It's worth bearing in mind that different actors will deal with working with greenscreen & tennis balls differently, depending on their individual experience and skill sets. In a way it's no different than some actors being unable to perform well in heavy make-up prosthetics or struggle with unscripted or overlapping dialogue. It's a skill.
If I had to hazard a guess, I'd say those actors that have extensive theatre experience are somewhat less likely to have difficulty than say those that came the TV/screen acting route. Indeed, IIRC the approach the Avatar mocap performers took was to treat the mo-cap stage a little bit like a theatre rehearsal.
I might go as far as saying that mo-cap is what finally made CG characters viable and able to fool the audience into forgetting they're not real. Compare Golem to say Jar-Jar, Watto, Taun We or the CG Yoda and you can tell the latter were (painstakingly) key-framed. The process is much closer to stop-motion animation puppetry and as such is *very* difficult to do convincingly. On top of that CG needs to read as physically real, while stop motion can get away with looking like puppets.
Since mo-cap became more common, it's treated less and less like animation and more and more like puppetry mixed with digital make-up prosthetics. It's increasingly allowing them to more able to bridge that treacherous uncanny valley.
It's also making it a lot easier for other actors to perform with CG characters if the other actor is there in the scene, with a performance capture rig.
Now back when they started making the prequels they didn't have any of that. Most of it was pioneered by Weta while ILM was still using trial and error to figure out what did and did not work. As is typical for this kind of thing, by the time they'd just started to get good at it, the project was over.
It's easy to sit back now and say this or that was a misstep because we have the benefit of hindsight. The first poor sods to beat a new path through an uncharted jungle are also the ones most likely to get cut to shreds or fall down a sinkhole in the process.
I recall an example of this trial and process in the very expensive Jar-Jar suit they had made for Ahmed Best to wear on-set for TPM. At the time they didn't really know when it came to the final process if they'd just be superimposing eye and lip movement or replacing the whole head. It turned out to be they needed manually animate the whole character and the suit was only really useful for lighting reference.
This is why in AotC, Taun We was just the voice actor on set, wearing a foamcore silhouette glued to a hard hat. If it were done today, it'd be a mo-cap rig.
