I have been a long time student of decision methods, and the concept of a 'Hobson's Choice' (i.e. take it or leave it) came up recently. The formal classification of different kinds of related decisions are described in this WP article.
Not to say this concept is often brought up when modeling decision trees or processes. It is much more invoked when describing observed and often bizarre human decisions. But could we tag a tree describing business process to make it more useful? Is this inherently what is done with Bayesian methods? Seems this should also be thought about when automating data science. Thoughts?