Bypassing anti-botting detection

Despite the initiatives taken by video game and anti-cheat developers, bot developers continued to innovate, producing increasingly sophisticated bots that continue to avoid detection by utilizing machine learning and other methods like noise creation.

What is noise creation?

The deliberate creation of randomized or irregular inputs to imitate human behavior and avoid detection by anti-bot systems is commonly referred to as “noise creation” in swipes and taps. By creating an unpredictably “natural” appearance for automated bot swipes and clicks, this noise is intended to lessen the possibility of being detected by anti-cheat or anti-bot algorithms. In order to tell the difference between human users and bots, input patterns are frequently examined by anti-bot systems. When it comes to interactions, people are more likely to have small flaws or randomness like irregular click timing or different swipe speeds from start to end of a swipe. Bots on the other hand often conduct actions in a repeatable, precise, and predictable manner. Noise is starting to be used by developers when creating aimbots in first-person shooter PC games (Figure 11), either using a more sophisticated movement that combines curving motions (Gaussian) to provide a more realistic and unpredictable path to the target, such from your crosshair to an enemy, or a smooth, regulated path between two points (Bezier).

Play sessions frequency and regularity

Because of a person’s varied schedule, real-world obligations, and other tasks, human players typically have inconsistent play pattern times. A person would need to step away from their PC to take breaks, go to work, eat, or go to school. Conversely, it can be concerning when bots operate at regular periods or durations. All-day active accounts may be tagged for odd behavior, particularly in the late evening and early morning hours when the majority of players are not online.

Figure 11: Sensory AI aimbot. Image taken from: https://www.elitepvpers.com/forum/valorant-trading/5171648-sensory-ai-aimbot-flickbot-triggerbot-rcs-raspberry-pi-private-slotted.html.