On the heels of my post yesterday about "luck" for batters and the inestimable natsstats' analysis of April pitch f/x data for the Nats' pitching staff, I thought I'd take a look at top-level pitching performance to see how much "luck" the Nats have been having. Warning: I'm going to be throwing around stats like xFIP, but I promise to explain what they mean. The shocking truth about Livan Hernandez, Tyler Clippard, and Tyler Walker after the jump...How do we define pitcher "luck"?
The problem with trying to figure out whether a pitcher is good is that a lot of stuff is out of his control. For instance, the defense plays a big role in whether he has a high ERA or a low ERA--fielders with limited range can let through a lot of balls without actually committing errors. I use a couple of ways of looking at this. The first is to track the defensive efficiency ratio (DER) for each pitcher. DER is the fraction of balls in play that the defense turns into outs. It's the flip side of BABIP for hitters; in fact, you calculate DER by taking the BABIP of opposing hitters and subtracting it from one. Like BABIP, DER shows you pitchers who are "lucky" with batted balls going right at fielders (high DER) and pitchers who are "unlucky" with ball that find gaps or go to the right of Cristian Guzman (low DER). League-average DER is about 0.700. If a pitcher has a DER over 700, he's lucky. If a pitcher has a DER under 700, he's unlucky.
The other way of taking defense out of analyzing the pitcher is fielding-independent pitching (FIP). This a stat that's scaled to look like ERA (league-average FIP = league-average ERA) and is based purely on things the pitcher can control: strikeouts, walks, and HRs. There's some amount of luck in the number of HRs a pitcher gives up, since HRs are generally a function of batter skill, not pitcher skill. A modification of FIP called xFIP ("expected" FIP) uses the league-average rate of HRs per fly ball in its calculation instead of the actual number of HRs a pitcher gives up. Anyhow, a pitcher is lucky if his ERA is less than his FIP or if his FIP is less than his xFIP (ie, he's pitching "better" than his Ks and BBs say he should, and he's not giving up dingers). A pitcher is unlucky if his ERA is more than his FIP or his FIP is greater than his xFIP (he's pitching "worse" than his peripheral stats and giving up dingers).
|DER > 700||DER < 700|
|ERA < FIP||ERA > FIP|
|FIP < xFIP||FIP > xFIP|
For fun, we'll also put up peripherals like strikeouts per 9 IP (K/9), walks per 9 IP (B/9), and homers per 9 IP (HR/9) if they're notable. Also, I'm only looking at pitchers with 10+ IP, so no discussion of the Dreadful Jasons.
Let's start with this season's surprises, both good and bad:
Livan Hernandez has a league-leading 0.87 ERA. Don't expect that to last: his DER is an unsustainable 820, compared to a career average of 690: there's lot of good glovework behind Livan. That sub-one ERA doesn't compare well to his 4.07 FIP (only 2.9 K/9), and that FIP is even lower than his 5.03 xFIP. Livo has only given up 0.5 HR/9 so far this season, about half his career rate. Enjoy the good performances while they last--we'll still get a lot of solid innings from "Mr. National," but don't expect more shutouts.
John Lannan hasn't been much of an "ace" so far with a 6.34 ERA. His ERA is slightly higher than his 5.79 FIP and 5.79 xFIP (Lannan has exactly league-average HR/FB), and his DER is a below-average 661, both hinting at slightly bad luck. But consider that Lannan's performance has previously defied his FIP: in 2008 and 2009 his ERA was nearly a run better than his FIP (and over a run better in 2007!), and his DER was consistently about 724 in 2007-9. Fans of defense-independent pitching stats point to this and say, "ah-ha! We told you Lannan wasn't good all along!" It seems like this is a combination of some bad luck, and maybe a loss of control (Lannan's walks are up and strikeouts are down this season). He's not getting weak grounders consistently this season--it's a topic for another post.
Scott Olsen has been a mildly pleasant surprise, managing a 4.35 ERA with 7.8 K/9 and 3.5 BB/9. That's actually a bit worse than his 3.78 FIP (3.97 xFIP) and with a somewhat unlucky 657 DER (career he's 694). If Scotty keeps pitching like he is, we might expect some regression toward better performance later in the season.
Craig Stammen has disappointed so far with a 6.75 ERA and only 3.9 K/9, although his low walk rate of 1.4 BB/9 is nice. His DER is a somewhat-unlucky 661 (he managed 718 last season), but his FIP is 4.24 (xFIP 4.00). Stammenator seems to be having some bad luck; pitching to contact may yet work out for him if he can keep the walks down.
Luis Atilano is something of a Cinderella story so far, with a 2.25 ERA despite only 1.5 K/9 and a worrying 3.8 BB/9. He's had the benefit of a 743 BABIP, and is defying his 4.02 FIP, which is further defying his 5.62 xFIP. Livan 2.0 will likely end up the same way, in a regression of home runs driving in walked baserunners. Enjoy the fairy tale while it lasts.
Overall, Nats' starters have a 695 DER, 5.33 ERA, 5.13 FIP and 4.73 xFIP. Keeping in mind that those figures include some disastrous Mock/Marquis starts, the Nats' starters have results that are average-luck-to-slightly-unlucky.
The fact that I'm able to refer to a "bullpen," instead of some derisive nickname for the collection of Nationals' relievers is a sign of improvement this year. Like the rotation, we've had some (probably) unsustainably-good surprises and some, uh, less good.
Tyler Clippard is on a tear. We walks a lot of batters (4.5 BB/9) but strikes out even more (11.6 K/9--more than Tim Lincecum). Sadly, this is on top of a 799 DER (what does this look like? Nyjer Morgan's diving catch of a sinking liner against the fish last Friday to save 3 runs = unsustainable DER). Clip's tiny 0.50 ERA is, of course, less than his still-respectable 2.38 FIP. The problem is that his xFIP is 3.89. Peaches is a fly ball pitcher who hasn't given up any home runs yet! He'll still be respectable as a high-leverage reliever, but expect to start seeing some blown saves and losses come up when balls finally start dropping for extra bases or leaving the yard in the 7th or 8th.
Matt Capps is a good performer that I didn't expect, racking up 10 saves out of 10 chances despite a high walk rate (4 BB/9). His miniscule 0.66 ERA actually comes with an average-to-slightly-unlucky 682 DER (career he has 700 DER) and is much better than his solid 3.17 FIP, which is better than his decent 4.22 xFIP. This is another case of being lucky with baserunners (he's stranded 96% this season, compared to 76% on his career) and dingers. But the dingers will come eventually, the baserunners will score, and the saves will be blown.
Is Brian Bruney terrible? Oh, my, yes! Sure, the 647 DER is unlucky, but that doesn't hurt as much as the 10.4 BB/9! His 6.23 ERA is slightly unlucky compared to his (admittedly poor) 5.87 FIP, but not compared to his 6.23 xFIP--just think, he could be giving up more home runs, too. Bruney isn't going to regress to better performance on his own; he needs to fix his pitching.
Miguel Batista is actually... lucky? His current 803 DER is second only to Livan's (compare to his career 698). Sure, his ERA is an ugly 5.71, but that's better than his 6.68 FIP, which is actually worse than his 6.08 xFIP! Let's hope Batista stays in junk time, because his regression to the mean is going to be ugly(er).
Tyler Walker has gotten clobbered by the jinx stick, with a 6.28 ERA despite an awesome 10.7 K/9 and 0.7 BB/9. His 661 DER is unlucky, but not as unlucky as his underperforming of his 4.56 FIP and 2.65 xFIP! This is only through 14 IP, but we can possibly look forward to solid league-average bullpen work from Walker Texas Reliever.
Overall, the bullpen has a 4.05 ERA, with a 717 DER, 4.68 FIP and 4.66 xFIP. That's on the lucky side of average, no doubt. This means there's some regression coming; the luck won't hold up all season, and the Nationals will lose more close games in extra innings.
What have we learned?
The lesson's the same as yesterday's post: enjoy 0.500 while it lasts. Clippard, Capps, Livan, and Luis are all going to go from world-beaters to merely-above-average as the season progress (and Atliano might become below-average to bad). Olsen might stabilize around league average, while Stammen might improve from bad to below-average. Lannan might improve from worse to bad (unless he fixes his control issues), but his performance defies statistics. Bruney and Batista are not going to get better (barring a transformative McChatty), while Tyler Walker might be a pleasant surprise later in the season. Of course, we may have new names to analyze in the bullpen and rotation soon enough...