My issue with the thought process is the general picking of uniform distributions to show random data that follows DK. Uniform distributions are kind of uninteresting in this case because they're pretty artificial. It's not like competence is actually measured in the range 0-100 in anything that actually matters.