Opposite Directions of Skills and Results

While perusing the league leaderboards last night I noticed that two former Phillies—Gavin Floyd and Kevin Millwood—are on the opposite ends of the ERA-FIP spectrum. Floyd’s -1.45 E-F is the largest negative discrepancy in the American League. Millwood, however, has a +1.31 differential, which ranks behind nobody other than Carlos Silva in his league. By looking solely at their standard barometers of W-L and ERA the seasons of these two pitchers can be very misleading.

Gavin Floyd, White Sox, 25 yrs old
10-6, 3.57 ERA
6.19 K/9, 3.86 BB/9, 1.24 WHIP, .223 BA, 1.31 HR/9
.237 BABIP, 69.7% LOB, 5.02 FIP

Kevin Millwood, Rangers, 33 yrs old
6-6, 5.40 ERA
6.73 K/9, 3.32 BB/9, 1.74 WHIP, .326 BA, 0.91 HR/9
.379 BABIP, 69.7% LOB, 4.09 FIP

Millwood has the higher K/9 as well as the lower BB/9 and HR/9. His Zito-esque WHIP and 5.40 ERA can largely be attributed to his .379 BABIP, a number 35 points ahead of closest competitor Livan Hernandez. Though his strand rate isn’t abnormally below average, Millwood has allowed plenty of baserunners thanks to no help from balls put in play against him. His 27.1% rate of line drives, which leads the league, could account for this; next closest is Jon Garland’s 24.2%. Over the past three years, when his BABIP was lower, he posted line drive rates of 20.6%-21.3%.

Floyd’s ERA is deceiving not just due to his controllable skills but also because he has actually allowed more runs to score than it would suggest. Floyd has allowed just 49 earned runs but 66 total runs; that’s 17 unearned runs that have scored against him not taken into account with his ERA. He also has the second lowest BABIP in the league, behind only Justin Duchsherer, and the seventh lowest line drive rate. While both he and Millwood have identical strand rates, Floyd has allowed much fewer runners to reach base thanks to an unsustainably low BABIP; Millwood’s has essentially been unsustainably high.

What happens when we plug these guys into the in-season Marcel?

Floyd: 11 GS, 4.85 FIP, 67 IP, 62 H, 25 BB, 46 K, 1.31 WHIP
Millwood: 11 GS, 4.04 FIP, 58 IP, 73 H, 21 BB, 44 K, 1.61 WHIP

All told, Floyd would end with an FIP of 4.96 and Millwood much lower at 4.07. Regardless, Floyd could win 14-15 games and look much better than he should whereas Millwood would have been the better pitcher in terms of controllable skills; his results, however, would be much worse. He can’t possibly keep up a .379 BABIP, just like Floyd can’t possibly sustain a .237, but it might be too late for their regression to make a truly significant impact on their overall seasonal lines. If anything it should still lessen their ERA-FIP differentials.

This is just another example of how W-L and ERA don’t necessarily do a pitcher justice. It seems like forever ago that Millwood was an important part of the Braves rotation, and toiling in Texas, owners of perhaps the worst rotation of the last fifteen or so years in 2007, hasn’t helped, but he has definitely been much better than his numbers suggest this year.

While it’s certainly possible for pitchers to outdo their FIP with ERA (see: Carlos Zambrano) I would tend to bet Floyd won’t fall into this category. Then again, I could be biased due to seeing him struggle for the Phillies early on. Floyd has been a key component of the first place White Sox, who quite possibly have the best rotation in the league (it’s either them or Toronto), but his success is hinged upon a ridiculously low BABIP. It won’t even out as the remainder of the season plays out but I would be hard-pressed to believe that, assuming he hovers around the average strand rate, when his BABIP regresses that his ERA will stay in the 3.57 range.