NRS.. PLEASE fix the input drop bug!!!

[webreg]

I've done some additional tests and I think I know whats up with special attacks. As you know, the game checks the state of all the inputs every 30 ms. Let us take Tom's clone for example. Down, back, 1. This is the order of the commands. You must first press down, then back, then 1. If you do it like that, you are not going to drop a special attack. If you ever press a direction and a command at about the same time you play roulette and the result seems to sometimes randomly screw up. Let me explain:

You can't press two buttons at the exactly same time reliably. You have to be accurate to about 1 millisecond. Neither the brain, nor the hands, not even the controller with his mechanic buttons is able to do that. So if you press them seemingly at the same time, they will be at least 5 ms more likely 10 ms or more apart from each other and you have no real control which one you press earlier. For you it seems identical, the PS3 disagrees. For the sake of simplicity let us assume they are pressed within 10 ms of each other.

Back to our 30ms interval at which the commands are evaluated. If you press "back" and "1" at the same time to do a clone you sometimes get the "1" in front of the "back". This is no problem as long as both inputs happen in the same collection cycle of 30ms but if it so happens, that the two inputs get separated into two evaluation cycles, the command will not work.

00ms:Press down
30ms:Collect commands
45ms:Press 1 (player thinks he pressed 1 + back at the same time)
55ms:Press back (player thinks he pressed 1 + back at the same time)
60ms: Collect commands
=clone

10ms:Press down
30ms:Collect commands
55ms:Press 1 (player thinks he pressed 1 + back at the same time)
60ms: Collect commands
65ms:Press back (player thinks he pressed 1 + back at the same time)
90ms: Collect commands
=no clone but something else

In both scenarios the player executed his command EXACTLY the same way but with a different results.

Summary:
- If you press a direction and a button at the same time they will be in a sequence and you can't reliably control the order.
- If the commands are in the wrong order and happen to get separated during evaluation, the command will not work.
- It is a combination of failed execution and "bad luck".
- Make sure you always press buttons after each other.

Theoretical calculation: You are a good player and can press both buttons within 5 ms of each other. Your muscle memory allows you to reliably press them in the right order 9 out of 10 times. Now, the chance of the 5ms falling over the edge of a 30ms cycle are about every fifth time. So about every 50 executions you might screw up your clone or whatever.

Show me a video of training mode, where the commands are listed correctly and the special move didn't work. In that case I shall be instantly convinced and will fall on my knees in awe.

 

[GGA soonk]

I've done some additional tests and I think I know whats up with special attacks. As you know, the game checks the state of all the inputs every 30 ms. Let us take Tom's clone for example. Down, back, 1. This is the order of the commands. You must first press down, then back, then 1. If you do it like that, you are not going to drop a special attack. If you ever press a direction and a command at about the same time you play roulette and the result seems to sometimes randomly screw up. Let me explain:

You can't press two buttons at the exactly same time reliably. You have to be accurate to about 1 millisecond. Neither the brain, nor the hands, not even the controller with his mechanic buttons is able to do that. So if you press them seemingly at the same time, they will be at least 5 ms more likely 10 ms or more apart from each other and you have no real control which one you press earlier. For you it seems identical, the PS3 disagrees. For the sake of simplicity let us assume they are pressed within 10 ms of each other.

Back to our 30ms interval at which the commands are evaluated. If you press "back" and "1" at the same time to do a clone you sometimes get the "1" in front of the "back". This is no problem as long as both inputs happen in the same collection cycle of 30ms but if it so happens, that the two inputs get separated into two evaluation cycles, the command will not work.

00ms:Press down
30ms:Collect commands
45ms:Press 1 (player thinks he pressed 1 + back at the same time)
55ms:Press back (player thinks he pressed 1 + back at the same time)
60ms: Collect commands
=clone

10ms:Press down
30ms:Collect commands
55ms:Press 1 (player thinks he pressed 1 + back at the same time)
60ms: Collect commands
65ms:Press back (player thinks he pressed 1 + back at the same time)
90ms: Collect commands
=no clone but something else

In both scenarios the player executed his command EXACTLY the same way but with a different results.

Summary:
- If you press a direction and a button at the same time they will be in a sequence and you can't reliably control the order.
- If the commands are in the wrong order and happen to get separated during evaluation, the command will not work.
- It is a combination of failed execution and "bad luck".
- Make sure you always press buttons after each other.

Theoretical calculation: You are a good player and can press both buttons within 5 ms of each other. Your muscle memory allows you to reliably press them in the right order 9 out of 10 times. Now, the chance of the 5ms falling over the edge of a 30ms cycle are about every fifth time. So about every 50 executions you might screw up your clone or whatever.

So the input bug isn't causing dropped specials, a mix of human execution(not necessarily error) and the way the game reads inputs is. That about right?

 

[webreg]

again, MKDC i used GL.. its the same engine.. i NEVER.. i mean 0.0% of the time got a standing 3 instead of DF3 or 1/B1 instead of DB1.

Would you mind explaining to me how I walk up and down through the 3D environment in MK9? How can I run with my opponent through a wall, go into an upclose infight or activate rage mode? Or to quote from an interview:

Mortal Kombat runs on a heavily-modified version of the Unreal Engine 3, similar to the engine used by its predecessor, Mortal Kombat vs. DC Universe.

 

[webreg]

So the input bug isn't causing dropped specials, a mix of human execution(not necessarily error) and the way the game reads inputs is. That about right?

Probably but I can't verify this because the theory contains random elements of human inaccuracy which I can't reproduce to be statistical relevant. If someone claims to to be able to reproduce this fairly regularly then I would very much welcome a video from training move with activated input display. I'm not a very good player so my own results (not based on computer assistance) are of no value.

However, the "input bug" that exists with down pokes definitely doesn't happen with special moves. This I tested extensively. There is a possibility for another, maybe related problem however, just not for the exactly same one. As I said, the know problem has to do with the game "ignoring/forgetting/reseting" the characters crouch status, not the input itself. The crouching state is not a prerequisite for special moves, only the down input itself and that works in any case. Come to think of it the Roll-Attack of Mileena could be affected but I doubt it because I suppose crouching is irrelevant only the input itself counts.

 

[Hitoshura]

ima quote someone if I may...

"INPUT BUG!!!!!"

 

[G4S MinotaurLord]

Theoretical calculation: You are a good player and can press both buttons within 5 ms of each other. Your muscle memory allows you to reliably press them in the right order 9 out of 10 times. Now, the chance of the 5ms falling over the edge of a 30ms cycle are about every fifth time. So about every 50 executions you might screw up your clone or whatever.

If this is the case, i congratulate you on pinpointing the issue. Would the ramifications of this be that a player could input moves slightly slower (i.e., not trying to press the last direction at the exact same time as the 1,2,3,4 button command) to have full move reliability? That you would be trading speed for accuracy? I don't think that's a healthy way to think about or play the game, but it would be a nice tidbit of information to know. It would also help Netherrealm address the problem directly in MK10 (theoretically).

Tom (or anyone else) - would you stagger your inputs (by pressing the x a little bit after the forward for freeze, for example) if you knew that you would be input-bug free?

 

[GUTTY333]

My thoughts on the input bug.

 

[webreg]

Would the ramifications of this be that a player could input moves slightly slower (i.e., not trying to press the last direction at the exact same time as the 1,2,3,4 button command) to have full move reliability? That you would be trading speed for accuracy?

If my tests and assumptions on the matter are correct, yes, absolutely. If you would really have to sacrifice speed is a matter of debate. Let me quote myself:

I suggest the method where you move your thumbs at the same time and speed but start with the thumb for the second input a bit further away.

 

[DanCock]

This existed before the negative edge change.

right i know this.. i dont get the response i wasnt saying it just started now, im just saying maybe from day one that negative edge could be causing the dropped inputs because its too busy saving other inputs so the recent inputs are getting ignored.. i honestly dont know just sharing ideas..

 

[GGA soonk]

right i know this.. i dont get the response i wasnt saying it just started now, im just saying maybe from day one that negative edge could be causing the dropped inputs because its too busy saving other inputs so the recent inputs are getting ignored.. i honestly dont know just sharing ideas..

What he's saying is that in the last patch they lowered the negative edge window and the input bug has remained the same, so it's unlikely negative edge is affecting the bug.

 

[Slips]

Found some new circumstances of when the glitch occurs. It seems anytime during a wake-up move this glitch becomes more prevalent. It becomes less of a notation of d~1, but more d+1.

I first noticed trying to duck jab Jax's wake-up dash punch. I knew I wasn't hitting down and then 1 which is the most known way to re-create the input glitch, but d+1, but kept getting a standing jab. Then I noticed that it happened on any wake-up attack he did, whether I was in his face or full screen away.

I tested to see if it was just jax but this happens with all characters. During wake-up attacks the input glitch kicks in hard.

The best way to see this for yourself is pick any character and have the cpu as Sindell. Put wake-up attacks set to 'on' and get the wake-up attack to where she flies in the air and suspends herself. During that whole time she's in the air, start doing d+1s. Half the time you'll get a duck jab, the other half you get a standing 1.

Hope this helps NRS find out wtf is going on with the madness.

 

[GGA soonk]

NRS knows what's going on, they just won't/can't fix it. Doesn't matter, deal with it until something better comes along.

 

[Slips]

NRS knows what's going on, they just won't/can't fix it. Doesn't matter, deal with it until something better comes along.

Is that something they've stated? I thought the whole point of this thread was to let them know so they could fix it.

 

[GGA soonk]

Is that something they've stated? I thought the whole point of this thread was to let them know so they could fix it.

Not officially, but yes.

 

[MKK hanzo]

Not officially, but yes.

I used to think like that in the Deception-Armaggedon era.

 

[GGA soonk]

I used to think like that in the Deception-Armaggedon era.

I don't know what that means.
Word of mouth is they aren't fixing it. Officially, we've let them know about the bug and they've said nothing. Take your pick.

 

[webreg]

Only Tom Brady claims it can't be fixed but he has never said/written where this info comes from so we have to assume it's his opinion.

 

[MKK hanzo]

I don't know what that means.
Word of mouth is they aren't fixing it. Officially, we've let them know about the bug and they've said nothing. Take your pick.

I mean that I "grew" accustomed to Midway shitting on their own work and dont caring about it. This patch thing is good but there are stuff from day one that they still havent adressed like the stance thing getting into combos lol.

 

[GGA soonk]

I see what you mean hanzo.

Webreg, I didn't hear it from TB, I heard it from some of the guys I spoke with at SB who heard it from the source. It's just the way it is. We appreciate your work on it, though.

 

[Shadowfire]

no, it is not my opinion.. it is the word i got. it cant be fixed so we should just deal with it and move on.

Nothing "can't be fixed". It's software. They're either choosing to fix it or they're not.

 

[Slips]

If it can't be fixed then the best thing we can do is find out how to do the bug in each of these scenarios so we know how to AVOID it.

We know when everything is neautral pressing down then hitting a button just after the bug has a good chance of kicking in.

I know during wake-up attacks (dont' know if its just special moves in general or not) d+input (d+1 hit at the same time) the bug kicks in a lot.

And after blocked strings it happens alot as well so we need to figure out the notation of the bug there as well.

Or at least we should.

 

[webreg]

I just saw the recorded stream from the console combat tournament. Brady was commentating and mentioned the "input bug" and then went on to play on stream. The two guys that were commentating had no clue and started using "input bug" to describe Kabals ability to do air blasts in fast succession and said Scarlet has that "input bug" as well because she can throw "instant daggers" very fast.

And this ladies and gentleman is, what happens, when you throw something about "bugs" out there without explaining it. It doesn't even matter anymore if there are ways to avoid it all that matters is that people know what they are talking about and this can only be achieved if sufficient knowledge is gained. Those two commentators have no experience or knowledge about the game so they can be forgiven but at least people who play MK9 should exactly know what it is about so that they can educate those who don't know. Otherwise people will start to trash the game because "it is full of bugs and unplayable".

 

[LEGEND]

way too many people are blaming the input bug in place of their own failed inputs now, Shits annoying (more so than the actual bug itself, which is incredibly rare)

 

[webreg]

dude, its over. we know what it is and why it is and what it effects and i got this from the horses mouth. it cannot be fixed so all we can do is play through it.

Yeah well, either you didn't read what I wrote or you are being a bit arrogant here. You mentioned it on stream to people who know nothing about it and they then commented ON YOUR MATCH, that you have an incredible hard time against Kabal because of it. And that doesn't bother you? Are you sure? Because the bottom line is, Sub Zero against Kabal is a perfectly fine matchup if not for the "input bug". At least that's what viewers were being told between the lines.

I know it won't be fixed and that's ok with me. No problem. But what we can do is either not mentioning it on streams permanently (as you do sir) or if we mention it, make sure we also explain it so that people watching know what it is about. So telling me to shut up about it while talking about it on streams seems a bit hypocritical, doesn't it?

 

[GodsLonelyman]

i am telling you that it happens ANYTIME down is involved for a fact. any special that has down in it gets effected. i am unsure about the F and B inputs as i havent tested it enough but i am positive on down.

its so random on specials, sometimes it almost never happens, sometimes it happens all the time.

again, MKDC i used GL.. its the same engine.. i NEVER.. i mean 0.0% of the time got a standing 3 instead of DF3 or 1/B1 instead of DB1. i find it hard to believe that now i magically miss the same inputs that i abused so much in the same fucking engine. if i have my own input execution issues why did i not have this problem on the same commands then?

for a fact, anything with down has a bug and this may extend to other inputs but i have not tested extensively on anything but down.

This. I've noticed lately when playing Noob when I input a shadow running man I get a standing 3 and when sometimes doing a shadow upknee as well. The upknee pisses me off because I eat a full combo off a jip.