The Smash Lab: What is it? + A contest to be the first Smash Researcher ever!

This is identical to the sticky in the General Brawl Discussion, so please discuss this over there: http://www.smashboards.com/showthread.php?t=205203



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.


Also, I'd like to announce a contest to become the first ever Smash Researcher! That's right, anybody here can become the first member... but that is if you can pass this test. So here we go:

King Dedede has a standing infinite downthrow on 5 characters in the game, but for 3 of those characters (Samus, Luigi, and Mario) his downthrow stales for some reason. You have to throw in grab punches in order to keep it unstale, and continue the infinite (see this video for more details on Dedede's infinite throws: http://www.youtube.com/watch?v=23UyrHXK36A ). So the challenge I'm setting forth for you all is to figure out how Dedede's downthrow stales when he does the infinite grab. What becomes stale? The speed of the regrab? The distance it sends them? The range of the grab? As far as I know, this question remains unanswered to this day, so even I have no idea how this works.

The first person to correctly discover exactly what stales in Dedede's downthrow will become the first accepted member of The Smash Lab! And after this is discovered, I will begin to accept/invite other members who I feel would work quite well as the first wave of Smash Researchers.

First post yayuihz.

Its the distance of which he sends them after the D throw.
The D throw sends them a farther distance away from DDD, the opposite of when other types of moves stale.

Grab punches refresh the D throw and return it to the original distance that the opponent is thrown.

The range and speed of a move are unaffected is never affected by moves becoming stale
Priority is only affected when a move follows the 10% rule. i.e. Ftilts,Smashes, grounded tornado.

This was found some time ago, I am surprised people forgot about it.

^If that's true, then it's pretty clear that Dedede's Dthrow was intended to be a chain throw. Why else would it act in the exact opposite way as other moves in such a way that it just *happens* to make the chain throw harder.

I think I have an answer to why Dedede's Dthrow will stop infiniting on those 3 specific characters (Samus, Luigi, Mario) after 5 throws. I think the characters actually travel a bit farther/are given more knockback than usual after the first 5 throws. It reminds me a bit of Falco's Dthrow; It traps most characters for about 5 throws until after a certain % threshold the character is given more knockback and can escape.

Falco's dthrow is different from Dedede's however in that his Dthrow will act more like a normal throw past 40% or so, because it's knockback will increase noticeably when the opponent has more damage where as Dedede's knockback stays virtually the same irregardless of %.

Edit: Shadowlink beat me to it (procrastinated posting this even though I had it typed lol.), but yeah it's the distance they are sent. It increases for some reason after the first 5 throws.

1.) you're all wrong

2-) This is actually the wrong thread

forget that part. Its so far only DDD' throw. Made an error when testing.
In anycase it does have to do with staling of DDD's Dthrow.

Or

it may be that DDD's D throw has very little growth in terms of knockback.
Combined with staling this could be the reason. *shrug*

DDDs D-throw can do it because Captain Falcon said so.


Seriously tho I have no idea but it probably has something to do with the knockback. And yes I did just look at the posts above mine to find that out so you should make them send the answers in pms or people can just copy it.

This is a really good idea tho this smash lab thingy.

Epic Idea.

Read the link on general brawl boards shadowlink. the distance the dthrow sends them is the exact same, but the path the enemy takes to get to that location changes. just imagine the direction the enemy takes follows a 1/square root function for a fresh dthrow, while they follow a linear path on a stale dthrow to end up at the exact same location at the same frame where they can first react out of hitstun... if that makes any sense lol

I earlier had thought that may be the case but I ruled it out earlier cause I figured the trajectory wouldn't change. The angle at which they are sent should remain the same if there is no DI being factored.
Eh guess the knockback changes trajectory without the DI.

meh *shrug*

I want a cookie.

In the future, things will get a little more technical in this thread, so I'll keep my eye on this.

Unless someone proves that the regrab starts taking longer (which doesn't make sense to me), or even that grab range of Dedede deminishes (but grabs themselves don't stale only throws) the only answer I can think of is an increase in the initial speed they are sent away at (and with that probably a slight increase in the distance traveled). Which would mean that Dedede Dthrow does indeed stale the other way around.

I'm pretty sure I've figured it out.

Decay seems to extend the IASA Frames of moves.

I originally brought up this hypothesis when I was compiling a list for Marths FThrow/DThrow to FSmash combos on each character. While testing, I started to notice that character weight had an effect on how long it takes for a character to throw another, and how quickly they can react after the throw was executed. If I were to perform a FThrow to FSmash on Jiggs, I would never tip because Marth would react too fast for Jiggs to be sent far enough for the proper spacing to tip. If I threw Bowser, it would noticeably be much longer during the process of the throw, as well as how long it took after the throw to react.

That led to my assumption that certain aspects of Brawl can have different effects on the frame rates on attacks. I went and tested to make sure that Dedede's DThrow did not extend or shorten a characters distance after throw after decay set in, as well that damage percent had no effect. The only thing that made sense was the ability to react after the throw, and it made me think back to Marths throw with heavier characters versus lighter characters.

Anyway. For this test, I used Bowser and King Dedede.

Grab Bowser with King Dedede when moves are fresh, and do a DThrow. Upon the DThrow, spam Standard A. If you notice, Bowser will get hit prior to him landing fairly easily. Now, if you were to DThrow him, let him settle his landing after the throw, and then use Standard A, you would still hit him. This displays that Bowser is in proper range to be hit with the Standard A regardless of percent.

Now, after decaying DThrow, try the same attempt. You will notice that King Dedede cannot hit Bowser immediately after the DThrow, regardless if you buffer the attack or not. Because there is no error involved for Shield, we can conclude that it is not mistiming. We know Bowser goes the same distance each throw regardless of percent. Bowsers animation also does not change during the throw regardless of percent or decay.

This can only indicate that decay is effecting King Dedede's ability to react after a certain move has been executed. I assume this holds true for all moves and characters, although I do not have frame rate equipment to prove such theories.

Try for yourselves. I am 99% positive this is the reason. It's decays effect on IASA Frames.

Also, my assumption on why it is so tight to do a regrab with Samus or a Jab follow up with Bowser is both a combination of there animations, the size of their hurt box, and because they are both heavy characters. During DThrow without decay, Dedede has just enough time to perform these moves, despite the effect of character weight on the move. However, after decay, it is just enough to extend the IASA Frame rate to a point where Dedede simply cannot react in time.

I'm far too lazy to be a researcher.

Sliq said:

I'm far too lazy to be a researcher.

Click to expand...

Haha, me too. This is an awesome idea though, I look forward to the findings.

great idea man just when i decide to start looking at the game differently and talk about frame data// look at WHY things are they way they act. i need all the info i can get.

will people be able to look at your findings in the smash lab board or just invited pple and ull make a threat about it elsewhere

Sliq said:

I'm far too lazy to be a researcher.

Click to expand...

Dude, not lazier than me. If I was the first Smash researcher, I would just ask other people to do my work, and then not even take the time to be evil and take credit for their work lol.

I wanna be a researcher just so i name the ATs hahaha.

fellow researcher: dude, that is an awesome technique you discovered!

me: yes i know. im calling it joeysmashing.

fellow researcher: what? but it makes no sense with the technique at all?

me: shut up its gonna be called joeysmashing.

we cannot reject Ho due to lack of evidence.

I thought M2K was the first Smash Researcher?

Ulevo said:

I'm pretty sure I've figured it out.

Decay seems to extend the IASA Frames of moves.

I originally brought up this hypothesis when I was compiling a list for Marths FThrow/DThrow to FSmash combos on each character. While testing, I started to notice that character weight had an effect on how long it takes for a character to throw another, and how quickly they can react after the throw was executed. If I were to perform a FThrow to FSmash on Jiggs, I would never tip because Marth would react too fast for Jiggs to be sent far enough for the proper spacing to tip. If I threw Bowser, it would noticeably be much longer during the process of the throw, as well as how long it took after the throw to react.

That led to my assumption that certain aspects of Brawl can have different effects on the frame rates on attacks. I went and tested to make sure that Dedede's DThrow did not extend or shorten a characters distance after throw after decay set in, as well that damage percent had no effect. The only thing that made sense was the ability to react after the throw, and it made me think back to Marths throw with heavier characters versus lighter characters.

Anyway. For this test, I used Bowser and King Dedede.

Grab Bowser with King Dedede when moves are fresh, and do a DThrow. Upon the DThrow, spam Standard A. If you notice, Bowser will get hit prior to him landing fairly easily. Now, if you were to DThrow him, let him settle his landing after the throw, and then use Standard A, you would still hit him. This displays that Bowser is in proper range to be hit with the Standard A regardless of percent.

Now, after decaying DThrow, try the same attempt. You will notice that King Dedede cannot hit Bowser immediately after the DThrow, regardless if you buffer the attack or not. Because there is no error involved for Shield, we can conclude that it is not mistiming. We know Bowser goes the same distance each throw regardless of percent. Bowsers animation also does not change during the throw regardless of percent or decay.

This can only indicate that decay is effecting King Dedede's ability to react after a certain move has been executed. I assume this holds true for all moves and characters, although I do not have frame rate equipment to prove such theories.

Try for yourselves. I am 99% positive this is the reason. It's decays effect on IASA Frames.

Also, my assumption on why it is so tight to do a regrab with Samus or a Jab follow up with Bowser is both a combination of there animations, the size of their hurt box, and because they are both heavy characters. During DThrow without decay, Dedede has just enough time to perform these moves, despite the effect of character weight on the move. However, after decay, it is just enough to extend the IASA Frame rate to a point where Dedede simply cannot react in time.

Click to expand...

Well, I did some testing on this, and either this is a case for only D3's grab or it is something else. I used MKs and Marth's dtilt and counted how many attacks it took before they hit a lucas standing on the opposite side of smasville. I then did the same, but with sandbags and characters standing in the middle of the stage. According to your theory it would take Marth/MK more dtilts to get to the other side because their IASA frames wouldn't begin until later, and thus they wouldn't move forward as far (Please correct me if this part is wrong) but they took just as many dtilts to get across, and know this wasn't in training (although I did do a trial in training to make sure the sandbags didn't effect the distance of their dtilts, which it didn't)

This is why I believe it is something different, and I think it is that the change doesn't happen to the cg victims horizontal velocity, but their vertical velocity. So the character is going the same speed horizontaly, but reach the ground quicker and can therefore escape the grab.

I am not too sure though...

JesiahTEG said:

First post yayuihz.

Click to expand...

Liar, that post was mine!

Big Red said:

Well, I did some testing on this, and either this is a case for only D3's grab or it is something else. I used MKs and Marth's dtilt and counted how many attacks it took before they hit a lucas standing on the opposite side of smasville. I then did the same, but with sandbags and characters standing in the middle of the stage. According to your theory it would take Marth/MK more dtilts to get to the other side because their IASA frames wouldn't begin until later, and thus they wouldn't move forward as far (Please correct me if this part is wrong) but they took just as many dtilts to get across, and know this wasn't in training (although I did do a trial in training to make sure the sandbags didn't effect the distance of their dtilts, which it didn't)

This is why I believe it is something different, and I think it is that the change doesn't happen to the cg victims horizontal velocity, but their vertical velocity. So the character is going the same speed horizontaly, but reach the ground quicker and can therefore escape the grab.

I am not too sure though...

Click to expand...

Clever method of testing, but you wound up counting not how much time is taken by a dtilt but how far the dtilt goes. If the IASA frame is being delayed, it should take measurably longer to do the dtilt spam - you therefore want to time how long it takes to traverse the stage. This becomes difficult, though, because you need to make sure the same number of hits are applied in each case of the test. A bit of messing around with variables yields the following possible test:

Set damage ratio very low
Use someone with a good short range juggling Utilt (Utilt is the easiest grounded move to buffer, apart from AAA which is useless for this testing)
Hit DDD or Bowser upward 10 times from 0%, timing how long it takes from the start until the last hit is scored.
Kill the dummy to reset health, repeat the experiment
Save replay, do the timing from there
See if there is a difference in time large enough to attribute to something other than human error.


Also, even if we follow the logic of the original proposal we're going on, it is not clear whether it is an increase in IASA time or in the attacker's hitstun that is being increased. Repeat the same experiment without actually hitting the dummy; if the result is still positive, then IASA is being delayed, if not and the previous experiment was positive, then it is hitstun; if neither is positive, then we need to return to the drawing board.

Crow! said:

Clever method of testing, but you wound up counting not how much time is taken by a dtilt but how far the dtilt goes.

Click to expand...

I figured that if the IASA frames came out later, than the dtilts wouldn't have moved as much distance. Since Marth/MK slide forward, then backward, so if you couldn't interrupt until later, he would have been further along in the backward slide motion. I am not sure if this would be true though, so I would need to look into that a bit more.

I also like your method, though, and I will see if I can do some testing with it some time tomorrow.

Crow! said:

Clever method of testing, but you wound up counting not how much time is taken by a dtilt but how far the dtilt goes. If the IASA frame is being delayed, it should take measurably longer to do the dtilt spam - you therefore want to time how long it takes to traverse the stage. This becomes difficult, though, because you need to make sure the same number of hits are applied in each case of the test. A bit of messing around with variables yields the following possible test:

Set damage ratio very low
Use someone with a good short range juggling Utilt (Utilt is the easiest grounded move to buffer, apart from AAA which is useless for this testing)
Hit DDD or Bowser upward 10 times from 0%, timing how long it takes from the start until the last hit is scored.
Kill the dummy to reset health, repeat the experiment
Save replay, do the timing from there
See if there is a difference in time large enough to attribute to something other than human error.


Also, even if we follow the logic of the original proposal we're going on, it is not clear whether it is an increase in IASA time or in the attacker's hitstun that is being increased. Repeat the same experiment without actually hitting the dummy; if the result is still positive, then IASA is being delayed, if not and the previous experiment was positive, then it is hitstun; if neither is positive, then we need to return to the drawing board.

Click to expand...

error will be encountered in the bolded section, you would have to hit frame perfectly when they come into range each time after they start bouncing out, i don't think any Utilts exist that will not knock you high enough so that you fall back down (on the 10th hit) before the previous Utilt animation finishes.

im gunna go head over and check out the correct thread now

Definitely checking this out to see if I can help at all.

J4pu said:

error will be encountered in the bolded section, you would have to hit frame perfectly when they come into range each time after they start bouncing out, i don't think any Utilts exist that will not knock you high enough so that you fall back down (on the 10th hit) before the previous Utilt animation finishes.

im gunna go head over and check out the correct thread now

Click to expand...

Lucario's u-tilt on giant mode in a walled-in cube in custom stage.

I just want the Cyan name tag...

DDD's CG is a true infinite on all 5 characters. The diminished throw only effects DDD before 50% when he isn't able to get a grab attack out. But once he can it's to death. Before that it's five standing grabs which lead into the chain grab or back throw which almost always adds up to 50 or so. You just can't start it directly from 0.

Cyan name plz.

Umbreon said:

DDD's CG is a true infinite on all 5 characters. The diminished throw only effects DDD before 50% when he isn't able to get a grab attack out. But once he can it's to death. Before that it's five standing grabs which lead into the chain grab or back throw which almost always adds up to 50 or so. You just can't start it directly from 0.

Cyan name plz.

Click to expand...

If u watched the vid, it shows all that, so it's kinda information that's old. Definitely some sort of diminish, because all the influences of decay (dying, refreshing, decaying) have been shown in the video. I think it has nothing to do with "delay", because the grab frame is exact (I think) according to the video. It's either speed (which I have more faith in, as I will explain below), or inability to grab. Here's why I think it's speed of grabbed character that's more of the reason. I used a stop watch (which had milliseconds compatable), and I started and stopped on three different scenarios:

-Grab (non-decayed), dthrow, and fail on regrab (control test) (stopped when thrown character stops moving)
-Grab (non-decayed), dthrow, and successfully regrabbed (stopped at time of grab)
-Grab (decayed), dthrow, and fail on regrab (once again, stopped when thrown character stops moving)
I did each 4 times, to have consistency. Here are the results:
-about .74 seconds on throw one (control)
-about .69 seconds on throw two (non-decay, regrab)
-about .69 seconds on throw three (decayed, failed regrab)

So basically, I saw a change of speed on the character moving, and moved away from danger in the same amount of time that D3 would have regrabbed. While some character hit/hurt box changes could influence the results, I think a major contribution is speed of character (as seen by the RESULTS).
-disclaimer: I only tested Samus, didn't test on Mario or Luigi's results, but could prolly find the same results.
Question: If you are a "researcher", do you have to devote LOTS (like LOTS of LOTS) of time to help? Or is it kind of a discussion/ more free flow thing?

The properties are the same as Falco's shine in Melee. After repeated hits, trajectory is increased by arbitrary code. This is what makes it inconsistent. Doing a grab attack simply resets the arbitrary code that says "don't let this happen".







I'm BSing.

When am I actually going to see Teal on the people that I know that deserve Teal?

Crow! said:

Clever method of testing, but you wound up counting not how much time is taken by a dtilt but how far the dtilt goes. If the IASA frame is being delayed, it should take measurably longer to do the dtilt spam - you therefore want to time how long it takes to traverse the stage. This becomes difficult, though, because you need to make sure the same number of hits are applied in each case of the test. A bit of messing around with variables yields the following possible test:

Set damage ratio very low
Use someone with a good short range juggling Utilt (Utilt is the easiest grounded move to buffer, apart from AAA which is useless for this testing)
Hit DDD or Bowser upward 10 times from 0%, timing how long it takes from the start until the last hit is scored.
Kill the dummy to reset health, repeat the experiment
Save replay, do the timing from there
See if there is a difference in time large enough to attribute to something other than human error.


Also, even if we follow the logic of the original proposal we're going on, it is not clear whether it is an increase in IASA time or in the attacker's hitstun that is being increased. Repeat the same experiment without actually hitting the dummy; if the result is still positive, then IASA is being delayed, if not and the previous experiment was positive, then it is hitstun; if neither is positive, then we need to return to the drawing board.

Click to expand...

If this is the case, couldn't you just time the number of dtilts you can get out in an arbitrary time frame - say 10 seconds - when fresh without hitting someone, then fully decay it and see how many come out in the same time frame?

Reading through all of these explanations, doing tests of my own, and watching a lot of D3 videos has led me to ask one question: What if this is just some code anomaly that is completely unique? This implies that either HAL wanted D3's Dthrow to be a chaingrab or that it is a glitch. I highly doubt it is an intended CG, as the game was tailored to be casual, not competitive.

Also, how would anyone figure this out if it was a unique move?

If you guys read the real thread for this in General Brawl Disc. you'd see that everything posted here has long since been proven wrong, and that the real reason for his dthrow acting as it is has already been figured out.

Since everyone on the Internet is lazy except a few, I'll be one of those few, and bring the result here:

scotu said:

Full Description of the DDD dthrow Fresh vs Stale:

When DDD dthrows his opponent, he hits them with his ***, and while the amount of hitlag changes depending on if it's stale or not, this doesn't make a difference, because once that hitlag is done, they're at the same point in the throw animation.

However, the frame that hitlag is done on this *** attack, the throwee is hit by the puff of wind generated by DDD's *** hitting the ground. This hit is responsible for the knockback that lets DDD chainthrow. If you look in training mode, this is the 2nd of 2 hits that DDD's dthrow deals, you can also see the percents change twice if you view it frame by frame.

This asswaft is a "projectile", and not part of DDD, so DDD does not recieve hitlag from this hit. HOWEVER, the Throwee DOES recieve hitlag. Moreover, the throwee recieves less hitlag from the stale waft, than the fresh waft. Since the throwee is in less hitlag from the stale waft in relation to the fresh waft, while the DDD recieves no difference (he does not receive hitlag from it regardless), the samus starts moving away from the DDD earlier in relation to the DDD's throwing animation, thus avoiding the infinite CG.

One more thing to note is that the throwee is pushed upward into the air during the waft hitlag (presumably by DDD's presence). The extra frame of hitlag caused by the waft being fresh allows the throwee to begin descending right before the hitlag wears off, giving them a lower launch angle than that of the stale waft (which doesn't get a frame to descend).

Click to expand...

Also note that on the topic of IASA frames, Kirk, myself, and several others have all simply counted the number of frames before IASA and it doesn't change based on if it's decayed or not.

*** waft. Brilliant.

Calciphoce said:

Reading through all of these explanations, doing tests of my own, and watching a lot of D3 videos has led me to ask one question: What if this is just some code anomaly that is completely unique? This implies that either HAL wanted D3's Dthrow to be a chaingrab or that it is a glitch. I highly doubt it is an intended CG, as the game was tailored to be casual, not competitive.

Also, how would anyone figure this out if it was a unique move?

Click to expand...

Word. It's a glitch.

NEW CHALLENGE. Find out why MK has an infinite cape. Maybe it has something to do with IASA frames.