VF5FS Wiki Movement

From my experience, all dashes are cancelable by an attack upon the dash's 1st frame (instant.) BUT, you can't buffer an attack before the dash comes out. There was a blog post on the front page explaining why/how it's possible.

A dash input will out-prioritize button inputs. So the button input(s) has to be on the literal 1st frame of the dash. It's nearly impossible to do consistently, but it's certainly doable.

Sarah's 44p and 44k execute at 10 and 13 frames respectively. They're pretty useful as punishers. But if I buffer the move, she'll just dash and attack. So I have to input the punch or kick button with very specific timimg.

 

EDIT

OK I got.

Cozby you are only talking about attacks that have a back dash as their directional inputs and about the additional frame you were referring about the neko's blog where they talked about those type of moves

What I'm talkig about instead is the ability to cancel a back dash in a move that does not contain any 44 inputs as its directions.
Like 44 > 6P
In these cases I'm now pretty sure you can cancel the back dash in the attack after 4 frames 1 frame (edit: I did an error during my recording)

 

I tested with Jean vs. Jean. Jean can punish with and (). It's hard, but possible. Realistically I would use either or as a punish in that case.

 

As you explained in your last post you are talking about another situation.

look at my previous post edited while you were writing your last post

 

Vanessa can punish Goh's 3KP with DS 3P, and 443P, and 663P. No point in doing that though right?

Do you mean cancel from a neutral state? Regardless, you still can cancel on the 1st frame, but it's just not humanly possible, because you have to take frames to move the stick to neutral, and into another direction. That's not including the time it takes to activate back dash.

There's always Hitbox!

 

Cozby your examples are about punishment situation when you can buffer the forward or backward input (naturally included in the command of the punisher move) during the blockstun or recovery in order to not waste any frame. Not the same situation. So yes he was talking about neutral state

 

Doesn't make a difference!

 

But the problem here cozby is how do you know that ? I meant where did you read that or how did you test that because the situation you were taking as examples are not related to the backdash cancelled itself since you're talking about punishment situations with moves with backdash command in their input ( which means possible buffer). Maybe when G0d3L will explain how he got his 4 frames we will understand better.

 

A dash is a dash, regardless of what you did before it.

OK so we all agree that dashes can be canceled on the 1st frame right?

New question is: From a neutral state, how long does it take an individual to input a dash, then cancel it with an attack with an input unrelated to the dash input?

The answer lies within you! It's specific to the person. Maybe their hands are cold; it might take the longer. Hitbox will help.

 

We all agree you state that without any proof.
By the way, it's true for forward dash ( as already discussed AND tested previously ), but we're talking about backdash here, that's huge difference.

My bad, I didn't see G0d3L already run the test here:
http://virtuafighter.com/threads/diy-frames-analysis.16734/page-2

Take a look.

 

K I'm done here lol.

 

This is not true.
663P is a commonly known technique called... 1 frame kill. It is called like that for a reason.
Also, there is a reason why 1FK is performed as 663P and not as 443P. And it is not because of spacing.

I stated it already several times but let me repeat myself:
NO
I know, you all saw how awesome hitbox is. On that youtube hitbox channel. Great. But you know, every car, every parfume, every dildo and every shit looks awesome... in an advert. And this channel in nothing but a advert of the company that is making and selling hitboxes.
... Oh what a surprise, but it looks SO GOOD ON THEIR VIDS AND IT IS ON YOUTUBE SO IT MUST BE TRUE!!!

Anyway, I am far from saying that hitbox is BAD or anything, but you know, there is a reason why there are many pro players using stick, several using gamepad but NOT A SINGLE ONE USING HITBOX. In the environment where you play for pretty nice money and where every, even small, advantage counts.

But yes, Shang and his awesome blog says that hitbox is bad so it must be true. Also, it looks so good on that hitbox youtube vids, right..?

 

It's not the same as 1fk. You cancel into the attack before the frame can be killed. If it were 1fk, it wouldn't punish Goh's 3kp.

 

EDIT I made an error during my recording that screwed my buffering. The post linked in the 6th point has been fixed.


Cozby, Kingo, Unicorn let's recap the situation and what is 100% confirmed because I think there are a few misunderstandings.
I know reading a WOT could be boring/takes a lot of time but in this case it's fundamental:

1) I'm not talking about cancelling any dash from neutral cause that it would involve frames for moving the stick or buttons to be pressed making it a situation where you cannot evalute how the cancelling system really works.
All the tests are done buffering the desired commands in the recovery of a precedent move

2) when I talk about dash cancel I'm talking about the ability to cancel a dash into a move that does not have a 66/44 inputs.
This is really important, because otherwise when using moves with 66/44 in it then the buffering system comes into play as we can read here in Neko's Dojo blog.
So if I want to test the cancel property of the back dash I will not use a move that has 44 as inputs (66 in case I want to study a forward dash)

3) a forward dash cancel into an attack (with no 66 as inputs) like 66 > 6P takes 1 frame as you can read here under the Attack forward dash cancel frame analysis section.
You can get it 100% of times and has nothing to do with the buffering mechanics of moves with 66/44/33/11 as inputs.

4) There are situation where even with moves with 66 inputs you have to enter the dash animation for 1 frame before the move can be executed like when you do with Kage 2/8 > 66P+K.
In this case you have to cancel the failed evade into a dash which will cancel itself after 1 frame in the attack 66P+K.
Keep always in mind that avoiding the 1 frame delay of, let's say, a bufferd Kage's 66P+K not in a match but in a dojo test is something really simple that can be achived 100% of times

5) even when you perfectly buffered a move that has the 44 inputs (so 0 frame delay and no back dash animation at all) if you get hit (obv on CH) by a move that has the property to cause a stagger on back dashing opponent you'll get the back dash stagger.

Let's take Kage vs Kage
Record Kage doing (everything on whiff!) 6P > 44P+K (perfectly bufferd; it's easy)
Play it and with Kage block the elbow (+4) and then do 3K.
You'll get a back dash stagger even if there wasn't any back dash at all.

6) Final point.
Now that we are on the same page (I hope) read this example and test it out in dojo.
If you want to do your own custom examples the important thing to keep in mind to be sure to have inputted a real back dash (even if invisible) is that hitting the dummy with a move with a back dash stagger property you'll get the stagger.


When I do these type of analysis I always do multiple tests with different settings to have a series of consistant results.

 

@G0d3L I did some testing on my own and my conclusion is that I was wrong and that you actually CAN cancel backdash after 1st frame.

Now what is the problem in testing: you can input your dash too fast so it goes on BEFORE buffer window and then it is ignored. I was trying to find any test where it will be possible to confirm i inputted this dash command IN the buffer window.

To do so, I decided to use move that can not be made from crouching position. For this purpose, I picked Lau and his 3P (15-frame). If you input it from crouching, you will get 2_3P (17f).
Then I searched for low move that is -16 on guard. Lions 426P works just fine.

Now the actual test is simple. After you block this low attack with Lau, you need to input dash command to stand up so you will be able to perform standing 3P. If you input this dash too early, it will be ignored and you will get 2_3P that can be easily distinguished from 3P.
Also, you have 16 frames to perform dash to cancel your crouch state and then to cancel this dah into 3P attack. If you will succeed with 1-frame cancel, you will get RC hit. If you will fail, Lion will block your attack (I recorded 426P - G)

1st I started with 663P. I got RC hit several times.
Then I continue with 443P. It took me a while to get used to the finger motion on HBX but then I was able to get RC as well.

CONCLUSION: I was wrong and it is possible to cancel both forward and back dash with attack after 1st frame just fine

 

Ehm...I discoverd I made a stupid error while recording my test: I didn't make whiff the moves while recording screwing up all my buffers

The fun part is that I always write in my tests to be sure to make whiff everything while recording.
And a few hours before I even made fun of Kingo that did the same exact mistake

I run another test, this time I whiffed everything , and got that you can cancel the back dash into an attack (that doesn't have 44 as inputs) after 1 frame.

I'm sorry guys (Cozby, Unicorn and Kingo) for this mess but hey as someone always say to me...I'm a scrub so please forgive me

I will edited the posts above adding (I will not delete anything) this info.

 

I just run your test and I can confirm it (it's quite esay even with a stick)

This test is good but from this you cannot get a single value because it can only be deduced that the time needed to cancel the back dash into an attack is =< 1 frame.

You need something else, while using the same recording, to delimit the lower part of the range.

 

So basically, the way your test works is right but because of one parameter, the results were screwed. Cozby were right but the way he tested things were confusing ( at least for me ), I ( still ) didn't understand the point of his "tests"...
And Unicorn found the "right" test.
We are all guilty here.

Now I hope we're at the same page.

PS: I still don't understand the part where Vanessa (44)3P punishes Goh's blocked 3KP to test that 1fr BD cancel.

 

as it is now it's not definitive but still a good cross proof

The important thing is that in the end we sorted it out

Mmmh...I don't understand too.
3KP is a mid-mid string that's -10 on block so I can't uderstand the purpose of 44 before 3P.
Anyway we got our answer so we can move on discovering new amazing stuff

 

Just run the same test with -15 low move (LIon's 43p will do). You will see that it is not possible to hit him with 3P for RC by any mean, both with 66 and 44.

With this second test, you will determinate that you can cancel your 44 or 66 dash in >0 frames.
As the result of the previous test was =<1, if you put them together, you will end up with 1 frame.