This is because DataDome employs Bézier curve fitting detection. If it detects that the mouse trajectory can be fitted to a Bézier curve, it will be directly flagged as invalid, causing the attempt to fail. Simply put, when DataDome detects that the movement is not from a real human user but rather a machine-generated "trajectory," the use of a Bézier curve becomes ineffective.
This article will analyze the reasons for this failure and provide better recommendations.
What is a Bézier Curve?
In mathematics, a Bézier curve is an important parametric curve in computer graphics, described using control points to define its shape. Bézier curves are often used to simulate human sliding behavior in an attempt to bypass slider CAPTCHA detection. They generate a sliding path to make the slider's trajectory appear more human-like.
How Does DataDome Detect Bézier Curves? (What Factors Lead to Failure?
- The mouse trajectory is fitted to a Bézier curve.
- The mouse trajectory is overly smooth and continuous. Real users exhibit imperfections, pauses, or overshoots – irregular variations.
- The speed variation is too uniform and does not match the speed patterns of a real user.
How to Better Simulate Real User Behavior?
How can one better simulate a "human"? In everyday life, when encountering a CAPTCHA, users occasionally make mistakes – such as overshooting the target, not sliding far enough, or moving too fast or too slow. Therefore, to better mimic a real user, the following points are essential:
- Incorporate a wider range of mouse trajectories, including scenarios where the user overshoots or undershoots the target.
- Avoid perfection; add slight jitter or noise to the curve to simulate natural hand tremors.
- Introduce appropriate pause times to simulate slower movement or user hesitation.
- Implement non-uniform speed, incorporating acceleration or deceleration patterns to simulate users sliding too quickly or suddenly slowing down.