I don't think this clip/graph is the best. The bigger tell, which is more evident from watching the demos all the way through, is the difference between Tarzul's natural sweep, mouse speed and aiming tendencies before lock and what happens as and after he locks on a target.
He doesn't always use the aimbot, and while he has good aim naturally, there are some things that stand out:
Tarzul aims cg center target naturally and in previous demos, but when he uses the lockbot he aims above the head; exactly the same way Lou demonstrated the lockbot aims in his rec.
All great cg aimers exhibit a logarithmic decrease in mouse speed as they center in on the target area. Tarzul's mouse "snags" or snaps to point on the target with high speed and zero over hang or aim compensation.
When the target is dead, the lock releases unnaturally and you can see an abrupt change in mouse speed and directional control.
Moving from one target to a target going the opposite direction at a different speed will always result in some minor fine aim compensation, even if the shooter is static. Even on altimors clip, you see tarzul has no fine aim or even mouse speed compensation in these situations.
Strafing and fast ground motions (from either the shooter or the target) causes natural aim over hang and compensation. When Tarzul is locked, even when he is strafing, there is no aim compensation period.
If someone wants to link all three demos, I will post the time signatures and points of interest. Also, just watch tarzul's older demos. His aiming didn't look like this previously.
Heres a good TA video that illustrates natural fine aim compensation:
http://www.youtube.com/watch?v=JkhwjGyOZS4