Researcher fingerprints CS cheaters via mouse and keyboard input patterns

A master's student at NTNU developed a fingerprinting system that identifies Counter-Strike players from mouse and keyboard input: mouse data matched unique players 100% of the time, keyboard data 98%. The method runs on a single 20GB slice of an Nvidia A100 GPU and links smurf accounts.

• Mouse dataset identified unique players 100% of the time, keyboard 98%
• Correlation between strangers' mouse and keyboard habits is just 0.11
• The system runs on a single 20GB slice of an Nvidia A100 GPU
• It links smurf accounts but breaks with shared accounts