#1: Doesn't matter if it is Zones with defined boundaries or fuzzy iLevels. If they are defined incorrectly from a poor or incomplete data set they are wrong. iLevels aren't magic.

#2: Not really sure how this applies to my question.

#3: Not sure how this applies either. It is well seen that the WKO model when given a limited number of performances can't accurately or logically determine performance possibilities.