Outliers and the Planning Fallacy

I follow John D. Cook’s blog and yesterday’s entry was very much related to my prior post relating to tracking/measuring blockers. The connection is this piece:

You may also decide that an extreme value is legitimate, but that you wish to exclude from your model. The latter should be done with fear and trembling, or at least with an explicit disclaimer.

So, even if you think an event is an extreme value (your one-time blocking issue), try to avoid excluding it from your data just because it’s extreme — you may find that it provides valuable insight down the road.

Leave a Reply

Your email address will not be published. Required fields are marked *