Some important points were left out of this analysis. The basic premise of matching up relievers to batting order ignores counter moves, such as the effect of pinch hitting and the randomness of the game, whereby bad hitters are still hitters who can take walks, get hit by pitches, run into one and hit it a long way, etc. It was a very one-sided, static analysis.
All the writer did was describe "how to burn your bullpen 101...in a game and over the course of the season." What I see being the difference in a successful bullpen year after year is its depth. It doesn't matter if you have one guy shouldering a heavy load, you are still going to need arms at some point in many games. The pecking order isn't really that big of a deal, because effectiveness will end up deciding that.
And that is the problem with most of this theory, it purports that there is a correct way to manage the bullpen and that smart people will eventually find the holy grail (a science), when really it is more an art based on the quality, durability, and depth of arms in the pen. Usually when you see a manager trying something "outside the box," it is a case of "necessity is the mother of invention" rather than the manager being smarter than everyone else.