Re: The thought presents itself...
If it's anything like the Falcon 9, the problem will be how deep you can throttle the engines.
Rocket engines can't throttle down indefinitely (much like an internal combustion engine has an idle speed). On a F9, the minimum throttle of one Merlin engine still generates more thrust than the entire stage weighs (since it's basically an empty coke can at the point).
Consequently the Falcon 9 cannot hover. The maneuvre that Harriers, F35s and helicopters pull of coming to a halt next to a carrier and then edging in and dropping down is not available to F9. Once it stops falling it will start to rise again. It's termed a suicide burn (or a "hover-slam") - you have to time it so your vertical speed equals zero at exactly the moment you touch the ground, and then you kill the engine before you lift off again.
If your vertical speed hits zero above the ground, then you either have to fly off and blow it up safely, or cut the engine and fall. You can't lower it down "under power". Unless you run this simulation quite high up, there won't be time to cut and relight the engines for the real landing. You pretty much guarantee you'll lose the prototype instead of probably losing the prototype.
I'm not sure about the exact thrust/weight ratios of StarShip/Raptor engines, but I suspect that if they tried to do a "simulated landing" at (say) 500ft, the rocket would then have to fly off and be destroyed anyway - they wouldn't be able to land it afterwards.