Better Know a Stat- Expected Fielding Independent Pitching (x-FIP)

For the third installment of the TBLA "Better Know a Stat" I am going to take a look at a stat that builds on Fielding Independent Pitching (FIP) and functions as a better predictor for future pitching performance.

From The Hardball Times:

Expected Fielding Independent Pitching is an experimental stat that adjusts FIP and "normalizes" the home run component. Research has shown that home runs allowed are pretty much a function of flyballs allowed and home park, so xFIP is based on the average number of home runs allowed per outfield fly. Theoretically, this should be a better predictor of a pitcher's future ERA.

x-FIP begins at the same point as the FIP, which is to say it eliminates the defensive contributions (or deficiencies) in place behind the pitcher. Strikeouts and walks are handled essentially the same way as they are individual matchups between the pitcher and the batter.  The third component of FIP, homeruns, is looked at differently in this model.  As the definition above states, x-FIP argues that home runs are largely a function of the number of flyballs surrendered by the pitcher and the park they play in.  The idea that home runs are largely impacted by the home park is not terribly revolutionary.  For years, people have been discounting the offensive production of sluggers who compete at Coors Field (regardless of how true that remains today, but that’s a different discussion for a different day.) 

As for home runs being a function of fly balls allowed, this seems to me to be rather intuitive as well. A pitcher can’t surrender a home run without surrendering a fly ball. The more fly balls you allow, the more opportunities for home runs.  As their was some good conversation regarding FIP in the previous column, where certain pitchers are "ground ball" pitchers, no pitchers in MLB consider themselves fly ball pitchers.   The more fly balls the pitcher surrenders, the more likely they are to surrender a home run.

Examining the three mainstays of the Dodger pitching rotation from 2009 (Kuroda missed too much time so I am not including him here) it finds that 2 of 3 pitchers had an x-FIP that outperformed their ERA, while the third pitcher had an x-FIP that was essentially identical to their ERA.  Clayton Kershaw had the largest delta of x-FIP to ERA amongst the starters.  Kershaw had the best ERA amongst starters on the Dodgers at 2.79 but posted a (still very good) x-FIP of 3.90.  In fact, Kershaw’s HR/FB ratio of 4.1% was the lowest in all of major league baseball (and as a result his BABIP of .274 was 11^th in MLB.) That being said, his x-FIP of 3.90 is rather good, and it’s built on a very strong K/9 of 9.74 (7^th in major league baseball.) It’s difficult to predict what 2010 will be like for Kershaw.  The delta from x-FIP to ERA suggests that he should regress towards an ERA of upper 3’s, but his age and the prospects for increased development should temper the regression to some degree. That being said, it’s pretty difficult to post more strikeouts per 9 innings than he did last year.  Clearly, expect the home runs to go up a bit in 2010 and the strikeouts should hopefully stay around the same level.  I expect Clayton to have a solid year, but not an ERA below 3.00 again.

Randy Wolf had a very strong year with the Dodgers, so strong that he was able to secure a three year contract with Milwaukee after the season.  Wolf’s ERA was solid at 3.23 but had an x-FIP of 4.17; the highest of the 3 Dodgers analyzed.  Unlike Kershaw, Wolf doesn’t strike out an insane number of batters (though his K/9 of 6.72 is still pretty good) nor did he get as lucky with the HR/FB ratio (HR/FB ratio of 9.2%) though there was a downward adjustment towards the league average.  Wolf was very fortunate with balls in play, as his BABIP of .257 was tied with Jarrod Washburn for best in MLB.  Wolf’s BB/9 of 2.44 is in the top third of the league.   Considering Wolf’s age and moving to Milwaukee from LA, I expect a significant regression for Randy.

And what is there to say about Chad Billingsley?  Remarkably his ERA of 4.03 is right on with his x-FIP of 4.04.  With this evidence it’s hard for even the most hardline Billingsley apologist (such as myself) to argue that Chad was particularly unlucky last year.  A 4.04 x-FIP is certainly not bad as it’s a bit above average for a starting pitcher in MLB in 2009.  For 2010, I do expect a better performance from Billingsley, as I believe a bit of his "struggles" were the result of the broken leg last offseason—but the x-FIP doesn’t suggest that his ERA of 4.03 was abnormal in 2009.

Lastly, I appreciate the debate that occurred in the FIP thread last week and I hope it is continued with this installment. X-FIP certainly should be debated, as it’s a very new statistic (even in this world of new statistics) and a lot of the comfort level for me was how intuitive it was.  When I read about it for the first time, it just all made sense for me. While certain batters possess only "warning track power" I would think a pitcher would approach every batter with the objective of limiting fly balls.  It’s not as though Tim Lincecum pitches differently to Russell Martin knowing that he doesn’t have the strength to hit the ball out of the park.  That being said, if any of the readers have strong evidence, particularly arguing against the correlation of fly balls and home runs, please share them.