I personally have difficulty believing in something that has so many moving parts. I tend to find simpler to be better. It really doesn't help that it's proprietary which means it's a black box and you don't really know how it works. That's not to say it doesn't have some value. I think the best way to use statistics is never to use 1 catch all but instead look at what their data suggests. There's a rather famous phenomena where you can have individuals guess at say the number of marbles in a jar or the weight of something. However accurate one individual may be, more often than not, if you take a large sample of people and use them instead to average their guesses you'll usually find that guess is more accurate than any one person. Obviously any model tries it's best to model what's happening but there's almost always going to be flaws. Using multiple aspects allows you to use the wisdom of them all.
I will say this though, the system seems to be overrating Arizona pitchers for some reason. Robbie Ray is #44, Brad Ziegler was #46, Rubby De La Rosa was #47, Zachary Godley was #81, and Zack Greinke was #90. Those examples strike me as fairly odd.