Complete Hitbox/Frame Data For Every Character (Now with KO Percents)

[KERO]

Decided to go ahead and analyze some of Greninja's substates. Here's what I believe some of them are:

unknown substate 0x23a Shadow Sneak back flip
Frame 7- 8: 10% 60b/102g (1428g*) 48° 1.3-Hitlag

unknown substate 0x23b Shadow Strike back flip
Frame 10-11: 15% 65b/102g (1938g*) 55° 1.8-Hitlag

unknown substate 0x23c Shadow Dash back flip
Frame 16-18: 6% 65b/90g (900g*) 80° 1.3-Hitlag

unknown substate 0x23e Shadow Sneak back kick
Frame 7- 8: 12% 60b/101g (1616g*) 48° 1.3-Hitlag

unknown substate 0x23f Shadow Strike back kick
Frame 10-11: 17% 65b/101g (2121g*) 55° 1.8-Hitlag

unknown substate 0x240 Shadow Dash back kick
Frame 16-18: 7% 65b/90g (990g*) 60° 1.3-Hitlag

unknown substate 0x24a Aerial Shadow Sneak back flip
Frame 7- 8: 10% 60b/102g (1428g*) 48° 1.3-Hitlag

unknown substate 0x24b Aerial Shadow Strike back flip
Frame 10-11: 15% 65b/102g (1938g*) 55° 1.8-Hitlag

unknown substate 0x24c Aerial Shadow Dash back flip
Frame 16-18: 6% 65b/90g (900g*) 80° 1.3-Hitlag

unknown substate 0x24e Aerial Shadow Sneak back kick
Frame 7- 8: 12% 60b/101g (1616g*) 48° 1.3-Hitlag

unknown substate 0x24f Aerial Shadow Strike back kick
Frame 10-11: 17% 65b/101g (2121g*) 55° 1.8-Hitlag

unknown substate 0x250 Aerial Shadow Dash back kick
Frame 16-18: 7% 65b/90g (990g*) 60° 1.3-Hitlag

Note: There seems to be no difference between the aerial and non-aerial Shadow specials.


unknown substate 0x26c <---- No idea about this one
Enables all ledge grabs on real frame 0

unknown substate 0x26e Substitute (counter window)
Enables state transition on real frame 8
Cancels state transition on real frame 30

unknown substate 0x26f Explosion Attack (intangibility not included?)
Frame 40-42: 10% 55b/100g (1400g*) 361° 1.5-Hitlag Fire
Frame 40-42: 8% 60b/100g (1200g*) 361° 1.5-Hitlag Fire

unknown substate 0x270 Substitute Ambush (counter window)
Enables state transition on real frame 11
Cancels state transition on real frame 41

unknown substate 0x276 Substitute (successful counter)
Frame 5- 4: 14%(+5) 50b/100g (1800g*) 90° 1.1-Hitlag (Upward)
Frame 5- 4: 13%(+5) 62b/100g (1700g*) 60° 1.1-Hitlag (Diagonal upward)
Frame 5- 4: 13%(+5) 62b/72g (1224g*) 25° 1.1-Hitlag (Diagonal downward)
Frame 5- 4: 11%(+5) 62b/100g (1500g*) 361° 1.1-Hitlag (Side-ways/no input)
Frame 5-19: 14%(+5) 50b/100g (1800g*) 45° 1.1-Hitlag Ground-Target-Only (Downward)
Frame 5-19: 14%(+5) 50b/100g (1800g*) 270° 1.1-Hitlag Aerial-Target-Only (Downward)

unknown substate 0x278 Substitute Ambush (successful counter)
Frame 5- 4: 18%(+23) 75b/65g (1430g*) 90° 1.1-Hitlag (Upward)
Frame 5- 4: 17%(+23) 85b/65g (1365g*) 60° 1.1-Hitlag (Diagonal upward)
Frame 5- 4: 17%(+23) 85b/55g (1155g*) 25° 1.1-Hitlag (Diagonal downward)
Frame 5- 4: 15%(+23) 85b/65g (1235g*) 361° 1.1-Hitlag (Side-ways/no input)
Frame 5-19: 18%(+23) 75b/65g (1430g*) 45° 1.1-Hitlag Ground-Target-Only (Downward)
Frame 5-19: 18%(+23) 75b/65g (1430g*) 270° 1.1-Hitlag Aerial-Target-Only (Downward)

unknown substate 0x286 Aerial Substitute? (counter window)
Enables state transition on real frame 8
Cancels state transition on real frame 30

unknown substate 0x28a Aerial Substitute? (successful counter)
Frame 5- 4: 14%(+5) 50b/100g (1800g*) 90° 1.1-Hitlag (Upward)
Frame 5- 4: 13%(+5) 62b/100g (1700g*) 60° 1.1-Hitlag (Diagonal upward)
Frame 5- 4: 13%(+5) 62b/72g (1224g*) 25° 1.1-Hitlag (Diagonal downward)
Frame 5- 4: 11%(+5) 62b/100g (1500g*) 361° 1.1-Hitlag (Side-ways/no input)
Frame 5-19: 14%(+5) 50b/100g (1800g*) 45° 1.1-Hitlag Ground-Target-Only (Downward)
Frame 5-19: 14%(+5) 50b/100g (1800g*) 270° 1.1-Hitlag Aerial-Target-Only Downward)

unknown substate 0x28c Aerial Substitute Ambush? (successful counter)
Frame 5- 4: 18%(+23) 75b/65g (1430g*) 90° 1.1-Hitlag (Upward)
Frame 5- 4: 17%(+23) 85b/65g (1365g*) 60° 1.1-Hitlag (Diagonal upward)
Frame 5- 4: 17%(+23) 85b/55g (1155g*) 25° 1.1-Hitlag (Diagonal downward))
Frame 5- 4: 15%(+23) 85b/65g (1235g*) 361° 1.1-Hitlag (Side-ways/no input)
Frame 5-19: 18%(+23) 75b/65g (1430g*) 45° 1.1-Hitlag Ground-Target-Only (Downward)
Frame 5-19: 18%(+23) 75b/65g (1430g*) 270° 1.1-Hitlag Aerial-Target-Only (Downward)

Note: Not sure why Substitute Ambush is lacking a second substate like Substitute has. If I had to guess, they may share the same aerial frame data, but I find that unlikely as Substitute Ambush does still seem to have a little slower start-up/more end lag.... Also, as far as aerial vs grounded goes, Substitute's counter window and counter hitboxes all act the same/have the same frame data regardless.

unknown substate 0x292 I want to say Single Shot Pump, but that is already covered under gekkouga_water
Frame 20-26: 5% 8b/40g (360g*) 361° 2.0-SDI

unknown substate 0x294 <--- This one utterly perplexes me. None of Greninja's moves are unblockable
Frame 8-10: 2% 100f/40w 90° 0.2-Hitlag Slash Unblockable
Frame 243-243: 15% 65b/160g (3040g*) 275° Slash Unblockable

unknown substate 0x298 <--- Not sure about this one either.
Frame 15-24: 2% 50b/50g (300g*) 22° 2.0-Hitlag Stun Unblockable

gekkouga_shuriken


unknown substate 0x8 Uncharged/Slightly Charged Water Shuriken
Frame 0- 1: 11% 20b/85g (1275g*) 50° Slash
Frame 0- 1: 3% 10b/85g (595g*) 50° Slash

unknown substate 0x9 Uncharged/Slightly Charged Stagnant Shuriken
Frame 0- 1: 1.2% 10b/70g (364g*) 50° 0.8-SDI Slash
Frame 0- 1: 1% 10b/70g (350g*) 50° 0.8-SDI Slash

unknown substate 0xa Uncharged Shifting Shuriken
Frame 0-15: 3% 90b/45g (315g*) 93° 1.5-Hitlag Slash
Frame 16-17: 3% 90b/45g (315g*) 98° 1.5-Hitlag Slash

unknown substate 0x10 Fully Charged Water Shuriken
Frame 0-28: [0.7%]x5 100f/50w 21° 0.15-Hitlag 0.6-SDI Slash
Frame 0-28: 0% 30b/85g (340g*) 50° 0.6-SDI Slash
Frame 30-31: 9% 40b/122g (1586g*) 60° 1.5-Hitlag Slash

unknown substate 0x11 Fully Charged Stagnant Shuriken
Frame 0-69: [2%]x9 10b/70g (420g*) 60° 0.6-SDI Slash
Frame 0-69: [1.5%]x9 10b/70g (385g*) 60° 0.8-SDI Slash
Frame 70-71: 3% 70b/100g (700g*) 45° 1.5-Hitlag Slash

unknown substate 0x12 Fully Charged Shifting Shuriken
Frame 0- 9: 3% 60b/70g (490g*) 110° Slash
Frame 10-19: 3% 60b/70g (490g*) 120° Slash
Frame 20-29: 3% 60b/70g (490g*) 130° Slash
Frame 30-31: 3% 60b/70g (490g*) 135° Slash

gekkouga_water

unknown substate 0x0 Hydro Pump (first hit)
Frame 1- 2: 2% 65b/100g (600g*) 58° 2.0-Hitlag Water Push

unknown substate 0x1 High-Capacity Pump (first hit)
Frame 1- 2: 0% 130f/90w 58° 2.0-Hitlag Water Push

unknown substate 0x2 Single-Shot Pump (first hit)
Frame 1- 2: 5% 55b/85g (765g*) 40° Water Push

unknown substate 0x4 Hydro Pump (second hit)
Frame 1- 2: 2% 65b/100g (600g*) 58° 2.0-Hitlag Water Push

unknown substate 0x5 High-Capacity Pump (second hit)
Frame 1- 2: 0% 130f/90w 58° 2.0-Hitlag Water Push

unknown substate 0x6 Single-Shot Pump (second hit)
Frame 1- 2: 5% 55b/85g (765g*) 40° Water Push

There seems to be no differences at all (frame and damage-wise) between the first and second hits/pumps of any of Greninja's up-bs.

Finally, there were some interesting substates that I'm about 95% sure relate to Water Shuriken in some way (the very first eight after Dthrow), but it's sort of hard to match them up.

EDIT: After some testing, I believe that Stagnant Shuriken is the slowest (the one that allows a state transition on frame 12 most likely) with Shifting Shuriken being the frame 9 one. All the frame six ones may just be the various Water Shuriken sizes. Then again, Shifting Shuriken has two distinct speeds. Not really quite sure what to make of this, but from testing, I do believe Stagnant is the slowest followed by Shifting.

If anyone could perhaps piece this together better than I am, we'd at least have somewhat complete data on Greninja's specials.

 

[Lavani]

Thanks @ KERO , I knew if I slacked off long enough someone else would get things done :^)

unknown substate 0x26c <---- No idea about this one
Enables all ledge grabs on real frame 0

I'd imagine this is just the non-water portion of upB, since they all share the same animation and afaik can grab the ledge at any point.

unknown substate 0x292 I want to say Single Shot Pump, but that is already covered under gekkouga_water
Frame 20-26: 5% 8b/40g (360g*) 361° 2.0-SDI

unknown substate 0x294 <--- This one utterly perplexes me. None of Greninja's moves are unblockable
Frame 8-10: 2% 100f/40w 90° 0.2-Hitlag Slash Unblockable
Frame 243-243: 15% 65b/160g (3040g*) 275° Slash Unblockable

unknown substate 0x298 <--- Not sure about this one either.
Frame 15-24: 2% 50b/50g (300g*) 22° 2.0-Hitlag Stun Unblockable

These are all related to the final smash. 298 is the tatami mat spawning, 292 is the launching flip, 294 is the midair slashes and finisher.

 

[KERO]

Thanks @ KERO , I knew if I slacked off long enough someone else would get things done :^)


I'd imagine this is just the non-water portion of upB, since they all share the same animation and afaik can grab the ledge at any point.


These are all related to the final smash. 298 is the tatami mat spawning, 292 is the launching flip, 294 is the midair slashes and finisher.

Okay, yeah, the Final Smash things make sense. However, I was testing the non-water portion on each up-b, and all three don't grab the ledge until the moment the water comes out (and even then I feel there's some delay as you can be right above the ledge, where you'd still grab it, up-b, and miss the ledge entirely).

 

[Lavani]

Hmm.

It's possible to grab the ledge before he shoots water with upB, but he does go through that initial animation first before he'll grab it I think.

I'm going to guess for now that 26B is the startup (no sfx) and 26C is the water firing part of upB.

 

[Big O]

I noticed that they changed the official name of DK's Down B 1 to Hand Slap and that I accidently typed donky for all the Storm Punch entries which made your parser skip them. This is an updated list with the fixes.

donkey 0x220 Cargo Ftoss
donkey 0x221 Cargo Btoss
donkey 0x222 Cargo Utoss
donkey 0x223 Cargo Dtoss

donkey 0x230 Giant Punch (ground, uncharged)
donkey 0x231 Lightning Punch (ground, uncharged)
donkey 0x232 Storm Punch (ground, uncharged)

donkey 0x234 Giant Punch (ground, charged)
donkey 0x235 Lightning Punch (ground, charged)
donkey 0x236 Storm Punch (ground, charged)

donkey 0x244 Giant Punch (air, uncharged)
donkey 0x245 Lightning Punch (air, uncharged)
donkey 0x246 Storm Punch (air, uncharged)

donkey 0x248 Giant Punch (air, charged)
donkey 0x249 Lightning Punch (air, charged)
donkey 0x24a Storm Punch (air, charged)

donkey 0x24c Headbutt (ground)
donkey 0x24d Jumping Headbutt (ground)
donkey 0x24e Stubborn Headbutt (ground)

donkey 0x250 Headbutt (air)
donkey 0x251 Jumping Headbutt (air)
donkey 0x252 Stubborn Headbutt (air)

donkey 0x254 Spinning Kong (ground)
donkey 0x255 Chopper Kong (ground)
donkey 0x256 Kong Cyclone (ground)

donkey 0x258 Spinning Kong (air)
donkey 0x259 Chopper Kong (air)
donkey 0x25a Kong Cyclone (air)

donkey 0x260 Hand Slap (ground, loop)
donkey 0x261 Focused Slap (ground, loop)
donkey 0x262 Hot Slap (ground, loop)

donkey 0x268 Hand Slap (air)
donkey 0x269 Focused Slap (air)
donkey 0x26a Hot Slap (air)

 

[Thinkaman]

Merry Christmas, have a table:

http://pastebin.com/NJ4Q5PzU

TODO:

• auto-filtering of extra files
• Data output overrides
  • MK/ZSS Jab
  • Kirby/Luigi dash attacks
    • Hits On
    • Total Damage
  • Peach f-smash
    • Hits On
    • Total Damage (3x)
  • Link grab/dash grab
  • Yoshi dair
    • Hits On
    • Total Damage
  • Fox/Falco shines
• Article spawning event detection
• Pikmin
• Specials (requires manual subaction flagging)
• Damage/Frame output mode
• Adjusted knockback output mode
• OoS HitsOn output mode
• Total damage auto-calc

 

[Alaya]

A few obvious Lucina ones:

Neutral B

unknown substate 0x21a
Frame 8- 9: 8.075%(+30) 45b/95g (1147g*) 361° Pierce


Up B's

unknown substate 0x27a
Frame 5- 6: 11% 70b/68g (1020g*) 74° 0.5-Hitlag Slash
Frame 5- 6: 11% 70b/68g (1020g*) 361° 0.5-Hitlag Slash
Frame 6-11: 7% 20b/90g (990g*) 74° Slash
Frame 6-11: 7% 20b/90g (990g*) 361° Slash
Frame 6-11: 6% 20b/90g (900g*) 74°
Enables state transition on real frame 5
Enables forward ledge grabs on real frame 7

unknown substate 0x27b
Frame 9-10: 7% 60b/50g (550g*) 30° 0.5-Hitlag Slash
Frame 9-10: 7% 60b/50g (550g*) 35° 0.5-Hitlag Slash
Frame 11-16: 12% 85b/60g (960g*) 45° 1.2-Hitlag Slash
Frame 11-16: 12% 85b/60g (960g*) 55° 1.2-Hitlag Slash
Frame 11-16: 6% 40b/60g (600g*) 55° 1.2-Hitlag
Enables state transition on real frame 10
Enables forward ledge grabs on real frame 14
unknown substate 0x27c
Enables state transition on real frame 15
Enables forward ledge grabs on real frame 17


Final Smash

unknown substate 0x295
Frame 2- 5: 60% 120b/120g (7680g*) 361° Slash Unblockable

 

[Meru.]

230

#0x61727163
def game_230():
AsynchronousTimer_0D0(Frame=11.000000, )
FrameSpeedMultiplier_0DF(1.500000, )
AsynchronousTimer_0D0(Frame=41.000000, )
FrameSpeedMultiplier_0DF(1.000000, )
End_196()

def gfx_230():
pass

def sound_230():
AsynchronousTimer_0D0(Frame=2.000000, )
unk_18E(0x11003B, )
End_196()

def expression_230():
pass

230 in Peach's folder is very likely Side B's (Peach Bomber) start-up. Comes out on frame 11, active for 30 frames (till 41).

20E
#0x41477b2f
def game_20E():
unk_0E0(0x1, )
unk_16A(0x1, 0x0, )
AsynchronousTimer_0D0(Frame=11.000000, )
unk_1C8(0x2, 0x0, 0x1, )
AsynchronousTimer_0D0(Frame=32.000000, )
unk_1C7(0x2, 0x0, 0x1, )
End_196()

def gfx_20E():
AsynchronousTimer_0D0(Frame=9.000000, )
unk_0CE(0x1000009, 0x0, -4.000000, 0x0, 0x0, 0x0, 0.000000, 0x0, 1.000000, 0.000000, 0x0, 0.000000, 0x0, 0x0, 0x0, 0x0, )
AsynchronousTimer_0D0(Frame=10.000000, )
unk_079(0x100008F, 0x19, 0x0, 3.800000, -0.500000, 0.000000, 0x0, 0x0, 1.400000, 0x1, )
End_196()

def sound_20E():
AsynchronousTimer_0D0(Frame=9.000000, )
unk_182(0x43, 0x11, )
End_196()

def expression_20E():
unk_203(0x0, )
unk_1D2(0x6, )
AsynchronousTimer_0D0(Frame=17.000000, )
End_196()

20E is Neutral B Toad's start-up. Comes out on frame 11, active till 32.

Not sure if this is helpful, probably not very much

 

[Thinkaman]

230 is flying peach bomber startup

Edit: Can anyone get me field data on the frame MK jab hits on?

 

[Alaya]

Merry Christmas, have a table:

http://pastebin.com/NJ4Q5PzU

TODO:

• auto-filtering of extra files
• Data output overrides
  • MK/ZSS Jab
  • Kirby/Luigi dash attacks
    • Hits On
    • Total Damage
  • Peach f-smash
    • Hits On
    • Total Damage (3x)
  • Link grab/dash grab
  • Yoshi dair
    • Hits On
    • Total Damage
  • Fox/Falco shines
• Article spawning event detection
• Pikmin
• Specials (requires manual subaction flagging)
• Damage/Frame output mode
• Adjusted knockback output mode
• OoS HitsOn output mode
• Total damage auto-calc

Very nice. Would be pretty cool to have this data in an excel file or something so that it could be easily sorted whenever you want to see who has the fastest/slowest of certain moves.

 

[Thinkaman]

I already can output to both comma-separated (for excel/google doc) and ascii. Ascii is just faster for use while we fix the data. I'll post a public google doc of the data once it is polished up and fully correct.

I currently need "HitsOn' (frame timing), "TotalDamage", and optionally "Duration" (total hitbox coverage life) of the following oddball non-special moves, which have to be collected manually:

metaknight 0x50 {}
miigunner 0x69 {}
murabito 0x56 {}
murabito 0x69 {}
murabito 0x6a {}
ness 0x63 {}
ness 0x65 {}
pikmin 0x5f {}
pikmin 0x63 {}
pikmin 0x65 {}
pikmin 0x69 {}
pikmin 0x6a {}
pikmin 0x6b {}
pikmin 0x6c {}
pikmin 0x72 {}
pikmin 0x73 {}
pikmin 0x74 {}
rockman 0x58 {"HitsOn": 2.0}
rockman 0x5f {}
rockman 0x6b {}
rockman 0x6c {}
yoshi 0x6c {"HitsOn": 16.0, "Duration": 28.0, "TotalDamage": 34.4}

Also, TotalDamage for any multi-state normal (jabs, MK f-tilt, Link/TL f-smash) are welcome

Edit:
Total Damage Table!

http://pastebin.com/GgwvXpdr

There should be plenty of errors, please note them. Repeating jabs and repeating jab finishers are not included.

 

[Big O]

I was testing out some MK stuff the other day. MK jab hits frame 7 as far as I can tell (in-game).

 

[Lavani]

I can confirm frame 7.

Considering he only has a rapid jab though, it wouldn't be related to 0x50.

 

[Thinkaman]

Considering he only has a rapid jab though, it wouldn't be related to 0x50.

Right, the chart's script is looking at 050 for every character, and MK's naturally comes up empty. He has to be handled as a special case.

The main data dump has been updated. It includes the rough draft max (total) damage values now, and the most up to date special strings. Fake grab/throw hitboxes should also have been completely filtered out for good.

Edit:
Updated damage table with pummel + throws

Edit2:
Updated damage table with separated jabs and repeating jabs (min length)

 

[Amazing Ampharos]

I've started gathering the hits on data for the remaining cases. Villager's fair hits on f10, bair hits on f13, and dash attack hits on f14. The last one was very hard to test due to the nature of dash attacks; it consistently did trade with DDD dsmash (f14), lost to Kirby fsmash (f13), and beat Mario fsmash (f15) though so I think that's a good sign it must be f14. I'll be moving on to Ness and Mega Man next, then figuring out if Mii size affects Mii frame data, and either way figuring out Mii Gunner's missing move after that leaving just Olimar to be a "fun" project.

EDIT: Ness usmash hits on f9 (but only from SUPER close range) and Ness dsmash hits on f10.

 

[Thinkaman]

Charts updated. Please review them before I export the data to sortable excel/google doc files.

http://pastebin.com/NJ4Q5PzU (Earliest Frame)
http://pastebin.com/GgwvXpdr (Total Damage)
http://pastebin.com/PzTjdtx8 (Full Data)

Edit:
Note that sometimes, the mechanically possible total damage is totally impractical. This is true of many dual-hit down-smashes; you CAN hit with both hits (think Ganondorf), but it all-but can't happen due to the knockback and angles in play.

This is why you see d-smash damages like the following:
captain: 32
diddy: 27
ike: 31
link: 29
littlemac: 24
lucina: 23
luigi: 30
mario/mariod: 22
marth: 29
MK: 17
miiswordsman: 27
ness: 22
pit/pitb: 24
reflet: 27 (surprisingly)
samus: 22
shulk: 50 (lololol)
yoshi: 24
zelda: 22

If you guys think it's a good idea, we can manually handle all of these as edge cases, inserting "correct" damage values. But, they ARE, mechanically, the true total damages of the moves...

Edit 2:
We might also want to change Villager F-smash, which is currently listed at its true max of 17% damage, which it only gains after falling a long distance. 99.99% of all bowling balls do 15%.

 

[Amazing Ampharos]

We're getting close to being able to have things really nice with the amount of info we have. I'm pretty sure Thinkaman decided on some sleep, but I decided to get frame data for the last character:

Olimar:

fsmash: f11
usmash: f12
dsmash: f10
fair: f7
bair: f10
uair: f9
dair: f9
stand grab: f10
dash grab: f10
pivot grab: f10

Fun Olimar Facts:

When Olimar's grab is broken by getting hit on the same frame you land the grab, the opponent does not take 3% damage as is the case when other characters grab.

Olimar's pikmin as aerials can be clashed out! Even better, since the pikmin are unique actors from Olimar, clashing with the pikmin doesn't prevent the hitbox from hitting Olimar directly. For instance, if Bowser and Olimar are directly on top of each other, Olimar uses uair, and Bowser uses bair, both will try to hit on frame 9. Bowser's bair and Olimar's uair will clash, rendering Olimar unable to hurt Bowser and Bowser unable to hurt the pikmin. The bair will then hit Olimar. Good character.

With that, we now have the basic start-up frame data for every "A" button normal and grab in the game. We also do have the ledge attacks and get-up attacks; I'll have to pester Thinkaman to include them in whatever he puts together to look nice as a release. Special moves are of course a mess, and in terms of studying match-ups, they are even more complicated even if we did have them all documented since you never quite know which moves your opponent will be packing. However, knowing what the fastest options are is very important so I've put together a table.

The following table is a summary of the fastest actions characters can do from a neutral standing position without using special moves (since you never know what special moves you'll be packing, and most are slower than the character's fastest normal option anyway). That is, I look at jabs, tilts, smashes, and standing grabs. Dash attack and other grabs are not included since starting up a run takes frames that make them never faster than standing in place and acting (though Wario in particular has a f4 dash attack and no other f4 grounded moves). If I don't indicate which move it is, jab is the exclusive fastest option that character has. If I do indicate moves, any moves in parenthesis are the fastest (if multiple, they are tied).

Frame 1:

Little Mac
Zero Suit Samus

Frame 2:

Dr. Mario
Falco
Fox
Luigi
Mario
Mega Man (Jab/Ftilt)
Mii Brawler
Peach
Pikachu

Frame 3:

Captain Falcon
Diddy Kong
Greninja
Kirby
Meta Knight (Dtilt)
Ness
R.O.B. (Jab/Dtilt)
Samus
Sheik
Sonic
Villager
Yoshi

Frame 4:

Bowser Jr. (Jab/Dtilt)
Charizard
Duck Hunt
Ike
Mr. Game & Watch
Olimar
Pac-Man
Robin
Wii Fit Trainer

Frame 5:

Dark Pit (Jab/Dsmash)
Donkey Kong (Jab/Utilt)
Jigglypuff
Lucina
Marth
Mii Gunner (Jab/Utilt)
Mii Swordfighter (Dtilt)
Pit (Jab/Dsmash)
Rosalina (Dtilt)
Shulk
Wario (Dtilt)
Zelda (Dtilt/Dsmash)

Frame 6:

King Dedede (Dtilt)
Lucario (Jab/Utilt/Grab)
Toon Link

Frame 7:

Bowser
Ganondorf (Grab)
Link
Palutena (Jab/Grab)

Do note that frame data for Mii Fighters is size invariant; all sizes of Miis have the exact same frame data so, yes, this DOES apply to any and all matches including them.

Special moves can change the equation. While I'd like to eventually have a full list of all special moves with properties that can act before the character's fastest normal, I can help out a bit now with a list of all of the moves I know with frame 1 invincibility or super armor with moves:

Invincibility (ground only):

Dr. Mario Up-2: Super Jump
Dr. Mario Up-3: Ol' One-Two
Mario Up-1: Super Jump Punch
Mario Up-2: Super Jump
Zero Suit Samus Down-1: Flip Jump
Zero Suit Samus Down-2: Shooting Star Flip Kick
Zero Suit Samus Down-3: Low Flip

Super armor (ground or air):

Charizard Down-3: Rock Hurl
Donkey Kong Side-3: Stubborn Headbutt
Wii Fit Trainer Down-2: Volatile Breathing

Shorter list than you expected? Marth, Lucina, and Bowser up specials are NOT frame 1 invincible contrary to popular belief, and yes, Mario and Dr. Mario are different in which offensive up-B variant has invincibility with the two of them being the only f1 invincible attacks I can find that will follow up the invincibility quickly with a hit. If you don't believe me, just get them to stand next to Zero Suit Samus in training mode, set frame advance, and have them up-B at the same time as ZSS uses her jab. They'll get hit every time while the characters who have actual invincibility will not. I also tested a variety of other moves I suspected might be invincible (G&W up specials, Mii Brawler down-2, Diddy side-2, Kirby up-3, Mii Swordfighter up-1, Pit up-1, Charizard up-3, Samus up-1, Bowser Jr. up-2, Falco side-2) and none of them were. I did not test every special move in the game as that would be excessive, and I don't have a way to view invincibility in the files like that. If anyone knows of any moves I've left out, if any exist, please let me know. And yes, Wii Fit Trainer's f1 armored Volatile Breathing does work and is every bit as goofy as you probably imagine it to be.

That's all I have for now; hopefully the information in this post will help everyone have a more complete understanding of some of the most important particulars of the frame data. Why this topic, IMO the most important topic on the board, has only three pages of posts I really don't know. Maybe some tables like we're starting to make (like the ones above) will have people notice "hey, frame data, that super gameplay critical thing every competitive player needs to look at" or something.

 

[TheReflexWonder]

For the Total Damage chart, does Smash 4 factor in a "fresh" damage bonus in normal matches, as they did in Brawl? I think it was 1.05x in Brawl.

As such, would a move dealing 8% break Little Mac's Neutral-B armor threshold on its first use?

 

[Thinkaman]

For the Total Damage chart, does Smash 4 factor in a "fresh" damage bonus in normal matches, as they did in Brawl? I think it was 1.05x in Brawl.

As such, would a move dealing 8% break Little Mac's Neutral-B armor threshold on its first use?

These are raw values from the data, as viewable in training mode. There are no external modifiers applied.

Also note that the table values are floored. For example, Palutena repeating jab is a minimum of 8.8 damage, which is shown as 8.

 

[Thinkaman]

Edit: Double post. Needed to notify someone, not sure if edits do that?

Now I ask because if Im looking for my quick options to utilize. Should I be taking WFT's grab as a frame 3 or 4 option? Or Frame 6?

Sorry for getting distracted and forgetting about this. The most likely explanation is that you recorded at 30 frames per second, as most capture devices are limited to.

 

[icraq]

The following table is a summary of the fastest actions characters can do from a neutral standing position without using special moves (since you never know what special moves you'll be packing, and most are slower than the character's fastest normal option anyway).

just a little nitpicking! it should be noted if rosalina has luma, utilt comes out on frame 2. luma's jab is frame 3. you may want to include one of those! or not, i'm not sure if luma counts for this.

 

[Thinkaman]

I'll manually make a special case that adds luma in under rosalina soon.

I might also move lucina to under marth.

Edit:
I haven't done those chat things yet, but I updated the full list to account for Change Hitbox Damage instructions. I also labeled Mii Brawler and Gunner.

Edit2:
Added several other updates, including many special labels for a variety of characters.

Edit3:
The full data now has approximate angle-adjusted KO percentages for all applicable hitboxes. It's important that you understand the assumptions in play for these estimates:

• Default Mii target
• Starting position on FD
• No stale moves, rage, or other modifiers

Because this is ground level, uairs will be severely underrepresented from their usual kill case.

Horizontal moves will also kill better or worse depending on stage position. And, obviously, spikes will kill magnitudes sooner off-stage. (I'm currently accounting for the stage-bounce, adjusting angle and knockback accordingly.)

 

[Thinkaman]

Sorry for double posting, but I just made a chart of all KO percentages--I'll quad post if I bloody feel like it.

I NEED YOUR HELP!

Adding automated special support to the charts requires 3 things for each special in the game:

• filename and subaction_index of strongest hitbox (with maximum charge, if applicable)
• if it is a multi-substate special (and the strongest hitbox isn't in the first substate), we need the confirmed frame the move first hits on (with any hitbox, with minimum charge if appliable)
  • Some of this can be gleaned from inspection of the data, but most require field testing
  • All projectiles are like this!
• if it is a multi-substate special with hitboxes across multiple subactions, we need the confirmed total max damage (with maximum charge, if applicable)

Currently we are only considering ground versions of specials.


Example 1:
I want to add data for Mario Fireball.

I know hitbox is in subaction mario_fireball 0x0. I check this list and confirm.

I compare the speed to other moves in Training Mode, by using the freeze feature and doing two moves at the same time, to see which is faster. I try difference moves and discover that Mario Fireball can first hit on frame 16.

I don't need to manually go add up the damage. (If say, Mario hit the enemy for 3% and then the fireball itself also did 5%, I'd report that the total max damage is 8%.)

So:
mario NB1 mario_fireball 0x0 {"HitsOn": 16.0}
(The "NB1" is "neutral B 1", side B 3 would be SB3, ect.)


Example 2:
I want to add data for Falcon Dive.

I know the powerful throw hitbox I want is in captain 0x23a, from looking at the data and the subaction list.

I compare the speed to other moves. Here I'm going to cheat because it's a simple grab: The "grabbing lunge" subaction of Falcon Dive grabs on frame 14. I'd typically go check this in game against some other frame 14 move to confirm.

Now I need to go add up the total max damage this move can do. Where possible, we should trust the data over the game. (Due to rounding) Falcon Dive does 5% on initial grab and 12% on throw, for 17% total.

So:
captain UB1 captain 0x23a {"HitsOn": 14.0, "TotalDamage": 17.0}

 

[Unknownkid]

I noticed you guys are missing the titles for Kirby's Special Move. So I did a cross reference with our Moveset Thread and here what I achieved. I made sure to verify with my fellow Star Warriors before posting here. The bolded names are the moves I am absolutely certain about.

unknown substate 0x212 (Inhale)
Frame 0- 1: Grab
Frame 0- 1: 0% 100f/10w 160° 0.0-Hitlag 0.0-SDI

unknown substate 0x213 (Ice Breath)
Frame 0- 1: 6% 50b/50g (500g*) 40° 1.2-SDI Ice
Frame 0- 1: 2.4% 100f/10w 60° 0.7-Hitlag 1.2-SDI Ice
Frame 0- 1: 2.4% 10b/100g (640g*) 40° 0.7-Hitlag 1.2-SDI Ice
Frame 0- 1: 1.2% 100f/10w 60° 0.7-Hitlag 1.2-SDI Ice

unknown substate 0x214 (Jumping Inhale)
Frame 0- 1: Grab

unknown substate 0x21a
unknown substate 0x21c
unknown substate 0x21e
unknown substate 0x220


unknown substate 0x232 (Inhale – Star Shot)
Frame 0- 1: 6% 0b/0g (0g*) 361°

unknown substate 0x234 (Jumping Inhale – Star Shot)
Frame 0- 1: 5% 0b/0g (0g*) 361°

unknown substate 0x25e
unknown substate 0x260

unknown substate 0x282 (Inhale – Swallow)
Frame 0- 1: 10% 0b/0g (0g*) 361°

unknown substate 0x284 (Jumping Inhale – Swallow)
Frame 0- 1: 8% 0b/0g (0g*) 361°

unknown substate 0x288 [All of these are Giant Hammer I think]
Begins super armor on real frame 45
unknown substate 0x294
Begins super armor on real frame 0
unknown substate 0x298
Begins super armor on real frame 0
unknown substate 0x29c
Begins super armor on real frame 0
unknown substate 0x2a0
Begins super armor on real frame 0
unknown substate 0x2a4
Begins super armor on real frame 0
unknown substate 0x2a8
Begins super armor on real frame 0
unknown substate 0x2ac
Begins super armor on real frame 0
unknown substate 0x2b0
Begins super armor on real frame 0

unknown substate 0x2ba (Uncharged Hammer Flip)
Frame 11-11: 19% 60b/70g (1610g*) 53° 1.2-Hitlag

unknown substate 0x2bb (Hammer Bash)
Frame 11-11: 21% 55b/70g (1750g*) 75° 1.2-Hitlag
Frame 11-11: 18% 55b/70g (1540g*) 75° 1.2-Hitlag

unknown substate 0x2bc (Uncharged Giant Hammer)
Frame 11-12: 22% 60b/70g (1820g*) 53° 1.5-Hitlag

unknown substate 0x2be (Aerial Uncharged Hammer Flip)
Frame 11-11: 15% 60b/70g (1330g*) 55°

unknown substate 0x2bf (Aerial Hammer Bash)
Frame 11-11: 4% 100f/20w 75°
Frame 25-25: 17% 55b/70g (1470g*) 75°
Frame 25-25: 15% 55b/70g (1330g*) 75°

unknown substate 0x2c2 (Fully Charged Hammer Flip)
Frame 11-11: 35% 60b/70g (2730g*) 361° 1.5-Hitlag Fire
Begins super armor on real frame 5
Ends super armor on real frame 14

unknown substate 0x2c4 (Fully Charged Giant Hammer)
Frame 11-13: 40% 60b/70g (3080g*) 361° 1.5-Hitlag Fire Unblockable

unknown substate 0x2c6 (Aerial Fully Charged Hammer Flip)
Frame 11-11: 27% 60b/70g (2170g*) 45° Fire


unknown substate 0x2c8 (Aerial Fully Charged Giant Hammer)
Frame 11-13: 32% 60b/70g (2520g*) 45° Fire Unblockable

unknown substate 0x2ce (Wave Cutter ?)
Frame 1- 2: 5% 100f/117w 85° 1.3-SDI Slash
Frame 1- 2: 5% 100f/117w 93° 1.3-SDI Slash
Frame 1- 2: 5% 100f/102w 90° 1.3-SDI Slash
Frame 1- 2: 5% 100f/102w 93° 1.3-SDI Slash
Frame 19-27: 2% 100f/100w 275° Slash
Frame 28-29: 2% 30b/180g (1080g*) 361° Slash


unknown substate 0x2d0 (Upper Cutter)
Frame 2- 3: 10% 45b/100g (1400g*) 60° Slash
Frame 4-11: 10% 45b/100g (1400g*) 70° Slash
Frame 12-13: 10% 45b/100g (1400g*) 80° Slash
Enables forward ledge grabs on real frame 4

unknown substate 0x2d7 (No clue)
Frame 0- 1: 5% 70b/100g (900g*) 361°

unknown substate 0x2ea (Stone)
Frame 0- 1: 14% 40b/40g (720g*) 0° Ground-Target-Only

unknown substate 0x2eb (Grounded Stone – Near the opponent)
Frame 0- 1: 6% 40b/40g (400g*) 0° Ground-Target-Only

unknown substate 0x2ee (Aerial Stone)
Frame 0- 1: 18% 65b/70g (1540g*) 70° 0.7-Hitlag


unknown substate 0x2ef (Grounded Stone)
Frame 0- 1: 12% 65b/80g (1280g*) 70° 0.6-Hitlag Bury Aerial-Target-Only
Frame 0- 1: 10% 50b/80g (1120g*) 70° 0.6-Hitlag Bury Ground-Target-Only

unknown substate 0x2f0 (Meteor Stone)
Frame 0-11: 12% 60b/100g (1600g*) 270° 0.7-Hitlag
Frame 12-21: 10% 50b/100g (1400g*) 270° 0.7-Hitlag
Frame 22-23: 8% 40b/100g (1200g*) 270° 0.7-Hitlag


unknown substate 0x2fa (Ultra Sword)
Frame 17-19: 3% 100f/30w 22° 0.0-Hitlag 0.0-SDI Stun Unblockable
Frame 54-57: 5% 70b/0g (0g*) 361° 0.5-Hitlag Slash Unblockable


unknown substate 0x2fb (Ultra Sword)
Frame 21-21: 2% 20b/100g (600g*) 35° 0.5-Hitlag Slash Unblockable
Frame 21-21: 2% 20b/100g (600g*) 145° 0.5-Hitlag Slash Unblockable
Frame 35-36: 2% 20b/100g (600g*) 35° 0.5-Hitlag Slash Unblockable
Frame 35-36: 2% 20b/100g (600g*) 145° 0.5-Hitlag Slash Unblockable
Frame 48-48: 2% 20b/100g (600g*) 35° 0.5-Hitlag Slash Unblockable
Frame 48-48: 2% 20b/100g (600g*) 145° 0.5-Hitlag Slash Unblockable
Frame 59-59: 2% 20b/100g (600g*) 35° 0.5-Hitlag Slash Unblockable
Frame 59-59: 2% 20b/100g (600g*) 145° 0.5-Hitlag Slash Unblockable
Frame 71-72: 2% 20b/100g (600g*) 35° 0.5-Hitlag Slash Unblockable
Frame 71-72: 2% 20b/100g (600g*) 145° 0.5-Hitlag Slash Unblockable
Frame 83-84: 2% 20b/100g (600g*) 35° 0.5-Hitlag Slash Unblockable
Frame 83-84: 2% 20b/100g (600g*) 145° 0.5-Hitlag Slash Unblockable
Frame 97-98: 2% 20b/100g (600g*) 35° 0.5-Hitlag Slash Unblockable
Frame 97-98: 2% 20b/100g (600g*) 145° 0.5-Hitlag Slash Unblockable
Frame 109-109: 2% 20b/100g (600g*) 35° 0.5-Hitlag Slash Unblockable
Frame 109-109: 2% 20b/100g (600g*) 145° 0.5-Hitlag Slash Unblockable
Frame 173-178: 16% 75b/100g (2000g*) 45° 0.5-Hitlag 1.3-SDI Slash Unblockable

 

[dragontamer]

@ Thinkaman

How were KO Percentages calculated? Dedede's FSmash surely KOs sooner than 655%, doesn't it? Pac-Man's Smashes also seem wrong to me.

I'm thinking that your script has automatically grabbed the wrong hitboxes. PacMan's FSmash and Dedede's FSmash have weak-hitboxes. I guess ~200% or so makes sense for PacMan's weak hitbox on the FSmash, but certainly not from the "strong" hitbox... which has a KO closer to 120% or so in my experience.

Ditto with PacMan's Nair, which is one of my typical KO moves. PacMan's Nair has a lingering hitbox that gets weaker the longer it is out... but I've definitly KOed with that move around 170% center stage.

-------------

And... for entirely selfish reasons... I'd also like to know what your formula was for taking damage, BKB, KBG (and whatever other variables) and turning them into KO%.

I'm going to assume that you used the Melee / Brawl Knockback formula. But a solid answer would be most appreciated. Thanks :-)

 

[Thinkaman]

def get_KO_percent(damage, cumulative_damage, base, growth, angle):
    target_distance_modifier = 1.0
    my_angle = angle
    if my_angle > 360:
        my_angle = 40
    elif my_angle > 180:
        my_angle = 360 - my_angle
        target_distance_modifier = target_distance_modifier / 0.8
    if my_angle > 90:
        my_angle = 180 - my_angle
    if my_angle < 15:
        my_angle = 30 - my_angle
    target_distance = 17900
    if my_angle >= 60:
        target_distance = 2.3204 * my_angle * my_angle - 409.35 * my_angle + 37360
    else:
        target_distance = 2.1303 * my_angle * my_angle - 35.851 * my_angle + 16013
    target_distance *= target_distance_modifier
    ko_value = ((target_distance - 100 * base)/(0.07 * growth) - 18) / (2 + damage)
    return ko_value - cumulative_damage

 

[dragontamer]

@ Thinkaman

Sorry to bother you once more, but... I have a question here.

target_distance = 2.3204 * my_angle * my_angle - 409.35 * my_angle + 37360

This line boggles me.

Personally, my methodology would have been to use the Pythagorian Theorem... or simulate things with right triangles. So it'd be simply sin(angle) for upwards KOs, and cos(angle) for sideways KOs.

I'm trying to figure out what formula you based this theory off of, but I don't recall any trig-formula with "angle^2", where angle is represented in degrees. I admit that my trig is rusty however... but I generally remember converting from degrees to radians before doing any such tricks like that. So I'd be interested in learning any tricks you're doing with raw degrees as well. Thanks :-)

 

[Thinkaman]

The two polynomials used are the best fit second-degree polynomial trendlines found in excel for distance required (expressed as knockback) to kill (starting position default miigunner) vertically and horizontally respectively.

Strict trigonometric functions will not work due to the way the game applies knockback force. A second-degree polynomial model is going to be the most accurate, and fits the collected data very tightly.

I'd show you my graphs, but they are really ugly.

Edit:
Also that inverse-sort bug for KO percentages in the chart has been fixed.

Edit:
Also the knockback formulas used is Brawl's, which I have found to be extraordinarily accurate. It might not be exactly correct, but the margin of error is extremely small if so.

 

[dragontamer]

The two polynomials used are the best fit second-degree polynomial trendlines found in excel for distance required to kill (starting position default miigunner) vertically and horizontally respectively.

Strict trigonometric functions will not work due to the way the game applies knockback force. A second-degree polynomial model is going to be the most accurate, and fits the collected data very tightly.

I'd show you my graphs, but they are really ugly.

Makes plenty of sense actually. I do have a preference for statistical approaches over pure theoretical ones... at least until the game code is reverse engineered and knockback code is fully explored.

I'll trust your data and keep the magic-formula.

 

[Big O]

How do you know the distance to the blastzone is 17900? Are you assuming Brawl distance values?

 

[Thinkaman]

How do you know the distance to the blastzone is 17900? Are you assuming Brawl distance values?

You'll note that I overwrite 17900 with the polynomial-generated value based on the angle.

17900 is just the approximate distance for a 361 (or 40) degree move.

I tested the kill %s on starting position default miigunner with a ton of different 361 degree moves. I compared the growth and base knockback values, and found the relative multiplier of the base (100) that gave a consistent distance (~17900) for all pairs of values.

 

[dragontamer]

Edit:
Also that inverse-sort bug for KO percentages in the chart has been fixed.

Lulz. King Dedede's FSmash % went from 600%+ to 63%.

I had a feeling it was something along those lines.

Edit:
Also the knockback formulas used is Brawl's, which I have found to be extraordinarily accurate. It might not be exactly correct, but the margin of error is extremely small if so.

I've also found the Brawl KB formula to be very accurate at representing this game. This was mostly a curiosity question. Thanks for clarification.

 

[Thinkaman]

Once you guys help me check this stuff and get more special moves in, I can easily export to excel/google docs for conditional formatting and easy sorting.

Edit: Fixed a special case typo where Mega Man f-smash was labeled as repeating jab.

Edit2: DK down-bs are probably a repeating hit subaction that follow a startup. Removed for now.

 

[Big O]

For clarification purposes, Did you have the Gunner set to stop or control with no inputs when testing the kill %'s? I'll probably get around to trying out some of this stuff later on and wanna make sure we are on the same page before I proceed.

 

[Thinkaman]

For clarification purposes, Did you have the Gunner set to stop or control with no inputs when testing the kill %'s? I'll probably get around to trying out some of this stuff later on and wanna make sure we are on the same page before I proceed.

On stop, because this was 3DS. I did all the kills multiple times to reduce the AI variance.

 

[Big O]

On stop, because this was 3DS. I did all the kills multiple times to reduce the AI variance.

Ah I see. 3DS stages have larger blastzones when compared to the Wii U version (at least for FD/Omega stages). Perhaps I will try reproducing what you have, but on Wii U stages and set to control with no inputs to minimize random influences.

 

[Thinkaman]

Ah I see. 3DS stages have smaller blastzones when compared to the Wii U version (at least for FD/Omega stages). Perhaps I will try reproducing what you have, but on Wii U stages and set to control with no inputs to minimize random influences.

The great thing is I could easily drop in a different formula, and regenerate all the data in seconds. Nothing is wasted.

 

[Tristan_win]

The KO chart is very confusing to me. I understand that each column is broken up by characters on the left and the moves used on the top but where did you get these numbers from?

 

[Thinkaman]

The KO chart is very confusing to me. I understand that each column is broken up by characters on the left and the moves used on the top but where did you get these numbers from?

Calculated from raw data, for default miigunner on starting position of 3DS FD.

Edit:
Fixed a bug that made dash attacks vanish.

Added a novelty Bonus Shield Damage Table. (Note: Unlike the damage chart this quick-and-dirty novelty does not tally multi-hit, only reports the single biggest source of bonus shield damage in the substate.)

For example, Palutena's multi-hit f-tilt and u-tilt actually do (+1) Shield damage on every hitbox, for (+3) or (+7) extra respectively. Her nair does (+8) total likewise.

Ness dash attack and fair similarly do (+3)/(+4) total as well.

Pound Blitz also does (+8) total.

As a reminder, aerial Sideways Pound does not do any extra shield damage, hits on frame 18, as is a dismal move.

Edit2:
Fixed a bug allocating one frame too few to multi-hit move division.

 

[dragontamer]

I think I'm seeing the theory of Palutena now. It seems like most of her staff attacks (tilt, aerials) are not only multi-hit and meaty... but also deal bonus shield damage.

In theory, that should make her a pretty decent shield-eating character. Theory of course...