1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
  2. Welcome to Smashboards, the world's largest Super Smash Brothers community! Over 250,000 Smash Bros. fans from around the world have come to discuss these great games in over 19 million posts!

    You are currently viewing our boards as a visitor. Click here to sign up right now and start on your path in the Smash community!

  3. Use the Smashboards Store to get awesome Smash stuff and support the site, like a Nintendo Controller or the Wii U - Gamecube adaptor ! Check out the inventory in our store and support Smashboards with your purchase today!

Peach data and notes

Discussion in 'Peach' started by Veril, Aug 19, 2011.

  1. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    I stopped playing Brawl (for the most part) and went back to melee Peach. Mostly this is due to the booty pop.

    Anyway, I've been doing testing every now and then with her and thought I would share since the hitbox/frame data thread needs moar science and eye melting walls of text. Expect a lot of random crap like

    Any turnip or item pull can follow any previous pull. So two stitch faces in a row is completely possible, though highly unlikely.



    Ratio of % to knockback decreases from top to bottom (other then the fixed tumble % moves) and left to right. Since I don't have the actual bkb and kbg values this isn't exact, but it's a good representation of relative knockback and thus hitstun relative to weight.

    Tumble requires 31+frames of hitstun. hitstun is derived from knockback, so really I'm measuring when moves produce a certain amount of knockback.
    [​IMG]

    Note: The tumble state has certain defining characteristics.
    • During tumble, a character colliding with the ground must bounce or tech. Prior to tumble characters will land on their feet.
    • The character is in hitstun for at least 31 frames.
    • Moves that hit directly upwards will not launch a grounded until they induce tumble. Ex. Mr. Saturn
    •*Characters in tumble cannot airdodge.
    •• though it doesn't specify in this chart, the % listed are those PRIOR to hit and unstale.

    Significant character and % dependent launch effects.
    When does a given move have an advantage on hit vs ___ character:
    --fresh or stale
    --with or without crouch canceling
    --based on DI and teching or lack thereof

    When will a move kill ____ character:
    --fresh or stale
    --with or without crouch canceling
    --with or without DI
    --From the center or sides of FD
    --From the center and sides of Battlefield as well as on the side and top platforms
    --ditto that for Yoshi's, FOD, all but Pokemon Stadium. F that noise

    When will a move launch/induce tumble/force techs/allow for a given followup? How much damage do moves different hitboxes deal? What are their trajectories? What will a phantom hit deal? How can one space to achieve a phantom hit?

    *FYI: I use the term hitbox to refer to a portion of a move with a given damage, trajectory, etc. Hitbubbles make up a hitbox, and share the same properties. For example: Peach's up-smash has a weak hitbox which contains 2 hitbubbles on either side of her.


    single move KO% list:
    Character (Move Name, hitbubble/box description, stage, location on stage) Fresh, no DI - Stale, no DI - Fresh, with DI - Fresh, Crouch Canceled, no DI
    *s=off the side, *t=off the top
    Bowser
    (Up Aerial, Strong Hitbox, FD, Center) 137 - 143 - 149 - 223
    Captain Falcon
    (Up Aerial, Strong Hitbox, FD, Center) 149 - 154 - 160 - ?
    Doctor Mario
    (Up Aerial, Strong Hitbox, FD, Center) 121 - ? - ?
    (Forward Throw, FD, Center) ? - ? - 149 - NA
    (Forward Throw, FD, Side) ? - ? - 134 - NA
    Donkey Kong
    (Up Aerial, Strong Hitbox, FD, Center) 143 - ? - 154 - ?
    Falco
    (Up Aerial, Strong Hitbox, FD, Center) 134 - ? - 144s, 147t
    Fox
    (Up Aerial, Strong Hitbox, FD, Center)126 - ? - 138s
    (d-tilt, FD, Center) 180 - ? - ?
    Ganon
    (Up Aerial, Strong Hitbox, FD, Center) 133 - ? - ? - ?
    Jigglypuff
    (Dash Attack, Strong Hitbox, FD, Center) 116 - ? - ? - ?
    (Forward Tilt, Strong 13% Hitbox, FD, Center) 113 - ? - ? - ?
    (Forward Tilt, Mid 11% Hitbox, FD, Center) 136 - ? - ? - ?
    (Forward Tilt, Weak 10% Hitbox, FD, Center) 149 - ? - ? - ?
    (Down Tilt, FD, Center) 119 - ? - 125 - 217
    (Up-smash uncharged, weak Hitbox, FD, Center) 159 - ? - ? - ?
    (Up-smash fully charged, weak Hitbox, FD, Center) 129 - ? - ? - ?
    (Up Aerial, Strong Hitbox, FD, Center) 88 bwahahaa - ? - ? - ?
    Marth
    (Up Aerial, Strong Hitbox, FD, Center) 119 - ? - ? - ?
    Sheik
    (Up Aerial, Strong Hitbox, FD, Center) 122 - ? - ? - ?


    Jabs: Theory-Fox/Falco can shine in between jab 1 and 2 when CCing at any %. I should check the % the second jab launches, and other character CC punishes. Against a wall its possible to continually jab characters at mid % (gotta test), though they should be able to SDI out. The first jab will combo into the second on all airborne characters and grounded characters except Fox. Why fox? Cause he's a *****.

    Jab 1
    •"reset" limits these characters to the "stand" action: Bowser, DK, Samus, Yoshi, Ganon, CF, Link, Mari/Luigi/Doc, Ness, Zelda/Sheik, Peach, ICs, Marth, MewTwo, Roy, and Young Link. Characters not listed still have the full set of getup options.
    •Damage and Knockback: 1.65-3%; weight dependent bkb, no growth.

    Jab 2
    •Damage: 1.1-2%, low growth. Induces tumble at high%.

    The first jab will not combo into the second jab if it is CCed, though further testing is needed to determine punishing options other than shine. It is also worth noting that the jab hitboxes are disjointed enough that many punishing options will at worst clash with the jab.


    Down-tilt: 6.6-12% damage with all hitboxes. The spike (elbow) and non-spike (hand) hitboxes seem to have the same effects on a grounded opponent. The non-spike hitbox has a different angle when used on a grounded vs. airborne opponent. Against a grounded opponent the trajectory is 90°, against an airborne opponent it is ~70°.

    Fresh d-tilt is impossible to punish by crouch canceling. Even before it starts inducing tumble, it has advantage on hit when CCed. Aerial to d-tilt is in no way a shield combo, though (following a FC aerial) it can safely pressure characters that have poor OOS options (BAD characters), ie substantial startup lag on jump/usmash/upb/grab.

    Combos with the d-tilt: forgive me for not testing with DI... or don't. >.>

    vs Fox: 0-20% d-tilt to jab/d-smash: depending on the timing you can get different results obv. Doing it as early as possible will hit them just before they land, which means they'll have a few additional lag frames and you should be able to get a grab in. If you've conditioned them to expect d-smash cheese they may reflexively be attempting SDI upwards, in which case just grab or dash attack and lol. If you wait a little longer for the jab I believe it can lead to a reset (I did the testing really late at night so my notes got kinda sloppy). Either way, follow up with ****. Or d-smash. :awesome:
    from roughly 20%-74% you can grab, f-tilt or uptilt. Somewhere in this range repeated d-tilts are possible. FC up-air works as well, but I didn't check the exact % where FCing up-air is best. Mix as desired. Yum!
    from 75%-87 up-tilt can reliably followup but f-tilt and grab don't. Up-air can work, but if they didn't DI, why not use the up-tilt instead and keep the up-air fresh?
    from 85-106% you can combo into the tipper hitbox of a charged up-smash and potentially land a KO! HAWT! Within this range it WILL kill without DI on FD, at 100+ iirc it will kill regardless of DI (on the up-smash). Pretty awesome, no? Charge it such that you hit fox as early as possible with the tip and it's a true combo.
    from 105%-140% you can FH up-air for the KO on FD without DI. Again, the upper range of this % will likely kill with DI.


    Up-Tilt:
    Because of its invincibility frames, up-tilt will beat Fox's FH dair. Thank you, magic weave.

    [​IMG]
    [COLLAPSE="Up-Tilt Hitboxes"][​IMG][/Collapse]


    Dash Attack: MMmmmmmm


    Neutral Aerial: The hitbox data thread only lists info for 2 hitboxes which is sad cause I've noted 3. There's the strong hitbox which is just the initial part of the move, that deals 7.7-14%. Then there are 2 weak hitboxes, a back/booty hitbox (4.95-9%, 1 hitbubble) and one for her hands (5.5-10%, 2 hitbubbles).


    Up Aerial is pretty freakin sweet. Its got three different hitboxes: the strongest one hits more or less vertically and does 7.7-14%, the middle hitbox does 6.5-12% and has a slightly more horizontal trajectory, then there's a 6.05-11% hitbox which hits at a 45ish angle if your opponent happens to stick their face up just barely into peach's nether regions... REJECTED! I want to be able to consistently land that but haven't figured out a way yet and its really character dependent (how easily you can land that hit). The middle hitbox is really massively inferior as a KO move compared to the strong one, but it allows for comboing at higher % than the stronger hit. They are pretty similar at low %, but the middle hitbox has craptastic knockback growth compared to the strong hit, so their differences become much more apparent at high %.


    Down-Smash: [​IMG]

    The equalizer. Pivot d-smash, dash cancel d-smash, DJ-land d-smash, vertical launcher -> dsmash, tech chase d-smash, wavedash OOS d-smash, d-smash the shield, FC fair/nair -> dsmash the shield, d-smash people recovering low, SDI out of d-air shine combos and d-smash, crouch cancel any approach by Mario and d-smash,... just don't d-smash Jiggs.


    [COLLAPSE="Magus CG data"]

    [/COLLAPSE]

    Out of shield options frame data stuff I need to organize... if they don't get pushed off the edge... if they do then none of this matters and you should just look at aerial startups and stuff. I think its more useful to look at the vulnerable frames with a defensive maneuver. This includes grabs since they give you a form of invincibility sorta. vf=vulnerable frames before invincibility/hitbox. I'm not gonna bother listing OOS options that have more than 14 vulnerable frames.

    Jump start: the number of frames between the input (frame 0) and the first frame a character is actually in the air. Pretty important for figuring out JC options OOS, aerials OOS, etc.
    Fox, ICs, Pichu, Pikachu, Sheik, Samus, Kirby: 3
    CF, Doc, Mario, Luigi, Ness, GW, Marth, Young Link: 4
    Peach!, Yoshi, DK, Falco, Jiggs, M2, Roy: 5
    Ganon, Zelda, Link: 6
    Bowser: 8 ew

    Vulnerable frames prior to invincibility from a roll:
    Jiggs: 1; Yoshi: None! Yoshi's entire roll duration is invincible; Everyone Else: 3

    Vulnerable frames prior to invincibility from a spotdodge:
    Bowser: 3; Zelda and CF: 2; Everyone else: 1

    Vulnerable frames prior to a shield-grab:
    Samus and Yoshi: 17; Zelda: 11; Link and Young Link: 10; Bowser: 8; DK, Ness: 7
    Everyone else: 6

    Bowser OOS: spotdodge or roll (vf=3), grab (vf=8+), up-b OOS (amazing). All his other options are baaaad.
    Captain Falcon OOS: spotdodge (vf=2), roll (vf=3), grab (vf=6+), uair (vf=9+), nair (vf=10+).
    Donkey Kong: spotdodge (vf=1), roll (vf=3), grab (vf=7+), uair (vf=10+), bair (vf=11+), up-smash (vf=13+), nair (vf=14+).


    Many thanks to:
    Strong Bad for clarifications on move staling and nair hitbox timing.
    KirbyKaze... stuff
    Magus for the Shieldstun and Hitlag formula, up-throw CG info, and Down-B percentages.
    Seanson for his hitbox program
    Stratocaster for the hitbox gifs
    Superdoodleman for frame data
    Sveet for showing me the d-smash phantom hits vs. the sandbag ;p
    VaNz for being Vanz
     
  2. Cia

    Cia
    Expand Collapse
    das kwl

    Joined:
    Feb 18, 2007
    Messages:
    8,231
    Location:
    Top of the Tier List

    Wow Veril. This is amazing and will answer a LOT of questions people have. Thanks a lot for taking the time to do this!

    :teemo:
     
  3. Fortress | Sveet

    Fortress | Sveet
    Expand Collapse
    ▀▄▀▄▀▄▀▄▀▄▀▄▀▄▀▄▀

    • Premium
    Joined:
    Dec 21, 2005
    Messages:
    16,258
    Location:
    Northern IL

    I give this thread a
    [​IMG]
     
  4. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    Thanks Sveet!

    Vanz, the tactical thread is a maze, if you have any data for me to check over or expand on which is mentioned in its dark recesses, could you post the links. I am afraid to wander too far into that thread.
     
  5. thesage

    thesage
    Expand Collapse
    Smash Hero

    Joined:
    Dec 26, 2005
    Messages:
    6,774
    Location:
    Arlington, Va
    3DS FC:
    4957-3743-1481
    I like this thing. Good job. What percents can we do Fc Nair chains on ppl?
     
  6. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    :bee:

    FC Nair chains are definitely something I want to test, but I don't know how I could accurately test something like that without it being in frame advance... er action replay? FC aerial followups are hard to ascertain by oneself without completely neglecting things like opponent SDI or stuff like CCing in between hits to shine. Just too hard for me to do one handed. So it'll be kinda lame west-coast DI testing. Too bad I know nothing about melee hacks :c

    ...

    hmm M2K had a weird method of checking frame data without frame advance. Maybe I could use that. Something with pausing and comparing timing with Samus's d-smash. Idk.

    I'm working on a table with all the shieldstun data that wasn't included in the hitbox and frame data thread. Though I'll probably pass out soon since I drove like 6 hours today and my butt/brain/everything is sore.
     
  7. eet

    eet
    Expand Collapse
    Smash Ace

    Joined:
    Oct 9, 2004
    Messages:
    608
    Location:
    |l|l|l|l|l|l|l|l|l|l|l|l|
    Awesome notes, thank you.
     
  8. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    Been finding F-throw KO% and saw something which surprised me. Correct me if I'm mistaken on this, but it appears that the f-throw trajectory is influenced by something like SDI. Basically with Jiggs I noticed that DIing up simply by inputing the DI as normal resulted in lower% KOs than if I was inputing DI as though it were smash DI, right after the first hit of the throw (which has 2 hits dealing 2 and 8% respectively). Basically the more smashlike DI resulted in a higher launch angle. I was fairly certain this existed in Brawl (didn't really tell people cause it was a trick I actually valued from a competitive standpoint), but didn't know about it in melee. Thoughts? Am I missing something obvious (that happens from time to time lol).

    "Goes back to watching sherlock holmes"
     
  9. ShroudedOne

    ShroudedOne
    Expand Collapse
    Smash Hero

    • Premium
    Joined:
    Mar 14, 2011
    Messages:
    5,492
    This is... a lot of stuff. I'll have to print it out and read it sometime. Thanks, Veril. As for the f-throw thing, it would make sense that the trajectory is influenced by SDI if the f-throw has two hits, right? By SDI'ing the two hits, you would be able to change the trajectory (I think). ASDI may also influence it, but I'm not sure.
     
  10. Strong Bad

    Strong Bad
    Expand Collapse
     

    • Moderator
    • Premium
    • Back Roomer
    Joined:
    Feb 27, 2008
    Messages:
    24,183
    Location:
    Oakland, CA

    uhhh do you not know how staling works? fully staled moves do *.55 damage, so jab would do 1.65% damage. same with pummels.
    on nair: frame 3 through 7 (or 8, i don't remember) the hitboxes all do 14. past that, two of them do 10 and the other one does 9.
    sometime when I'm not busy I'll check everything else.
     
  11. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    My area of expertise was brawl, so I'm kinda relearning physics/mechanics for melee. Its kinda frustrating but fun cause I like learning new stuff. I did not know what the move decay, though I already checked how it affects shieldstun/hitlag as per Magus's excel sheet.


    Noted. I wasn't gonna hit with a stale move 100 times to find out the .05% difference ;p

    Though 10 times was fine (now I feel stupid). Idk, the cutoff for pathetic geekery is a pretty arbitrary line lol.
    Thanks for the frame data behind that :D much appreciated. Source?

    I figured out the damage for all the nair hitboxes but not the timing since I don't have action replay. Trying to get into melee hacking but its daunting and I'd rather just check CC, KO and combo stuff.

    Word.

    I'm gonna fix the damage minimums to account for the staleness multiplier I now have.


    You should wait till the semester starts if you're gonna print it out. I'll probably slow down a lot around the start of September. Then again, I might be motivated by the Vassar smashfests or just have data from them. We shall see.
     
  12. Strong Bad

    Strong Bad
    Expand Collapse
     

    • Moderator
    • Premium
    • Back Roomer
    Joined:
    Feb 27, 2008
    Messages:
    24,183
    Location:
    Oakland, CA

    You mean my excel sheet?
    It is 10.
    1 - (.09 + .08 + .07 + .... + .01) = 0.55
    the character files
    You have a hacked Wii, right? That's how I run AR. Hit me up on AIM or something and I'll help you out.
    Melee keeps track of decimal percentages and calculates knockback based on the percent after the hit (staleness is irrelevant if you keep track of the percent after the hit, though with stale moves you can get an absolute % before the hit that will kill regardless of staling). You'd need to use training mode and major abuse of the stale moves system to calculate decimal %'s though (it's what Magus and I have done on occasion~).

    Also
    You can't SDI Peach's F-Throw since the hitbox occurs before you're released. Jiggs' hitbox hits after the throw release so you're not held and can thus SDI.
    but SDI doesn't change your launch trajectory in either game... <_<
     
    Kadano likes this.
  13. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    That's not what I meant. I'd get a move to max staleness and then reset the opponents damage counter and hit with it 10 times to figure out the damage to the tenths place. Thus how I figured out 1.6 (the .05 doesn't register from 10 hits). 16% damage / 10 hits = 1.6


    My bad, I meant Magus's formula as shown in the sheet that I did not realize you made.


    I just got the melee iso tonight so I could... be confused by it. I wish I had a text dump of the character files that would be way easier for me to go through.

    That would be awesome. I'm not up for it tonight, but perhaps sometime later in the week. Either that or I'll ask Cape this weekend. Again, the help is really appreciated.

    Ah, I remember Brawl has knockback being calculated after the hit also. "more notes"


    I'm not sure I agree with this but definitely have spent too much time not working on smash stuff to be up for debating this point (that SDI allows for a greater change of trajectory than DI alone).

    What?

    Could you clarify that. Either I can or can't influence survival % on the f-throw with SDI. That's the point that actually matters, because I threw that puffball enough times to notice that there was a very distinctly SDI-like effect from a very distinctly SDI-like input which resulted in KO% a few points higher than obtained through normal survival DI.

    I don't want or need you to type up a primer on throw mechanics. Really. Just explain the bold stuff. You mean Jiggs weight pushes the release point back before the other f-throw hitbox (which is still part of the f-throw... and not something Jiggs creates so the "Jiggs' hitbox hits" statement confuses me).

    In brawl pretty much every multi-hit throw was weight independent (other than stuff involving articles iirc, like the spacies lasers). I guess its different in melee, I'll have to look at Magus's throw data thread (oh boy).
     
  14. Strong Bad

    Strong Bad
    Expand Collapse
     

    • Moderator
    • Premium
    • Back Roomer
    Joined:
    Feb 27, 2008
    Messages:
    24,183
    Location:
    Oakland, CA

    You can't SDI Peach's F-Throw, regardless of what character you are.
    You can SDI Jigglypuff's F-Throw, regardless of what character you are.

    Was a bit ambiguous there, sorry.
     
  15. Cave Dweller

    Cave Dweller
    Expand Collapse
    Smash Apprentice

    Joined:
    Sep 5, 2010
    Messages:
    157
    Location:
    Pittsburgh, PA
    Yeah thanks man, this kicks ***. STICKY THIS.
     
  16. darkatma

    darkatma
    Expand Collapse
    Smash Hero

    Joined:
    Dec 10, 2005
    Messages:
    5,745
    Location:
    St Louis, Missouri/Fremont, CA
    The first hit of peach's jab has set knockback I believe.. or something similar.. but it's weak enough to CC even at 999%. Meaning fox can always shine between jab 1 and 2.
     
  17. KirbyKaze

    KirbyKaze
    Expand Collapse
    Smash Legend

    Joined:
    Nov 18, 2007
    Messages:
    17,678
    Location:
    Spiral Mountain
    1st jab is always strong enough to give a FFer their full reset repertoire (roll both directions, rise, and rising attack).

    For this reason, you can't "reset" in the same way that Jigglypuff can. They're never forced into a "rise only" situation after jab; they always have the other options available. Consequently, at low percents, I think it's better to just down smash again or FC Nair > see what they do > react accordingly (since FC Nair > block often beats rising attack, and can sometimes still let you chase rolls). That said, jab resets are sexy so if you use them instead of a convoluted thing (esp. at mid percent when they might procure you a good grab) then nobody will judge. Just don't anticipate a simple rise and prepare for the unexpected.





    edit:

    If most characters CC your D-tilt at 0% and they get put into a "reeling back" or "flipping over" animation that keeps them on their feet but still in stun (think Falco's Dair at low percents or when you CC Falcon's stomp) you can combo to D-smash from that DIRECTLY with good timing.

    On FFers, I still think it's better to D-smash after a D-tilt for a low percent range (not sure what range) just because jabbing is ****ing awful because of her set (low) KB and stun on it. If they are prepared to SDI up for D-smash there's no reason why they won't flick up after the jab (as a result of how SDI works) and jump whatever you try to follow the jab with. And given how craptastic Peach is at following mobile jumps like Fox's and Falco's (especially at low percent) there's no point in jabbing. I think down smash or letting them land and going for the tech chase are both superior choices.
     
  18. eet

    eet
    Expand Collapse
    Smash Ace

    Joined:
    Oct 9, 2004
    Messages:
    608
    Location:
    |l|l|l|l|l|l|l|l|l|l|l|l|
    Can this be the new technical thread? The other one has 4 pages of fabulous(ly annoying)
     
  19. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    Yeah the post launch pre-techable state that occurs with ccs is pretty harsh. I list the percents where that occurs for falcon when he ccs the d-tilt and would have for fox except he goes right from non-launch to forced tech.

    Your wall of text was delicious but the ipad keyboard is balls and I'm at Cape's house without my computer :(

    I will have a better response when/if I can borrow his laptop while he's at work.

    Have my babies?

    :phone:
     
  20. Strong Bad

    Strong Bad
    Expand Collapse
     

    • Moderator
    • Premium
    • Back Roomer
    Joined:
    Feb 27, 2008
    Messages:
    24,183
    Location:
    Oakland, CA

    kirbykaze is mine, veril, back off.
     
  21. Wake

    Wake
    Expand Collapse
    Smash Master

    Joined:
    Sep 27, 2010
    Messages:
    3,191
    Location:
    Thank you Based Mimi.
    You lucky guy you.
     
  22. KirbyKaze

    KirbyKaze
    Expand Collapse
    Smash Legend

    Joined:
    Nov 18, 2007
    Messages:
    17,678
    Location:
    Spiral Mountain
    Only if DippnDots lets me out of the chicken cage :p
     
  23. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    OK, home again. Driving back from PA to NY right after the hurricane was... an adventure. The sort of adventure that has you saying things like "oh crap I just drove over downed power lines and boy am I glad they weren't live" and "that tree wasn't there last week" while navigating a mystery route because the major highways all had portions shut down. In short, good mother****in times!

    Visit to Cape's was fun and we talked shop about melee and brawl modding, worked out, etc. It really jogged my memory to get grilled about data for days ;p Plus I got to see BionicSonic and talk about planking (god what a terrible thing) and why Shield SDI can in no way shape or form beat it, even with silly frame perfect testing+diddy dair (or anything for that matter). Ah reasons not to play brawl lol

    I digress. As is my way...

    Some points on topic :p

    1. Peach's d-smash is not safe on CC vs. all characters at low % (as is, does not have advantage on CC). People just assume it is cause bazillion damage omg wtf. VaNz, if you're curious and read this, shoot me a PM and I'll decide if its something I should post. I already told Cape and he was like, "Really... you can do THAT! Makes sense I guess."

    2. ASDI can influence trajectory by virtue of the bounce/tech state canceling vertical momentum.

    3. The flipping over state I was referring to is not the same as the stationary pre-launch state. Pre-launch occurs with all characters I've tested so far when the d-tilt is CCed at 0. "Flipping" occurs at a % before teching in place is possible but after the pre-launch stationary percentage range. There's no term for it as far as I'm aware but it looks like a backflip vs. pre-launch which is just them standing in place and flinching. Example: Fox and Jigglypuff don't "flip" on CCed d-tilt at all, they skip that state and go straight to forced tech. I'm pretty sure its something that occurs with heavier characters primarily.


    No school today cause of the hurricane so I'm working on the d-tilt CC effect table.

    We just have to eliminate 2 of the 16 guys. After we break him out of the chicken cage, of course.

    Peach board you so strange.
     
  24. Strong Bad

    Strong Bad
    Expand Collapse
     

    • Moderator
    • Premium
    • Back Roomer
    Joined:
    Feb 27, 2008
    Messages:
    24,183
    Location:
    Oakland, CA

    get me some hitstun values for these thresholds and I can probably figure this mechanic out. I already have a hypothesis for what's going on but need some data to prove it. Since you're already grinding out this stuff do it for a char or two and I can most likely calculate it for the rest.
     
  25. Magus420

    Magus420
    Expand Collapse
    Smash Master

    Joined:
    Dec 13, 2003
    Messages:
    4,541
    Location:
    Close to Trenton, NJ Posts: 4,071
    Smashing the stick has no effect on trajectory DI in either game (having worked with and modified the assembly code in memory that is run during those routines directly). When you're getting different results smashing it vs holding it's because you are having the stick in a slightly different position on each of those tests. When trajectory DI is to be read, in melee it finds the distance perpendicularly of the default trajectory's slope to the stick's current position, squares that distance, then multiplies that by 0.314159 radians (18 degrees) resulting in the amount the default angle is to be changed by your trajectory DI. Since the edges of the controller can get up to about a distance of 1 (varies by the controller itself and the angle in question) the strongest trajectory DI in melee is around ±18° on all attacks when input full distance perpendicularly. Odd angles like Sheik's f-air can't get a full distance of 1 perpendicularly for the most part because it's inbetween controller notches so they usually can't be DI'd the full 18, but more like 16.5-17.5 or so.

    In brawl it's the same formula but does not square it (so a sub-par angle is stronger than it'd be in melee). Also in brawl the stick reaches directional input caps of 1/-1 well before the max range of motion of the stick, so something like diagonally up+right can reach a stick position close to 1,1 (instead of closer to 0.707,0.707 like melee). The strongest trajectory DI is around ±18° on attacks with default trajectories in cardinal directions (perp distance can reach up to 1 as it too would be cardinal), and about ±24° on ordinal trajectories (can get perp distance up to around 1.333).

    Pound (and other attack) spike in brawl is when you are hit in the air in that game you can hit floors with ASDI like melee (when hit off the floor ASDI does not trigger a floor collision unlike melee). When you are hit just off the ground, SDI back into the ground, ASDI + trajectory DI into the ground you hit the ground (like melee CCing and instant ground teching). It's not 'SDI knockback inversion'. The floor bounce/tech actions cancel vertical KB, and you are ASDIing towards the floor far enough that it plus the KB away from the floor still has you connect with the floor on the 1st frame, triggering the stage collision.

    The trajectory DI in those vids of mine are within the normal range (I don't 'SDI' the trajectory DI either). I am trajectory DIing as low as possible by holding it (usually down+away), and also ASDIing with c-stick directly into the floor along with it (usually straight down). Using c-stick for ASDI lets you ASDI in a different direction than you are holding for trajectory DI.
     
    Kadano likes this.
  26. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    Woooord. Thanks Magus. The part about floor bounce / teching canceling vertical knockback is good to know.
     
  27. Magus420

    Magus420
    Expand Collapse
    Smash Master

    Joined:
    Dec 13, 2003
    Messages:
    4,541
    Location:
    Close to Trenton, NJ Posts: 4,071
    That is only true for characters that are Falco/Pikachu's weight or lighter. Everyone heavier than them never have the option of a roll or getup attack and are always forced into standing up on a reset. Her Jab1 is set KB yes, but it's set KB that is affected by weight and the KB threshold on it to allow doing those other things falls right between those weight classes.

    117 - Bowser
    114 - Donkey Kong
    110 - Samus
    109 - Ganondorf
    108 - Yoshi
    104 - Captain Falcon / Link
    100 - Dr Mario / Luigi / Mario
    94 - Ness
    90 - Peach / Sheik / Zelda
    88 - Ice Climbers
    87 - Marth
    85 - Mewtwo / Roy / Young Link

    80 - Falco / Pikachu
    75 - Fox
    70 - Kirby
    60 - Jigglypuff / Mr Game & Watch
    55 - Pichu



    Also, it's not just FFers that have a harder time DIing up out of a reset like most people would think. It's based entirely on fall acceleration and not top speed, so ignoring how easily they can get into the situation, Bowser and Ganon actually have a more difficult time escaping the reset than CF, since they all have a gravity of 0.13 and CF is the lightest.

    [​IMG]

    I explain this chart more and how resets work in general here: http://www.smashboards.com/showthread.php?p=12815207

    Since the jab is so strong compared to stuff like Jigglypuff's jab and Sheik's at very low damage though it will easily pop most of the lighter characters on that list up onto their feet without even DIing up (DIing towards is enough for Roy to land on his feet since he's light and grav is a little lower than some of those). Someone like Ganon would need to actually be DIing up since he's heavier and has higher gravity.



    The flip some characters do on high amounts of grounded hitstun but before causing tumble (tumble happens when hitstun is 31+ frames) is just an animation thing and doesn't mean anything special compared to a character with a non-flipping animation.
     
    Kadano likes this.
  28. Cia

    Cia
    Expand Collapse
    das kwl

    Joined:
    Feb 18, 2007
    Messages:
    8,231
    Location:
    Top of the Tier List

    Magus and Veril data in one thread?!!

    YESSSSSSSSss!!
     
  29. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    Magus data is better. He's some kinda I don't even know... math ninja. I try though ;p and I'll definitely be looking over his info and trying to extrapolate Peach...jab stuff... from it. Gotta find your throw thread and nab the Peach data from that as well :D

    This is clearly a nexus of geekery though. KirbyKaze, Magus, Veril... needs Cape. Maybe he'll name-search and share his turnip recovery tricks.

    The flipping animation is really funny looking, even if it is nothing special. Ganon does it over a wider % range than anyone else for some reason, at least when CCing Peach's d-tilt. Ganon is an acrobat, clearly.

    Tested CC d-tilt effects, but the tests are assuming they are holding straight down to CC rather than down and to the side. So as a result I've got the launch % with ASDI down (its still called ASDI even if its not done with the C-stick, right? I'm not really up to date on my melee terminology.). I noticed that with the stick down and to the side, the launch % are different and assume this is due to different ASDI. Like if I d-tilt say Ness at 102 and the stick is straight down, he'll stay grounded and bounce or tech in place, but if I have the stick down and to the side, he'll still CC (as shown by knockback differences with non-CC) but he'll get launched. Got like 5 more characters to test than I can finish the CC d-tilt table (though I might want to test with the CC with the stick diagonal rather than straight down to remove ASDI effects).

    Ugh school >.< 60 page syllabus for clinical rotations! WTF! At least I already know all the pharmacology stuff we're going over this semester.
     
  30. ShroudedOne

    ShroudedOne
    Expand Collapse
    Smash Hero

    • Premium
    Joined:
    Mar 14, 2011
    Messages:
    5,492
    ...
    ...
    ...
    ...

    My head hurts. But I suppose Magus's posts will do that to you.
     
  31. Strong Bad

    Strong Bad
    Expand Collapse
     

    • Moderator
    • Premium
    • Back Roomer
    Joined:
    Feb 27, 2008
    Messages:
    24,183
    Location:
    Oakland, CA

    there is a space in my name, Veril
     
  32. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    Some stuff I tested, and am putting into a chart, but the chart isn't done yet so here's some idea of what I'm working on:

    Lowest to highest % inducers of tumble based on what I've tested so far (fresh, if you want to find the % stale... math, I'll explain if someone really wants to know, anyway... using Bowser as he is the heaviest and tumbles at the highest %):
    1. 0%: Fair, all throws but forward, d-tilt spike, d-tilt ground hitbox.
    •. 18%: 12 damage up-tilt hitbox.
    •. 24%: 13 damage f-tilt hitbox
    •. 27%: f-throw, strong nair
    •. 34%: 14 damage upair hitbox
    •. 41%: anti-air d-tilt non-spike hitbox
    •. 119%: Jabx2 both hits

    did testing for a bunch of other characters but I'll post that when I've got it neat and tidy.... ish

    :phone:
     
  33. Composeur

    Composeur
    Expand Collapse
    Smash Ace

    Joined:
    Nov 10, 2007
    Messages:
    518
    This thread should be stickied. Really nice job, Veril. Anyone see what the implications from this might be?

    Also, Veril, what is the exact deal with her dtilt "spike" on the edge? Can it be done consistently, teched/cancelled in some other way?
     
  34. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    The utility of the down tilt spike is situational. Whether you can reliably hit a character sweetspotting the edge with their recovery depends on the stage and their character. You've gotta be right up on the edge to hit them with the spike rather than the non-spike. You have to be really close to the edge cause the recovering character has to hit Peach's elbow. The non-spike is on her hands, below and in front of the spike-hitbox. Against phantasm it isn't practical to try and hit with the spike. Most up-b moves can sweetspot without the spike hitting them on FD. If they mess up the sweetspot and get spiked: if the stage has a solid ledge, they can SDI and tech at any% or meteor cancel, which can save most characters at low %. Bowser would die but you'll be getting hit by his up-b before you can land the spikebox.

    Its usually better to hit recovering characters with the d-tilt non-spike. Depending on % and DI you can combo into nair/bair/fair any DI at low to mid%, another d-tilt/f-smash/d-smash at lower % or vs FFers, or turnips (aerial b-toss, smash tosses, w/e) vs. anyone with DI away. Once the opponent hits high % it isn't really practical to use d-tilt unless its to spike and the stage is conducive to this. The d-tilt spike becomes somewhat better on Brinstar or Kongo Jungle where characters can be hit through the bottom of the stage if their hurtbox pokes through. Lava combos and the barrel cannon can set this up. Situational as anything, but hilarious and awesome.

    tl;dr: d-tilt spike is usually bad. d-tilt non-spike vs recoveries is ok but there are probably better options. D-tilt on stage is pimp as f***, doesn't care if they CC, reliable combos.... <3 onstage d-tilt.



    Brinstar really is amazing. Wouldn't it be cool if VaNz wrote about Brinstar?! I think so.
     
  35. Composeur

    Composeur
    Expand Collapse
    Smash Ace

    Joined:
    Nov 10, 2007
    Messages:
    518
    I second the statement about VaNz. *cough*

    Wow, thank you so much for the data and analysis, Veril. I will definitely put dtilt "spike" (referred to by this even though it can be meteor cancelled?! What a gip) out of my head except on crazy counterpick stages (where are you, Mute City? waahh).

    Another question, if you have time. This one's something extremely pertinent to my gameplay that I've asked on the forums before, but you're very methodical and I think that would be very beneficial in this case. I want to know the mechanics of breaking out of tumble, and what the best ways to go about it might be in various practical situations. I know mashing A (or b or x/y) and wiggling the control stick are the two ways to break out, but I don't know which is "stronger" or if they can be done during the stun period to buffer a break out of tumble or if they don't help until stun is over. It seems like that could have implications for how you might DI in the aftermath of some high-stun moves like falco's dair. It becomes even trickier when trying to airdodge/waveland out of tumble because specific timing is very important there. I often find myself wondering if it's even theoretically effective to DI and tech, DI and nair, DI, mash and nair, or mash and waveland, or something else in certain circumstances (like dthrows of various characters, just to name one).
     
  36. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    That's a really good question, but I've got a test in 20 min. No real point in cramming and I'm not that nervous but... Nicotine time. I'll try to answer all of that when I get home.

    :phone:
     
  37. standardtoaster

    standardtoaster
    Expand Collapse
    Tubacabra

    • Moderator
    • Premium
    Joined:
    Nov 26, 2009
    Messages:
    9,258
    Location:
    Eau Claire, Wisconsin
    Yo veril, find out peach pluck chances. kthx
     
  38. Veril

    Veril
    Expand Collapse
    Frame Savant

    Joined:
    Jun 20, 2008
    Messages:
    3,062
    Location:
    Kent Lakes, New York

    Too bad we can't rely on VaNz namesearching himself. There are a lot of people who can be summoned to threads by just typing their name, but he isn't one of them afaik.

    Thanks for the thanks!

    Ah yes, Mute City... yeah that stage is pretty **** if its legal. I should have included that. D-tilt spike is more useful there than on the neutrals. Probably its best on Brinstar, followed by Mute City, then Kongo Jungle.

    Spike is easier to write than meteor smash. If only it were the other way around, and "true spikes" were called meteor smashes. If only...

    OK, there are parts of that question which I'll need to use the debug menu to answer completely. Wiggling the control stick and other inputs will reduce the time you are in the tumble state. You can do this during hitstun. This doesn't change the amount of hitstun you suffer, but it does allow you to airdodge sooner after the end of hitstun because you won't be in tumble. Afaik it has no impact on how soon you can use an aerial attack, those aren't limited by tumble. I'll have to check if specials are restricted by tumble.

    The number of frames of tumble removed per input is something idk... yet. "adds to list"

    My problem is that my copy of melee is stuck in a wii that will not eject disks and is not hacked. My wii, which IS hacked, is lacking a melee disk. le sad

    Mash -> nair is unnecessary because you can already aerial out of tumble (iirc). The list of options you've got on hit would probably include:
    • DI+tech • DI+wiggle+waveland • DI+nair • DI+double jump ->waveland • DI+double jump land • DI+wiggle while holding jump -> float • DI+SDI like a fiend ->nair

    Obviously depends on the situation, move hitbox properties, your %, the opponents habits, etc. I'm not really at the point where I'd be confident telling anyone the best option for combo-DI vs whatever. I have good survival DI but my combo DI needs lots of work.

    Sorry if I didn't really answer your question fully. I need to be able to use the debug menu / frame advance. Is it cool if I call it frame advance? I have been forever and I'm slow to change lol.

    ok
     
  39. ShroudedOne

    ShroudedOne
    Expand Collapse
    Smash Hero

    • Premium
    Joined:
    Mar 14, 2011
    Messages:
    5,492
    So, I have a question, if it's not too much trouble.

    Veril, what are the exact percents that Peach can reliably begin a CG on Fox/Falco/Falcon, and be able to maintain that CG? As in, if I up throw Fox at a certain percent, I want my throw animation to end, and for me to be able to dash after him on the earliest possible frame, before he hits the ground.

    So, I'm looking for the earliest percents to start one. I already have a good idea of the caps. Thanks in advance.
     
  40. Composeur

    Composeur
    Expand Collapse
    Smash Ace

    Joined:
    Nov 10, 2007
    Messages:
    518
    I am also curious about what % the (guaranteed) CG starts on the FFers!

    Thanks for the response so far. It's really good to know that mashing can be worth doing during hitstun. In that case, it would be most efficient to actually do full rotations of the control stick rather than wiggle it, right? As I understand it, it is the most effective way to mash, but it's just dangerous to do because if you get caught doing it once the hitstun is over then you waste your second jump (and also because you can't DI if you're doing that).

    Hmm, I really thought mashing A helped made the nair come out of tumble sooner, but I guess the reason that it helps is just because it makes it more likely that it will come out as early as possible when you're not sure of the timing. I actually mash A and Z alternately so I can do it faster.
     

Share This Page

Users Viewing Thread (Users: 0, Guests: 0)

We know you don't like ads
Why not buy Premium?