Each factor won't be a clear binary but a continuum between the two. Your answers will depend on mood, sure, but will also give a result on that continuum and with a margin for error.
You might've got ENTP consistently. I've had both INTP and ESTJ in fairly quick succession without any intention to mislead. The idea that the test is anywhere near perfect, foolproof and likely to produce absolutely consistent results for all appears unsupported by anecdotal evidence here.
Absolutely. I just wanted to provide an anecdotal counterexample, given that my experience was opposite to the one of the original commenter.
Actual psychologists have studied Myers Briggs and found that the likelihood of getting a different result on the exam after taking it more than once is about fifty-fifty, which is pretty much the definition of pseudo-science.
I didn't claim otherwise. I simply provided one data point that was different from the one expressed by the person I was replying to. I didn't claim the test was scientific or accurate. (Note: Precise != Accurate. And it was only precise in my case, not in general.)
That's interesting, I had the opposite experience. I must have taken the test 5 times over the years, and I got ENTP every time. I didn't expect it to be the case, but I was surprised by its precision.