Special Option Selects: Buffer Window, Command Interpreter and Special State

I'm in agreement with @Myke here.

 

I'm sorry for not having posted it already but I have been very busy and the video proof for this tech is a nightmare (I challange you all to do it yourself) but if this was the case I wouldn't have claimed to know an OS for this purpose.

Here's the proof it's not like you wrote:

 

Indeed. The until elaborate further was the part I was agreeing with aswell.

 

INSTRUCTIONS FOR TRICKIES

Spoiler

DON'T BOTHER
TOO HARD

The original version that is found in the video posted by Nero (and than by Myke in this thread) follow these rules:

1. Perform a Fall Recovery
   
2. During the input buffer you input a move
3. You start holding before the Fall Recovery have ended

This is the simplest form for this OS but it's limited to only moves

So in order to get around this limitation I reverse engineered it and found that the general rules for this OS are:

1. Perform a Fall Recovery
   
2. During the input buffer you input your combination of buttons
   
3. You start holding before the Fall Recovery have ended

This version of the OS works with every buttons but it doesn't work with directional inputs (evade, dash, crouch dash etc).

For having it works with directional inputs too (evade, dash, crouch dash etc) you need to hold a combination of (i.e. ) in order to get the movement option otherwise the single button will override it:

1. Perform a Fall Recovery
   
2. During the input buffer you input your combination of buttons or any directional inputs (evade, dash, crouch dash etc)
   
3. You start holding before the Fall Recovery have ended

The downside of this version is that is not reliable due to strict timing.

Infact the steps have to be perfomed in less than 9 frames (12 frames of buffer minus 3 frames where you can only guard) while starting to input the first part on a visual clue of the Fall Recovery's animation.

So in reality you have way less than 9 frames (most likely 4-5 frames) making it quite strict and being based on a visual clue it becomes quite unreliable (compared to the other versions).


For those who wants the complete version of this OS you can add a LTE at the end:

1. Perform a Fall Recovery
   
2. During the input buffer you input your combination of buttons or any directional inputs (evade, dash, crouch dash etc)
   
3. You start holding before the Fall Recovery have ended
4. After the 3 frames where you can only guard has been passed let go of while still holding
5. Immediately after press and hold // to perform a LTE

TL;DR

Spoiler

It's strict and unreliable compared to the other verrsions.

How to perform it:

1. Perform a Fall Recovery
   
2. During the input buffer you input your combination of buttons or any directional inputs (evade, dash, crouch dash etc)
   
3. You start holding before the Fall Recovery have ended
4. After the 3 frames where you can only guard has been passed let go of while still holding
5. Immediately after press and hold // to perform a LTE

 

fyi I was able to pull it off about 80 percent of the time after some practice. Not too hard overall.

 

As you can see just having the inputs doesn't help much because it's so unreliable.

And being so unreliable it could led to think that it doesn't work.

So in order to demonstrate that it works you need to record the dummy doing this version of the OS.


I challange you all recording the dummy doing:

1. input // in a five frames window in order to perform a perfect ukemi after a knockdown of your choice
   
2. input your combination of buttons during a window that you cannot relate to any visual clue in a 2-3 frames window
   
3. start holding in a 2-3 frames window and before the Fall Recovery have ended

In Nero's video you can see the inputs (as I wrote) but you cannot be sure the OS has been proven because the timing for all the steps are not the same (frame wise) in the two cases.

Without all of this (the video proof and the explanation) the discussion could have been a mess...and in fact it has been.

And to top it off it is needed a summary for which OS is better to use depending on the situation.


As you can see all these processes do take a lot of time and having been quite busy in this last period I had some difficulties finding time to dedicate to this project.

Still I'm sorry if I kept you waiting.

 

You thinking that you're able to pull it off 80% of the time it means you didn't fully understand how it works.

And most important you don't know which systematic error is affecting the result of your data set.



And no disrespect at all but it was you that said personally to me that the OS in the OP was hard when in reality I told you it was quite easy.

Having said that after I studied this version of the OS for quite some time and discovered why it's objectively strict and unreliable you come up with "Not too hard overall"?

ok...

 

I mean, I won't be using this in a fight even with an 80% effectiveness rate. I never said the P+K OS was OP. We literally haven't talked about it. Don't put words in my mouth. The OP tech was hard to do until I practiced it and it became easier. Let's not rehash an old already resolved conversation.

 

wut?
lost you here

I just put your statement in a context to gauge its value.

Nothing more nothing less.

And you stil thinking you're getting it 80% better look into that or your next step could be joining the Flat Earth Society
jk

 

My point is, my ability to use the tech is not part of this thread discussion. It's irrelevant.

 

it's not your ability per se that we are discussing now but it's the reliability of the OS.

Your claim is heavily in contrast with my findings and the process behind this OS so taking your 80% success rate as true it means you're marking my results as false.

And that's more than ok as long as you have a valid proof.

But without a proper proof to corroborate it it's just a personal opinion.

You're more than welcome to share your thoughts but keep in mind that this thread it's meant to be an "objectively and proofed thread" and not an "opinion thread" with nothing solid behind those opinions.

I want to stress out that I don't have any beef with you and I don't want to drag this any longer.

 

I thought the point of the thread was how to do the thing. Reliability isn't really the question here. Simply, how is it done and what are the parameters to do it.

I too with not to have beef.

 

The value and usage of an OS is like any other technique in a fighting game and it's based on:

1. Situations during which you can use it
   
2. level of execution required
3. risk/reward
4. reliability (not skill related)

Take for example a reversal in SF3 Third Strike, the window for it it's 2 frames.

In SF4 instead the window for a reversal is 5 frames.


Let's take a look at the differences in a reversal shoryuken in these two games.

1. the usage is the same, on wakeup as an invincible reversal
   
2. the motion to get a regular shoryuken is the same in both games (beside the motion to get an auto correct shoryuken on crossup in SF4) so no differences in level of execution required
3. with no meter to FADC in SF4 the risk/reward is the same in both games
4. the main difference is in the reliability due to the huge difference in the reversal window. You can practice how much you want but with a 2 frames window you cannot go over a certain success rate no matter how godlike it's your execution. That's why the same players didn't do as much reversal shoryuken in SF3 as compared to the amount they did in SF4.

So reliability is a core aspect of an OS.

 

I thought he meant overpowered indeed.

 

So can you actually enter command attacks using this tech? Anyway, I'm really have a hard time applying this. The window to enter everything seems quite small to me.