DIY frames analysis

You can just watch the framedisplay (the advanced one) to rule that out.

 

I think you are doing something wrong with your recording.

Pick Goh vs Sarah.
Set the cpu to "block and attack with a mid".
With Goh do 6P (really close to the dummy) > 44 and then try to guard.
You will get hit on CH.
There's almost no pushback after Goh's 6P.
The only mid you can backdash out of it is a perfectly buffered Akira's 66P.

Other charachters can back dash after their mid poke and make whiff the opponent mid poke.

sorry but he doesn't have the 6P string where he can chose to go back by inputting 4?

 

a move can change its total frames of animation when on hit/guard.
That's why if you record something while hittig the opponent than it can lead to strange stuff if you play it but this time you block.

The data we have about a move's total frames is only for a whiffed move.
If we can know the total frames of a blocked move that leaves +1 on block than we can use your test to demonstrate that you can back dash cancel into an attack after 1 frame if the total frames before the 2P is <= move's total frames on guard.

But there's that if.

 

That's what I found out, you must be at tip reach of 6P and not to close to your opponent.
Yes Akira' s got the 6P but it's a i17 ( which oddly seems quicker than goh's 4K ^^")

 

I meant using the input display to see/calculate on which frame you press 2p. This way we can rule out a manual delay, since we can identify the actual frame on which 2p was inputted (put in?).

I dont really wanna write everything down here cos im using my phone, but using lau I can use 2pk to get +1 and using a clean 2p will beat the opps 2p. Using 44 to delay will make us clash. To confirm the 1 frame delay is because of the 44, I input the 2p on the 36th frame, which gave me the aforementioned results.

 

I know what you meant.

Your test is good and very reasonably true.
Still from a theorical point of view it leaves some questions open.
It's not that I like to be a pain in the ass, it's just that I'm trying to use to calculate all these frames analysis stuff only tests that don't have any weaknesses, even the smallest ones.
The same happened with Unicorn: he found a test (for the very same thing) but wasn't a perfect demonstration; then he coupled it with an other test to make it bullet proof.

So about your test I totally got what you meant about the frame data in the adavance input display in the dojo.
Problem is that to calculate if you input your 2P before the animation's end you are using the value found in the frame data, here on VFDC, which is referred only to whiffed moves.
It's almost impossible that the animation of a blocked 2P+K could be (47-36=) 11 frames less than a whiffed one but we don't know for sure how many frames it is.
We are only 100% sure about the total animation's frames of a whiffed move.
So you need a way to exactly calculate how many frames is the total animation of that move when it's blocked.
With that value if you hit 2P before it and you got the trade than you got your bullet proof demonstration.

 

I see your point, but in that case you can calculate the point where the game doesnt accept your 2p input as buffered (so the input would be too early) and accept that as your 10 frame starting point, right? Might also be a way to find out how the total frames of connected moves change.

 

that's really cool!
nice (theorical) job

it's a trial and error process but in the end you will get the info you need.

For your test keep in mind that the input buffer is 12 frames

 

Right, 12, that makes sense actually. I'll give it a go right now.

 

So with Lau, what I found is that on the 37th frame after you input 2p+k and it hits the opps block, 2p will buffer. On the 36th it will not.

This would mean that I can safely buffer it until the 48th frame and still beat out 2p, and I should clash at frame 49. I have tested this and both assumptions are correct.

The command list tells me that 2p+k has 47 total frames, so I guess I should take a frame off to account for the frame in which the buttons were pressed to activate 2p+k. This means 2p+k does not lose total frames regardless of whether it connects or whiffs.

Now, for the 44 buffer, I have to input the first 4 before the 35th frame (I took a frame off, see above), and the second one afterwards, as we know a dash+command buffered perfectly inside 12 frames never loses a frame. Results are as predicted: backdash is cancellable after the first frame.

What is interesting though is that it seems not all moves have their total frames altered if they connect!

EDIT; did a quick check with lau 46p. You can buffer another attack from the 30th to the 41st frame. Meaning total frames are 41. Buffer an attack on the 29th frame doesnt work. Again this is consistent with the command list data. So now the question seems to become; which moves change their total frames on hit/guard?

 

I am really curious about this, because I never read before this page about total frames changing depending of the guard/hit/counterhit.

I always thought that was the hitstun that was varying to cover for the difference of advantage.

 

Kage's 2K is longer on whiff and shorter on block.
In this moment I don't exactly remember how much is the difference

 

God I love this thread.

 

I am sorry for ressurecting old thread, but actually I am in request for a frame analysis to be done yet I really do not know how to do it.
I am really interested in know how struggling out of stagger really works:
1) does 6K or P+K or anything similar counted as 1 or 2 inputs?
2) how about K~6K~K input or some others; for example G~PG~PKG~KG~G? How many inputs this is in reality worth of?
3) is for example 123 thrated input-wise same as 426?
4) is there struggling-wise any difference between 426 and 41236?
5) I presses something - lets say G - during the green-stick animation and I was holding it still when the stick changed to red. Does it count as an input into unstaggering-count or not?

 

Let's see here. Not sure if I can break it down properly but I did run some tests. I picked a light stagger with Wolf 62K then P. It appears it takes 4 inputs shake the stagger. So I can do stuff like GKPG or 6KPG

1 If you don't press them down exctly on the same frame then it counts as 2 but if you press them down on the same frame then it counts as one. I tested this by using a P+K bind and I couldn't block the jab with 3 inputs having P+K bind in there. 6~P+K~G did not work, 6~4~P+K~G did.

2 I guess you are asking if releasing directions or buttons counts? No, for example 65PG counts just 3 inputs and I get hit I still need 65KPG.

4, I recorded 87412 and I didn't duck the jab so it counts only 3 inputs for struggling in there. If however I do K~87412 I duck the jab so 4 inputs. Whether or not it counts it as 84x2x or 8x4x2 is a bit confusing for me because of the following one.

3. This one confuses me after testing. 2123 did no duck the jab even though I thought it should count 4 inputs. Neither did 32123 (I managed to hit straight at the diagonal for a couple of times). Adding a single button input to either one does duck it. I haven't got a clue here. I'm pretty sure it wasn't my timing because of multiple tests. Dunno anymore.

Oh, I'm also a bit curious now as to whether or not 6P on the same frame counts as one input or 2. I'm starting to think its just 1.

 

Thank you a lot @Combolammas !!!

The 2) was actually what you answered in your 1) post. I was asking if I hold 6, then add K and then release 6, how many inputs this will be. 1 or 2 or 3?

About the directions, I think you may try the 4) test with k~7412 and see if it ducks or not. From what you posted, it seems that 32123 is counted as 1x 2, 1x4 and 1x6 input (so 3 in total), what makes some weird sence as you never "release" the down button actually.

did you managed to test 5)?

 

Oh, I didn't notice that #5. I kinda did since I tried that #3 by starting it with holding 6 at the beginning at first but it didn't change a thing for it. 632123 didn't duck if the 6 was held before red stick. Adding a button input there did.

Also for #2 if you hold 6 down before red stick and press K you get 1 input and if you press 6 down during red stick it counts as an input and the following K counts as an input.

 

THX Myke, I was looking for something like it but I was checking only FS sdection, all dojo forum

 

ok one last question - is there any way how to learn how much inputs is needed for specific stagger; or how to calculate how much is needed?