I've got no problems dropping pitcher PA, I suppose, but if you really want to drill it down, you'll want to adjust for all opposition PAs. For example, Odorizzi had the 3 worst offenses in the AL as divisional opponents and got to accumulate 10 starts against them; Wheeler only had the Marlins as such a team in his division, for 3 starts. Not to mention adjustments for parks, defense, etc. Over 700-800+ PAs, there are all kind of factors that might be as big as ~50 pitcher PAs. FWIW, bWAR already includes some adjustment for opponent offenses, "RA9opp", although I think it's just on a team level rather than PA level. And it looks like, based on that, and park and defense, Odorizzi beats Wheeler in bWAR 3.6 to 3.5 for 2019 despite one fewer start and 36 fewer innings. So you might be right to give a slight advantage to Odorizzi. (Although Wheeler has a higher bWAR -- 4.1 -- listed in his summary at the top of the page, but I suspect the table value is most up-to-date.)