FIP, xFIP and what to do with HR/FB

Update: I have run a calculation on some players an they can be seen here.

There has been more discussion on the uses of FIP and xFIP at FanGraphs and a response by Derek Carty at THT Fantasy. I agree that FIP has it's limitations, but we don't have to judge a pitcher by FIP alone and we do have access to the HR/FB to access a pitchers actual numbers. FIP is just an assistance and should not be looked at as a full prediction alone. My problem right now with xFIP is the assumption of a league average.

It's well understood that a home park as well as league can effect a pitchers HR/FB and many teams will be effected by these results to have much lower HR/FB numbers. Several teams have maintained HR/FB as a team that beat the league average by a full standard deviation. I will show the data here for 2006-2009 by team to see what teams maintain certain levels.
The column under the year is the amount of change from the year before (all made positive to find the average change). There is a lot of data here, but you can see there is some change from year to year. There is regression, but there is also changes due to stadium changes, free agency, etc.

Let's look at the teams that might be better predicted by using FIP instead of xFIP. This may not be the best answer, but these teams have shown an ability to control their pitchers xFIP.

The San Francisco Giants are by far the most consistent at maintaining a HR/FB one full standard deviation from the league average. They have also not had large changes in HR/FB with a change from year to year of no more than the league average. The Kansas City Royals have also maintained a strong HR/FB year to year, but that has been hidden by poor pitching in general. The Padres also have been strong at keeping a solid HR/FB, but not as strong as the others.

Another team to watch is the New York Mets who have had the largest change of any single team in the past 4 years. The new stadium has caused nearly a 50% lower rate of HR/FB for Met pitchers. This maybe an exaggeration of the new stadium effect, but it still looks like we will have another team beating the league average year in and year out.

We already know that xFIP shows better results at predicting ERA than some of the other choices, but it looks like this data could suggest there is more that should be accounted for. I think the best calculation using what we have is to take xFIP and replace the league HR/FB factor with a team HR/FB factor. I might try to pull this together if I can get a hand, effectively adding park and league factors to the use of xFIP.

Tags: FIP, Baseball, Projection