Pseudo R-Square

This is just a quick  plug for a good blog post by Brendan Halpin on Pseudo R-Squares. I don't think I've ever had any answer, never mind a good answer, when I've asked paper writers, seminar presenters etc why they have reported a pseudo R -Square measure? Yet still it turns up again and again and ritual dummheit builds on ritual dummheit.

Come to think of it I've not seen much intelligent use of plain vanilla R-Square in the context of linear models unless the problem at hand has a clear prediction flavour to it.