Most Reach detection methods are simply based off of running a ray trace from the attacking player (attacker)'s cursor position (middle of the screen) to the attacked player (defender)'s entity model. Sometimes this can be thrown off because if you have a higher ping, you're not sending your own position update packets as often as the defender, so the server may have you placed at, say, x 10, z 10, y10, when on your screen you are actually at x 10.4, z 9.8, y 10, but since your client hasn't sent the position update packet yet the server doesn't know that you're legitimately at those co-ordinates and so the ray trace would return an inaccurate measurement. So, most anti-cheats basically have to take the last position packet sent, then do some vector math basically trying to estimate where the attacker actually is client-side, then finally compare that value against the next real position value you send.
At least, that's how I would imagine most reach checks work, because that's how I would write one if I was to write a reach check. Keeping that in mind, you should really just always be using a very very slight reach level, especially if you get good ping since in that case you send your position updates more often so anti-cheat measurements will be more accurate against you.
You could use a combination of BeanerNet and Reach to get around that to an extent, at least theoretically. You'd have to do your own testing to see if that would work.
At least, that's how I would imagine most reach checks work, because that's how I would write one if I was to write a reach check. Keeping that in mind, you should really just always be using a very very slight reach level, especially if you get good ping since in that case you send your position updates more often so anti-cheat measurements will be more accurate against you.
You could use a combination of BeanerNet and Reach to get around that to an extent, at least theoretically. You'd have to do your own testing to see if that would work.
