I've attached a very simplified LTSPICE model.
The LED is modelled as having a forward voltage of 250 V, an on-resistance of 100 Ohms and an active circuit limiting the forward current to 200 mA. You would need to tweak this to match your LED and drive. Run the model and plot V(line,neutral) and I(I1).
I find that if I put a capacitor in series then about 4u7 will shift the conduction earlier by around 1.2 ms, which is only 22 degrees. A smaller capacitor shifts it further but also significantly limits the LED current. A series inductor of around 500 mH (huge!) will shift it about 1.5 ms, around 27 degrees, the other way before the LED current starts to drop.
If I change the model to be closer to a 110 V device (say 130 V forward voltage, 50 Ohms on, 400 mA limiting circuit) then I can put in greater X:R ratio before the LED current starts to drop and I can get phase shifts of around 90 degrees each way but there is a strong danger of just blowing up the active current source feeding a 110 V from 230 V like this. Essentially you must have enough series reactance to drop it out of regulation or as soon as the current source starts to cut in it will increase the voltage drop and be destroyed (unless they use the high voltage components in the lower voltage models).
TLDR: You can simulate it. You can't phase shift 230 V ones enough to matter without seriously dimming them, and whilst you can phase shift 110 V ones enough you can't allow the driver to operate without risking destruction so the only thing you're gaining over a driverless device is the rectifier.