OSBot's mouse movement is easily detected

[Juggles]

9 minutes ago, asdttt said:

Yup and like I said, it doesn't contain any flaws from my testing. That's why I recommended it from any other "randomization" from inside the Mouse API. Still though, they compare large/small/medium movements against different checks, so you still should probably include some small/medium randomization that doesn't include OSBot's API (Assuming you don't have an API method that doesn't produce the same "endpoint" flaw I've discussed).

Even keeping the mouse perfectly still is better then moving it randomly around with OSBot's mouse API. 

I'm not responding to you. You wont even bother reading anything I've said or testing yourself. Mouse is flawed, get over it

You've posted a problem but I don't see any solution. I want to test your proposed solution myself and see if what you're saying is true. 

But the only way to truly test your hypothesis is to run n=1000+

500 with default and 500 with your proposed movements. Even 1000 bots might not be large enough of a sample size. 

[asdttt]

3 minutes ago, Juggles said:

You've posted a problem but I don't see any solution. I want to test your proposed solution myself and see if what you're saying is true. 

But the only way to truly test your hypothesis is to run n=1000+

500 with default and 500 with your proposed movements. Even 1000 bots might not be large enough of a sample size. 

Solution is to implement better mouse movement that doesn't produce the same flaws (Although using OSBot's mouse + your own mover is fine too). I can't hand you all my own mouse mover or then I'd have to be worried about Jagex finding flaws in my own mouse mover and banning my ass. 

Just take what I've said, and fix it yourself. You're all knowledgeable with Java, and all have scripting experience. And maybe hopefully they'll implement better/different mouse movement on OSBot and then we'll only have to worry about the other ban  checks. 

[asdttt]

6 minutes ago, Malcolm said:

The only thing I am saying is that you like us are guessing.

You don't know what is analyzed about the mouse although I agree it is sent and it is analyzed.

Let me put this plain and simple for you.

Is the mouse detected? Probably. (I wont say yes and I won't say no because I don't actually know for sure but it is more than likely that it is detected and analyzed)

Do you know how to fix it? No,

Do I know how to fix it? No.

Do the developers know how to fix it? No.

Do the employees at Jagex know how to fix the bot detection issues? Yes, because they are the only ones who know how it works and they are the only ones who know a workaround it.

Now, unless Jagex's bot detection methods are leaked nobody knows how to fix anything.

This is the only point I was trying to make to you. I am not denying any detection.

 If you would of read my post, you'd be fully aware that OSBot's mouse is flawed. Stop arguing and wasting my time. I've literally told you the same thing with different wording many times. Not only that, you've done NOTHING to test what I'm sampling, yet you are very quick to deny everything. Evidence for denying? None. 

Can YOU, or anyone else detect OSBot based on mouse delta? YES. Look at my sample and tell me that looks human. Does Jagex detect this huge flaw? In my study, YES. Could they detect this huge flaw? YES, VERY VERY EASILY. Are you wasting my time? YESS!

 

Edit:

Do you know how to fix it? Yes

Do I know how to fix it? I'd hope so

Do the developers know how to fix it? Absolutely.. It's not that difficult. Patrick's a smart guy

Edited April 7, 2019 by asdttt

[Molly]

13 minutes ago, Malcolm said:

The only thing I am saying is that you like us are guessing.

You don't know what is analyzed about the mouse although I agree it is sent and it is analyzed.

Let me put this plain and simple for you.

Is the mouse detected? Probably. (I wont say yes and I won't say no because I don't actually know for sure but it is more than likely that it is detected and analyzed)

Do you know how to fix it? No,

Do I know how to fix it? No.

Do the developers know how to fix it? No.

Do the employees at Jagex know how to fix the bot detection issues? Yes, because they are the only ones who know how it works and they are the only ones who know a workaround it.

Now, unless Jagex's bot detection methods are leaked nobody knows how to fix anything.

This is the only point I was trying to make to you. I am not denying any detection.

I'm not sure it's fair to say nobody knows how to fix this. A straightforward solution would be to record a couple thousand mouse paths of varying distances and base your script's mouse movements off these. Example: I need to move my mouse 70 pixels, I grab from my thousands of human mouse paths some that are around 70 pixels, say between 40-100. I grab one of those paths, stretch it or shrink it, add some noise to it and use that path to move the mouse.

The downside to this is it requires a fair amount of work recording that many mouse paths and realistically is easier to do and probably better for not being "detected" if done by each scripter individually for their own scripts.

Edited April 7, 2019 by Molly

[asdttt]

7 minutes ago, Molly said:

I'm not sure it's fair to say nobody knows how to fix this. A straightforward solution would be to record a couple thousand mouse paths of varying distances and base your script's mouse movements off these. Example: I need to move my mouse 70 pixels, I grab from my thousands of human mouse paths some that are around 70 pixels, say between 40-100. I grab one of those paths, add some noise to it and use that path to move the mouse.

The downside to this is it requires a fair amount of work recording that many mouse paths and realistically is easier to do and probably better for not being "detected" if done by each scripter individually for their own scripts.

It's not easy to code a mouse mover, but you'd only need to account for a few factors (If you want me to show proof, I'll DM you the code). 

Factors are: Reaction time, reaction time variation, mouse speed, mouse speed variation, mouse step variation, deviation (Very important, mind you, the human wrist cannot easily move the mouse without deviation), noise, and OVER-move (When you move  your mouse, but accidentally go over a button and have to return to it). Edit2: (All of these should also be editable by scripters. A 12 year old would have different reaction time and mouse speed compared to a pro CS:GO player for instance)

edit: Not denying your solution, that'd also work. Just adding on that you can make a realistic mouse mover with code

Edited April 7, 2019 by asdttt

[asdttt]

7 minutes ago, Malcolm said:

What have I denied? Please elaborate?

^ I certainly haven't denied this.

Idk everything you're saying is dumb so I just assume you are denying everything lmao. You literally just said the OSBot developers can't fix the flaw. Like whaaaat..? 

[Molly]

8 minutes ago, asdttt said:

It's not easy to code a mouse mover, but you'd only need to account for a few factors (If you want me to show proof, I'll DM you the code). 

Factors are: Reaction time, reaction time variation, mouse speed, mouse speed variation, mouse step variation, deviation (Very important, mind you, the human wrist cannot easily move the mouse without deviation), noise, and OVER-move (When you move  your mouse, but accidentally go over a button and have to return to it). Edit2: (All of these should also be editable by scripters. A 12 year old would have different reaction time and mouse speed compared to a pro CS:GO player for instance)

edit: Not denying your solution, that'd also work. Just adding on that you can make a realistic mouse mover with code

Oh yeah you can for sure make it better than the barebones example I gave, was just trying to illustrate the point that this isn’t something without a solution that’s impossible to figure out.

[Naked]

6 minutes ago, Malcolm said:

No, I have not denied anything.

I am telling you that the only thing we can do is guess what the issues are.

Can the flaws be looked at and the mouse movements be changed? Yes.

Will these changes solve anything? I understand that you are claiming that it does and dude I am all for this. If it is possible to make an "undetectable bot" it would be incredible really, but, we don't know if the changes will solve anything. We are guessing and it is trial and error.

Does it hurt to try? Absolutely not.

I am not trying to say you haven't found anything. Just like you I am pointing out flaws in your analysis. The biggest flaw being your sample size.

Like I suggested you can be far beyond 2-3 standard deviations and your sample potentially doesn't accurately represent the mean. I am not saying it doesn't.

I certainly don't disagree with anything that you're doing here.

What I do disagree with is

• Calling out developers in a negative way
• Calling out scripters in a negative way

 

What I would like to see is you run hundreds or thousands of bots and then statistically analyze it and return with your findings because it is very possible that once you run it on a larger scale that you end up with the same ban rate as the regular mouse.

Hard to test at scale when you're gray

 

Correlation doesn't imply causation blah blah blah small scale tests mean nothing.

[asdttt]

13 minutes ago, Malcolm said:

No, I have not denied anything.

I am telling you that the only thing we can do is guess what the issues are.

Can the flaws be looked at and the mouse movements be changed? Yes.

Will these changes solve anything? I understand that you are claiming that it does and dude I am all for this. If it is possible to make an "undetectable bot" it would be incredible really, but, we don't know if the changes will solve anything. We are guessing and it is trial and error.

Does it hurt to try? Absolutely not.

I am not trying to say you haven't found anything. Just like you I am pointing out flaws in your analysis. The biggest flaw being your sample size.

Like I suggested you can be far beyond 2-3 standard deviations and your sample potentially doesn't accurately represent the mean. I am not saying it doesn't.

I certainly don't disagree with anything that you're doing here.

What I do disagree with is

• Calling out developers in a negative way
• Calling out scripters in a negative way

 

What I would like to see is you run hundreds or thousands of bots and then statistically analyze it and return with your findings because it is very possible that once you run it on a larger scale that you end up with the same ban rate as the regular mouse.

 

 

 

How many bots do you think I've tested with...? 5...? No.. Well over a hundred. How many miner bots am I using right now? 3, simply because it uses lots of ram/cpu. Although when I'm not using my PC, about 10-13 max

Let's move onto the actual subject, and then hopefully this argument will be over. I sampled the same data as Jagex, on the same interval. I found that not only is OSBot's mouse movement very robotic, but it also has constant flaws. I posted the sample comparing it to a humans mouse movement so everyone could see the major flaw. 

Now onto my "sample size" of banrates after changing my mouse movement. Been running miner bots, each for 6  hours a day, each on it's own paid proxy IP, each at the same exact mine. How many bans have I gotten? 0. How much profit? Surprisingly not that much because iron is fucking worthless. 

Minerbot results (Not including my other bots scripts): 

Before implementing new mouse movement: 56/56 bans - Each bot lasting exactly 1 day

After implementing new mouse movement: 0/12 bans (Only have 12 accounts because there's no need to make more)

 

Now banrate samples ARE NOT to be taken as evidence, but rather reference. These can very easily be inaccurate for all I know. 

What you CAN however take as factual evidence is the samples of OSBot's mouse movements. That's consistent, and you can see for yourself the output. 

Edited April 8, 2019 by asdttt

[Impensus]

4 minutes ago, asdttt said:

How many bots do you think I've tested with...? 5...? No.. Well over a hundred. How many miner bots am I using right now? 3, simply because it uses lots of ram/cpu. Although when I'm not using my PC, about 10-13 max

Let's move onto the actual subject, and then hopefully this argument will be over. I sampled the same data as Jagex, on the same interval. I found that not only is OSBot's mouse movement very robotic, but it also has constant flaws. I posted the sample comparing it to a humans mouse movement so everyone could see the major flaw. 

Now onto my "sample size" of banrates after changing my mouse movement. Been running miner bots, each for 6  hours a day, each on it's own paid proxy IP, each at the same exact mine. How many bans have I gotten? 0. How much profit? Surprisingly not that much because iron is fucking worthless. 

Minerbot results (Not including my other bots scripts): 

Before implementing new mouse movement: 56/56 bans - Each bot lasting exactly 1 day

After implementing new mouse movement: 0/12 bans (Only have 12 accounts because there's no need to make more)

 

Now banrate samples ARE NOT to be taken as evidence, but rather reference. These can very easily be inaccurate for all I know. 

What you CAN however take as factual evidence is the samples of OSBot's mouse movements. That's consistent, and you can see for yourself the output. 

How are you using 3 bots at once when greys are limited to 2 bots per client?

[asdttt]

1 minute ago, Impensus said:

How are you using 3 bots at once when greys are limited to 2 bots per client?

Can't elaborate without getting banned off here

[asdttt]

Just now, Malcolm said:

So you have multiple accounts avoiding VIP?

For all you know I have an alt with a paid subscription. Don't worry about it. It's for the best

[Naked]

1 minute ago, asdttt said:

For all you know I have an alt with a paid subscription. Don't worry about it. It's for the best

Alt accounts are against the rules too

Edited April 8, 2019 by Naked

[asdttt]

Just now, Malcolm said:

Multiple accounts are against the rules.

It's my brother's account obviously MonkaS

[Naked]

1 minute ago, asdttt said:

It's my brother's account obviously MonkaS

So you're sharing accounts? Also against the rules.