Input Bug: First results of a computer assisted analysis

[aj1701]

Sent it to Hec and Paulo

Actually I think that's even better Thanks!

Now, the big question is, is it a bug?Or was it thought that way by NRS for any reason whatsoever? <--- That's the interesting part.

Well, even if they did do it intentionally, its pretty clear users aren't happy, so "its not a bug its a feature" likely won't fly.

Besides I fail to see why they'd want it like that? D+1 instant = instance poke, D+delay+1 = poke, D+halfDelay + 1 = stand up and jab? I fail to see how that would be useful, especially given the window is so small.

 

[eskuAdradit0]

Actually I think that's even better Thanks!



Well, even if they did do it intentionally, its pretty clear users aren't happy, so "its not a bug its a feature" likely won't fly.

Besides I fail to see why they'd want it like that? D+1 instant = instance poke, D+delay+1 = poke, D+halfDelay + 1 = stand up and jab? I fail to see how that would be useful, especially given the window is so small.

Well I can't disagree with your statement since I'm not happy with it either as it affects my gameplay too. NRS should adequate this to the players who love this game the most, but maybe it isn't a bug and it was just made for...I don't know, helping the newbie-player? Just like it could be with the up button so you don't get random neutral jump + 4 instead of Mileena's U+4. Seeing that MK9 is a simplistic fighting game, it may have it's (good or bad) reasoning behind.

We still have to wait for NRS' response to this.

 

[Panque]

I thought NRS was aiming to make this a relevant fighter. If thats the case, this game should work like one. I don't care if I pressed the poke too fast, even if the character can't duck instantly, the character should take it's time to duck and then poke, not just randomly stop ducking in the middle of the animation and give me a standing attack, this is ridiculous! I hope Paulo sees this.

 

[aj1701]

I thought NRS was aiming to make this a relevant fighter. If thats the case, this game should work like one. I don't care if I pressed the poke too fast, even if the character can't duck instantly, the character should take it's time to duck and then poke, not just randomly stop ducking in the middle of the animation and give me a standing attack, this is ridiculous! I hope Paulo sees this.

Storms sent this thread to them

 

[GNG Iniquity]

Or was it thought that way by NRS for any reason whatsoever? <--- That's the interesting part.

I highly doubt it, I don't think Cyrax's 81% 1 bar resets were thought out. We shouldn't have assumed that these bugs were merely character related. I really don't feel that adequate testing was put into this game, clearly, but now that we're finding bugs THIS big...it's a problem.

 

[Robotic]

I think this can be explained by the presence of "up+3" commands in some characters.

Also, can you test "down,up" (teleport) commands, just to see if "up" activates during that timeframe in which the down jab command is randomly ignored?

This happens with up commands as well and has been happening since the beginning of the game. I've been talking about this for months and I'm very happy folks like op and soonk who know how to post video have inquired further to get to the bottom of this for the community.

Anyone who does instant air anything will have suffered from this before but may not have realized it and thought they just fucked up somehow.

 

[Bildslash]

Wow! Great job getting to the bottom of this.

 

[kandehbar]

PUT THIS ON THE FRONT PAGE

 

[Big Pampering]

Reallly nicw work, this will finally prove it to all these people who are in denial. God I love this, I hope these people burn for what they said , about the bug not existing

 

[webreg]

Sent it to Hec and Paulo

Thanks for the support and attention on the matter. I think it is still a bit early to involve the devs because there are still a lot of questions left unanswered. I've spent only three hours testing this and they spent a lot of money and manpower doing the same prior to the release of the game. Since my hand was now somewhat forced I made a video showing the glitch. Unfortunately due to the short time I had to make this the quality is abyssal but I think you can see what I'm trying to show.

YouTube Text: Hello there. I was in a hurry and I'm currently lacking the proper equipment to make a video of adequate quality. The whole thing turned out to be so small you can't even read the script so I'll explain:

First I show that my notebook is capable of exact and reliable control of my PS3 with the help of a combo script. The actual poking demonstration is cut in 4 segments of 15 pokes that are different only in the milliseconds between "down" and "1". The times in order are 70ms, 85ms, 120ms, 135ms. All tests within the range of 85-120 show this problem. Everything earlier or later works fine.

This is one of the scripts (I have a loop script that executes each of the four scripts five times, so 4x5x3=60 pokes):

 

[webreg]

Now, let me be clear on this. I have only tested "down 1" on three different characters (Scorpion, Smoke, Liu Kang), "down 3" with Smoke and "back 3" with Liu Kang. While the "down 3" behaves exactly the same way as "down 1" does I wasn't able to get the "back 3" to fail like that. All of this is no conclusive proof that it does or doesn't happens with command other than "down 1".

As of now there is no evidence whatsoever to claim that the "input bug" affects anything other than "down 1". We can talk about what these tests might or might not tell us regarding special moves, block strings and other aspects of the game but at the end of the day that's just speculation and not a fact. Talking about it is encouraged and speculation can always be the beginning of great discoveries but I urge everyone to make the distinction between fact and speculation because mixing these two together will only bring confusion and distraction.

Please be reasonable and show the people who work at NRS the respect they deserve. Show them that we are the mature audience MK9 was made for and that we can speak with each other on the same level. Show them, that we are reasonable customers but that we in turn also demand the respect we deserve. They should come to the conclusion that we can handle the truth and that an open and candid communication about problems with the game is the best way to go if you want to keep your fans happy and loyal.

 

[aj1701]

[MENTION=4121]webreg[/MENTION], great stuff. Anecdotally, I know I've had this happen for d+2 as well... instead of upper cut i get standing 2. I don't recall 3 or 4 having a problem, but I don't use them nearly as much (Kitana d1 FTW!)

 

[webreg]

Yes, it seems to affect all down pokes. I think the most important question now is if it also effects special moves that start with a down command. This will be the next thing I'll try to test out because Tom Brady claims it happened to him while doing clones with SZ. This is possible if a directional command after a "down" is handled the same way an action button is. The problem with special attacks is the negative edge that can skew the tests but fortunately I can eliminate that. So order of business:

- Confirm/refute that the inconsistency happens with all down pokes
- Confirm/refute that the left and right commands are affected by the same inconsistency (up works very differently)
- Confirm/refute that only action buttons are affected (not right or left) after a down command by using special movies

I guess this will take up to two hours of work if everything works smoothly. So while I very much appreciate requests please understand, that my time is limited and that I probably won't be able to answer additional questions "on the fly" even though I'll try.

 

[GNG Iniquity]

This will be the next thing I'll try to test out because Tom Brady claims it happened to him while doing clones with SZ. This is possible if a directional command after a "down" is handled the same way an action button is. The problem with special attacks is the negative edge that can skew the tests but fortunately I can eliminate that. So order of business:

I've certainly noticed the drop on SZ's clone and there's certainly been quite a few instances where it felt as if JC's nut punch dropped as well, though that's BD 1. Still, it involves the D1...

 

[Check]

in my opinion over the past 10 years of studying MK games..........

........its not really a input bug. its the transitions from high to low past the midblock frame. it looks like you are going past the midblock frame but you arent. this also effects neutralizing your block (even though you tap up before being hit, you still get hit by a mid sometimes)

in fact, now that i think about it, the midblock frame might of possibly been removed, and that might be what is causing the problem also (if you transition from high to low, while you are transitioning it is possible to block a mid low and a low high, i know fuzzy guarding is still in, but i've never seen the mid low)

this is something they've known about since before the game was released. i'll just leave it at that....

later.

 

[CY MasterHavik]

When you think there is a bug. Who do you call? CHECK!*Plays the Ghostbuster theme song*

Anyways, I have been trying to get the bug. But everything work fine for me.

 

[webreg]

[MENTION=25]Check[/MENTION]
I guess the chances that NRS already knows about this are quite high yes but ignoring the players that ask for a clarification or at least an acknowledgement doesn't strike me as a wise decision because this is an inconsistency that makes no sense, happens regularly and can only be avoided by players if they change their style of play. I wouldn't say anything if there would be a time window right after pressing the first button where the command won't work (no instant down pokes) but the blind spot is in the middle somewhere. Faster and slower execution works but if you hit that particular windows of about 30 ms right in the middle your attack might simply not work. We can argue what to call it and I agree, that "bug" might not be the correct terminology but the way this works makes not a lot of sense from a players perspective.

 

[webreg]

Alright, some additional stuff from the last round of testing. Please note that I'm using "1" to represent all three available pokes. The following section should help you make the distinction between the "bug" we are talking about and simply screwing up the input.

How to do a down poke reliably?

• "Down" must be pressed first. If you press "down" and "1" at the same time you might press "1" a few milliseconds earlier and it won't work. There seems to be no tolerance in MK9 for the wrong order of commands however small. I suggest the method where you move your thumbs at the same time but start with the thumb for the "1" a bit further away. I guess this might be the most commons execution error by top players that try to do "instant down pokes".
• "1" must be pressed while you are holding "down". If the game registers that you have released "down" even a few milliseconds before pressing "1" it might not work. So instantly tapping both buttons freakishly fast might let you run into this problem.
• The game needs at least 20ms to 30ms to register button presses. This is between holding down the button and letting it go. I don't know if this has even to do with MK9 or if this is just the way the PS3 works but from a practical perspective it doesn't matter who is responsible. I have a very hard time reproducing this manually though because even with a notebook keyboard this is unbelievable fast. So it might not be an issue.

Now, even if you follow this advice you might not get your down poke and that's what the inconsistency we discuss here is about. If you press your poke button "1,3,4" somewhere between 80ms and 130ms after pressing "down" the game gives you a standing instead of a crouching attack.


Some tidbits about special attacks

• For a special attack with three button commands (almost everything in MK9) there is an input window of about 240 milliseconds and it doesn't matter where you waste that time. If you do a "down back 1" command it doesn't matter if you spend 200ms of time between the "down" and "back" as long as you do the "1" within the next 40ms.
• I have not yet been able to reproduce the inconsistency that exists with down pokes with special moves. I'm quite certain, that the exactly same problem does not exists there.
• Probably common knowledge but you don't have to let go of any of your buttons for special attacks to work. Simply pressing them down in the correct order is enough. Not even "back" and "forward" despite the inability to do that with normal controllers.

All in all I still think, that a lot of people are blowing the "down poke inconsistency" out or proportion by trying to link that to random execution errors. A lot of rumor, claims and legends are spun currently and it seems on streams the "input bug" is omnipresent and held accountable for every input missed by tournament players. We should not forget, that dropping inputs in fighting games is a common sight and that the lex parsimoniae also applies here.

I haven't tested anything in this regard but my guess is that missed special moves are most often the result of missing the first input due to not being at a neutral position or being within a portion of a string that can't be canceled. So if you input "down back 3" after any kind of recovery the "down" might be pressed to early and all you get is the "back 3". This is speculative however and will most likely turn out not to be entirely true.

 

[Prinz]

I agree. It seems that in this game all movements of the characters have frames of start-up and recovery, even ducking, jumping, directions, etc. There is no way to input an instant down jab because down has to be pressed before the jab button. Also I agree on special attack inputing. Unless they are input after a cancelable move/string they can not be done instantly because the animation before has recovery frames (or, better said, recovering to neutral). However the window of inconsistency you found with "down" is interesting. It is so tempting to research

Oh yeah, I have read CHECK's post a couple of times, didn't understand much, but there might be a point there with something that should have been in that window of inconsistency that has gone or been removed.

 

[GGA soonk]

Were you or will you be able to test a prepatch version and post the results? The input bug will not be fixed from what I understand, but I'd like to know if it was being covered up this whole time with negative edge.

 

[aj1701]

All in all I still think, that a lot of people are blowing the "down poke inconsistency" out or proportion by trying to link that to random execution errors.

I'm not so sure it is blown out of proportion. I know I fuck up moves all the time, I'm missing one of the directionals for a special, I hit the wrong attack button, whatever. I can deal with that. I can't deal when I can see for myself the game knows I input the right thing because the character reacted, but randomly says "eh fuck you, you're gonna stand for this attack."

There's no logical reason for there to be a window in the middle of the processing where you get a standing 1 where either end of the window you get what you really wanted. That's a pretty big deal.

There is no way to input an instant down jab because down has to be pressed before the jab button. [snip]
Oh yeah, I have read CHECK's post a couple of times, didn't understand much, but there might be a point there with something that should have been in that window of inconsistency that has gone or been removed.

That's the thing, if you hit down and then 1 within 75ms you actually do get the poke to come out. If it were "you must have 125ms beween pressing down and 1 for a poke to come out" I think everything would be fine. There's a spllit second you can build in to your execution. But we have a sitution where you must avoid a window of 1 to 2 frames... earlier or later and your find, but hit that window..

I also have no idea what Check was getting at

 

[SilverKeyMan]

I have not yet been able to reproduce the inconsistency that exists with down pokes with special moves. I'm quite certain, that the exactly same problem does not exists there.

Webreg, thanks for clarifying exactly what the input bug is, and more importantly, what it is not.

 

[charlieonline]

Interesting stuff Webreg.

 

[Trini_Bwoi]

Okay so is it safe to assume that the problem also happens 80ms-130ms after recovering from crouched blockstun/hitstun?

 

[circle masher]

Interesting find Webreg. And explains to a degree why in training mode sometimes I could press up to jump, have the up arrow appear in my on-screen inputs, yet I didn't jump.