DRA, cFIP, and the Cubs Rotation Going Forward

Would it surprise you to learn that Chicago Cubs starters have been among the best in baseball in 2015? If so, it shouldn’t. Their 3.38 staff ERA ranks fifth in the game. Their 3.22 FIP is second, as is their strikeout rate. They have walked fewer batters, as a percentage of the total they have faced, than any team save the Nationals and the two New York squads. And, perhaps most impressively, their total fWAR derived from starting pitchers, 11.1, comes in behind only the Dodgers’ 11.5. Those guys, as I’m sure you know, have Clayton Kershaw and Zack Greinke fronting their operation. All in all, it’s impressive, and comes, by the way, despite the fifth starter spot being a bit of a disaster zone for Chicago: Dallas Beeler, Clayton Richard, Tsuyoshi Wada, and Travis Wood are not names that will linger long in the franchise annals.

Given that the starters have shown notable success already, and that the front office has a stated desire to add pitching before the deadline—one particularly juicy rumor from Ken Rosenthal, yesterday, has the team targeting Detroit’s David Price—I thought I’d take this opportunity to very briefly examine the performance coming from the top four spots in the rotation using BP’s new statistics, DRA- and cFIP. Taken together, these statistics should give us a sense of the degree to which each pitcher’s performance to date has been deserved, and the approximate level of performance we should expect from them going forward, before any potential deadline additions. Join me, if you will.

Let’s begin with the basic numbers, then break them down:

The first thing you should take away from this table is that the numbers look good. That much is expected: when the whole looks good, the parts generally do as well. (This is not true, by the way, of bratwurst.) So, onwards, to the relationship between ERA and FIP. In general, the way to interpret an ERA that runs higher than a FIP—which is a statistic that measures only walks, strikeouts, and hit by pitches—is to say that the pitcher in question is getting a bit unlucky; in other words, that they’re doing everything they can to get outs, the hits just aren’t falling their way. Here’s the Cubs’ pitchers, ordered by ERA-FIP differential:

By this interpretation, at least, Hendricks and Lester have gotten a little bit of bad luck, Hammel’s gotten a bit of good luck, and Arrieta has just about broken even. But, obviously, this isn’t the whole story. Just watching the Cubs regularly, you can see that Hendricks’ stuff (and, to some degree, 2015’s version of Lester) isn’t always located with the precision that Arrieta’s or Hammel’s is. That means that, sure, some of the discrepancy here might be luck, but it also might be pitchers missing their spots and leaving pitches that are easy to hit within easy reach of big-league hitters. It’s partly because FIP can’t capture these effects too well on its own that my much-smarter colleagues at BP recently invented DRA, which stands for Deserved Run Average. It aggregates a huge list of factors (you can read about them here) to come up with a pretty good estimator of how many runs a pitcher should have allowed, after luck is stripped out. It’s set to the same scale as ERA, so the basic interpretation is the same as FIP: if the ERA is higher, the pitcher has probably been unlucky. Here’s the ERA-DRA board:

Here, everyone comes out looking lucky. Given the way DRA is calculated, that’s probably the result of two main things: (a) the Cubs having a pretty good defense this year and (b) any exceptional performance, which the Cubs have achieved this year, usually being the product of over-performing true talent level. What’s more interesting—to me, at least—is that the order has switched. Whereas in FIP’s estimation the luck ran down from Hammel to Arrieta to Lester to Hendricks, in DRA’s estimate the luck has run Lester to Hammel to Arrieta to Hendricks. The only point of agreement between the two is that Hendricks has been the least lucky of the four. That makes sense to me: when Hendricks misses, his fringy stuff doesn’t allow those misses to go unnoticed. On the areas of discrepancy between the two? I’d tend towards DRA’s guess: overdetermination is often a problem, but here I think it’s pretty clear that DRA is a better estimator of past performance.

All right. Enough with past performance in general: the basic message there is that the Cubs have been various degrees of good, and Kyle Hendricks is probably due for a little regression going forward. What’s going on with everyone else, then? That’s where cFIP comes in real handy. This is another BP product, courtesy of Jonathan Judge, which unlike DRA (which is descriptive) is intended to be prescriptive, and predict—to whatever degree that such prediction is possible—the future. Let’s take a look at the chart for cFIP:

Well, that’s nice. Arrieta is predicted to be 24 percent better than league average; indeed, all of the Cubs’ top four pitchers are expected to be above average. Hendricks, as should probably be expected given the discussion above, is likely to be the worst of the bunch, but if the Cubs get performance six percent better than league-average out of their fourth starter, they’ll be very happy indeed. Even happier, actually, if they manage to add a starter like Price (or Johnny Cueto, Cole Hamels, or Tyson Ross) and bump Hendricks down to the fifth spot. Until then, they’ll have to be satisfied with having one of the best rotations in baseball. That’s not something that’s been said of them often lately, and it’s nice to say it now.

Lead photo courtesy of Jason Getz-USA TODAY Sports