Just "armchair guessing" based on "questions that have not been tested for thoroughness or reliability"?
Personality psychologists working in the respectable districts of the field have been validating typologies by means of psychometric analysis of suitably large samples for many years now. McCrae & Costa (the leading Big Five psychologists) long ago acknowledged that the MBTI's dichotomies (unlike Jung's original concepts) basically passed muster by modern standards, and a very large meta-review and supplemental study in 2003 found that the MBTI was more or less on a par with the Big Five in the validity and reliability departments.
And there's more discussion of those kinds of issues on
this TC Wiki page if you're interested.
And speaking of large-sample studies, and for what it's worth,
here's a study of 5,700 gifted adolescents where the self-selection ratios for the types (i.e., the ratio of their percentage among the gifted population to their percentage of the general population) were as follows:
INTP 3.4
INTJ 2.87
INFP 2.68
INFJ 2.67
ENTP 2.32
ENFP 2.03
ENTJ 1.49
ENFJ 1.26
ISTJ 0.99
ISTP 0.78
ESTP 0.49
ISFJ 0.40
ISFP 0.40
ESFP 0.28
ESTJ 0.26
ESFJ 0.24
That's what
validity looks like, ZNP — and validity is what separates the respectable districts of the personality psychology field from astrology. The way the 16 types fall along the spectrum in that study strongly indicates that an N preference has the strongest influence on whether a child will be "gifted" in one or more of the ways (including IQ) defined in the studies included in that meta-review, and introversion has a significant (but less strong) influence.