Once upon a time, in the early days of WHYGAVS, there was a pretty heated debate between myself and some of the commenters about Jason Bay's value on defense. This was before really advanced defensive metrics were widely available on the internet and I was arguing that while Bay had a weak arm and wasn't the fastest cat out in left, that he had generally good positioning and ran good routes and that made him a good outfielder. The commenters, lead by Gavin (who I think still reads but doesn't comment as much, as the Pirates have worn his spirit down) argued that I was wrong; that Bay sucked. As the more advanced stats became available, it seemed more and more clear that the commenters were right. Metrics like UZR, PMR, and Plus/Minus aren't perfect, but when they all agree that a player is a bad fielder it's a pretty good indication that he's a bad fielder. Still the functional difference in defense from a good defender to a bad one is some times hard to detect just from watching games.
Enter Zach Duke. Duke was clearly an improved pitcher in 2009; he struck out more batters while dropping his walk rate and leaving his home run rate virtually unchanged. His FIP dropped from 4.46 to 4.29, his xFIP dropped from 4.81 to 4.42, even his line drive rate dipped a bit (see this post for an explanation of FIP and xFIP, if you're unfamiliar with the terms). We'll talk about why this improvement happened later this week (charts and graphs!), but today I want to put the focus elsewhere. Everything that Duke did in 2009 pointed to less than half a run improvement per nine innings, yet his ERA dropped from 4.82 to 4.06. Why did this happen?
The explanation itself is tied up in the idea that for the most part, pitchers have very little control over the balls put in play against them. This is basic stuff for the sabermetric types around here, but if you're not familiar with the concept, it's very straightforward. Most pitchers tend to allow a batting average of about .290 on balls that are put in play but not hit over the fence (that's Batting Average on Balls In Play, BABIP). Hitters fluctuate; fast guys like Nyjer Morgan or line drive hitters can consistently be above .290 while other hitters can consistently be below.
One fun exception to this shorthand rule for pitchers is bad defenses. Prior to 2009, the Pirates' defense was generally bad and Zach Duke's BABIPs from 2006 through 2008 were .327, .360, and .327. Ian Snell's in the same span were .322, .313, and .358. Paul Maholm's were .325, .315, and .289. It's notable, at least to me, that Maholm's numbers are lower (while Duke and Snell spiked high, he had a reverse-spike low in 2008), because he's really the only ground ball pitcher of the group and the Pirates infield defense in that span was generally better than their outfield defense.
In 2009, though, the Pirates replaced Bay with Nyjer Morgan and Lastings Milledge, two players who rated excellently according to UZR. They mostly replaced Nate McLouth (who the defensive metrics were unkind to, despite his gold glove) with Andrew McCutchen, and they replaced Xavier Nady (who was slightly above average, according to UZR) with Brandon Moss and Garrett Jones (Moss grades out very well defensively while Jones was below average but not awful). As a result, Duke's BABIP dropped to .296. In fact, of the three pitchers who pitched the most for the Pirates in 2009 (Duke, Ohlendorf, and Paul Maholm), their BABIPs are directly related to their flyball percentage. Ohlendorf had the highest FB% at 42.1%, his BABIP was .265. Duke's FB% was 32.8, his BABIP was .296. Maholm's FB% was 29.9%, his BABIP was .310.
Now, please keep in mind that I am not arguing that there's a linear relationship between fly ball percentage and BABIP; not on the 2009 Pittsburgh Pirates or ever. But what I think is clear is that a big part of Duke's turnaround in 2009 (and Ross Ohlendorf's breakout season) is due to a huge improvement in outfield play (the infield defense improved too, but I don't think that it was nearly as dramatic an improvement as people think; I'll take a closer look at this in the near future). That's important to consider when trying to predict if Duke can duplicate his 2009 numbers; Milledge and McCutchen will be back roaming left and center field at PNC Park, so while Duke may come back towards his periperhals a bit, it's not likely that he'll drop back down to pre-2009 levels when it comes to ERA and WHIP because the defense behind him will likely still be good.