You’d compare all the telemetry you can from a sample of non cheater users and compare that to the cheater. Chances are the cheat is better than a non cheat user because thats the entire point of cheating. It will probably manifest in stuff like accuracy for aimbots, how fast they train onto a target they see, if they put their reticle on another user through a wall and how often that happens, how they move or jump, etc. theres bound to be some significant difference in some of these or other metric compared to non cheaters. And if there is no significant difference, maybe the cheaters don’t have the edge they thought they did after all and you don’t need to worry about it.