"When you defer to formulaic movies, you have a lot of "lets put this person and that person in a movie and see what happens"..."
You seem a bit caught up with this 'formulaic' approach, but I guess so are America's artistic producers. The primary problem with this approach is art cannot be, and never has been, quantified. A similar approach is often taken in the automotive sector, where sample groups and surveys are taken to decide how to style a car. The result; incredibly bland, uninspired vehicles nearly indistinguishable from past vehicles, current competitors and sales flops if the bones can't sell the car.
Hollywood, using a formulaic approach as you suggest, may be caught in this same cycle of unoriginality that churns out uninspired and unartistic rehashes indistinguishable, unremarkable and unable to sell in the absence of 'good bones'. For movies, these bones are story, acting and directing - three traits nearly absent in Hollywood today.
A similar move away from making quality art, and instead looking at formulaic approaches began after the music boom of the 90's. Attempts were made to mix rock, hiphop, country and electronica into a formula appealing to the largest focus groups. The result; music now sucks and the dramatic decline in global sales is proof. If not for the boom in digital sales that allows buyers to pick that one-hit from a crap artist worth listening to, sales would be even lower.
Using a formulaic approach to art, regardless of the median, has always and repeatedly been a failure. The only reason it continues to exist is as the only means for the unartistic to attempt to imitate the artistic, but the outcome is always unartistic crap. In the same way sports statistics are used by those that didn't watch, or don't know what to look for, to try and quantify what 'good' is, but if teams/players are selected as 'good' based on stats it will usually be wrong.
That's the difference between theory (formulaic) and reality - the proof in the pudding.