I know, let’s use the outlier as a baseline
You will find this very difficult to believe, but I recently had a discussion with a person working at another company. This person is responsible for the statistical analysis of the work that other parts of the organization is to perform. Those looking for the information come to this person to make some sense out of the historical record.
In the recent history of the work, there was an event, in which hour estimates were made for the work, in the middle of the work, it became necessary to close the work up as best possible and stop since a hurricane was in bound. It was not possible to keep the project going given this storm so the work is prematurely closed out best possible to allow all to reach safety. Those on the team know this antecedent event and the subsequent consequences to the effectivity work.
Given the knowledge that this last point is an outlier, and we know the reason for it being an outlier, you would think the choice for the next effort would exclude this value, as it does not represent what is likely to happen, or at least happen in the way that truly meets the project objective. Yet, the outlier is the number the team wanted to use even though this number and situation do not represent typical performance.
What would make a team choose this? This value is lower than all previous values, but represent an anomaly from previous performance.