What's new

Video example of frame data changes on crouch vs. standing

BillStickers

Do not touch me again.

The video above demonstrates frame data inconsistencies depending on standing/crouching state. Contrary to popular belief, character size is not what causes these issues.

What does the video show?
The video shows D'vorah performing her b3 on Mileena, immediately followed by her standing 2. If done perfectly, there is no gap between these two normals on block (hence, no ability to armor).

When Mileena stands, the input data shows me trying to input her EX roll (and either remaining blocking or getting hit). This is the expected behavior.

When Mileena crouches, however, she is able to EX roll after blocking b3.

Why is this important?
The screenshot below shows that NRS developers are not aware of this issue. We need to inform NRS of any other instances of inconsistent frame data in hopes that these issues get resolved. I suspect that this is a larger issue with NRS's collision detection, and I'm hoping that serious consideration is given to reworking the system to remove variable block advantage, which seems to have existed in some form in MK9, Injustice, and MKX(L).

 

Akromaniac27

Ready to lose your head?
A better way to test this is setting the AI to just do a reversal....that way you're not going through a series of trials to be frame perfect. That, and do it between crouching blocking and standing. As far as I could tell, I genuinely haven't really seen this happen at all for a while.

You're using a high hitting move after a blocked move with push back. Because of it being a high AND travel time, crouching can and in this case, did, create a 1f gap to get out armor. Frame data as a whole is far from NRS's strong suit, like FAR, but the block advantage/disadvantage inconsistencies aren't really there since there are multiple variables, hit property being one of the bigger ones, present.

What I find stupid though is the whole, add a frame of advantage and/or taking away a frame of disadvantage if the opponent does anything but a reversal. As well as non existent recovery frames, especially on whiffed jump ins or air attacks in general
 
@BillStickers I think to prove that character size doesn't matter here you have to give an example of a blockstring on at least normal hitbox character (unlike Milena).
I know for sure that with Quan Chi prior to XL patch when I performed his old Rune Trap (which was basicly connecting a 10 frame 1 after a +14 Ex Rune on block) on low blocking small hitbox characters s1 wouldn't connect as it is a high. Although highs do connect if it is a true blockstring. It wasn't a frame data issue because when you did a small Run cancel before the s1 after the EX Rune it would jail. Run cancel xx 1 takes more frames than just s1 but it jailed, so the issue was the hitbox indeed.
Is D'Vorah's s2 a high???
 

BillStickers

Do not touch me again.
A better way to test this is setting the AI to just do a reversal....that way you're not going through a series of trials to be frame perfect. That, and do it between crouching blocking and standing. As far as I could tell, I genuinely haven't really seen this happen at all for a while.

You're using a high hitting move after a blocked move with push back. Because of it being a high AND travel time, crouching can and in this case, did, create a 1f gap to get out armor. Frame data as a whole is far from NRS's strong suit, like FAR, but the block advantage/disadvantage inconsistencies aren't really there since there are multiple variables, hit property being one of the bigger ones, present.

What I find stupid though is the whole, add a frame of advantage and/or taking away a frame of disadvantage if the opponent does anything but a reversal. As well as non existent recovery frames, especially on whiffed jump ins or air attacks in general
D'vorah's moves are recorded in this video; I'm controlling Mileena. Setting the AI to "reversal" is an inaccurate way of determining what will happen in an actual match.

Notice here that I attempt both frame-perfect and mashing reversals. If you don't think this exact scenario happens all the time (go to any MKX stream and hear people saying "WHAT?! I'm plus <x>, how are you getting out??"), you're probably assuming it's another factor like lag or inaccurate execution.

The bigger point is that for the *same* attacks, with the *same* characters, in the *same* distance, different frame data results depending on standing or crouching.
 

16 Bit

Mash d+1~Cat Claws
Premium Supporter
NetherRealm Studios
The block advantage on b+3 isn't changing. The time it takes for D'Vorah's standing 2 to reach is changing. This could be caused by things like different pushback or different blocking staggering reactions. When standing Mileena goes through the stumble animation, which may be causing her hurtbox to be close enough to be hit by a faster active frame from D'Vorah's standing 2.
 

Trauma_and_Pain

Filthy Casual
The block advantage on b+3 isn't changing. The time it takes for D'Vorah's standing 2 to reach is changing. This could be caused by things like different pushback or different blocking staggering reactions. When standing Mileena goes through the stumble animation, which may be causing her hurtbox to be close enough to be hit by a faster active frame from D'Vorah's standing 2.
What this man said ^. Crouching changes the location of a character's hurtbox, which changes the moment blockstun begins relative to the attacker's recovery.
 

BillStickers

Do not touch me again.
When standing Mileena goes through the stumble animation, which may be causing her hurtbox to be close enough to be hit by a faster active frame from D'Vorah's standing 2.
This is the *cause* of the frame data changing; it's two parts of the same issue. Frame data itself is a derived metric that is partly based on character animations, so if Mileena's hurtbox causes a follow-up attack to hit a frame later, then the block advantage in crouching state is indeed one frame less.

To your point, though, we're in agreement that this *isn't* a rule built into the game engine, it's just a side effect of model-based collision regions used in the game. I think, then, that this is further proof that the model-based collision system is fundamentally broken and should be changed for future NRS games, especially with new features like dynamic block animations.
 

Trauma_and_Pain

Filthy Casual
To your point, though, we're in agreement that this *isn't* a rule built into the game engine, it's just a side effect of model-based collision regions used in the game. I think, then, that this is further proof that the model-based collision system is fundamentally broken and should be changed for future NRS games, especially with new features like dynamic block animations.
I don't see how this is proof that the system is broken. More like an opinion.

It's something that players have to account for on a situational basis, but I don't see anything wrong with it. I think it's fun uncovering the tech involved with all this. My character for example used to (not sure if it's still possible) be able to meaty the second hit in one of our strings by intentionally whiffing the first hit. The meaty enabled us to string into an attack we otherwise could not string into, much like your example in the video here.

I think if it allows overpowered examples, then those examples could be broken on an individual basis. I don't agree that the whole system is broken.
 

16 Bit

Mash d+1~Cat Claws
Premium Supporter
NetherRealm Studios
This is the *cause* of the frame data changing; it's two parts of the same issue. Frame data itself is a derived metric that is partly based on character animations, so if Mileena's hurtbox causes a follow-up attack to hit a frame later, then the block advantage in crouching state is indeed one frame less.

To your point, though, we're in agreement that this *isn't* a rule built into the game engine, it's just a side effect of model-based collision regions used in the game. I think, then, that this is further proof that the model-based collision system is fundamentally broken and should be changed for future NRS games, especially with new features like dynamic block animations.

No, the block advantage has not changed. The block advantage is the same, regardless of what D'Vorah or Mileena do.

If you have a move that has 10 frames of start up but you're spaced out far away the move could end up only hitting on the 12th frame. If you had previously used a move that left you at +10 then the opponent has a gap to armor through it. The move being +10 did not change.
 

Trauma_and_Pain

Filthy Casual
No, the block advantage has not changed. The block advantage is the same, regardless of what D'Vorah or Mileena do.

If you have a move that has 10 frames of start up but you're spaced out far away the move could end up only hitting on the 12th frame. If you had previously used a move that left you at +10 then the opponent has a gap to armor through it. The move being +10 did not change.
I think you're misunderstanding his claim a bit. As an example, if the move is +10, you could hit the opponent on active frame 1 when close, or active frame 4 when far away. If you hit them on active frame 4, the stun the opponent experiences begins 3 frames later and so ends 3 frames later, meaning you are at 3 frames more advantage if you use it that way, since the attacker's recovery doesn't change. That's does increase the block advantage and I think that's what the OP is complaining about.
 

16 Bit

Mash d+1~Cat Claws
Premium Supporter
NetherRealm Studios
Well I don't think he means the initial b+3 is hitting at a later frame because he says "if Mileena's hurtbox causes a follow-up attack to hit a frame later, then the block advantage in crouching state is indeed one frame less."

Regardless of if it's block stun timing being different because it's hitting on a later active frame or a follow up hitting on a later active frame there's nothing weird happening in this video.
 

MK FIGHTER

Knife Fight
Well I don't think he means the initial b+3 is hitting at a later frame because he says "if Mileena's hurtbox causes a follow-up attack to hit a frame later, then the block advantage in crouching state is indeed one frame less."

Regardless of if it's block stun timing being different because it's hitting on a later active frame or a follow up hitting on a later active frame there's nothing weird happening in this video.
Hi Steve. I've watched you on the kombat kast numerous times.
 
is this similar to smoke being able to link b1?(overhead) into smoke bomb on crouching opponents but not standing opponents.
 
This is the *cause* of the frame data changing; it's two parts of the same issue. Frame data itself is a derived metric that is partly based on character animations, so if Mileena's hurtbox causes a follow-up attack to hit a frame later, then the block advantage in crouching state is indeed one frame less.

To your point, though, we're in agreement that this *isn't* a rule built into the game engine, it's just a side effect of model-based collision regions used in the game. I think, then, that this is further proof that the model-based collision system is fundamentally broken and should be changed for future NRS games, especially with new features like dynamic block animations.
Frame data isn't changing! It is Melina's hurtbox. Try doing this on a crouch blocking Raiden. It will jail him because he has a larger hurtbox. Block advantage in crouching state isn't one frame less. But some moves connect later into their active frames on low hurtbox characters.