Official Frame/Hitbox Data Repository

[rPSIvysaur]

Yes, albeit I was also planning on reformatting a page and seeing how it goes for myself.

As to why every character has every field. I copy paste'd the format from one. Go ahead and get rid of any unneccsary rows or columns you need to.

 

[Amazing Ampharos]

Okay, I'm mostly done with G&W; my only problem is jabs. Before I get into the problem areas, let me discuss what I did do. The main thing I did that isn't kosher is remove the animation ending data and merge it with IASA, only including the data from Brawlbox if there wasn't an IASA. This has no real disadvantage to gameplay unless you want to know what cool poses you can look at when. I also properly included data for all of G&W's multi-hitting normals and durations, and I added special fields that will be nearly G&W exclusive for landing hitbox data. I removed a lot of rows and generally opted for more columns. I removed every column that didn't apply to G&W, including auto-cancel (after) which probably does apply to the majority of the cast.

Some things I just had to do in frame advance. I thought I had to do aerial landing lag in frame advance and actually did it, but then I realized I'm a doofus and knew this. Unfortunately, your guide on the front page doesn't seem to cover this. The actual way to find this data is to go into the attributes tab and check 0x140 through 0x150. I did this by hand and got values that agree with this and with the G&W boards historic data. The value in BrawlBox is a nonsense value and doesn't even correspond to any animation that I can tell. The only other thing is that I documented in frame advance how G&W moves during utilt and usmash, which affects how he can hit. I don't expect anyone else to go this far, but G&W's movements made it too easy, and I had to use frame advance to check other stuff anyway to make sure I was reading numbers correctly and had absolutely no one frame errors.

Jabs are a big issue. I don't know what you mean by jab loop and jab flow. Here is what I do know. G&W has a jab 1 and a rapid jab and no other jabs. If G&W does a jab1, he needs to press A one more time to begin a rapid jab. If he presses it on frame 6 or earlier, he'll do the rapid jab at the earliest possible moment. If he presses it on frame 7 or later, he'll do the rapid jab later than possible by as many frames as he waits. He will always do a rapid jab if he presses A before his jab 1 completes except if he presses it on the very last frame before the IASA in which case he will do another jab 1 immediately (but this is basically doing the input before his first free frame, which is how I'm assuming the game looks at it). None of the bits seem to match up to this, and he doesn't even have a bit 22 in any of his jabs. As per his rapid jab, it's just weird. He keeps hitting in an infinite loop with only RA-Bit[25] being set and two undocumented functions (01000000 and 01020000) seemingly having the potential to break it. I don't really know how it works, and my experience making Bbrawl ill prepared me for this (the only jab we really messed with frame-wise was Sheik's which I don't think was quite like this). I could take all the data I need from frame advance kinda, but it would be better if we had some good knowledge back here about how jabs actually work, especially everything involving rapid jabs. Does anyone have any knowledge that could help here?

Also a few values I notice. All three of G&W's smashes have values of 61 in Brawlbox for charge time which means they can be held for a full second (for some reason, Brawlbox's values tell you the frame you are free in the same way IASA works, not the total number of frames in the animation; I tested this in frame advance). I notice in the attributes that G&W has 3 frames of landing lag when he's not doing an aerial. Are these universal values? I didn't document either of them on that page since I assume the first one is and think the second one *probably* is. Information anyone?

In any case, I also included an *extensive* notes section on G&W's page, which I'd advise everyone to include on every page. Maybe you won't have that much to say, but that is a method for different people to communicate conventions and observations and data gathering methods to prepare for the possibility that someone is doing something wrong and it has to be corrected.

Anyway, I plan on doing other characters like I did G&W and was hoping my reformat of G&W would provide a good general guideline for other characters. Thoughts on my approach and solutions to the jab situation would really help me move forward.

Also, I corrected his name so if you are having trouble finding him, remember he is "Mr. Game & Watch".

 

[Veril]

I never used the smashlab data format. ;p

Mine is a whole hell of a lot better for including large amounts of information and nearly everything AA mentioned is somewhere in the data I've already posted on google docs (you know, the other file that actually contains the majority of the data uploaded). "Makes note to update"

I know a lot about jabs (having obsessed over them for a while), and even setup an excel sheet to calculate the block advantage permutations for all of them AA we should talk, I'm free tomorrow. I herd u like math

 

[rPSIvysaur]

Is there a time where both of you can get on AIM? I'd like to do a group chat with both of you for formatting the spreadsheet.

 

[Veril]

I'm down for anytime in the next 2 hours or tomorrow if I'm given advance notice.

 

[phi1ny3]

What are the subaction names for the banana slip/the options that come out? Are they under "slip"? They seem to be more like trip frames b/c of the "slipdash" subaction.

 

[rPSIvysaur]

Alright everyone, I set up a new spreadsheet: https://spreadsheets1.google.com/ccc?key=tu82p13WaUgYL-_LZrWcPOw&hl=en#gid=1 (IDK if link will work, but it's in the Smash Lab folder in Google Docs)

When you start a character, just copy the format spreadsheet and rename it and alphabetically place it.

I'll probably be revising the OP for correctness.

 

[KayLo!]

What's IBC? Something to do with super armor, or there just not a column for that?

 

[rPSIvysaur]

IBC is invincible bone collisions. Don't worry about it, it's hardly on any moves (unless your name is Bowser, lol) but you should realize it when you see it.

Don't forget to reference my PSA Frames vs. Ingame Frames. But make sure you record it as in game frames (f1, f2, etc.) rather than PSA inputs (I0, I1, etc.)

 

[KayLo!]

Okay. I'll start filling in some of Zelda's stuff a little later tonight.

 

[Luxor]

Back to work :/

I was confused by what a few of the lines on Fox's U/Dsmash meant at first, thanks for clarifying that it's IBC.

 

[Amazing Ampharos]

I just did Charizard, another character who doesn't quite follow the guide for jabs. I think he just uses a differently numbered bit for something, but I seriously am thinking that for the sake of doing things right that I'm going to have to do all jabs by hand in frame advance. So far G&W and Charizard are going to require that, and since Ivysaur has a rapid jab, I know he's going to be "fun" too.

Charge frame should be added to the format sheet; I realized while working on Charizard that we forgot it (sorry for not bringing it up when we talked about it rPSI; I didn't notice then). In any case, I have that data appropriately gathered for Charizard, though I did work under the assumption that the numeric patterns I established while working on G&W were also valid on Charizard (it would be insanity if it were otherwise so it's all good).

There's also another issue Charizard raises with several moves (fsmash, fair, bair). He has hitboxes that don't disappear but do have new hitboxes made on top of them that cover them completely and have higher hitbox priority. To explain what this means, I need to explain how the whole hitbox ID system works. So when a character gets hit by more than one hitbox at a time, whichever hitbox has the higher ID is the one that "counts". For instance, if Marth's fair has hitbox ID 0 for the base of the sword and 1 for the tipper and he hits Bowser very directly so both parts touch Bowser at the same time, the tipper is the one that counts because it has a higher hitbox priority. Once any hitbox hits, hitboxes from the same actor (this means counting direct hitboxes from characters and hitboxes on articles separately) cannot harm the victim until a new sub-action is begun or a terminate collisions occurs which will then allow new hitboxes to hit. If the hitbox ID is in the next group up in terms of 256 (this makes sense looking at an example like any of those three moves on Charizard), the hitboxes can hit even if the previous ones hit on previous frames, but they aren't allowed to hit on the same frame the way hitboxes from different actors could hit the same victim at the same time (pretty common in FFA and teams) but seem to flag as off if they try to. The issue on these three moves on Charizard, and doubtless several other characters, is that they function as multi-hit moves by making new hitboxes with ID numbers in the next group of 256 without deleting the previous hitboxes. These overlap the other ones mostly (completely? That's hard to establish) and effectively make it so they can't hit, but the other hitboxes still exist until the Terminate Collisions at the end of the move and could maybe hit if there's not 100% perfect overlap. I also just kinda made educated guesses at how this 256 case is handled based off how the game works, and if you hacked it up and made unusual cases, maybe I'm wrong about some subtle details. The main point here is that they're multi-hit moves for which there are no frames at all on which they aren't hitting in-between the hits (much unlike with, say, G&W's multi-hit moves which leave small windows). This makes recording the duration unusual since each set of hits has progressively "less" duration though, naturally, you aren't going to be able to hit with the earlier hits once the later ones come out to play.

I also changed how I was recording IBC to be more communicative and hopefully generally superior. I'll fix G&W's usmash to the new format before going to bed. On that note, bones are named as hexidemical values; I hope no enterprising person is converting to decimal on those. G&W's usmash is bone invincible on bone 2B for a while, but Charizard's bones all happen to not involve letters (but I still recorded as hex values). I just hope everyone is clear on that.

 

[Luxor]

Thanks for clarifying how that works, AA. I was a bit confused on that myself.

Will update Fox soonish.

 

[KayLo!]

So..... just to make sure I'm doing this right/like everyone else for multihit moves.....

Under "hitframes," we write only the first frame each hitbox is out, and under "duration," we write how long each hitbox lasts, right? So for Zelda's jab, I'd put 11, 13, 15 under hitframes and 1, 1, 1 under duration?

EDIT: Meh, I went ahead and filled in some stuff..... if anything looks wrong, lemme know.

 

[rPSIvysaur]

So..... just to make sure I'm doing this right/like everyone else for multihit moves.....

Under "hitframes," we write only the first frame each hitbox is out, and under "duration," we write how long each hitbox lasts, right? So for Zelda's jab, I'd put 11, 13, 15 under hitframes and 1, 1, 1 under duration?

EDIT: Meh, I went ahead and filled in some stuff..... if anything looks wrong, lemme know.

Right on the nose KayLo.

 

[KayLo!]

When/where're we doing data for specials?

 

[Mr. Escalator]

What characters are needing data still/not claimed?
I still need to edit Ness to AA's format, but that shouldn't take long.

 

[rPSIvysaur]

OP has been updated. *phew*

 

[Pikaville]

I guess in that case I will take Yoshi or MK

 

[rPSIvysaur]

https://spreadsheets4.google.com/ccc?key=tr1fWAXWPMLDSY66j8Eymaw&hl=en#gid=0

Alright, that's the sign-up sheet. Please fill out your name for the character. Do not mess with my highlighting

 

[EverAlert]

What do the colors mean?

Also sif AA is doing ICs, I totes claimed them first page, in fact doing them right now. D:

 

[rPSIvysaur]

What do the colors mean?

Also sif AA is doing ICs, I totes claimed them first page, in fact doing them right now. D:

You should talk with AA about ICs. Colors are for me, it helps me track the progress.

 

[EverAlert]

kinda troublesome, but ok then. :/

 

[Amazing Ampharos]

e_alert, the main thing about ICs you need to know is that Nana is completely separate from Popo in terms of her .pac logic and could potentially have completely different frame data. You'll want to make a second character page for her. If you're already doing it, I'm going to assume you are familiar enough with your .pac reader of choice. Open SA 2 makes finding Nana's offset fairly easy otherwise; if you're using PSA, you need to know how to navigate subroutines manually (I almost always use PSA so it's easy for me to see ICs as a really, really big pain). Luckily, since we aren't doing special moves, you don't have to handle the cases of the united moves, and if you just model them as two completely different characters, you should be mostly fine as I think about it.

 

[EverAlert]

This'll be kinda redundant given the VM exchange, but to show here I'm following along:

Sure thing, I got it covered! :3

 

[Mr. Escalator]

I've grabbed Dedede in the sign up sheet. It's left white, just so rPSI can have the honors of highlighting it red.

Should start working on it later today.

EDIT: Mostly done with Dedede. Left some things blank that I'll get to work on when people decide what we're doing with jabs/when I figure out the rest. I started porting Ness over, but I've only started. I will finish later.

 

[Amazing Ampharos]

So, here's what I've figured about about jabs so far...

Bit 16 is set for the jab flow as we've defined it; it leads to the next jab. Bit 17 is a jab loop bit that goes to the next jab. Bit 22 is a jab loop bit that does jab1 again (even if the jab it is done from isn't jab 1, such as Squirtle holding A upon whiffing jab2). Bit 20 is somehow involved in leading into rapid jabs. If you choose to press A a second time instead of holding it, you can do jabs on whiff from the jab flow point, not just the jab loop point. Pressing A too early though won't buffer your next jab, but you can press it that "one frame" early in order to get it with perfect timing. Pressing A a second time always goes to the next jab, never repeats the first jab.

This should allow very thorough documentation of all non-rapid jabs (that is, all jabs in no way related to rapid jabs). Rapid jabs are odd, and I still don't know why Ivysaur is absolutely compelled to do rapid jab after jab2...

 

[A2ZOMG]

This sounds very interesting. Can you somehow explain how Falcon's Jabs work? I know Falcon is able to loop Jab3 into Jab1 and it actually combos (at specific percents since it's a growth rate based hit). It involves holding the A button until Jab3, releasing, and then pressing the A button again. Or you can do two Jabs, delay, and hold the A button again.

 

[Luxor]

So it's more or less gentleman-->Jab1?

 

[Mr. Escalator]

I've finished Ness/Dedede/Ike so far. Well, I've finished all but Jab loops/flows/rapids (which haven't be agreed upon yet how we are gonna go about them), as well as the last three columns on each.

I found an interesting thing, which I've already told AA/rPSI. Ness' Fsmash is the only smash in the game that charges for more/less than 60 frames. It charges for 59 frames. Check brawlbox yourself! I also checked frame by frame, and it checked out.

Also, I really want to format everyone's page like mine, but I won't~

 

[Amazing Ampharos]

I imagine Falcon's jab3 works like Squirtle's jab2. It has bit 22 set which causes the next jab if you hold A to be jab1. Like I said, bit 22 isn't unique to jab1s (though it's most common on them). I haven't looked at it in specific though.

Sorry for being lazy about figuring out rapid jabs; they're a pain, and I'm kinda sick of this business for today. Tomorrow if I get around to it I'll work that out for real and then maybe do DK or something (I've finished with G&W/Squirtle/Ivysaur/Charizard outside of jab business).

EDIT: I was bored so I did DK. That puts me at 5 done other than jab business which I still intend to look at more tomorrow.

 

[Veril]

RABit 16 = if hit and button held OR if button press goto next jab in sequence
RABit 17 = if button held go to next jab in sequence (doesn't require hit or release and press to transition)
RABit 20 = as RA16 except transitions into the rapid jab
RABit 22 = if button press goto jabx1

that's how they work afaik


Not only did I already do the jab analysis but iirc I gave it to rPSI!

rPSI we should talk...... I guess tomorrow night cause that's the next time I'll be at all free. Microbiology and trying to get another jab are keeping me maaad busy.

 

[rPSIvysaur]

Actually you didn't Veril. Albeit, I probably would have followed the variables in Fighter.pac soon, when I get a large amount of free time. That thing is gdlk.

 

[hotgarbage]

Wheeee I've finally found some time to contribute.

I've signed up for Peach; it'll be a good way for me to transition into THE NEW AGE of frame data gettin', as I can cross-check it with my old Peach frame data thread (and fix up some inevitable hitbox duration errors while I'm at it).

Yap

Spoiler

OAS2 is sexy

 

[A2ZOMG]

I'm thinking that the reason why Falcon's Jab3 is able to cancel back into Jab1 is actually related to Bit 16

There is no Bit 22 on Falcon's Jab3. Instead there is a Bit 16 placed after the Bit 20 that is valid starting frame 12.

And I don't quite understand if this is related to what I'm talking about, but after a 32 frame Asynchronous timer, Falcon's Jab3 is interruptable. I wonder if this is why delaying Falcon's Jab sequence is necessary for this technique.

I think I'll test this extensively to see if there is any correlation with what I'm saying.

Plus Pit's Jabs don't make sense to me.

 

[Luxor]

What conventions are we using for damage? Only listing the strongest hitbox/last one for multihits? No data for the middle or weak hits? Also, for autocancel frames are we listing inclusively or exclusively?
Oh, and A2Z, if it's interruptible after 32 frames you can do anything after that, including Jab1. You're just waiting for the current jab flow to end and starting a new one if you do it that late.

Oh, and I looked at Pit's. With Veril's bit data right ^ there, it's a breeze, mostly. Basically his jab can go jab1-->jab2-->rapidjab OR jab1-->jab2-->jab3, and there's some extra details about whether you just hold A on whiff or whether it hits or you mash A again.

Other other edit: finished reentering Fox. It's faster just to reread it all from OSA again =/ all my other work was a waste, apart from teaching me what to do in the first place.

 

[A2ZOMG]

New little things I tested and confirmed on Falcon's stale Jabs.

Releasing the A button for 21 frames after Jab3 causes you to cancel into Jab1 if you press the A button again

Releasing the A button for 18 frames after Jab2 and pressing the A button again here causes Falcon to do a Jab3 that is immediately canceled into Jab1 (time between hits is 15 frames, Jabs were stale after testing)

EDIT: Hitlag (which is damage dependent right?) has nothing to do with this.

 

[Luxor]

Jab2 has IASA after 18 frames. I would expect it to just Jab1, then, but because the bit is still set? IDK.

If you could describe what you saw as one of Veril's input maps it would be much easier for us to grasp what's going on. For example, I'm not sure if you mean 21 frames after input or after hit or what. Input maps have no such ambiguity.

 

[A2ZOMG]

To clarify, I held the A button until the Jabs actually connected. So Jab3 starts up in 6 frames. I would be holding A until the 6th (the first active) frame and then release the A button for said amount of frames.

In that specific situation, holding the A button til Jab 3, waiting 21 frames after impact, and then pressing the A button again resulted in a Jab that connected 4 frames after input.

Another odd input sequence that works:

After first active frame of Jab2, release A button for the next 13 frames, activate Jab3, release A button for 5 frames after the first active frame of Jab3, press A before proceeding to the 6th frame to initiate Jab1 (happens in 4 frames after pressing A)

 

[rPSIvysaur]

*COUGH* make a project thread about jab links for the public */COUGH*