What's Really Behind the Twins BABIP?

Over the past few days, there have been a number of great discussions about the Twins pitchers across the Twins blogosphere. After some discussion about what starting pitchers the Twins should target during Twinkie Town's "Trade Deadline Week" (a week early), Luke in MN provided an excellent rebuttal of the conventional wisdom by arguing that The Twins have the best starters in the AL. The core of Luke's argument focused on the starting rotation's fielding independent pitching numbers, noting Francisco Liriano is #1 in MLB (and it's not even close), Carl Pavano is top ten, etc. But the pitching staff doesn't get the credit they deserve because high BABIP numbers for Liriano, Scott Baker and others has resulted in ERAs that are significantly greater than xFIP. I've argued that while Liriano's high BABIP is due to luck. Baker's is due to a very high line drive rate. Yesterday, Aaron Gleeman provided another explanation in his Twins Notes that outfield defense (Twins rank 24th in MLB according to UZR) is a major reason that Liriano's BABIP stood at a league high .357. Naturally, I wanted to test our theories using my "Total Run Accounting" software. Is Liriano's high BABIP really due to our outfielders' inability to catch his fly balls? Is this consistent across all five of the starters? In general, how are the Twins starters BABIP relative to their "Expected" BABIP, which I discussed in far more detail here in "Another Way of Looking at BABIP" in May. 

The answers might surprise you. More after the jump...

First of all, let's look at the Twins starting rotation (with Nick Blackburn instead of Brian Duensing, since he has so many more starts) and how their batted balls break down by type. Note: My numbers below may differ from what you find on Fangraphs because (a) my data is only through July 19th, downloading MLB's data takes some time, (b) I use data directly from MLB's Gameday site, they specify ground ball, fly ball, pop fly, line drive and bunt, and (c) I don't remove sacrifice bunts from the equation.

From my previous examination of BABIP, I determined the "expected" BABIP for each type of batted ball.

We can apply these "expected" BABIP numbers to each pitcher's actual batted ball breakdown to determine their own overall "expected" BABIP (eBABIP).

A few numbers jump out at me here. First, Liriano has been even better than I thought at inducing ground balls, nearly 2.5 to 1 when you remove the pop flies. Second, both Slowey and Baker have been very good at inducing pop flies, which helps to counter their high line drive rates and keep their eBABIP down. Finally, guess who has given up the fewest line drives and has the lowest eBABIP? Mr. out of the rotation Nick Blackburn. Go figure.

Now look at each pitcher's ERA compared to xFIP and BABIP to eBABIP.

As you can see, only Pavano has outperformed his xFIP, likely due to a significantly lower than expected BABIP. On the other hand, everyone else's BABIP is much higher than expected. I have to admit that I was wrong about Baker's BABIP being driven by his line drive rate, he has nearly the same gap relative to expectations that Liriano has. I noticed the same overall BABIP effect for the 2009 season, but attributed to a well below average defense, as evidenced by the Twins overall BABIP by type for the 2009 season:

Here are the 2010 numbers for the Twins pitching staff:

What jumps out here is a much improved infield defense (thank you J.J. Hardy, Orlando Hudson and Danny Valencia), but the second worst outfield defense in baseball. This may be partially due to more balls falling in front of the fence rather than flying out of Target Field, but when you put Delmon Young (albeit improved over 2008 and 2009), Michael Cuddyer and Jason Kubel out there every day, that's not a very good outfield defense.

To Aaron Gleeman's point, he appears to be correct, fly balls are a driver for the high BABIPs. But one final point regarding BABIP and fly balls. For whatever reason (likely luck / sample size), not every pitcher has had issues with fly balls. Remember, across all of baseball the expected BABIP for fly balls is .171, and overall for the Twins it is much higher, .220.

Wow. Just compare Liriano to Pavano on fly balls. Liriano has been extremely unlucky on fly balls, his .313 BABIP is highest in the Majors, well above the second most unlucky pitcher...Scott Baker! Among non-Twins, Chad Billingsley is third highest at .286. How valuable is outfield defense again? On the other side, Pavano is one of the luckiest pitchers in baseball at .132 BABIP on fly balls. Ted Lilly is by far the lowest at .058, but Pavano is still well below average, though not in the bottom 40 across all starters.

In summary, Carl Pavano has been really good so far, but evidence indicates that he's been lucky, which may come back to haunt him at some point. Francisco Liriano and the other starters have been relatively unlucky, although outfield defense certainly appears to have played a role in that "luck".