Vectoring: The replacement to Directional Influence in Smash 4

[Jaxas]

Attention to everyone in the name debate​

This is an important topic and one that we need to reach a consensus on. However, we need to talk about this in the thread that's designed for it, not here.
http://smashboards.com/threads/the-di-vi-naming-debate.368986/page-6
http://smashboards.com/threads/the-di-vi-naming-debate.368986/page-6

 

[Conda]

EDIT: I didn't realize we're shifting all name-debate posts to the other thread. Please don't respond to this, as it'll be the last I talk about this topic in here.


You are following in Zipzo's footsteps in not addressing the argument.

1. DI is to vectoring as walking is to running.
2. No one in their right mind would treat running as a "kind of walking."
3. We shouldn't treat vectoring as a "kind of DI."

Pick one of the two premises and explain why you disagree with it.

I did write a long response to your post (if you don't believe me, see how long my posts in this and other threads are). But I did not want to extend the pointless discussion on 'naming' as it is off topic and I respect the mods.
I also erased it because I feel it diluted the only important and vital point that is worth trying to communicate, which is below:

"'Vectoring' accurately points out the math that is in the mechanic. That's great. That is worth absolutely nothing for what mechanic names are actually used for in the competitive scene. Mathematical technical descriptiveness does not make it useful at all as a name, for how names are actually used most frequently in the community - how-to tutorial videos, and competitive tournaments at streamed venues. Those matter more than this forum's snowglobe discussion on what name has the most technicality in it."

 

[κomıc]

I'm still going to call it Directional Influence. It is Smash 4's take on it.

 

[Fachewachewa]

Hello everybody, first post here. (disclaimer : I’m french, I hope I don’t make too much mistakes...)

I’ve been reading this thread for some times now, and you gathered good data on VI regarding vertical KOs, but not so much on horizontal KOs. So I grabbed my brother’s 3DS and ran some tests. (and seriously, this name debate needs to stop on this thread, I’m gonna arbitrarily call it VI here, don‘t focus on that)

INTRODUCTION

So, we know there is 8 directions we can input to modify knockback. We roughly know how VI works with vertical knockback, but there one thing that troubles me : even though Strong Bad said down VI does not affect horizontal KO, around the game’s release, a lot of people talked about “down DI”. This technique seemed effective, but does not seem to fit with our actual theory [by MadKraken].

Plus, Raykz mentioned how he survived to a side smash with pikachu at 180% by VIing down, and how we might have a little problem with our current model.

METHODS

I’m using the demo on a 3DS and a 3DS XL.
“KOed” character was Link at differents %.
Link was standing on the edge of FD Battlefield (just before the oh-gawd-im-gonna-fall animation)
KO moves used : Villager’s Tree, Pikachu’s side smash, Mario’s side smash (both fully charged)
Each KO move was used fresh by SDing
“KOer” characters were at 0% to avoid any rage effect.
VI is inputted this way : for down VI by crouching before the hit, for side VI by dashing just before the hit.
I then try to jump / move away from the right blast zone out of hitstun (as VI is not registered after hitlag).

RESULTS

First off, I noticed VI was not taken into account while taunting. I don’t know how Strong Bad did, but I didn’t manage to use tants effectively. Taunt are an easy way to avoid movement while inputting a direction, so this made the test a little harder (especially for up VI).

About Villager’s Tree and Pikachu’s side smash : the knockback is diagonal, and Link is sent flying in the upper corner, resulting in a diagonal KO/explosion. (btw, are we sure corners are true corners and not slanted corners ?)
So, as I couldn’t determine if Link died by crossing the right blast zone or the upper blast zone, I decided to use Mario’s side smash, which results in an horizontal KO/explosion.

I still got some results for the Tree :
without VI, Link is KOed at exactly 58%
with down VI, Link is KOed between 75% and 81% (survived at 75%)
Which gives us about 20% difference.

For Mario’s side smash :
without VI, Link is KOed at exactly 45%
with down VI, Link is KOed at exactly 64%
with back VI, Link is KOed around 66/67%
(for this, Link might have survived 1-2% because he moved a little before being hit)
Difference : 19% / 22%

I tried up VI and up + back VI, but this was not easy to do alone, I can just tell you that Link was KOed at 54% with up VI, and 66% with up + back VI, but I haven’t tested more values.
(Up VI sent him flying higher, and maybe faster. This could mean Link being KOed before 45%, but I didn’t test VI towards bast zones.)

DISCUSSION

From previous tests, ItsRainingGravy showed a survival difference of 30% to 42% for a vertical KO with different levels of Rage. Then, Jaxas showed a survival difference of 39%, again, with a vertical KO.
Here, I found differences of approximately 20%.
Thus, my first proposition is that VI is able to help survive about 20% in general, the differences of 10 to 20% from vertical KO might be due to gravity. I don’t think this is really important and all this might also be influenced by the moves used, but it’s worth mentioning.

Now, the big problem here was the “square” model. With the previous model, down VI should not be useful for an horizontal KO. We clearly saw that it is not true. Here is my theory :
I think the square model is still true. It works well for vertical KOs, but not for horizontal ones. The reason is there's no true horizontal KO (at least, not from the ground in the demo ?). So yeah, knockback is diagonal, and we need to integrate that within the square model.

With that in mind, we then need to consider there is only 8 directions for VI, without differences between differents diagonals (this needs to be tested with vertical knockback). This way, VIing down is considered by the game as a diagonal, relative to the knockback received : VI back and down.

If this is really the case, VIing down is an effective way to survive, unless you need a high recovery, or the move that hits you sends you flying horizontally.


Last thing : as I said earlier, VIing during a taunt did not affect horizontal knockback, this may be why Strong Bad did not notice any change with Bowser’s side smash. Still, it seems to work with vertical knockback, and I have no idea why (I didn’t test this). This may be due to a form of crouch canceling, but I had almost the same results with side VI, so I don’t really think that’s the case.

Anyway, the best way to have a clear answer might just be to try with throws. I feel a little dumb to not have thought of this before, might test this tomorrow.


What do you think ?

 

[Jaxas]

Hello everybody, first post here. (disclaimer : I’m french, I hope I don’t make too much mistakes...)

I’ve been reading this thread for some times now, and you gathered good data on VI regarding vertical KOs, but not so much on horizontal KOs. So I grabbed my brother’s 3DS and ran some tests. (and seriously, this name debate needs to stop on this thread, I’m gonna arbitrarily call it VI here, don‘t focus on that)

INTRODUCTION

So, we know there is 8 directions we can input to modify knockback. We roughly know how VI works with vertical knockback, but there one thing that troubles me : even though Strong Bad said down VI does not affect horizontal KO, around the game’s release, a lot of people talked about “down DI”. This technique seemed effective, but does not seem to fit with our actual theory [by MadKraken].

Plus, Raykz mentioned how he survived to a side smash with pikachu at 180% by VIing down, and how we might have a little problem with our current model.

METHODS

I’m using the demo on a 3DS and a 3DS XL.
“KOed” character was Link at differents %.
Link was standing on the edge of FD Battlefield (just before the oh-gawd-im-gonna-fall animation)
KO moves used : Villager’s Tree, Pikachu’s side smash, Mario’s side smash (both fully charged)
Each KO move was used fresh by SDing
“KOer” characters were at 0% to avoid any rage effect.
VI is inputted this way : for down VI by crouching before the hit, for side VI by dashing just before the hit.
I then try to jump / move away from the right blast zone out of hitstun (as VI is not registered after hitlag).

RESULTS

First off, I noticed VI was not taken into account while taunting. I don’t know how Strong Bad did, but I didn’t manage to use tants effectively. Taunt are an easy way to avoid movement while inputting a direction, so this made the test a little harder (especially for up VI).

About Villager’s Tree and Pikachu’s side smash : the knockback is diagonal, and Link is sent flying in the upper corner, resulting in a diagonal KO/explosion. (btw, are we sure corners are true corners and not slanted corners ?)
So, as I couldn’t determine if Link died by crossing the right blast zone or the upper blast zone, I decided to use Mario’s side smash, which results in an horizontal KO/explosion.

I still got some results for the Tree :
without VI, Link is KOed at exactly 58%
with down VI, Link is KOed between 75% and 81% (survived at 75%)
Which gives us about 20% difference.

For Mario’s side smash :
without VI, Link is KOed at exactly 45%
with down VI, Link is KOed at exactly 64%
with back VI, Link is KOed around 66/67%
(for this, Link might have survived 1-2% because he moved a little before being hit)
Difference : 19% / 22%

I tried up VI and up + back VI, but this was not easy to do alone, I can just tell you that Link was KOed at 54% with up VI, and 66% with up + back VI, but I haven’t tested more values.
(Up VI sent him flying higher, and maybe faster. This could mean Link being KOed before 45%, but I didn’t test VI towards bast zones.)

DISCUSSION

From previous tests, ItsRainingGravy showed a survival difference of 30% to 42% for a vertical KO with different levels of Rage. Then, Jaxas showed a survival difference of 39%, again, with a vertical KO.
Here, I found differences of approximately 20%.
Thus, my first proposition is that VI is able to help survive about 20% in general, the differences of 10 to 20% from vertical KO might be due to gravity. I don’t think this is really important and all this might also be influenced by the moves used, but it’s worth mentioning.

Now, the big problem here was the “square” model. With the previous model, down VI should not be useful for an horizontal KO. We clearly saw that it is not true. Here is my theory :
I think the square model is still true. It works well for vertical KOs, but not for horizontal ones. The reason is there's no true horizontal KO (at least, not from the ground in the demo ?). So yeah, knockback is diagonal, and we need to integrate that within the square model.

With that in mind, we then need to consider there is only 8 directions for VI, without differences between differents diagonals (this needs to be tested with vertical knockback). This way, VIing down is considered by the game as a diagonal, relative to the knockback received : VI back and down.

If this is really the case, VIing down is an effective way to survive, unless you need a high recovery, or the move that hits you sends you flying horizontally.


Last thing : as I said earlier, VIing during a taunt did not affect horizontal knockback, this may be why Strong Bad did not notice any change with Bowser’s side smash. Still, it seems to work with vertical knockback, and I have no idea why (I didn’t test this). This may be due to a form of crouch canceling, but I had almost the same results with side VI, so I don’t really think that’s the case.

Anyway, the best way to have a clear answer might just be to try with throws. I feel a little dumb to not have thought of this before, might test this tomorrow.


What do you think ?

I'll be over at a friends house later this evening where we'll have 3 copies of the demo, so I'll be able to help with testing as well. I probably won't get that much done, since it's the day before a tournament and we're all going to be practicing a ton for that, but still.

Anyways, great work! This looks really good; also, it's really weird how Taunts affect VI...
I'm gonna add this to my Vectoring/VI/KI info gathering thread, along with most of the other info I've found.
I'll let you know what I find when I run my tests, anything specific I should test? We may even be able to get video recordings, since one of my friends has a stream 3DS.

 

[Venks]

Vectoring isn't as accurate as it has been described.

Personally I think the numbers were off because he was using Jigglypuff who is the lightest character in the game and didn't test this out with other characters and attacks.

Testing Little Mac's KO Punch against Jigglypuff I discovered that by "Vectoring" down she could live up to 18%. And just like Strong Bad said Vectoring to the left did not alter what percentage Jigglypuff was hit into the up blast zone compared to not vectoring at all.

However when I was testing the KO Punch against Mario I noticed that "vectoring" down always caused him to be hit into the upper blast zone at 30%. So then I tested out Mario's "no vectoring" percentage and found it to be 24%. Out of a mere whim I decided to try out "vectoring" to the left and discovered that Mario did not reach the top blast zone at 24%. Further testing then showed that "vectoring" down-left allowed Mario so survive from the KO Punch until he reached 32%.

Now this doesn't sound like much because of the small percentage gaps when concerning the KO Punch, but the same applies to all attacks and thus the gaps will be wider for other attacks. This also means the term "Vectoring" isn't actually a proper term for Hitstun Shuffling in Smash 4. Vectoring does not account for the extra survival for diagonal inputs.

We need more people to test out Directional Influence in Smash 4. Use diagonal inputs when being hit with a strong attack and you should survive longer.

 

[Chiroz]

Vectoring isn't as accurate as it has been described.

Personally I think the numbers were off because he was using Jigglypuff who is the lightest character in the game and didn't test this out with other characters and attacks.

Testing Little Mac's KO Punch against Jigglypuff I discovered that by "Vectoring" down she could live up to 18%. And just like Strong Bad said Vectoring to the left did not alter what percentage Jigglypuff was hit into the up blast zone compared to not vectoring at all.

However when I was testing the KO Punch against Mario I noticed that "vectoring" down always caused him to be hit into the upper blast zone at 30%. So then I tested out Mario's "no vectoring" percentage and found it to be 24%. Out of a mere whim I decided to try out "vectoring" to the left and discovered that Mario did not reach the top blast zone at 24%. Further testing then showed that "vectoring" down-left allowed Mario so survive from the KO Punch until he reached 32%.

Now this doesn't sound like much because of the small percentage gaps when concerning the KO Punch, but the same applies to all attacks and thus the gaps will be wider for other attacks. This also means the term "Vectoring" isn't actually a proper term for Hitstun Shuffling in Smash 4. Vectoring does not account for the extra survival for diagonal inputs.

We need more people to test out Directional Influence in Smash 4. Use diagonal inputs when being hit with a strong attack and you should survive longer.

I had a huge hunch that it wasn't actually a square, but what you are describing is even a weirder shape than a square or a circle. Could you test this with Bowser using a weak kill move (such as Greninja's Up-Smash or something)? We need to have a big difference in the %s to understand what's going on accurately. Using the heaviest character and killing him with the weakest move (that can kill from the top) will be the best way to test this.

 

[Venks]

I had a huge hunch that it wasn't actually a square, but what you are describing is even a weirder shape than a square or a circle. Could you test this with Bowser using a weak kill move (such as Greninja's Up-Smash or something)? We need to have a big difference in the %s to understand what's going on accurately. Using the heaviest character and killing him with the weakest move (that can kill from the top) will be the best way to test this.

I tested it out with Mega Man's down smash. DI or whatever down and to the left/right kept Mario alive for only an additional 1%.
DI/Vector Down: 100% KOs
DI/Vector Down-Left: 100% Does not KO.
DI/Vector Down-left: 101% KOs

I'm pretty disappointed in this discovery as it doesn't have that large of an effect as I thought it would. I started to think that maybe I just DI'd wrong, but I tested this 10 times over and continued to get the same results. And for some characters against some attacks it is 2% instead of 1%, but I've yet to see it ever go higher than that.

 

[Tagxy]

I tested it out with Mega Man's down smash. DI or whatever down and to the left/right kept Mario alive for only an additional 1%.
DI/Vector Down: 100% KOs
DI/Vector Down-Left: 100% Does not KO.
DI/Vector Down-left: 101% KOs

I'm pretty disappointed in this discovery as it doesn't have that large of an effect as I thought it would. I started to think that maybe I just DI'd wrong, but I tested this 10 times over and continued to get the same results. And for some characters against some attacks it is 2% instead of 1%, but I've yet to see it ever go higher than that.

This is the one Im curious about:
For Mario's side smash :
without VI, Link is KOed at exactly 45%
with down VI, Link is KOed at exactly 64%
with back VI, Link is KOed around 66/67%
(for this, Link might have survived 1-2% because he moved a little before being hit)
Difference : 19% / 22%

Could you test more side-attacks and DI inputs?

 

[Venks]

This is the one Im curious about:
For Mario's side smash :
without VI, Link is KOed at exactly 45%
with down VI, Link is KOed at exactly 64%
with back VI, Link is KOed around 66/67%
(for this, Link might have survived 1-2% because he moved a little before being hit)
Difference : 19% / 22%

Could you test more side-attacks and DI inputs?

Not a problem.

Little Mac's Straight Lunge hitting Mario right at the edge of Final Destination:

Without DI/VI: Mario is KO'd at 80%
With down DI/VI: Mario is KO'd at 86%
With back DI/VI: Mario is KO'd at 90%
With down-back DI/VI: Mario is KO'd at 92%
Crouching before the hit: Mario is KO'd at 118%

Oddly enough this only works with down-back and not with up-back so again something else is in effect besides vectoring. And of course it's worth noting that crouching before an attack lands drastically reduces the amount of knock back once again like it did in 64.

 

[Empyrean]

So many different results from different testers. This is becoming confusing at this point, but is VI's potency move-dependant or something?

 

[Deleted member 245254]

So many different results from different testers. This is becoming confusing at this point, but is VI's potency move-dependant or something?

I think one of the points Venks is trying to make is that "VI" is an inappropriate name for the mechanic due to the obviously unexplored depth of the mechanic, regardless.

It's not likely move dependent, there are already many factors affecting DI such as the "rage" effect, and the new mechanics we clearly have yet to fully understand, making it move dependent would probably make results a lot more inconsistent across characters than they have been during tests.

 

[Jaxas]

And of course it's worth noting that crouching before an attack lands drastically reduces the amount of knock back once again like it did in 64.

Are you sure about this? If so, then my previous info was incorrect as I was crouching prior to being hit for ease of controlling 2 3DS by myself. I'll retest when I can if needed, please let me know.

 

[Chiroz]

Are you sure about this? If so, then my previous info was incorrect as I was crouching prior to being hit for ease of controlling 2 3DS by myself. I'll retest when I can if needed, please let me know.

I don't know how correct he is but remember when I said I survived a Pikachu fresh F-Smash at 180% just with downwards VI (no horizontal VI)? Well, I was crouching before it hit me.

Brawl is the only game without crouch reducing your knockback,
crouch cancelling could very well be back.

 

[Locuan]

So after reading through the thread I want to point out my observations and see if I am correct.

Assumptions from what I read:

1. Any move that sends you upwards means hold downward, any horizontal direction (assuming kill %), will still KO.
2. Since falling speed is a factor, holding up and toward the stage in this case should be better. However if that is the case, and if Vector Response holds true, you would still add Vertical motion to the vector and send you further up.
3. The "vectoring", "vector influence", or Vector Response as I like to call it, will always be less than the magnitude of the vector that sent you skyward.

Alright and from that I begin.

Two dimensional Vectors have x and y components. For example, A = (x, y). x corresponds to horizontal motion while y corresponds to vertical motion.

Breaking down vectors unto their components, lets assume A is the vector that sent you skyward. B will be the Vector Response you implemented. C will be the resultant vector, what you actually observe in the game. C being the resultant vector means that the components of A and the components of B will be added and correspond to the components of C.

The first example will include implementing DI, or a direction perpendicularly opposite to the direction you were sent (opposite x, same y). If A = (3, 2), B = (-3, 2), C = (0, 4). So basically, if you DI in Smash 4, you will be adding horizontal motion unto yourself. In other words, you will fly further upwards. This would also correspond to Strong Bad's observations in the original post of 100+20 = 120 etc. Now since the Vector Response will always be less than the original vector that sent you skyward the actual values would be more like A = (3, 2), B = (-1, 1), C = (2, 3), but the observation still holds true. For this reason, I stand by the argument that we should not call it DI since they have completely different mechanics.

The second example will tackle the second assumption I made initially and reinforces the first example as well. So lets say in this case you are sent in a direction corresponding to A = (-2, 2), and you decide to Vector Response towards the stage but upwards with B = (1, 1), the resulting vector would be C = (-1, 3). Therefore, you cut your horizontal motion but added one unit of motion vertically. This brings me to believe that we have to be very careful when deciding our Vector Responses. While we will be limiting the amount they sent us horizontally we can still be KO'd upwards if we have a lot of % already on us.

The third example will correspond to the vector A corresponding to a direction of (3, 2). In this case we will test with a Vector Response that would be in the same horizontal direction but with an opposite vertical motion; B = (3, -1). Here C = (6, 1). While we did cut our vertical motion we added 3 units in our horizontal motion which sent us further off in that direction.

The fourth example will be what I call "Perfect Vector Response" or "Optimal Vector Response". Assuming no fall speed, then the optimal response to a vector A = (3, 2) would be B = (-3, -2). Your Vector Response should be directly opposite to the direction you were sent. C = (0, 0) and both motions would be cut off. However, since we know that we will never be able to pull that off due to game mechanics B would actually be more along the lines of B = (-3/2, -1). The resulting vector would be C = (3/2, 1). Here we effectively cut both x and y, horizontal and vertical motion units by half.

So from what I read from the original post and the subsequent explanations by others I believe that in essence Vectoring, Vector Influence or Vector Response basically works exactly the same as with Basic Vector Operations.

As a question to Strong Bad or anyone who may know, am I correct in my assumptions or is there something that I am missing?

This is very interesting and I would like to get into the mathematics a bit more.

 

[Chiroz]

So after reading through the thread I want to point out my observations and see if I am correct.

Assumptions from what I read:

1. Any move that sends you upwards means hold downward, any horizontal direction (assuming kill %), will still KO.
2. Since falling speed is a factor, holding up and toward the stage in this case should be better. However if that is the case, and if Vector Response holds true, you would still add Vertical motion to the vector and send you further up.
3. The "vectoring", "vector influence", or Vector Response as I like to call it, will always be less than the magnitude of the vector that sent you skyward.

Alright and from that I begin.

Two dimensional Vectors have x and y components. For example, A = (x, y). x corresponds to horizontal motion while y corresponds to vertical motion.
Breaking down vectors unto their components, lets assume A is the vector that sent you skyward. B will be the Vector Response you implemented. C will be the resultant vector, what you actually observe in the game. C being the resultant vector means that the components of A and the components of B will be added and correspond to the components of C.
The first example will include implementing DI, or a direction perpendicularly opposite to the direction you were sent (opposite x, same y). If A = (3, 2), B = (-3, 2), C = (0, 4). So basically, if you DI in Smash 4, you will be adding horizontal motion unto yourself. In other words, you will fly further upwards. This would also correspond to Strong Bad's observations in the original post of 100+20 = 120 etc. Now since the Vector Response will always be less than the original vector that sent you skyward the actual values would be more like A = (3, 2), B = (-1, 1), C = (2, 3), but the observation still holds true. For this reason, I stand by the argument that we should not call it DI since they have completely different mechanics.
The second example will tackle the second assumption I made initially and reinforces the first example as well. So lets say in this case you are sent in a direction corresponding to A = (-2, 2), and you decide to Vector Response towards the stage but upwards with B = (1, 1), the resulting vector would be C = (-1, 3). Therefore, you cut your horizontal motion but added one unit of motion vertically. This brings me to believe that we have to be very careful when deciding our Vector Responses. While we will be limiting the amount they sent us horizontally we can still be KO'd upwards if we have a lot of % already on us.
The third example will correspond to the vector A corresponding to a direction of (3, 2). In this case we will test with a Vector Response that would be in the same horizontal direction but with an opposite vertical motion; B = (3, -1). Here C = (6, 1). While we did cut our vertical motion we added 3 units in our horizontal motion which sent us further off in that direction.
The fourth example will be what I call "Perfect Vector Response" or "Optimal Vector Response". Assuming no fall speed, then the optimal response to a vector A = (3, 2) would be B = (-3, -2). Your Vector Response should be directly opposite to the direction you were sent. C = (0, 0) and both motions would be cut off. However, since we know that we will never be able to pull that off due to game mechanics B would actually be more along the lines of B = (-3/2, -1). The resulting vector would be C = (3/2, 1). Here we effectively cut both x and y, horizontal and vertical motion units by half.
So from what I read from the original post and the subsequent explanations by others I believe that in essence Vectoring, Vector Influence or Vector Response basically works exactly the same as with Basic Vector Operations.

As a question to Strong Bad or anyone who may know, am I correct in my assumptions or is there something that I am missing?

This is very interesting and I would like to get into the mathematics a bit more.

From what I've gathered in the past days, assumption 1 is incorrect. From everyone's but the original finder's tests, horizontal inputs DOES affect knockback distance and can save you. I know Strong Bad discovered VI but I am more inclined to go with the other 6 testers just because of sheer numbers.

 

[Jaxas]

From what I've gathered in the past days, assumption 1 is incorrect. From everyone's but the original finder's tests, horizontal inputs DOES affect knockback distance and can save you. I know Strong Bad discovered VI but I am more inclined to go with the other 6 testes just because of sheer numbers.

I believe that @ Fachewachewa figured out that only vertical VI/KI/whatever is read during a taunt, and not horizontal.

Smash 4 has been having some weird things with mechanics like that; the numbers displayed in the waiting room are all rounded for example, and it still applies the freshness bonus but doesn't stale moves...

Anyways, if people could make a list of things for me to test today that would be great. I'm going to be running the Melee Salem Smashfest today, and one of my friends with a 3DS will be there so we can test things; I won't have internet though.

 

[Locuan]

From what I've gathered in the past days, assumption 1 is incorrect. From everyone's but the original finder's tests, horizontal inputs DOES affect knockback distance and can save you. I know Strong Bad discovered VI but I am more inclined to go with the other 6 testers just because of sheer numbers.

If that is the case, and your observations are indeed correct, then I believe what affects the distance you travel is determined by the magnitude of the vector aside from one particular component over the other.

Edit: Apparently I like being redundant.

 

[Chiroz]

I believe that @ Fachewachewa figured out that only vertical VI/KI/whatever is read during a taunt, and not horizontal.

Smash 4 has been having some weird things with mechanics like that; the numbers displayed in the waiting room are all rounded for example, and it still applies the freshness bonus but doesn't stale moves...

Anyways, if people could make a list of things for me to test today that would be great. I'm going to be running the Melee Salem Smashfest today, and one of my friends with a 3DS will be there so we can test things; I won't have internet though.

Well, horizontal VI was definitely being read in Strong Bad's video because you can clearly see Jigglypuff heading to the right after being hit. I think the problem was the way the testing was done.

He showed that Jiggz died at 45% without VI. He then showed she could survive at 51% with downwards VI. He lastly showed she cannot survive at 47% with horizontal VI.

There's a lot of problems with this test, we don't know if Jiggz could die earlier without VI because he doesn't show that she can survive at 44%, so technically she could be dying as early as 35% or something. Second we don't know if horizontal VI does change anything by this test because Jiggz is incredibly light and horizontal VI could have been affecting his knockback just not enough because of how strong the knockback was

(Note: All of these %'s are before pre-hit).



If you could, could you test the same example as Strong Bad?

Check what is the exact earliest % Jiggz dies from a fully charged G&W Up-Smash without VI.
Check if she can survive at 1 extra % with horizontal VI.
Check what is the exact latest % Jiggz can survive with downwards VI.

If 2 is true then check what is the exact % she can survive up to with horizontal VI.
If 2 is true then check what is the exact latest % Jiggz can die with diagonal VI (downwards + horizontal).



These 5 tests should definitely clear up is VI is just adding a vector or if there is multiplication (or changing angle like DI) involved. This way we can test if Strong Bad's first post is correct.

 

[Jaxas]

Well, horizontal VI was definitely being read in Strong Bad's video because you can clearly see Jigglypuff heading to the right after being hit. I think the problem was the way the testing was done.

He showed that Jiggz died at 45% without VI. He then showed she could survive at 51% with downwards VI. He lastly showed she cannot survive at 47% with horizontal VI.

There's a lot of problems with this test, we don't know if Jiggz could die earlier without VI because he doesn't show that she can survive at 44%, so technically she could be dying as early as 35% or something. Second we don't know if horizontal VI does change anything by this test because Jiggz is incredibly light and horizontal VI could have been affecting his knockback just not enough because of how strong the knockback was

(Note: All of these %'s are before pre-hit).



If you could, could you test the same example as Strong Bad?

Check what is the exact earliest % Jiggz dies from a fully charged G&W Up-Smash without VI.
Check if she can survive at 1 extra % with horizontal VI.
Check what is the exact latest % Jiggz can survive with downwards VI.

If 2 is true then check what is the exact % she can survive up to with horizontal VI.
If 2 is true then check what is the exact latest % Jiggz can die with diagonal VI (downwards + horizontal).



These 5 tests should definitely clear up is VI is just adding a vector or if there is multiplication (or changing angle like DI) involved. This way we can test if Strong Bad's first post is correct.

Unfortunately I only have the demo, so no Jiggz/G&W. I can try Pika being launched by Mario or something, though

 

[Chiroz]

Unfortunately I only have the demo, so no Jiggz/G&W. I can try Pika being launched by Mario or something, though

Well, yea, try those same tests with Megaman and Mario (Megaman because he is the heaviest of the demo). Remember to keep the Up-Smash fresh.

 

[Locuan]

These 5 tests should definitely clear up is VI is just adding a vector or if there is multiplication (or changing angle like DI) involved. This way we can test if Strong Bad's first post is correct.

Exactly, this is what is important. The underlying mathematical operation is what has to be understood. Additionally, while something might seem simple, the mechanics can be completely different. Algorithms can bring a lot of surprises.

I guess the best example to enforce this point would be Newton's concepts that were later corrected by Einstein. While Newton's observations hold true if we consider low speeds, in reality they are approximations to a more complex natural algorithm as speeds get close to those of light.

 

[Tagxy]

Not a problem.

Little Mac's Straight Lunge hitting Mario right at the edge of Final Destination:

Without DI/VI: Mario is KO'd at 80%
With down DI/VI: Mario is KO'd at 86%
With back DI/VI: Mario is KO'd at 90%
With down-back DI/VI: Mario is KO'd at 92%
Crouching before the hit: Mario is KO'd at 118%

Oddly enough this only works with down-back and not with up-back so again something else is in effect besides vectoring. And of course it's worth noting that crouching before an attack lands drastically reduces the amount of knock back once again like it did in 64.

Woooooooooooooooow

I think its safe to say things are definitely very different then we expected. Side and upward trajectories will be completely different, and down seems to be a powerful input at least as far as it affects hitstun. Good work!

Everyone should definitely be taking a look at the quoted post and the one on marios fsmash too.

 

[Locuan]

With down-back DI/VI: Mario is KO'd at 92%

By down-back do you mean the direction completely opposite to where your character was launched?

 

[Venks]

By down-back do you mean the direction completely opposite to where your character was launched?

Mario is being launched into the left blast zone. So in that situation I was holding down-right.

 

[Locuan]

Mario is being launched into the left blast zone. So in that situation I was holding down-right.

​

Like that? If so, then my analysis above should correlate to the data you posted.​

 

[Venks]

​

Like that? If so, then my analysis above should correlate to the data you posted.​

This is what I'm doing for maximum survival.

The KO Punch is easier to point out because it's a much more straight line. You would think that down-right DI/VI would be better than down-left DI, but my testing found the results to be the same. And of course both down-left and down-right are superior to pure down.

 

[Locuan]

You would think that down-right DI/VI would be better than down-left DI, but my testing found the results to be the same. And of course both down-left and down-right are superior to pure down.

Alright, interesting. Basically, as you say in the Little Mac's KO Punch case you would think that down-right would be better than down-left.

For instance, using that example you provided but with numbers:

If Mario Vector = (-1, 6) and your VI = (-3, -2) then the resultant vector would be (-4, 4). While you are effectively cutting your vertical motion you should be adding horizontal motion to your trajectory. Therefore, one would assume that down-left would be the better choice because it would cut motion on both axis. If you say that both yield same results then it should be safe to assume that vector sum operations take place on the vertical axis, yet something more complex happens regarding the horizontal motion. This is why considering "vectoring" or anything regarding vectors is a bit confusing to me. Vectors have both a direction and magnitude. To be truly a "vectoring, vector influence" thing then both properties of the vector should take place.

 

[Fachewachewa]

You guys did really good progress ! Well... at least we know that we don't know at all how "VI" works now :D
To sum up, "VI" is not about vectors, and crouch canceling is back. Right ?

I haven't been able to do more test, but I'll try today with throws. It's a detail, but i'd like to know what I did wrong (or not ?) when I used taunts. Also, I'll try with sub lethal knockback to see how different degrees of diagonals really influence knockback.


EDIT : Okay, did some quick testing, and as I expected, I was totally wrong about taunts, and what I first saw was juste the absence of crouch canceling. I don't know what happend with "side VI", but this time I managed to have result. So, with megaman's fsmash on pikachu, in the center of FD battlefield, I have similar results as Venks.

Pikachu is KOed at 72%
With Crouch Cancelling : KOed between 93 and 95%
With Taunt > Down VI : KOed between 73 and 76%
With Taunt > Back VI : KOed at 78%
I also tried one Taunt > VI Down+Back and one Taunt > Up VI at 76% and Pikachu was KOed.

So yeah, "down VI" is a thing for horizontal KO, but it's not as effective as back VI. Crouch Canceling is obviously the best solution here. You can see diagonal VI is not better than just back VI in this case, but it may be a little more effective than down VI, as I didn't test the exact KO % for each condition.
So, I don't know if it's move-dependant or if I did something wrong with the stick. I noticed pikachu going up with my "back VI" so maybe it's Up+Back / Back. But I'm pretty sure 45° diagonals are less effective in this case.

I also tried differents angles of side VI from an up throw but it didn't seems to modify knockback, with only two possibilities : diagonal and side. But I may be entierly wrong here, it's a quick observation, and this would need a lot more testing.


Anyone noticed differences between 45° diagonals and 20/30° diagonals ?

 

[Venks]

You guys did really good progress ! Well... at least we know that we don't know at all how "VI" works now :D
To sum up, "VI" is not about vectors, and crouch canceling is back. Right ?

I haven't been able to do more test, but I'll try today with throws. It's a detail, but i'd like to know what I did wrong (or not ?) when I used taunts. Also, I'll try with sub lethal knockback to see how different degrees of diagonals really influence knockback.


EDIT : Okay, did some quick testing, and as I expected, I was totally wrong about taunts, and what I first saw was juste the absence of crouch canceling. I don't know what happend with "side VI", but this time I managed to have result. So, with megaman's fsmash on pikachu, in the center of FD battlefield, I have similar results as Venks.

Pikachu is KOed at 72%
With Crouch Cancelling : KOed between 93 and 95%
With Taunt > Down VI : KOed between 73 and 76%
With Taunt > Back VI : KOed at 78%
I also tried one Taunt > VI Down+Back and one Taunt > Up VI at 76% and Pikachu was KOed.

So yeah, "down VI" is a thing for horizontal KO, but it's not as effective as back VI. Crouch Canceling is obviously the best solution here. You can see diagonal VI is not better than just back VI in this case, but it may be a little more effective than down VI, as I didn't test the exact KO % for each condition.
So, I don't know if it's move-dependant or if I did something wrong with the stick. I noticed pikachu going up with my "back VI" so maybe it's Up+Back / Back. But I'm pretty sure 45° diagonals are less effective in this case.

I also tried differents angles of side VI from an up throw but it didn't seems to modify knockback, with only two possibilities : diagonal and side. But I may be entierly wrong here, it's a quick observation, and this would need a lot more testing.


Anyone noticed differences between 45° diagonals and 20/30° diagonals ?

Really good stuff man. Personally I've noticed diagonals seem to only have a very slight influence advantage on KO percentage compared to "optimal" vertical or horizontal DI/VI. I don't have a second 3DS on me at the moment so I can't duplicate your findings for now, but I'll definitely give it a try when I can.

But as for 45° compared to 20/30°, I personally haven't seen a difference.

 

[Locuan]

Really good stuff man. Personally I've noticed diagonals seem to only have a very slight influence advantage on KO percentage compared to "optimal" vertical or horizontal DI/VI. I don't have a second 3DS on me at the moment so I can't duplicate your findings for now, but I'll definitely give it a try when I can.

But as for 45° compared to 20/30°, I personally haven't seen a difference.

So direction is not a factor when considering VI. If you guys can do multiple tests regarding different directions and the approximate distance you were sent and represent it graphically with your corresponding VI, I could try and see if the magnitude of the vector is the one that has the most influence on how this works.

Something like this:

Control group:

Sent 135 degrees no VI
Sent 30 degrees no VI
etc.

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

Test Group:

Sent 135 degrees with VI
Sent 30 degrees with VI
etc.

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

Instead of saying, survived with X% or Y% difference, try to graphically represent the difference in magnitude (how far sent) from the vectors.

This could help out to rule out some of the things Vectoring isn't.

 

[Smooth Criminal]

I thought I'd be able to keep up with you big brains.

I was wrong.

Keep up the good work, guys! This is some interesting stuff.

Smooth Criminal

 

[Chiroz]

You guys did really good progress ! Well... at least we know that we don't know at all how "VI" works now :D
To sum up, "VI" is not about vectors, and crouch canceling is back. Right ?

I haven't been able to do more test, but I'll try today with throws. It's a detail, but i'd like to know what I did wrong (or not ?) when I used taunts. Also, I'll try with sub lethal knockback to see how different degrees of diagonals really influence knockback.


EDIT : Okay, did some quick testing, and as I expected, I was totally wrong about taunts, and what I first saw was juste the absence of crouch canceling. I don't know what happend with "side VI", but this time I managed to have result. So, with megaman's fsmash on pikachu, in the center of FD battlefield, I have similar results as Venks.

Pikachu is KOed at 72%
With Crouch Cancelling : KOed between 93 and 95%
With Taunt > Down VI : KOed between 73 and 76%
With Taunt > Back VI : KOed at 78%
I also tried one Taunt > VI Down+Back and one Taunt > Up VI at 76% and Pikachu was KOed.

So yeah, "down VI" is a thing for horizontal KO, but it's not as effective as back VI. Crouch Canceling is obviously the best solution here. You can see diagonal VI is not better than just back VI in this case, but it may be a little more effective than down VI, as I didn't test the exact KO % for each condition.
So, I don't know if it's move-dependant or if I did something wrong with the stick. I noticed pikachu going up with my "back VI" so maybe it's Up+Back / Back. But I'm pretty sure 45° diagonals are less effective in this case.

I also tried differents angles of side VI from an up throw but it didn't seems to modify knockback, with only two possibilities : diagonal and side. But I may be entierly wrong here, it's a quick observation, and this would need a lot more testing.


Anyone noticed differences between 45° diagonals and 20/30° diagonals ?

So diagonal VI isn't really more effective than both like Strong Bad suggested. Also your % with taunting is very similar to the % Strong Bad had when he tested.

From what it seems VI is extremely weak, much weaker than DI ever was but Crouch Canceling is back and that's extremely strong.




Edit: Is it possible to check if Side-VI allows you to live longer from an Up-Smash?

Also could you try and check if upwards VI kills you earlier? This is very, very important to know.

 

[Gold 2.0]

To think I thought that this existed in all the other Smash games (still very much a newbie) xD. But I'm still learning the basics so I can't really give an educated response but I'm curious as to how much of an impact this will have on competitive play.

 

[Tristan_win]

If there is still any doubt if crouch canceling is back or not I found video proof.

It's pretty obvious due to the nature of the video that they were demonstrating crouch canceling.

 

[Signia]

Could you guys stop with that obnoxious colored text? I'm sure it's great analysis and all but it's too much of a pain to read.

 

[Chiroz]

If there is still any doubt if crouch canceling is back or not I found video proof.

It's pretty obvious due to the nature of the video that they were demonstrating crouch canceling.

Thanks for that. It seems my theories from about 9 days ago was right.



From the tests we've seen (including Strong Bad's which did not have crouch cancelling as he was taunting when he was hit) VI has about a 14-15% survival difference. Meaning that if you died at 130% you will probably only survive until about 150%, if you died at 45% (like Jiggz in the video) you will probably die around 51% (like in the video). This is actually slightly less of an effect than what DI had in Melee and Brawl.

It seems that the things that has been "breaking combos" and helping people survive to 180-200% is a really potent version of crouch cancel. Which going by the tests done by our members in this thread has an influence of about 20-30% more survival. Meaning that if you survived until 130% you will probably survive to about 160-173% (If you add correct VI to this then you get even higher numbers!).


The good thing about crouch cancelling is that it's much harder to do and you can't do it while in hitstun or end lag which means its use is much less global. Plus if you are hit by a horizontal attack it's going to be hard to crouch cancel and also VI correctly as the timing will be very strict (While if you are hit by a vertical attack it will be fairly easy to VI correctly if you were able to crouch cancel).


At least crouch cancelling is not as ******** as in Melee where it allowed you to get a free hit on the opponent at low %s.

Could you guys stop with that obnoxious colored text? I'm sure it's great analysis and all but it's too much of a pain to read.

Is my color the obnoxious one?

 

[Tristan_win]

I wouldn't call it obnoxious personally but on a black background it can be somewhat hard to see if you have ****ty eyes or when your sleepy. Maybe just a shade brighter would be better if it's really bugging people.
example
example or example

 

[Chiroz]

Hopefully this is better?

@ Tristan_win
@ Signia

 

[Gidy]

TL;DR, Oh look crouch cancelling is back :D

Have we decided if VI breaks combos or not?