Super Smash Bros. 64 Frame Data & Info

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#1

Thino

Smash Master
Joined
Apr 7, 2006
Messages
4,845
Location
Mountain View, CA
#2
Attacker Hit Lag-
This is the amount of time added to an attack when it hits a character. An attack which hits a character is slower than one which does not.
For example, Mario's jab is 19 frames long. If Mario jabs a character the 'attacker hit lag' value is added to the attack duration. In this case it is 19 + 4 [attack duration + attacker hit lag] From now on attacker hit lag may be referred to as AHL.
Is this what is commonly called Freeze Frames?
 

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#3
Maybe. But that term shouldn't be around. It might be relevant to SSBM somehow.

Edit: I mean the term itself doesn't give an accurate description. There are some frames where the attacker is 'frozen' but AHL =/= 'freeze frames'. I think its called freeze frames because it sounds cool. Excuse the pun.
 

KESHIKI

Smash Cadet
Joined
Jan 29, 2006
Messages
54
Location
Japan
#5
Just a suggestion.
It would probably be better if you can divide what you call attack duration into three parts

1. Frames between pressing button and occurence of hitbox (Start-up)
2. Frames between the occurence of the hitbox and dissapearance of the hitbox (Hitbox duration)
3. Frames between the disappearance of the hitbox and the point that the character can move (I have no idea how it's called in English)


The reason is because they give us so much more information about the use of moves.



e.g.
Let's compare Ness' U-tilt (NUT) and Kirby's U-tilt (KUT).

Their attack durations are 36F and 19F according to your data, which makes NUT look pretty bad. And their start-ups are the same (6F).
However, NUT has a much longer hitbox duration than KUT as you can understand intuitively.
This means that NUT works better in a typical situation like this.
Suppose there is Kirby trying to do D-air directly from above you, and you are trying to counter him.
then, NUT is much easier to counter than KUT since the longer the hitbox duration is, the more chance the move is going to hit.
(If you are insanely good at timing KUT against Kirby's D-air, nvm XD)



god. . . I need to work on my horrible English. . .
If someone can tell me how should those 3 parts be called, it would be greatly appreaciated...
 

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#7
Just a suggestion.
It would probably be better if you can divide what you call attack duration into three parts

1. Frames between pressing button and occurence of hitbox (Start-up)
2. Frames between the occurence of the hitbox and dissapearance of the hitbox (Hitbox duration)
3. Frames between the disappearance of the hitbox and the point that the character can move (I have no idea how it's called in English)


The reason is because they give us so much more information about the use of moves.

e.g.
Let's compare Ness' U-tilt (NUT) and Kirby's U-tilt (KUT).

Their attack durations are 36F and 19F according to your data, which makes NUT look pretty bad. And their start-ups are the same (6F).
However, NUT has a much longer hitbox duration than KUT as you can understand intuitively.
This means that NUT works better in a typical situation like this.
Suppose there is Kirby trying to do D-air directly from above you, and you are trying to counter him.
then, NUT is much easier to counter than KUT since the longer the hitbox duration is, the more chance the move is going to hit.
(If you are insanely good at timing KUT against Kirby's D-air, nvm XD)
1. Startup is already in another table. I will add this to the current table.

2. Hitbox duration is impossible to do sadly, I could only guess... It's possible in SSBM to view hitboxes using AR (debug mode) so its obvious when the hitbox appears and disappears. In SSB its not possible to view the hitbox. So it would only be my estimation, which could lead to errors.

3. Again I cannot do this as I cannot see the hitbox. I could only guess when the hitbox disappears. I could perform very complex tests but it would take an extraordinary long time, time which I don't have and still would not be 100% accurate.

"This means that NUT works better in a typical situation like this.
Suppose there is Kirby trying to do D-air directly from above you, and you are trying to counter him.
then, NUT is much easier to counter than KUT since the longer the hitbox duration is, the more chance the move is going to hit."


I think you are wrong about this. This has more to do with priority than how long the hitbox is out for. For example, if a low priority move has a very long hitbox duration, it wont matter because any other move has much more priority than it.

"Their attack durations are 36F and 19F according to your data, which makes NUT look pretty bad. And their start-ups are the same (6F). "
I don't think it makes NUT look bad. It justs shows the fact that it's slower than Kirby's up-tilt. It goes without saying that a 'slow' move does not mean bad move.

You are right that it will give more information about the moves. I just don't know how to measure hitbox duration as I cannot see hitboxes.
Thanks for the detailed post.
If you have some idea of how to measure hitbox duration then I'd like to hear it ^^
 

KESHIKI

Smash Cadet
Joined
Jan 29, 2006
Messages
54
Location
Japan
#8
1. Startup is already in another table. I will add this to the current table.

2. Hitbox duration is impossible to do sadly, I could only guess... It's possible in SSBM to view hitboxes using AR (debug mode) so its obvious when the hitbox appears and disappears. In SSB its not possible to view the hitbox. So it would only be my estimation, which could lead to errors.

3. Again I cannot do this as I cannot see the hitbox. I could only guess when the hitbox disappears. I could perform very complex tests but it would take an extraordinary long time, time which I don't have and still would not be 100% accurate.

"This means that NUT works better in a typical situation like this.
Suppose there is Kirby trying to do D-air directly from above you, and you are trying to counter him.
then, NUT is much easier to counter than KUT since the longer the hitbox duration is, the more chance the move is going to hit."


I think you are wrong about this. This has more to do with priority than how long the hitbox is out for. For example, if a low priority move has a very long hitbox duration, it wont matter because any other move has much more priority than it.

"Their attack durations are 36F and 19F according to your data, which makes NUT look pretty bad. And their start-ups are the same (6F). "
I don't think it makes NUT look bad. It justs shows the fact that it's slower than Kirby's up-tilt. It goes without saying that a 'slow' move does not mean bad move.

You are right that it will give more information about the moves. I just don't know how to measure hitbox duration as I cannot see hitboxes.
Thanks for the detailed post.
If you have some idea of how to measure hitbox duration then I'd like to hear it ^^

Ohh I see. Now I understand that measuring hitbox is very hard. I was wondering why no one does it.


"I think you are wrong about this. This has more to do with priority than how long the hitbox is out for. For example, if a low priority move has a very long hitbox duration, it wont matter because any other move has much more priority than it."

My bad that I didn't make it clear but, that is why I chose NUT and KUT
(both moves have a similar priority which is ridiculously strong, meaning that KUT and NUT can definitely beat Kirby's D-air if you get the timing)
Under this condition, longer hitbox duration makes players get the right timing much easier.


"I don't think it makes NUT look bad. It justs shows the fact that it's slower than Kirby's up-tilt. It goes without saying that a 'slow' move does not mean bad move."

yeah right, I guess I can't generalize like that.
But the fact that NUT is slower than KUT should mean something in practice.
For instance, someone missed KUT or NUT, and you try to grab him immediately.
it is hard to do it against KUT because you don't have enough time to go for a grab after you watched him missing it.
But it is relatively easier to do it against NUT because his ness is still in a process of U-tilting.
 

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#9
Yah yah I agree with everything you are saying but the data is meant to only show speed.

You are talking about moves which are "better" or "worse" etc which is not really what I'm trying to show. Since that is opinion and not fact.

But the fact that NUT is slower than KUT should mean something in practice.

As you say, in practice. My data is not trying to show what is better or worse. It is just giving raw values to characters attacks according to speed.
Of course, speed is only a small part of what makes up a good attack. I will try to find a way to get data such as hitbox duration and priority etc etc.
 

cmasterchoe

Smash Journeyman
Joined
Oct 31, 2006
Messages
238
Location
Silver Spring, MD
#10
again, ant-d continues to amaze with his data. Can't wait for the smash attack and priority data! I wonder if there really is a thing such as priority, or if its just one attack generating a hitbox faster than the other so it "seems" as if it always supersedes the other?? (or maybe thats just what priority is... :p)
 

Lord Kirby

Smash Journeyman
Joined
Nov 4, 2006
Messages
255
#12
why isnt there a "rating" on the smashes? oh and thino, nope.

fastest...
fsmash-kirby
usmash-pika
dsmash-mario, luigi, pika

slowest...
fsmash-falcon
usmash-samus
dsmash-dk

use this one for smash ratings!
 

moogle

Smash Ace
Joined
Mar 20, 2004
Messages
601
Location
Huntsville, AL
#15
Anyone know if there's a gameshark code to show hitboxes in SSB64?

SSBM has one with action replay. >_> It allows them to figure out things like this about Peach's upsmash:


Credit to RyokoYaksa for the picture. From this thread.
 

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#17
I've been viewing the memory addresses (essentially the rules that the game uses) and I've found the locations of some quite interesting things.

The game keeps a frame accurate count of the move you just did. So I no longer have to count manually. I can just do the move and view the maximum number of that address.

It's possible that there is also an address which shows when the hit box begins and when it ends. KESHIKI's dream could come true after all.
 

Thino

Smash Master
Joined
Apr 7, 2006
Messages
4,845
Location
Mountain View, CA
#18
It's possible that there is also an address which shows when the hit box begins and when it ends. KESHIKI's dream could come true after all.
you could do this with TAS .
what would be really interesting is seeing the location of those hitboxes , but I guess thats you were aactually talking about
 

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#19
I'm not talking about location. It's impossible with TAS, would take years.

I finally found the address which shows when a hitbox starts and ends. So, I can add all this into frame data now.

This is so I don't forget: Hitbox 0028B5E3 Hyrule, Training Mode. Duration 002888BF, Training Mode.
 

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#21
true or false. on or off. 1 or 0.

I messed about with this and made gameshark codes that can alter any moves hitbox or startup time so that it has more priority or even makes it harder for the opponent to DI.

No these codes wont be released to the public, kaillera would be more messed up =)
 

Thino

Smash Master
Joined
Apr 7, 2006
Messages
4,845
Location
Mountain View, CA
#26
**** that kirby fsmash ! people knew it was broken but not so much
so ness' usmash is as fast as pika's...plus the least lag at the end... hum I often looked down on it , should use it more often
 

Thino

Smash Master
Joined
Apr 7, 2006
Messages
4,845
Location
Mountain View, CA
#28
Good job on smash moves.
Ness dsmash is awesome... when the opponent is behind you, but in front of him is cool too when players try to shield it . they often end up letting it off too early
 
Joined
Apr 2, 2006
Messages
538
Location
East Canada
#30
This stuff is fairly cool, although obviously the locations of the attack hitboxes on each frame need to be taken into account before saying stuff like "X's usmash is better than Y's".

For instance, Mario's usmash hitbox just sort of "appears" above his head, whereas Fox's starts at his feet and has to travel upwards. So even if Fox's usmash hitbox started on frame 1 (obviously not the case >_<), maybe Mario's would be "faster" for hitting opponents directly above.

Also, is there a difference in how long it takes the various hitboxes to lose power? Like, most character's bairs or whatnot are much weaker when you hit with the last few frames of them. What's the system for that?
 

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#31
I'll be looking into power, reaction time, multiplier effect etc after I've done the basic frame data.

I know nothing about those things so far, as I haven't looked at the memory addresses which govern them.

Edit:

There's a value which shows the amount of % (damage) an attack will do. This is interesting as you can see the change when you do an attack which changes in %. Eg. The beginning of Fox's N-air compared to the end. Now that I've found this, it will be easier to find the 'hitbox countdown' if such a thing exists.

So to answer your question:
When I say 'Power' I mean percentage. Power and percentage are actually different, but I'll pretend they are the same here.

Its is actually the other way around. The FIRST few frames have a lot of power and the rest have less. So the majority of frames are not powerful.

I'll take Fox for example. When fox performs Nair, Fair or Bair the first 4 frames are the most powerful. On the 5th frame, a new value is allocated, which is of course lower. As you know, many other moves are likes this, not just aerials. (up tilt, up smash etc).
Each character is different.

All this will be added to frame data once the basics are complete.

Edit:

Jab:
http://spreadsheets.google.com/pub?key=p2VcZS6YDXFb7PVbX5CriRw

Aerial Neutral A:
http://spreadsheets.google.com/pub?key=p2VcZS6YDXFaK66m5CVACMw

Aerial Back A:
http://spreadsheets.google.com/pub?key=p2VcZS6YDXFZQwheuqUuHAA

Added 'strong frames'

Edit:

Some titles such as "hitbox ends on" and "% degrades on" will be hidden later. They are only included to do the necessary calculations for other values...

Edit:

Aerial Forward A-
http://spreadsheets.google.com/pub?key=p2VcZS6YDXFb6EWFdvgpCGw

Edit:

Taunt-
http://spreadsheets.google.com/pub?key=p2VcZS6YDXFYdDRSHw5O8Tg

Edit:

Roll-
http://spreadsheets.google.com/pub?key=p2VcZS6YDXFbBFD-PNXXpFw
 

Neowizard

Smash Rookie
Joined
Sep 4, 2007
Messages
2
#32
ant-d, I just had to tell you, and I mean HAD to tell you (that's the first reason I chose to register here), You're doing to great work! You deserve some more then a thank you, but I've got nothing better. So...Thanks a million!
 

ant-d

Smash Lord
Joined
Jun 3, 2005
Messages
1,314
Location
London, England
#34
Now that I have better ways of measuring the data I'm going to update most of this soon. And fix mistakes (HEX to DEC conversions mostly)

I'll add more useful information after the basics are done. Also I'll do data for (J) and (E) versions and somehow make an easy way to compare them with each other.

There is already frame data for the (J) version, but it's not done in the same way as mine, so I think it's better to have a standard.
 
Top Bottom