I have devised a way to demonstrate heavy aim

[Cursed Lemon]

Holy shit! You still exist?!?!?! I just brought up the name cursed lemon yesterday to someone and was wondering where you been at all these years. I remember you making videos for h2 and h3 back in the day about their problems.

 

Kind of funny that those same issues still exist to this day, huh? 

 

For sure, I've got a few ideas of my own.  I think something that oscillates like so:

 

1. Start at center point

2. Deflect 1/8 to the left

3. Deflect 1/8 to the right to return to center

 

Would be a great test, because even without the machinima coordinates turned on, anything "off" would be extremely apparent.  It's also an easy test to perform on the older installments as well.

 

Some other things I'd like to try once I have the circuit built to handle both X and Y axes is implementing an ogre twitch that sits there and spins the reticle in a circle super fast, something else that any delay or latency would be extremely apparent on.

 

Also, I have multiple control pins on my arduino, so I could also integrate automatic start/B button pressing to see whether that "corrects" anything.

 

Heckin' right. Let us know everything that you find. 

[Killmachine]

Heckin' right. Let us know everything that you find.

* h*ckin

 

Please watch your profanity

[Cursed Lemon]

* h*ckin

 

Please watch your profanity

 

Frick you, buddy.

[Killmachine]

Frick you, buddy.

* fr*ck

 

Final warning or I'm gonna have to call the admins

[Mib2347]

Hey guys, just wanted to give an update.

 

Got the arduino up last night, ran a few LED blinking tests, it's all pretty simple stuff. Got the controller apart and started just doing some shorting of the look stick potentiometer leads just to recreate what @@Mib2347 did. I'm hoping tonight I can get wires soldered from the controller to the breadboard and I can at least get a step up on what he did, because I'll be able to turn the stick turn on and off, with timing control, but only at full tilt. My digital potentiometers are supposed to arrive tomorrow, and it shouldn't be much work to get the circuit built to have full control and we can perform all sorts of fun tests .

 

I am so glad you came along!

 

I think you'll be able to get further than me so I think now you're our best hope of getting some good results. I won't have time to do anything until the weekend probably anyway so I can't wait to see what you come up with.

[automatousbeing]

I am so glad you came along!

 

I think you'll be able to get further than me so I think now you're our best hope of getting some good results. I won't have time to do anything until the weekend probably anyway so I can't wait to see what you come up with.

 

Hey man, just standing on the shoulders of giants here, you were the catalyst.  I've sorta had these ideas bouncing around in my head forever, but I never really sat down and wondered if it'd be as simple as it is.  I'm still giving you most of the credit 

[Wu Ip Man]

I'll make a list for possible mechanism alterations. Feel free to add to the list of possibilities. There are certainly more. Or if you can think of a reason to rule any of them out please say so.

 

1.  There is a delay on the initial input
    
2. There is a trimming somewhere after the input begins, and the input for present points are dropped... requiring a reinitialization from the last known point. This would effectively keep the input at a point on the curve longer than it should and keep the acceleration static. Though this one is impossible unless there is a contingency in place for input.
    
3. Each joystick axis point is put in a que. Under stress the input holds at that point instead of being dropped. Pretty much equivalent to 2, except no drop.
    
4. Each point on the sensitivity curve has a particular amount of acceleration that applies to it determined by the logarithm used that is dependent on acceleration chosen. There are a few ways this one could work I think
    

• When the sensitivity is chosen, a reference table is created for the joystick position and acceleration. The game could incorrectly reference these numbers under stress.
• Or instead of a table is simply calculates the accel using the computed formula. It sounds ridiculous saying a computer would calculate this wrong, but perhaps there is a variable affecting this computation.
  
  
  
  As far as the experiments are concerned. We need MIB or automatous to test his set up with inner deadzone to see if it bypasses the inner of the game or if it requires part of an inner dead zone to be registered. If he gets a different set of points that will be largely in our favor.
  
  The next step needs to be determining whether there is an issue with initiation when it is heavy. If there is not, we can compare middle points. As an example if there are 100 points from start to finish, we can compare the amount of time it takes to go from point 40 to 60. Or we can see if the beginning of frame 2 and frame 3 have equivalent points or not.
  
  We could run the same simulation on his map with 8 ppl as well. Since there is only one spawn people would just have to kill themselves until he spawns.  

[Mib2347]

Hey man, just standing on the shoulders of giants here, you were the catalyst. I've sorta had these ideas bouncing around in my head forever, but I never really sat down and wondered if it'd be as simple as it is. I'm still giving you most of the credit

Well thanks that's very nice of you to say. We're all a community and we all want the same thing. No one would bother if we didn't all believe that the aiming is messed up in some way. For a long, long time it was Wu Ip Man digging up stuff alone so I'm very pleased to see more and more and more people able to contribute.

 

Some of the things you've said are really exciting and even if they show my method to be flawed I think the setup you've got going will expose everything good or bad.

[automatousbeing]

Some of the things you've said are really exciting and even if they show my method to be flawed I think the setup you've got going will expose everything good or bad.

 

Agreed, my real goal is to just make sure we're being as objective as possible, which means proving to the naysayers and the "variable rumble pattern" audience that that isn't what's happening here.  Who knows, maybe we'll confirm all that, but either way we'll have a better understanding which is what really matters.

[TryHardFan]

Frick you, buddy.

Fudge you too.

[Killmachine]

Guys chill we're all friends here

[Wu Ip Man]

Kind of funny that those same issues still exist to this day, huh? 

 

 

Heckin' right. Let us know everything that you find. 

Haha yeah.

 

Well thanks that's very nice of you to say. We're all a community and we all want the same thing. No one would bother if we didn't all believe that the aiming is messed up in some way. For a long, long time it was Wu Ip Man digging up stuff alone so I'm very pleased to see more and more and more people able to contribute.

 

Some of the things you've said are really exciting and even if they show my method to be flawed I think the setup you've got going will expose everything good or bad.

I'm glad everyone is working together to solve this. Your post definitely created a much more active community on the issue. 

 

Even if there are variable rumble patterns, I'm willing to bet in all the weapons that are available that at least one of them would have a consistent rumble pattern. The method is useful regardless since we can test it out with people in the game. If it can't bypass the inner we will know if there is an initial delay. Even with varying rumble patterns, the points are consistent. If heavy aim changes them with added people and the points change we still win. Between both outcomes we still have a reference. We still have video that can be scrutinized. There are so many things we can learn regardless of how the method functions that it is a win win for us in so many ways.

 

 

[Cursed Lemon]

Just thinking, with respect to ruling out rumble durations, couldn't we test diagonal aiming? 

 

If the horizontal and vertical dimensions don't follow the same pattern, then that would mean the rumble has nothing to do with it. 

[Mib2347]

 

 

As far as the experiments are concerned. We need MIB or automatous to test his set up with inner deadzone to see if it bypasses the inner of the game or if it requires part of an inner dead zone to be registered. If he gets a different set of points that will be largely in our favor.

 

 

I did that quickly with inner and outer deadzones and I can't see any difference.

 

 

http://xboxdvr.com/gamer/Mib%20HCS/video/29295165

 

If you'd like a longer video with specific settings then let me know. I can squeeze in a bit of time.

[Apoll0]

Hey guys, just wanted to give an update.

 

Got the arduino up last night, ran a few LED blinking tests, it's all pretty simple stuff.  Got the controller apart and started just doing some shorting of the look stick potentiometer leads just to recreate what @@Mib2347 did.  I'm hoping tonight I can get wires soldered from the controller to the breadboard and I can at least get a step up on what he did, because I'll be able to turn the stick turn on and off, with timing control, but only at full tilt.  My digital potentiometers are supposed to arrive tomorrow, and it shouldn't be much work to get the circuit built to have full control and we can perform all sorts of fun tests  .

 

 

 

Taking the term "Halo scientists" to a whole new level

 

Just wanted to say thank you to @@Mib2347 and @@automatousbeing

 

Seriously guys.  This is amazing.

 

I actually have a boner right now.  No joke.

[automatousbeing]

Couldn't help myself over lunch...

 

We're in h*ckin' business now boys:

 

[Killmachine]

+1 just for h*ckin

[Mib2347]

Couldn't help myself over lunch...

 

We're in h*ckin' business now boys:

 

https://www.youtube.com/watch?v=Ws18oM9o0K8

Excellent stuff man. I genuinely cannot wait to see what you find.

[xSociety]

I did that quickly with inner and outer deadzones and I can't see any difference.

 

 

http://xboxdvr.com/gamer/Mib%20HCS/video/29295165

 

If you'd like a longer video with specific settings then let me know. I can squeeze in a bit of time.

 

 

Couldn't help myself over lunch...

 

We're in h*ckin' business now boys:

From all of us Halo fans:

 

[Devaneaux]

Couldn't help myself over lunch...

 

We're in h*ckin' business now boys:

 

 

This is way better than the original setup since now we don't have to worry about what kind of signal is entering the relay. As long as all of the signals are discrete pulses of a certain length, then this works. 

 

Question, have you double-checked the signal going into the relay from the Arduino using an oscilloscope to make sure that we have discrete pulses of a certain length?

 

Also, this is how I feel you should run your tests going forward to address the points that we've seen over the past year:

 

1) Test this setup with different games: We're not going to get the same kind of precision we can see in H5 with the machinima controls, but looking at CE and maybe something like H4 would be important just to test the distances.

 

2) Try different look sensitivities: This is going to be the main variable in your tests between games. You're going to have to find a look speed that is as similar between games as possible. This is a challenge that people here might be able to help with, but as it stands, it's the biggest issue in determining tests.

 

3) Load the dedicated server with a bunch of CPU taxing actions: Have someone forge a map with all sorts of explosions, projectiles, or other things being spawned in continuously and a teleporter to take you near the stuff that is taxing the server and then move as far away as possible from the stuff that was spawned in. This should help figure out the CPU and Server related stuff that we've been talking about for a year. 

 

4) Run enough trials to get a decent statistical sample: I can't stress this enough. In order to be as objective as possible, we need to run lengthy trials with the Arduino relay running so we can get enough information. Then through some simple statistical analysis, we can derive error %'s in aiming based on either CPU taxing, or just general inconsistency with the aiming. 

 

That's what I came up with for now for the tests that should be conducted in using this method to test the aiming. Of course, what you decide to do is totally up to you, but I feel all of this information I just presented is important in really being thorough with this and not being half-assed like this community can sometimes be with its proof/reasoning. As is the same in bug reporting, the more precise and in-depth you are in describing a situation, the better a developer can fix the problem.  

[Apoll0]

This whole thread is really putting my nerd-cred in perspective.

[automatousbeing]

This is way better than the original setup since now we don't have to worry about what kind of signal is entering the relay. As long as all of the signals are discrete pulses of a certain length, then this works. 

 

Question, have you double-checked the signal going into the relay from the Arduino using an oscilloscope to make sure that we have discrete pulses of a certain length?

 

*snip of all good things*

 

I haven't yet, I don't have a scope at home, but I'm sure my boss would let me use ours at work if I brought the whole setup in.

 

As for everything else, all good suggestions, few things I've thought of and some I haven't.  I'm hoping to get things soldered up tonight so I can at least get some initial testing going.

 

At the very least, being able to do a similar CE test like Mib2347 did should confirm that my pulses are consistent, and then I can trust the circuit when I switch over to H5 and run the same test, since this time we won't be using two different inputs.

[automatousbeing]

This whole thread is really putting my nerd-cred in perspective.

 

I fully intend to do a slower walk through video once I have everything together to explain how it all works, and also how relatively simple the whole process is, as well as how just about everyone could recreate what I've done with about $50 and a handful of online tutorials.  It's certainly not rocket science.

[MultiLockOn]

I love you.

[xSociety]

343, just patch in the BETA aiming system so these fine people don't have to waste so much time figuring out just how bad you fucked up.