[INTRODUCTION] Introduction to Cheating
-
Introduction
This is a thread that tells you about everything there is to know in the cheating community, this is the ultimate introduction for both beginners, intermediates, and everyone else.
In this thread, we will cover:-
Minecraft's history of both hacked clients and anticheats
-
How anticheats work
-
How hacked clients work
-
How to use LiquidBounce
-
The best free & paid hacked clients (Best by test)
-
How to contribute to LiquidBounce
-
How to add simple speed modes
1. How anticheats work
In this section, I will tell you about how the checks that I know of work, and how to bypass them. This will also help amateur config, script, and hacked client developers.
Aim Assist Checks
- Rotation Speed Consistency - Checks for 75-100% consistent speed.
It can get bypassed with ease by using random enough (60 min - 100 max, for example) Rotation Speed values.
- Rotation Speed - Checks for fast rotations.
It can get bypassed by reducing rotation speed, though it also can get bypassed by using random enough rotation speed (Recommended)
- Rotation Accuracy - Checks for rotation accuracy, aka Aim Accuracy that is either too good or too bad to be legit.
It can get bypassed by using RandomCenter and other methods, that are too hard for me to explain
(I recommend adding RandomCenterStrength, because the other checks will haunt you).But in a nutshell, those other methods fully simulate a player's aiming behavior, and are 100x better than RandomCenter, though you have to possibly spend years to make those methods work.
- Advanced Deep Learning Aim Heuristics - Machine Learning checks, that compare cheater and legit player behavior, to stop any aim assist bypasses (Though this creates a vicious cycle and may falsely flag legit players if not enough data is given; and I'll tell you all about that right now).
It can get bypassed by a Machine Learning aim assist or killaura, that fully copies legit behavior. This would need alot of legit players' data, thus, if someone hacked Lunar Client, and got all the legit aiming behavior it needed, it would work quite well. The 'virus' would need to check if any ghost clients got injected, though, otherwise the results can be possibly not as good as they could've been. Anyway, anti-cheats would catch up and so would the hacked clients.
- Aim Shakiness - Checks for extraordinarily shaky aim. For example, it would flag you if you did 500 360s in a second, since it's not legit at all.
It can get bypassed by making the RandomCenterStrength lower, and turning on the Micronized option + turning up the MicronizedStrength.
- Sensitivity / GCD Flaw exploit - Checks for the aim GCD not matching the GCD for everything else.
It can get bypassed by using a Client that doesn't have the bug, or a Client that has the bug, but has an easy way to patch it (Example: Old b73 builds before crossversion had this issue, though it could be fixed with CzechHek's KillAuraPatch).
-
-
im not sure why this is the ultimate anything when you have wrote a small part of one section based on your observations
-
@Gabriel What anti cheats use machine learning.
-
@Hunter171 If you're talking about non-advanced heuristics that are taught with 2 legit players and 10 cheaters, there are many. Just search up "Machine-Learning anticheat" on Google, you'll find it right away. I've found one open-source ML aura detector, though not as advanced as what I described. Here's the GitHub link: https://github.com/Nova41/SnowLeopard
-
SnowLeopard is ass, any normal good heuristic check is better
-
@Hunter171 I'm aware of that, it only even gets statistics from the tick that the cheater / legit player hits in.
-
@Gabriel There are no good anticheats using ML
-
@Hunter171 Of course there aren't. Since both hacked client developers and anticheat developers aren't working together (Why should they, after all?), it's hard to make a proper ML anticheat.
-
@Gabriel and not worth it at all...
-
how would you make unsupervised advanced deep machine learning aim detection heuristics with hacked client developers' big structured data?
