in how the move operates.

 

[KayLo!]

Found itttt. Thanks guys.

@A2: Yeah, the damage is listed as 1, but that's not a big deal I guess. I'll go in-game to test the things I can't figure out.

 

[rPSIvysaur]

Since the damage on Din's Fire scales, 1 is the base damage. =/

 

[KayLo!]

One small question: knockback's units are mph, right? Making knockback growth mph per percent of damage?

& one bigger question: How do collisions attached to bone 0 work? They sometimes have ridiculous offset number(s), like on Zelda's nair (X offset for one collision is 71576.79), whereas other times the offsets are fairly normal-looking (ex: Pika's fsmash).

 

[A2ZOMG]

Yeeeeeah, those X offsets really baffle me. I think those huge numbers represent something fundamental that I'm not aware of.

 

[Indigo Jeans]

Bone 0 is a temporary/"prop" bone, correct?

 

[rPSIvysaur]

Those "huge" offsets are just negative numbers.

 

[KayLo!]

Those "huge" offsets are just negative numbers.

Okay, and what does that mean?

Also, I'm assuming that since nobody corrected me on the knockback stuff, I'm right about it. =X

 

[rPSIvysaur]

When it's like 7400 or w/e it is, it is usually a small negative number, but it interprets it as a larger number instead of smaller. It is annoying, I had the conversion thing, but I forgot it. =/

There is no real measurement of knockback that is actually reliable in any form.

 

[A2ZOMG]

PSA 1.3 is gay when a negative number comes up. Stick to OSA2 and PSA 0.2.1

One of the things I eventually need to get my head around is how hitbox axis placement interacts with moving bones. And I don't yet understand exactly how Z axis displacement works.

 

[Kinzer]

I want to post this here before I go off making a new thread, calling it something new, find out it isn't, and making myself look like an idiot.

Could anybody tell me how/why Espy managed to go through the platform when he got grab-released?