The Smash Lab: What is it?

[SamuraiPanda]

As some of you may have noticed, there is a new sub-board in the General Brawl Discussion called "The Smash Lab." That is a brand new area that only members, called "Smash Researchers" who have a cyan name such as my own right now, that I either invite or accept after they've applied via the group memberships tab in the User CP.

Many of you must be wondering what the Smash Lab is, so let me fill you in on what I'm hoping to do with it. Here are some of the goals I hope to fulfill through the Smash Lab:

- Making "official" names for ATs we've discovered
- Making "official" frame data we find for characters
- Discovering new techniques
- Learning new codes for hacking the game, as well as new projects for modding the game in general
- Gathering "official" numbers and data for a wide variety of things such as character weight or length of glide tosses.
- Learning more about the game wherever and whenever possible (i.e. researching reasons for why some DI work better than others, figuring out why some ATs work although we have no idea, etc.)

And more.


As you can see, I'm hoping to make The Smash Lab into the most knowledgeable place on the boards, and we fully intend on sharing the knowledge with all of you whenever we discover it. We'll try to figure out things like if Wario's Dthrow really is an infinite against some characters, or if there is some way to alter the properties of specific moves through hacking the game. Yes, The Smash Lab will include Brawl hacking as one of the the many different aspects of the game we'd like to study. As of right now our purpose is fairly undefined, but I hope, as the new leader of this group, to make The Smash Lab into something incredible. But of course, I'll need the help of dedicated Smashers like those here to make this possible.

 

[JacobWins]

Nifty stuff

Some of the boards already have some frame data posted.

 

[Boofy!]

im on the case =]

 

[Browny]

omg im on it

----

OK heres my video

http://tinypic.com/player.php?v=2res1sm&s=4


Simply put, I cannot find a single frame increase in DDD's standing throw, no matter whether it is staled or not, done from the infinite, or buffered. even the exact number of frames (in my video editing im counting roughly 65 for a single grab) its even the same, exactly 65 frames for every grab shown

at the end youll see me comparing the grab range for a standing grab, fully staled and fresh. No matter how many times i test, the range is exactly the same...

stale grab range
http://i35.tinypic.com/2czvyx.png
fresh grab range
http://i37.tinypic.com/nzj3td.png


ok moar stuff. I think the only thing that gets stale from the dthrow is the initial launch distance, but it gets stale in reverse, it goes furthers. heres 2 pictures, both shot on the very last possible frame before ddd can buffer a regrab. as you can see, the staled throw has a much more noticeable distance from mario than the fresh one

fresh
http://i36.tinypic.com/zurv5u.jpg
stale
http://i33.tinypic.com/10crq6u.jpg

I can confirm that DDD's buffered grab time does not change also, as does the dthrow animation length. measured from a whole range of initial points of the dthrow animation, it was always relatively the same frame time for the beginning of the regrab, stale and fresh
if its not that... then i give up lol


tl;dr (copied from my page 2 post)

The following things are all constant whether the grab or throw respectively has been staled.

range of DDD's standing grab
speed of his standing grab
frame at which ddd can first buffer a standing regrab after a dthrow
speed of ddd's buffered standing regrab after a dthrow is the same as his standing grab
range of ddd's buffered standing grab after a dthrow
hitstun frames after a dthrow
total knockback distance from a dthrow

the only thing that changes is the angle of knockback of DDD's victims. it changes progressively as the move stales, launching at increasingly higher angles as the dthrow becomes weaker. however the distance they slide once touching the ground exactly counters the extra distance they fly from the dthrow

 

[Blitzmidfielder]

I'll do my best.

 

[shadowtroop]

The damage stales. If thats correct.

 

[Vyse]

Urgh, want to participate.

But want to do Traineeship work.

MUST.JOHN.GAH

 

[~ Gheb ~]

I'm interested.

If you care ybout frame rate data you should invite JimboCav. He finds them out for all characters (He already did them for Wolf, Bowser, MK and apparently Yoshi).

Question: Do we have to send you the answer via PM?

 

[Blitzmidfielder]

I have my theory finished, I need your go-ahead on how/where to post it.

 

[Levitas]

I think I might have to do some testing today... Yay capture devices.

 

[Napilopez]

I realllyyyy really like the idea of this smash lab. I had hunch it was going in this direction, and I'm glad to see it is.

Having "Official" numbers much better than having several different people test out data and such on their own with different methods and stuff. Standards make communication of princiciples and ideas so much easier.

I also like the idea of figuring out why certain ATs work: it could help understand ATs in contect and perhaps even instigate the discoveries of new ones.

Imma give an example about my main: Why can Sonic cancel his SideB with a shield at any point before the charge is full? Why does he have to do it before a full charge, and why can it only be cancelled with a shield? What if there were a way to cancel into different moves?

Or Spinshot: How can sonic travel near running speed in the air instantly? Sonic has great airspeed, yet horrible air acceleration, so this is a glitch I'm interested in learning about.

And thats just with my char, I can't wait to see what you guys will find. The prospect of some of the most knowledgeable and clever smashers working on this is indeed an exciting one.

I wish I had my Wii in college, I would definitely like to be a part of this. I'm pretty smart and clever, with a big immagination!(All thanks to you Disney). Physics major ftw =P Heck, I feel being part of this could even aid me as some sort of reasoning experience, as I intend to be a research scientist.

Hopefully I could do some testing when I get to go home.

I like how you started it off with a puzzle from the get go SP, nice XD.

Since I can't do anything now, I guess some suggestions for those who can test.

Initiating Theorycrafft mode:

First of all, test this in training mode. How is the staling affected? Is there any diffence at all? Since moves don't stale in training mode, it would be interesting to see if the 5 regrab rule applied in training mode.

Now some things to note:

I seriously doubt the range of the grab is hampered by the multiple regrabs. It just seems so unlikely that the actual grab hitbox would become smaller after 5 regrabs. Still possible though

At first I though it had to do something wth percentages, but clearly after watching the video, the percentage does not affect the amount of regrabs that can be done.

Perhaps someone should investigate why the regrab works with DK and Bowser differently than the other chars... Why does it work with only those 5 chars anyways? It seems like an odd assorment of characters except for DK and Bowser who are the 2 largest Chars.

This is just simple observation, but from watching the video, the 6th grab did seem to come out a tiny bit slower than the other grabs, alhough that might just be my brain playing tricks on me.

Is it possible that moves actually stale in attack speed as well? Lol, knowing Sakurai... I wouldn't be surprised.

If it were something range/speed related, it would explain why its easier to regrab DK and bowser, as they are larger characters.

Ahh I really wanna be a part of thisss, I love theory craftt and experiment craft =P Curse my lack of wii in college!

 

[Blitzmidfielder]

Perhaps someone should investigate why the regrab works with DK and Bowser differently than the other chars... Why does it work with only those 5 chars anyways? It seems like an odd assorment of characters except for DK and Bowser who are the 2 largest Chars.

This is just simple observation, but from watching the video, the 6th grab did seem to come out a tiny bit slower than the other grabs, alhough that might just be my brain playing tricks on me.

If my theory's accepted, then I've already addressed those.

 

[~ Gheb ~]

@djbrowny

I thought the same. When you look at your video and mark the places where Mario drops the first time and the fifth time you'll see a notable difference. However, why does this not affect his normal CG then?
I did single dthrow against Mario and he always landed exactly at the same spot

 

[Browny]

imo...

the final distance mario ends up is the same, but there is a noticeable difference in his hitstun animation immediately after the dthrow. while he flies further from a stale dthrow, he skids to a stop quicker

 

[shadowtroop]

I could have sworn he flew in the air differently...

 

[Vyse]

Argh you guys are too good.
Why must I commit myself to real-world responsibility!
-_-

Either way, djBrowny's findings would convince me.

So what you're saying is that the character is knocked back quicker at first when dthrow becomes stale (why the regrab fails) but slows to a stop at the same distance quicker than before?

 

[Browny]

TBH i think its quite obvious to anyone whos tried it in slow-mo will notice marios different animation from a stale dthrow

however i think my super-slow frame data (which i didnt post... its 25x slower on my video editing program ) pretty much proves that the speed of his grab, range of standing grab, quickest frame he can buffer it and dthrow animation length are all constant with the only thing left being a change on marios side

@ vyse
thats the only explanation i can come up with at the moment

 

[~ Gheb ~]

So is it merely coincidence that the dthrow is always "overstaled" when doing it for the fifth time? I think your theory is spot on but it doesn't answer all the question. Let's assume your theory is correct...that means that the stale dthrow does have the same knockback but still doesn't?

 

[Browny]

the overall knockback is the same, just the angle at which it sends you seems to change over time

I can liken this to G&W's dthrow. not all characters have the same animation. while floaties like MK are left with a seriously tiny frame window to avoid a dsmash, most heavies can easily tech it and roll away in time with no trouble. obviously this isnt the same thing as this is taking into account characters fall speeds but its the only other aspect of the game i can think which changes a lot depending on which character is thrown.

actually Sonics usmash spike is another similar thing. for a certain % range on characters (it varies per one) it can either send you straight up, or is a powerful spike. DI by the enemy can only affect it at the limits of the range at where it ceases to spike them, however when the spike % range begins, it does it pretty much every time. Staling the usmash also has an effect on the % range

 

[Vyse]

To restate dj's theory:
On the fifth throw, it has the same knockback distance-wise, but the method that the thrown character takes to go from A to B is different once staled in such a way that the character slides out of regrab distance faster than Dedede can regrab, but comes to a stop at the same point.

What disproves that though?
*strokes chin*

 

[~ Gheb ~]

To restate dj's theory:
On the fifth throw, it has the same knockback distance-wise, but the method that the thrown character takes to go from A to B is different once staled in such a way that the character slides out of regrab distance faster than Dedede can regrab, but comes to a stop at the same point.

What disproves that though?
*strokes chin*

Actually the angle always changes but only at the fifth throw the angle is already too low for a regrab. As long as King Dedede jabs before he uses the dthrow for the fifth time in a row, the dthrow will send him in a angle where D3 can just make it.

This makes sense indeed.

Edit: I can actually confirm that this is almost certainly true. If you do a CG to Mario (a nromal one) you'll see that after a few dthrows Mario will get knocked back at a very sharp angle. The distance remains the same. This is why he can escape the infinite if dedede doesn't pummel but he still can be CG'd regularily. Good job dj.

 

[Napilopez]

Good Job djbrowny, I think you are a great candidate for the Smash Lab of goodyness!

But just for some personal clarification, this change in angle is caused by change in hitstun? The lack of hitstun in the staled move is causing the initial trajectory to change slightly? And this is why at the fifth grab it doesnt work?

 

[Browny]

i like where this thread is going :D

 

[Vyse]

Actually the angle always changes but only at the fifth throw the angle is already too low for a regrab.

I totally meant to say that the staling affects every throw but figured it was implied

It'd be good to figure out all the other little nuances though.

 

[Adapt]

****... I was gonna work on this, and it was already pretty much solved
I can't test this, but I can theorize the reason because I did some stale move research.

Knockback and Hitstun increase with damage. Hitstun has 2 parts, hitstun 1 where you cannot do anything, and hitstun 2 where you can airdodge or use air attacks but not use b-moves or jump.

This is the essence of combos. When damage increases, both hitstun and knockback increase. Hitstun does not appear to be linear (knockback is). Situational combos (only at certain percents) happen when hitstun increases enough that the move is fast enough to hit again before hitstun ends. Once knockback increases enough, you get hit too far, and you cannot connect the second hit because the character is out of range.

When a move stales, the only things that appear to change are the damage, knockback and hitstun of the move. All of these decrease. Damage can stale by almost 50% while knockback is not nearly that bad. Hitstun is extremely hard to measure, but it is related to knockback.

Can anyone confirm whether you can perform normal DI during hitstun 1? If you can't I can see why it's possible to escape. You are on the ground so you can't actually air attack or airdodge.

Since Dedede's dthrow has so little knockback growth, a fresh throw will always be in the range where hitstun is enough that a regrab is possible, but if you stale the throw, the hitstun decreases enough that Mario/Luigi/Samus can DI away from dedede. Since Bowser and DK are so **** big, they cannot DI far enough away from the regrab range during hitstun 2.



Also Panda:
When will you be accepting applications through the usercp?

 

[Browny]

1st line post #19

and what other nuances vyse? i cant think of anything left to prove? whatever it is i wont be able to do more testing till tomorrow, really need to do study lol

 

[Adapt]

I have to doubt that djbrowny.
I'm pretty sure all moves stale. Unfortunately I can't prove that until I go home. I'll let you know my tests in a few hours

Maybe D3's throw is an exception...

 

[Browny]

stale how? i can be 100% sure nothing related to grab range, speed of the grab, dthrow animation length and frames at which DDD can buffer a regrab, change if an attack is stale or not. And i didnt say the actual knockback doesnt stale, only the animation (therefore direction) of the knockback applied to DDD's victim is changed. I only posted a very short version of my actual video which goes into all that stuff, but its all constant so i didnt put it in.

A summary;

The following things are all constant whether the grab or throw respectively has been staled. these have all been confirmed with frame data

range of DDD's standing grab
speed of his standing grab
frame at which ddd can first buffer a standing regrab after a dthrow
speed of ddd's buffered standing regrab after a dthrow is the same as his standing grab
range of ddd's buffered standing grab after a dthrow
hitstun frames after a dthrow
total knockback distance from a dthrow

the only thing that changes is the angle of knockback of DDD's victims. it changes progressively as the move stales, launching at increasingly higher angles as the dthrow becomes weaker. however the distance they slide once touching the ground exactly counters the extra distance they fly from the dthrow

 

[Napilopez]

Can anyone test this in training mode, out of curiosity? Is the infinite possible on Mario Luigi and Samus on training mode without pummels?

 

[Adapt]

stale how? i can be 100% sure nothing related to grab range, speed of the grab, dthrow animation length and frames at which DDD can buffer a regrab, change if an attack is stale or not. And i didnt say the actual knockback doesnt stale, only the animation (therefore direction) of the knockback applied to DDD's victim is changed. I only posted a very short version of my actual video which goes into all that stuff, but its all constant so i didnt put it in.

Ok I agree with you except I don't know about the angle... I have never tested that. Everything I said has to do with the way the character moves after they get thrown

 

[kr3wman]

DDD's D-throw, when ti get stales, gets its hitstun reduced like any other move.

Since DDD smacks the people on the floor, the hitstun is applied after getting hit on it.

Basically, it does something like this.

Since, in the first five throws there's more hitstun and knockback, the characters will get sent at a relatively higher angle then when the move gets stale.

Something like this :

(DDD being on the right)


O-X Dedede (X) Grabs Mario(They're on the ground, they're not flying, )
___________________
X Dedede does the D-throw
_____________O


now, look at this more closely (Imagine a Zooming feature being applied in reverse so you see things from father away or something)

Fresh

0 - < - < - < - X
) -
))))) -
)))))))) -
))))))))))) -
_______________________________________________

See the distance? Mario is sent higher because of the trajectory of the Knockback after the throw


Stale



O < - < - < - < - < X
))) -
)))))))) -
))))))))))))))) -
))))))))))))))))))))))) -
_______________________________________________



It loses it's knockback ''pop-up'' distance when it gets stale.





Or something like that.

EDIT : and of course it didn't work properly. ****it.

EDIT2 : There ya go.

Basically, knockback gets reduced, angle is less steep, lower angle means more distance between Mario-DDD, DDD can't grab him back.

 

[Browny]

might wanna read the last 2 pages first, would save you some time lol

 

[kr3wman]

might wanna read the last 2 pages first, would save you some time lol

lol.

lol.

lol.

lol.

Shut up.

j/k j/k j/k.

Basically, as soon as I saw the Chaingrab infinite video I knew it was something like this. I'm amazed at the fact that people didn't know.

 

[OmegaXXII]

Smash Researcher heh, well then I'll get to this right away after I come back.

 

[Red Arremer]

Aw crud.

I would've loved to check that out, but I'm at work. -.-

 

[MuBa]

As we all know that Brawl records 9 moves and keeps on staling if the same attack is used repetitively.

For the infinites, we have 3 characters that have a staling situation where they will be able to get out of the infinite. But in order for the infinite to be maintained you have to do a grab attack so the D-Throw totally won't get stale.

The CPU will record the following actions:

1. Grab Punch
2.DownThrow
3. Grab Punch
4. DownThrow
5. Grab Punch
6. DownThrow
7. Grab Punch
8. DownThrow
9. Grab Punch

After the 9nth input the last hit will be deleted and replenish the knockback and hitstun according to the previous number of attacks recorded.

So after a certain number of consecutive throws (or any attack) the move loses the knockback and hitstun to maintain the infinite.

And that's why Dedede requires to do a grab punch before a DThrow: To prevent the Dthrow from become so stale and give just enough knockback and hitstun to maintain the infinite.

I'll get deeper into it later on.

 

[Famous]

It also stops Snake from dropping his grenade...

I have no input for D3's staleness as of yet.

 

[Niko_K]

As we all know that Brawl records 9 moves and keeps on staling if the same attack is used repetitively.

For the infinites, we have 3 characters that have a staling situation where they will be able to get out of the infinite. But in order for the infinite to be maintained you have to do a grab attack so the D-Throw totally won't get stale.

The CPU will record the following actions:

1. Grab Punch
2.DownThrow
3. Grab Punch
4. DownThrow
5. Grab Punch
6. DownThrow
7. Grab Punch
8. DownThrow
9. Grab Punch

After the 9nth input the last hit will be deleted and replenish the knockback and hitstun according to the previous number of attacks recorded.

So after a certain number of consecutive throws (or any attack) the move loses the knockback and hitstun to maintain the infinite.

And that's why Dedede requires to do a grab punch before a DThrow: To prevent the Dthrow from become so stale and give just enough knockback and hitstun to maintain the infinite.

I'll get deeper into it later on.

You're an ******* you know that?? As soon as I was about to press reply you beat me to it.

This is EXACTLY what happens.

 

[kr3wman]

You're an ******* you know that?? As soon as I was about to press reply you beat me to it.

This is EXACTLY what happens.

And it's exactly what Samurai said in his first post.

He understands how you want to keep it relatively fresh for the chaingrab to work. People do it so that the vertical knockback doesn't decrease too much and increases the distance between DDD and whoever he is grabbing/Dthrowing.

 

[SamuraiPanda]

I think DJ's theory is interesting, and is most definitely the strongest theory thus far. I'm still not convinced that is exactly what happens, though. This definitely needs more testing. What happens against other characters that can't be infinited when D3 uses his staled downthrow? Does this only affect D3 or all characters? Does this mean that it may be possible to somehow DI out of his normal CG if its staled?

There are tons of unanswered questions left with this, and don't forget to look at the big picture, too. This most likely isn't just an isolated case with D3, so whatever we figure out here could potentially be applied to other characters as well.