Measuring mouse movement is possible after each shot
Each frame.
Yeah that's possible too, but I believe this might lead to even further false positives. Not to say that it should be completely ignored though. It should be looked into for sure. Although it might not seem as easy as it looks. It's been quite a while since I last used assembly but I've already reverse engineered some sections of the aimbot and ran it through a debugger. There were some "protections" put in place by the creator that disabled the aimbot the moment a debugger was attached to the process. It didn't took me long to bypass these "protections" and a more experienced assembly user could probably have done it even faster. Unfortunately my free time is getting even shorter as of late, so I haven't looked into it anymore yet. But what I mean by not being as easy as it looks is that the aimlock doesn't exactly lock into the player the moment their nametag is on the screen. The mouse/crosshair needs to be relatively close to the enemy players' nametag for it to lock into the player. This was obviously made this way to be harder for others to notice and to be harder for scripts that track sudden mouse movements to detect aswell. But the hard part here might not be detecting suspicious mouse movements, but instead keeping the amount of false positives to a minimum. Because detecting it becomes useless if the amount of false positives becomes so high that you can't differentiate the clean players from the hackers.
You know, combining my idea with what Vito said I believe it would make a strong case against it.
If this script is to be made it should be looking for patterns (especially the case with the lead aim variant) because like I said, no machine is capable of replicating a human's way of moving/aiming a mouse and vice versa.
The key is still to completely reverse engineer the aimlock and uncover how exactly it works and moves the cursor. Once you've done this you can most likely find a if not more patterns in relation to its movement and easily isolate and ban the users of it. This in combination with maybe opponent's position and hit registration might make it even stronger.
Resources should not be a problem if servers are not experiencing performance issues because of it (pakis with PCs that are not capable of playing this game at 20 FPS should not be playing the game to begin with) and I believe the impact is small to nonexistent.
I'm more than happy and willing to test this with you and I believe I have the resources to replicate any player's movement from very low sensitivity up to very high standards like mine with 16000 DPI.