Shoot for the Moon and You Will Never Miss: Characterizing and Detecting Aimbots in Online Games

摘要

We have designed and implemented an accurate and robust server-side aimbot detector, AimDetect. AimDetect utilizes a set of novel features to model the behavioral patterns of aimbots and honest players, which stem from both the automation of aimbots and the blatant nature of cheaters. These features are then fed into a two-step classifier, which first classifies each kill event as from either honest players or aimbots using SVM and then employs TRW to compute a comprehensive anomaly score. Our evaluation results show that AimDetect is able to capture almost all of the aimbots with quite low false positive rate.