This is where statistics are used. Not 100%, but preferably upwards of 70%. Or lower than 30%. Either suggests other than simple chance. As you move into the edges (<10% or >90%), then you have more reliable evidence.
I'm not comfortable with this. I'm much more comfortable with saying you're looking at different aspects of the personality.Perhaps it displays what could be termed as the actualised self where as the MBTI is more the raw self? (strictly ignoring the imbalance between the two systesm and other such intricacies)