Over on ENWorld, there is a interview between Morrus and Jason Bulhman (lead designer of the Pathfinder RPG). He, like several others including Mike Mearls, points to various parts of d20 design as more an "art than science".
The d20 engine has been in use since 2000. It's been stretched, it's been turned on it's head, it's been blown up and rebuilt. Part of Paizo's mission should be to move past that miserable statement. D20 fans have rated and compared classes for literally years. At this point there should be a clear science and if there isn't, then a central goal of the Advanced Class Guide should be to implement it, no merely present what appears at first blush to be pre-built gestalt classes
I think Paizo and in particular Bulhman do excellent, flavorful work, but c'mon let's move past hind behind these kinds of statements.
I won't lie, one of the reasons I love 13th Age is because they aren't hiding behind "art" to make a great game. Heck, even with my problems with DnD 4e and the direction DnD Next took, there is a "science" to what they are doing.
Part of the article touches on whether Paizo imagined they'd ever be the at top of the heap and on RPG.Net right now is an article asking if they, instead of WotC, would be better tenders to DnD.
Well, one, let's not pretend that Paizo has faced any competition from WotC in three years (and I don't think Paizo is pretending that). Make no mistake, DnD Next will be the leader in the RPG industry when it launches. As good a company as Paizo is, it's time to stop hiding behind DnD 3.75 and at the same time pretend their leading the industry. Make positive changes to the d20 engine and don't talk to us the same way WotC does. Innovate, even if it it is with an eye toward the DnD 3.5 and Pathfinder fan base.