Not too long ago this interesting article was posted which talked about how the park adjustment used in the calculation of fWAR for pitchers currently uses the basic park factor rather than a FIP-specific park adjustment for pitcher fWAR. Since fWAR for pitchers is calculated off of FIP rather than ERA, it makes sense that the park correction used should also be fielding independent (basic park factors are based off of total runs, a fielding dependent stat).

The theory was appealing to me, especially since it helped explain why my favorite team, the Mets, ended up dead last in MLB pitcher fWAR despite being ranked 17th in the majors in FIP (9th NL), 14th in the majors in xFIP (9th NL), and 9th in the majors in ERA (6th NL). Citi Field is a pitcher's park, but I didn't buy that the Mets staff was truly the worst in the league - especially since their 2014 team fWAR ended up a full 9.4 wins below their Pythagorean win total, suggesting that either the Mets scored a lot of lucky runs or that fWAR wasn't giving our team the credit it deserved. Something seemed fishy to me about the way fangraphs was calculating fWAR for our pitchers, so I decided to do some spreadsheet-ing.

I looked at the 2014 season here to see how well fWAR did at correlating with team outcomes, and then compared it to how they would have done if they used the FIP park adjustment suggested in Noah's article on fangraphs. The "fWar - actual/pythag same sign?" column is checking whether those values are the same sign as the "delta-Park Factor" column. That is checking whether the change in park factor from basic to FIP would move the pitching fWAR for that team towards or away from the actual/pythag record.

The conclusion:

-Out of 30 MLB teams, 26 would have had a different park adjustment using an FIP park factor rather than the basic park factor.

-Out of those 26 teams, 18 would have had their team fWAR adjusted in the correct direction towards their pythagorean record by using an FIP park factor rather than the basic park factor.

-Out of those 26 teams, 20 would have had their team fWAR adjusted in the correct direction towards their actual record by using an FIP park factor rather than the basic park factor.