Isn't it allowed when trying to figure somebody's type that you make your own conclusions because people can be wrong abou their types if they have not studied the subject? Even people who have studied the subject can be wrong about their type. Wishing to be ENTJ doesn't make you one. I didn't change to be ESTJ even if I one time wanted to be one.
It's hard to tell. I have been through this enough that I no longer discuss members' types in open forum (or avoid it like the plague, at least).
There are people who I think aren't as experienced with type, based on their online assessment of themselves and others, and their claims are not accurate; but when you do start arguing with people, yes, one of the key points is, "Is this person expressing their normal self online?" (i.e., are my perceptions picking up representative data?) and "Is it hubris to think that you know someone better than they think they know themselves?"
It is rather a gray area. People aren't always their complete selves online. And it does seem prideful at times to think you might read someone better than they read themselves. But you cannot assume people understand the system or know themselves well; if this was true, the need for therapists would shrink a great deal.
As an example, my ISTJ engineering FIL took the MBTI and came out ENFJ. If you'd meet him in person, you'd just have to laugh, because he is a
stereotypical ISTJ. But either the test didn't ask the right questions, or he got hung up on particular internal details rather than overall patterns of behavior. After he met with the work counselor, he agreed with the ISTJ. But it was just absurd... and yet he tested himself and came up with this ridiculous answer... and he's a very smart man. He's just not very good at all with intrapersonals.
It is sort of a matter of seeing what trends in one's perceptions and choices are important and which ones are not.