All interesting theory but...........
2 things are happening -
1/ The cyclic induced PIO. Friction up the cyclic as per instructions in the Llama RFM to the point where it is quite difficult to move. You'll get used to it. The controls are obviously fixed to the frame and if the mast/transmission start to oscillate it subsequently moves the controls more. The better the balance of the MR the better the hover as well. Don't stir the pot. With friction on you'll soon tire of trying to.
I have had a 355 hover for 45 seconds or more without touching the cyclic. This is with .02 IPS.
2/ The wake from the MR OGE misses the tail rotor due to contraction and when IGE moves across the tail rotor and causes translation lift in the TR but not consistently. This causes a yaw roll couple and gets worse the closer you are to the ground. This can be proved by hovering in a tail wind. Keep it straight with your feet. The wake around the cabin is minimal.
The strake on the B2 was to create a Coanda effect to help the tail rotor by destroying lift on the right side. I doubt there is much Coanda effect without it as per B, D. Interesting the B3 doesn't have it.
The tank sloshing is limited to earlier unbaffled tanks and will NOT be noticed in the hover unless your a stick stirrer, only in the cruise at about 60-70%.
Try it. Works for me, steady as a rock.