DIY frames analysis

When I saw the tools that we have in the dojo (frame data, recording, etc) I started messing with them and having fun to come up with my own results.
Most of them are well known others not (i.e the frame analysis of the fuzzy guard and the crouch fuzzy guard).
I'm not a pro like Noodalls and everything I found is not tool assisted so there could be errors but I'd like to share and discuss them.

When recording try to make whiff every move you are inputting to avoid buffer problems and input your commands asap.

Low and Standing status frame analysis (for attacks)

1P Lion
2P Eileen

Record Eileen doing 9K, 2G_ (only a tap down, almost invisible)
With Lion, at blank point, block the mid kick leaving you +15, then do:

1) 1K (16 frames low kick)
2) 1P (17 frames low punch)
3) 3K (17 frames mid kick)
4) 4P (16 frames mid punch)

If done correctly you'll get:

1) low guard
2) low hit
3) standing guard
4) mid hit

From this we can find that by pressing down you are consider crouched (for blows but not for grabs, not fully crouched) from the first frame and you regain the standing position (in this case for the grabs you are always standing up) after an other 1 frame has passed from when you have left the standing position & put the stick back in neutral position.


Fuzzy Guard frames analysis

1P Wolf
2P Kage

Record with Kage 46K, fuzzy guard.
46K on block leaves Kage -9.

With Wolf do:

1) Burning Hammer (16 frames throw)
2) 6P (16 frames mid punch)
3) 4P (15 frames high punch)
4) 46P+G (15 frames catch)

and if you have done it correctly you will get these results:

1) Burning Hammer whiffs
2) standing guard
3) the high punch connects on his second active frame resulting in a 16 frame high punch guarded
4) the catch connects

Because the Burning Hammer starts in 16 frames and Kage is -9 after the blocked knee Kage can perform a fuzzy guard to avoid the grab but not the catch that is 1 frame faster.
Kage is also able to guard a 16 frame mid.
How could he be low and standing at the same frame?
The answer is linked to previous result in the "Low and Standing frame analysis (for attacks)".

There are 3 possible states:
- standing (grabs and high blows connect)
- semi crouching (high blows whiff while grabs connect)
- fully crouching (high blow and grabs whiff)

Suppose you are in one of these three states, guarding, and that an attack is going to connect the very next frame. You'll be able to guard any low/mid attack pressing 2/5 at the very last frame (obviuosly while guarding too).

A possible explanation could be explained by three cases (Burning Hammer/6P/4P).

Case 1
frame 0 - Wolf starts Burning Hammer
frame 9 - Kage is standing and inputs (commands are executed one frame later)
frame 10-15 - Kage is semi-crouching (high blows whiff, grabs connect) and still inputting (7 frames total)
frame 15 - No attack is going to connect the next frame
frame 16 - Kage is now inputting only but he is executing the frame 15 command () achieving a full-crouching position avoiding the Burning Hammer

Case 2
frame 0 - Wolf starts 16 frame mid pumch 6P
frame 9 - Kage is standing and inputs (commands are executed one frame later)
frame 10-15 - Kage is semi-crouching (high blows whiff, grabs connect) and still inputting (7 frames total)
frame 15 - An attack is going to connect the next frame
frame 16 - Kage is now inputting only and goes from semi-crouching to standing (15th frame has not been executed) and blocks the mid

Case 3
frame 0 - Wolf starts 15 frame high punch 4P
frame 9 - Kage is standing and inputs (commands are executed one frame later)
frame 10-14 - Kage is semi-crouching (high blows whiff, grabs connect) and still inputting (6 frames total) and the next frame the high punch becames active
frame 15 - Kage is semi-crouching, stil inputting (7 frames total) so there's no transition to standing position (needed only pressed) and the high punch whiffs on its first active frame but has a second active frame that could connect on the next frame
frame 16 - Kage is now inputting only and goes from semi-crouching to standing (15th frame has not been executed) and blocks the high punch on its second active frame


Crouch Fuzzy Guard frames analysis

1P Wolf
2P Jacky

Record Jacky doing 2K+GK,G,crouch fuzzy guard
2K+GK on hit leaves him -11 and G is to end the string.

With Wolf do:

1) Burning Hammer (16 frames)
2) 3K (18 frames mid kick)
3) 6K (17 frames knee)
4) K (17 frames high kick)

and if you have done it correctly you will get these results:

1)Burning Hammer whiffs
2)the mid kick is blocked
3)the knee hits
4)the high kick connects on its second active frame resulting in a 18 frame kick being blocked

Because the Burning Hammer starts in 16 frames and Jacky is -11 after the second hit of the Lighting Kick Jacky can perform a crouch fuzzy guard to avoid the grab.
To verify how much frames it takes to regain a standing position from this type of crouched position (fully crouched) we use 2 mid blow and 1 high.
The 17 frames knee hits while the 18 frames mid kick is blocked.
This means that you need (18-11=)7 frames to perform the fastest crouch fuzzy guard and knowing that we can crouch (fully crouched to make whiff the grab) in 5 frames we regain the standing position in (7-5=)2 frames.
The high kick is there to have an other proof of this value.
Wolf's K executes in 17 frames but with Jacky not having regained the standing position on the 17th frame (it's the 6th frame after the start of the crouch fuzzy) it will connect on the 18th frame (K's second active frame) when Jacky is finally standing up (it's the 7th frame from the start of the crouch fuzzy).

The crouch fuzzy guard does not have the ability to go from one of three states (standing, semi-crouching, fully crouching) to an other one because it's a type of dash cancel that has its own animation.

Failed Evade frame analysis

1P Kage
2P Kage

With 2P Kage record 2P, 2/8, G.
With 1P Kage eat the 2P (+4 on hit) and enter K+G (22 frames).
If done correclty you'll be able to block the mid kick after your failed evade meaning that you can guard after (4+22=)26 frames.
Swapping K+G with 6P+K (21 frames) you will not able to guard in time meaning that a failed evade animation lasts 25 frames.


Dash guard cancel frame analysis

1P Kage
2P Kage

With 2P Kage record K, 66G.
Input the G after the forward dash asap.
With 1P Kage block the K and immediately input P.
If done correclty 2P Kage will block the 1P Kage's P.
Being 2P Kage -7 when his K is blocked and 1P Kage's P executing in 11 frames this means that you can guard canel out of a forward dash after 4 frames.
If you swap K with 2K+G you'll be -8 on block and you will not able to guard cancel after the forward dash confirming the value just found.


Evade Forward Dash Guard frames analysis

1P Kage
2P Kage

Record 2P Kage doing 6P, 6P.
With 1P kage eat the first elbow that leavs +0 and asap enter 2/8, 66, G (or 2/8, G_, 66).
Being +0 the 1P Kage's evade fails (you must enter the evade AFTER the start of the opposing's move).
If done correctly you'll be able to guard the second elbow.
The elbow starts in 14 frames so we can do an EFDG in 14 frames.
Swapping the second 6P with 4P (13 frames) you will not able to guard in time confirming the value just found.


Evade Crouch Dash Guard frames analysis

1P Kage
2P Kage

Record 2P Kage doing 6P, P+K.
With 1P kage eat the first elbow that leavs +0 and asap enter 2/8, 33, G (or 2/8, G_, 33).
Being +0 the 1P Kage's evade fails (you must enter the evade AFTER the start of the opposing's move).
If done correctly you'll be able to guard the second blow.
The P+K starts in 17 frames so we can do an ECDG in 17 frames.
Swapping P+K with 3K (16 frames) you will not able to guard in time confirming the value just found.


Failed Evade Cancel frames analysis

An EFDG takes 14 frames.
A dash guard cancel takes 4 frames.
So the minimum time after which we can cancel out of a failed evade is 14-4=10 frames.

An ECDG takes 17 frames.
A crouch dash guard cancel takes 7 frames.
So the minimum time after which we can cancel out of a failed evade is 17-7=10 frames.

As check proof we can see that we found the same value in both cases.


Attack forward dash cancel frame analysis

The tricky part here is to make sure the forward dash comes out and use an attack that does not overlap with the dash input.

1P Kage
2P Kage

Record 2P Kage doing 3K, 66, 4P.
With the 1P Kage guard the 3K (-6 on block) then enter:

1) 66P+K (20 frames mid punch)
2) 6P+K (21 frames mid punch)

If you record the 2P input correctly you'll get:

1) 1P hits 2P
2) 2P hits 1P

From the second result we can find that the duration of 66, 4P < 15 frames (21-6).
While from the first result we get that the duration of 66,4P >= 14 frames (20-6).
Knowing that 66P+K does more damage then 4P (18 vs 17) in the first case 1P hits 2P due to priority.
So we can find that the minimum amount of frames to cancel out a forward dash with an attack is 14-13=1 frame.

Crouch backdash frame analysis

1P Kage
2P Kage

Set the cpu (2P Kage) to guard and throw.
Then with 1P Kage from the distance at the start of the dojo (and matches) do a little step forward (to gain more ground but letting the K whiff), KP (-4 o block) and then input:

1) 11 (CBD)
2) 44 (BD)

If done correclty we'll get:

1) the grabs whiffs
2) the throw connects

Because the BD covers more space than the CBD we can conclude that in the first case the grab whiffs not due to the distance and that CBD takes 6 (10-4) frames to gain a fully crouched position.

Now record 2P Kage doing 6P (-4 onn block) and then enter 11, G.
With 1P Kage guard the elbow ad then do:

1) 6P+K (21 frames mid punch)
2) K+G (22 frames mid kick)

If done correctly you'll get:

1) 2P Kage is hit by the punch
2) 2P Kage guards the kick on its second active (frame resulting in a 23 frames kick)

Because K+G hits on its second active frame when the opposing character is not in a fully standing position (i.e. semi crouching/ fully crouching) we deduce that after a CBD we regain a fully standing position and able to guard after 19 frames (23-4) has passed making the CBD's animation lasts 18 frames.

With these value we can calculate that 13 (19-6) frames are needed to pass from a fully crouching position to fully standing one.

 

One thing I'd like to discuss is the crouch fuzzy guard.
In Leonard McCoy's blog about fuzzy guard option select

http://gohnotes.wordpress.com/2012/07/11/tech-guide-fuzzy-guard-and-its-hidden-attack-option-select/

is said that after getting in the crouch position (fully crouch to make the grab whiff) you regain the standing position in 0 frames which is not consinstant with my result of 2 frames instead.
Any ideas?

 

You are totally right.
Doing some more work I found that in the tests about the standing/low status and the fuzzy guard (the crouch fuzzy guard results seem correct) I had a delay of 1 frame due to a not perfect execution (it's been a pain in the ass compared to the other tests).
When I went to update the first post you have already pointed it
It turned out it seems to be a 3 types of states (standing, semi-crouching, full-crouching).
Now everything seem consistant but I'd like hear other opinions.

 

Just a consideration for who doesn't want to go through the WOT of the first post but it seems there is an option select in the fuzzy guard: at the seventh frame you can avoid grabs and guard mids too.
Obviously this doesn't come up because usually you fuzzy guard from -1 to -3 to avoid 10 frames grabs and guard 13-14 frames mid.
So it's quite useless nonetheless it's interesting to know.

A little bump for my previous question:

 

On the Evade Forward Dash Guard Frame analysis, at the end you meant 4p (13 frames) not 3p(no 3p for kage)

 

Maaan I always write it wrong...dunno why..
thanks for the heads up Ytpme

 

Perfect Tech Roll (PTR) frame analysis

Thanks to programmable controller and 60 fps videos we already know that the window to perform a just ukemi is 5 frames and that as soon as you get up there are 3 frames where you are able only to block/not block high/low.

Using the tools of the dojoi I was able to obtain those numbers plus some infos about the hurtbox of who is rising.

PTR Hurtbox analysis

1P Kage
2P Kage

Set the CPU on PTR (both directions are good).
1P Kage performs any juggle with the ! ender.
After that if 1P Kage immediatly does:

1) (22 frames low kick)
2) (22 frames mid kick)
3) (23 frames mid kick)
4) (25 frames mid kick)
5) (24 frames high kick)

then we'll get:

1) hits on the second active frame (23)
2) hits on the third active frame (24)
3) hits on the second active frame (24)
4) hits on the first active frame (25)
5) hits on the third active frame (26)

from these values we can deduce that the hurtbox starts to appear:

- low on the first frame
- mid from the second frame
- high from the fourth frame

Frame analysis of the inability window after a PTR

1P Kage
2P Kage

Set the CPU on PTR towards 1P Kage' belly.
1P Kage does any juggle with the ! ender.
After that if 1P Kageimmediatly does:

1) (22 frames low kick)
2) (22 frames mid kick)
3) (23 frames mid kick)
4) (25 frames mid kick)
5) (24 frames high kick)

then we'll get:

1) hits on the second active frame (23)
2) hits on the third active frame (24)
3) hits on the second active frame (24)
4) hits on the first active frame (25)
5) succesful evade

Because the dummy performs the PTR on the very first possible frame we can deduce that for 3 frames, after the hurtbox has started to reappear, you can only block/not block high/low.

Just Ukemi window frame analysis

1P Kage
2P Kage

Record 2P Kage doing >!> (whiffed, 27 frames)> (13 frames).
With 1P Kage try to just ukemi, after the knockdown, on the very last possible frame and then do (11 frames).
If done correctly (it could take several attempts) 2P Kage will counter hit 1P Kage (priority).
Then if we swap the with (14 frames) we'll "always" have (until somebody will find a counter example ) 1P Kage hitting 2P Kage.
To obtain the window for a just ukemi we have to calcualte:

27 (whiffed P) + 13 (4P) = 22 (knockdown) + 11 (P) + X (window for a just ukemi) + 3 (frames of the inability widow) - 1 (value to normalize the counter for the just ukemi window)

X = 5

So the window to perform a just ukemi is 5 frames.

 

So does that mean using a meaty attack with 4 active frames wont' be enough to cover all the tech roll timings possible then? Or are there 5 frames for Just ukemi and then 3 frames after that where you can "lazy" tech roll?

 

You need at least 3 active frames (and perfect timing) to hit last frame of the "lockup" frames after earliest TR possible and 1st "lockup" frame after slowest possible ukemi.

 

EDIT

correct an error on all "charts" where the high portion of the hurtbox would appear on the 5th frame instead of the 4th.
It doesn't change anything on the analysis, everything still stand as before.

First of all I want to remid that there's a big difference between a quick rise (QR) in place and a tech roll: the frames of limited options for a QR are 20!

Now the number of active frames you depends on your attack's level (low/mid/high).

You have five frames where you can just ukemi and for each of this five cases when your hurtbox starts to reappear you have 3 frames where your options are limited to just block/not block high/low.
Suppose that your opponent will Tech Roll ( TR ) on the very first frame possible, then he will have limited options ( L ) for the first three frames when he'll wake up, with his hurtbox starting to reappear, and after that he will be free to do anything ( F ).
Linked to this we need also to keep in mind which part of the hurtbox will start to reappear first and which will follow.

Here's a "chart" (the best I was able to get... ) to sum up the situation:

Frames | 1 | 2 | 3 | 4 |
TR 1st fr | L | L | L | F |
Hurtbox |low|mid|mid|high|

This means that if your move is not active at least on the third frame after his hurtbox had reappeared then he can evade it.

Now, keeping in mind the hurtbox process, let's add to the "chart" all the other four cases ( x = invulnerable ) :

Frames | 1 | 2| 3| 4 | 5| 6 | 7| 8|
TR 1st fr | L | L | L | F |
TR 2nd fr | x | L | L | L | F |
TR 3rd fr | x | x | L | L | L | F |
TR 4th fr | x | x | x | L | L | L | F |
TR 5th fr | x | x | x | x | L | L | L | F |

So considering all the five cases for a just ukemi we have a possible window of limited options of seven frames.

At this point will enter in play the level of your attack.

The purpose of a meaty is to limit your opponent's options in order to get a better mixup in terms of total risk/total reward.

Obviously it's better to not wast active frames of your moves of choice covering the first two frames of limited options of the opponent.

In virtua fighter the best attack level is the mid and that's the one the will limit the most of our oppoent's options, so will start first with that.

Now it's when enters in play the reapearing process of the hurtbox after a just ukemi.
Because we used a mid then the first possible frame that our move will connect is on the second of the three of each of the five possible cases of just ukemi.
The last one is the third.

To shorten the number of active frames necessary to avoid that the opponent could evade our blow (if it's linear or semicircular) our mid has to be active:

1) on the second frame, of the first of the five possible cases of just ukemi, when the mid part of the opponent's hurtbox will reappear; i.e. the 3rd one of the seven total

2) on the first frame, of the last of the five possible cases of just ukemi, when the mid part of the opponent's hurtbox will reappear; i.e. the 6th one of the seven total

Frames | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
TR 1st fr | L | L | L | F |
Hurtbox |low|mid|mid|high|
TR 5th fr | x | x | x | x | L | L | L | F |
Hurtbox | x | x | x | x |low|mid|mid|high|

So in case of a mid blow, perfectly timed, we need 4 active frames to cover all of the five possbile just ukemi.

Now let's look at the case of low meaty.

For this case we need to look in deatil at the reappearing of the three hurtbox's parts (low/mid/high).
Take a look a this chart ( + = present, - = absent )

Frames | 1 | 2 | 3 | 4
Low part | + | + | + | +
Mid part | - | + | + | +
High part | - | - | - | +

To shorten the number of active frames necessary to avoid that the opponent could evade our blow (if it's linear or semicircular) our low has to be active:

1) on the last frame, of the first of the five possible cases of just ukemi, when the low part of the opponent's hurtbox will reappear; i.e. the 3rd one of the seven total

2) on the first frame, of the last of the five possible cases of just ukemi, when the low part of the opponent's hurtbox will reappear; i.e. the 5th one of the seven total

Frames | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
TR 1st fr | L | L | L | F
Hurtbox |low|mid|mid|high|
TR 5th fr | x | x | x | x | L | L | L | F |
Hurtbox | x | x | x | x |low|mid|mid|high|

So in case of a low blow, perfectly timed, we need 3 active frames to cover all of the five possbile just ukemi.

Now the case of a high meaty.

Because the high portion of the hurtbox will reappear on the 4th frame corresponding at the first frame of when our opponent will regain all his options (i.e. evading/attacking) there can not be a high meaty (only for TR; on QR you can do a high meaty).

Keep in mind that beside doing attacks that goes under highs immediatly our opponent can also simply duck it.


Then there's the special case of the throws.

Even if you timed a throw leaving the opponent with not enough frames to avoid by ducking with a simple or a the opponent can make the throw whiff by simply crouching, and stay crouched, before the first frame of when he'll regain all his options.

The only case when the opponent can not to do anything except yutori to avoid a throw is for a well timed low throw during a tech roll ukemi.

 

Ah, now I understand why Kage's is meaty whatever your tech roll is perfect or not.

Nice write up Mr the Scientist.

Spoiler

but you're still my bitch

 

lol @ you two.

 

Low Throw frame analysis

Aoi vs Kage

Record Kage doing P > 33G (you need to do it asap).
With Aoi after getting hit by P (-5) immediately do a low throw.
If done correctly the low throw will connect.
Because Aoi's low throw will start as soon as Kage reaches his fully crouching positing at the 5th frame of his cfg we can get that a low throw' startup is >= (5-5) = 0

Record Kage doing P > 1P > 33G (you need to do it asap).
With Aoi block the first P (only there for timing purpose) than mash something to get hit by 1P on CH leaving you at -6 and then do a low throw.
If done correctly the low throw will miss.

From this we get that a low throw will not connect on the 6th frame of a cfg.
Combining this result with the previous one we get that a buffered low throw' startup is 0 frames.

Record Kage doing 2P+K > 33G.
With Aoi block the mid punch (-5) and then do a low throw.
If done correctly the low throw will connect.

Record Kage doing 3K > 33G.
With Aoi block the side kick (-6) and then do a low throw.
If done correctly the low throw will not connect.

From these two examples we can only deduce that low throws could have:
- 5 active frames & connect even on a not fully crouching opponent OR
- 10 active frames & connect only on a fully crouching opponent.

To solve this undetermined point we need some more data.

Aoi vs Aoi
Record Aoi doing PK > low throw
With Aoi get hit by PK (-4) and then immediately crouch (2_G).
If done correctly the low throw will not connect.
Because a simple fuzzy guard takes 7 frames to reach a fully crouching position and Aoi will only be fully crouching after (4 + 7 =) 11 frames, one frame more of the largest possible window of 10 active frames, we get that a low throw will not connect on a not fully crouching opponent.

Aoi vs Aoi
Record Aoi doing P+K > 2_G.
Play it and with Aoi block the mid punch (+3) and then do immediately a low throw.
If done correctly the low throw will connect.
Keeping in mind that a low throw will not connect on a not fully crouching opponent from this example we get that because the first frame that the dummy Aoi will reach a fully crouching position is at the 10th frame of Aoi's low throw the total active frames of a low throw are 10.


Total animation length.

Aoi vs Kage
Record Kage doing 3K on block > 9K+G (30 frames)
With Aoi fter blocking the side kick (+6) do a low throw and then guard.
If done correctly you will be unable to guard Kage's attack.

Record Kage doing K > 9K+G (30 frames
With Aoi after blocking the high kick (+7) do a low throw and then guard.
If done correctly you will be able to guard Kage's attack.
So the total animation frames of a low throw is = 6 + 30 = 36 frames.

So in the end a low throw
- starts in 0 frames if buffered
- has 10 active frames (plus the initial frame 0)
- connects only on a fully crouching opponent
- has a total animation of 36 frames

 

All this infos are really cool and usefull. However I have a question about the oki setups. I'm sure many of you know that there are plenty of ways to setup low throws when someone tech roll on the side. However it seems that the low throw doesn't enter as a "meaty" since it should have 10 active frames. Is there something else involved here or am I missing something?

 

Hey, thanks for the work, bi...buddy.
It took me a while to understand that part:

Then I realized the high kick is -7

Btw, what DIY stands for ?

Spoiler

My godel, I would have hated to have you as an annoying math teacher

 

I believe it is because you are throw-invincible during TR recovery and once this recovery ends, if you buffer any attack/movement, you will "cancel" your crouching state instantly.

At least this is how I understand TR works

 

You just anticipate what I'm going to study right now.
The problem about explaining how it works is that it's a very particular situation (froced crouching at varoius frames disadvantages) and find enough cases to study will be tough.
My feeling is that even if the low throw starts in 0 frames there are still 10 frames where the opponent can attack out of it.

don't worry mash...mate <3
next time I will a simpler version for your impaired understanding capabilities

Do It Yourself

Oh boy, if I have been your math teacher you would have not just hated me my dear little annoying scrub

 

Great stuff! I'll go for the Kage VS wolf stuff just for fun in dojo.

Let's do a "frame analysis" of this post ; what does it means kingo?
Your incapacity to use your 10 fingers and your brain at the same time...
-Like translating DIY and typing on your keyboard.
-Using tools to repair your controller without blowing all the other stuff inside.
-Playing VF without mashing?

Ben 1 - Kingo 0

 

Ok I was quite lucky to find the right cases to prove my above theory.

Again, when recording something make everything whiff (do some backdashes at the beginning of your recordings)

Wolf vs Shun

Record Shun doing 1P > PPP (just mash jabs after 1P)
With Wolf block standing the special low leaving you at +10 then do a low throw.
If everything was done correctly Shun should't be able to mash his way out of the low throw.

Aoi vs Aoi

Record Aoi doing 1P > G_
With Aoi block the low punch (+9) and during the blockstun let go the stick (or whatever are you using) to the netrual position and asap input a low throw.
If everything was done correctly the low throw should connect.

Record Aoi doing 1P > PPP (just mash jabs after 1P)
With Aoi block the low punch (+9) and during the blockstun let go the stick (or whatever are you using) to the netrual position and input asap a low throw.
If everything was done correctly the dummy Aoi should able to mash his way out of the low throw.

From these three cases we get that even if a low throw starts in 0 frames it's not guaranteed if there are less than 10 frames of advantage because the opponent can avoid it by throwing out an attack.

Basically the 0 frames startup and the 10 active frames are something "necessery" for beating fuzzy & crouch fuzzy guard (in VF staying too much crouching isn't a very good idea...RIGHT KINGO?!?!?!? ), their animation is 2 frames longer (36) compared to the standing ones (34) but to be "guaranteed" they they need at least 10 frames of advantage like for the standing ones.


Gotta look if it's possible to make a perfectly timed low throw on a tech roll where the opponent can not make his way out by attacking.
But for this guys I need some help by those of you who use charachters with low throws: I need your low throw setups on TR possibly not involving doing something based on visual clues (when the opponents start to TR, etc...)
Whiff moves, 1 frame kill, KG cancel, ECD > attack etc are more than welcome.