Lets use this thread to post 3dm files with cases where the “Blend surface” tool fails in a weird way, so that the developers could analyze the results and use this data for further improvements of the tool. the thread could be used to improve the “Match surface” tool, as well. I notice many similarities in the errors that both tools produce.
I will start with these simple input surfaces that produce quite interesting blend surfaces in-between, especially the bottom where the distribution of the control points is not logical. 
Funny results with Blend surface.3dm (151.6 KB)