There's an intrinsic problem with TVC, which is the TVC reliability. One of the major reasons the F-22s did 2D TVC instead of 3D TVC was simple; their 2D TVC broke down way less often than American 3D TVC designs. Likewise, if you ask about Indian Su-30MKIs, one of the first things that comes up is how often the TVC or the engines has to be overhauled due to lock-ups.
The problem is, when it comes to reliability, you can't prove reliability without testing it; i.e, the TVC systems have to be tested to failure, preferably in real situations, and both the pilots and maintenance crews need to know what to do when the TVC fails.
So 6 TVC J-20s makes sense (and we'll see pictures before we know whether Minnie Chan is full of it again) to test and improve the reliability of the TVC system.
===
And likewise, if Chinese TVC nozzles are detachable, it might make sense to have J-20s operate both with and without TVC because of the difference in maintenance requirements.
===
I'd argue that TVC is an important step forward for J-20s. Canard stealth fighters, as I've stated before, really benefit from TVC as TVC allows them to lock the canard control surface to reduce RCS. And as mentioned before, this is one step forward to a finless mod of the J-20 that can beat current American 5th gens in terms of pure RCS stealth.