Abstract:
"Counter-Strike 2" player magga_ completed a master's thesis at the Norwegian University of Science and Technology, studying how to identify different accounts used by the same person through operating habits. In a test sample of more than a thousand players, this method identified 13 sets of fan accounts known to the researchers.
This study focuses on a limitation of account bans: after cheaters lose their current account, they may still re-register or purchase an account to return to the game. The study used ordinary game records originally saved by the game to extract personal features from mouse movements and keyboard inputs, including repeated actions such as quick turns to aim, emergency stops, gun pressing, and throwing objects.
In the sample test described in the paper, the accuracy of mouse movement analysis in identifying users was 100%, and the accuracy of keyboard key patterns was 98%. After merging the two types of indicators, the system identified 13 groups of fan accounts known to researchers. Additional backup accounts were discovered when the author checked some of the matches that were initially believed to be false.
These results apply only to current testing conditions, and the method has not yet been tested on Valve's large player base. Researchers also pointed out that multiple family members taking turns using the same computer may make judgment more difficult. He suggested that behavioral characteristics should first be used to prompt manual review rather than directly as a basis for banning, and hoped that Valve or FACEIT would assist in conducting actual environment testing.
Comments