Psychology and predictive power are two things you should never use in the same sentence if you don't want to give psychologists an existential crisis. The mind is a chaotic system (as in chaos theory), to draw a parallel to physics the best psychology can do is say "this double pendulum will only swing within a certain radius": True, no doubt, but also somewhat unsatisfactory.

None of this will ever be "You're an ESTP therefore I predict you will be a base jumper". I mean if you look at base jumpers there's a fuckton of ESTPs there but even more of them are bus drivers or something completely random. Predicting, in this context, is more of the statistical kind. More like predicting someone's sex by their height, with the additional twist that we have to prove the existence of different sexes by there being a bimodal distribution in height.

Both CT and MBTI share the same root in Jungian function theory but they use different definitions, the CT ones are very precise, example: Sensing. MBTI as well as other systems have a very, very hard time actually agreeing on what type someone is, to a large degree I'd say those imprecise definitions are the cause. Even "vibe-typing" based on CT definitions and vultology will have more cross-examiner agreement than going with the formal MBTI tests.

But that doesn't mean that those definitions are miles apart -- as you see in the Sensing example, they do have some correspondence. Which means that while the type matrix between the two isn't identical, it's also not completely different. As such CT holding up to scrutiny would also mean MBTI holding up to scrutiny to the degree that it doesn't stray from CT too much.

Which is pretty much all just to say that comparing MBTI to astrology is kinda off-base. MBTI might be problematic in a hell a lot of ways but it's also not claiming that Mercury being in retrograde at the moment of your birth has any impact on you, that's just plain aphysical, MBTI makes no such claims. MBTI might be, as formulated, unfalsifiable, whereas astrology is right-out disproven.