The Road They’ve Traveled: George Plimpton, DRA Run Values, and the Cubs Rotation

One of my favorite baseball books is George Plimpton’s 1961 classic, Out of My League. Plimpton, the founding executive editor of the Paris Review—and an enormous baseball fan—took it upon himself to pitch batting practice to the 1958 National League All-Stars a few hours before they took on the Junior Circuit at Yankee Stadium that year. Plimpton had hoped to face the American League lineup as well, but he tired after his run through the first nine hitters, and was relieved by Ralph Houk. Out of My League, a slim book written in elegant and loving language, is Plimpton’s record of his experience on that hot July day. I love the book because it expresses in language of crystalline clarity what we too often forget: that major-league baseball is very, very hard. If you put me on a major-league mound, for example, I’d do no better than Plimpton, and quite likely a great deal worse. And I wouldn’t even have to face Willie Mays at the height of his powers.

But the fact that baseball is hard to play, in general, does not mean that it is equally hard for every player in every year. This is especially true for pitchers: some get lucky and face weaker batters, in cooler weather, and in ballparks more suited to their stuff and command. Some have excellent defense behind them, and umpires behind the plate who are willing to call the low strike on the outside corner. And some, of course, don’t have any of those things, and the fates conspire against them to put them in the worst possible situation every time they run out to toe the rubber. It’s this simple fact—that major-league pitchers do not ply their craft in consistent conditions—and the problems of pitcher evaluation that it engendered, that led my colleagues Harry Pavlidis, Jonathan Judge, and Dan Turkenkopf to develop Deserved Run Average (DRA) earlier this year.

The idea behind DRA is simple: it attempts to present—on the same scale as RA9 (runs allowed per nine innings)—the rate at which a given pitcher deserved to give up runs. It wouldn’t make sense, for example, to value Pitcher A (RA9: 2.30) and Pitcher B (RA9: 2.30) equally, if Pitcher B pitched in the NL Central and faced the Cubs, Cardinals, and Pirates consistently, while Pitcher A threw in the NL East and faced the Phillies, Marlins, and Braves on a regular basis. DRA adjusts for this difference in degree of difficulty, and assigns more credit to Pitcher B (DRA: 2.10) than to Pitcher A (DRA: 3.80). I’ve made up the numbers here to illustrate the point as clearly as possible, but—as will become obvious shortly—there’s a great deal of real-world value in understanding how the concept works. In one sentence: DRA is like RA9, but with the number of runs a pitcher deserved to give up, rather than the number of runs they actually gave up.

And DRA has been a revelation; I use it in almost all of my pieces. Until now, however, it hasn’t been entirely clear to the viewing public exactly how DRA is calculated. Pavlidis, Judge, and Turkenkopf were pretty upfront about what factors they considered in calculating the stat—temperature, umpiring, catcher framing, fielding, batter strength, batter handedness, home/away adjustments, pitcher role, and type of innings pitched—but they did not immediately break down each pitcher’s DRA into these individual components. They can’t be blamed for this—it’s a lot of work to run the numbers, and a lot of math to work out exactly what to run, in the first place—but the point remains: until recently, it was possible to see which pitchers had particularly good DRAs, or particularly bad ones, but how they got there was something of a mystery. Was it really hot when they pitched? Did they always pitch on the road? Were their catchers bad at framing? We just didn’t know. We do now.

***

I’m about to throw a couple of tables at you, and they’re going to contain a lot of numbers. Before I do that, it’s only fair that I explain to you what they’re all about. Let’s use a sample line—from a real pitcher—to test some things out. Here’s that line:

 The first thing to understand is that every number is expressed in terms of runs. The value of -2.33 in the ‘Framing’ column, for example, means that this pitcher gained 2.33 runs over the average from the value of his catcher framing (it’s expressed in the negative because pitchers are trying to prevent runs). So this pitcher, at the moment, has gained 2.33 runs over the average of his peers, because he’s had the luxury of pitching to catchers who are particularly good at framing his pitches. Neither RA9 nor ERA record that. DRA does, and debits his overall value because of the runs he’s gained from something outside of his control: catcher framing.

Let’s run through a few of the other columns: ‘Fielding’ is just the number of runs a pitcher has gained or lost by virtue of his defense; this pitcher, for example, has had a somewhat poorer defense than his peers, to the tune of 1.11 runs, and so will be credited with that amount of runs when all is said and done. ‘Temp’ is the run value the pitcher has gained or lost due to weather conditions (pitching is generally harder the warmer it gets). ‘LH-Stad’ and ‘RH-Stad’ are the runs the pitcher has gained or lost due to facing left-handed or right-handed batters in various parks (you don’t want to face righty pull hitters in Fenway, for example). ‘Inning’ is an adjustment for the innings a pitcher pitches in—it’s harder to pitch the farther into a game you pitch—and ‘Role’ has to do with starting or relieving. ‘H/A’ is home and away, which is fairly self-explanatory. Add ‘em all up, and you get ‘ADJ-R’, which is the adjustment you have to make from actual runs allowed—which is as easy to discover as reading the box score—to get to the deserved runs allowed. A negative value in ADJ-R means the pitcher has had it easier, overall, than his peers; a positive number means he’s had it harder. Add the ADJ-R to the Runs Allowed (not in the table), and you get ‘Value’.*

Phew! Got all that? I apologize for putting it all in one big block paragraph, but there’s no better way to do it. One final point, before we move on: a really high ADJ-R value—which means that the pitcher has faced really adverse conditions—still won’t compensate for really terrible pitching. You could throw me out there on a hot August day, against the 1927 Yankees, with Benito Santiago behind the plate, and the fact that I get lit up would still have more to do with the fact that I’m a terrible pitcher than to do with the fact that I faced really terrible pitching conditions. You have to pitch well to have a low DRA, and you have to pitch poorly to get a high DRA. The difference—and the interesting stuff—is on the margins. When you’re comparing me to Jake Arrieta, DRA is not very helpful. When you’re comparing Jason Hammel to Jake Arrieta, it’s extremely helpful.

***

Alright, enough of that. Let’s get down to the good stuff! Here’s some data:

A few things jump out at me immediately. First, all four Cubs starters have enjoyed pitching to catchers that are, on the whole, helpful to their pitching. That’s a credit to the front office, who carefully sought out catchers (in Miguel Montero and David Ross) who could help the team improve in this category, and it’s also of course a credit to Montero and Ross themselves. Secondly, it’s interesting to me that all four pitchers are benefiting from pitching to lefties—none more so than Arrieta, who seems to have worked something out with respect to left-handed hitters (possibly his cutter?). All four starters, meanwhile, get DRA credit for being starters (with positive values in the ‘Role’ column) and all four get slightly debited for the temperature environment in which they pitch—it was unusually cold at Wrigley this spring and early summer. Right off the bat, you can see why this statistic is interesting: it quantifies things we already knew, and allows us to relate them to one another.

Other interesting notes lie in the differences between the pitchers: Lester, for example, gets credit for facing a great number of difficult righties, while the others are all debited. (This, too, shows up in the data: 77 percent of the batters Lester has faced have been opposite-handed, higher than any other starter discussed; they’re all at 45 percent) Conversely, he gets a debit for his Home/Away pitching, while all the others are credited (this can be explained by the high percentage of his starts at home: 16 there versus eight on the road). On the whole, meanwhile, all the pitchers except Hendricks have had it easier this year than their peers (they all have negative ADJ-R values), and even Hendricks has a positive adjustment of just 0.22 runs.

Why bother looking at all of this? Well, because it helps us to better understand the conditions that pitchers face when they get to the mound. It can help us to illustrate why Fangraphs, which doesn’t use DRA, thinks Arrieta and Jon Lester have been pitchers with approximately 1.5 wins of value between them (4.5 fWAR vs. 3.0), while BP thinks Arrieta (4.7 WARP) has been about three and a half wins better than Lester (1.3). BP knows—while Fangraphs does not consider—that Lester, despite having baseline results only marginally worse than Arrieta’s, has been aided by his framing, his defense, and his location—all more so than Arrieta. This is meaningful—deeply meaningful—and it’s worth looking into in detail in the weeks and months to come.

***

I promised, near the top of this monstrosity of a no-pretty-pictures-at-all article, that the pitcher whose line I used to illustrate the column values was an actual major-league pitcher. And indeed he is. In fact, he’s the Cubs’ own Jason Hammel. I include the line I presented earlier—Hammel’s, from 2015—and his previous three years here:

This is where this article will end, because this chart is a good, final example of the utility of this type of data. Check out the extraordinary value of framing on Hammel’s numbers! Last year, Hammel pitched to poor-framing catchers, who cost him about 1.27 runs relative to his peers; this year, that value is -2.33. That’s a big difference. And while his role, his number of innings, the temperature, and a whole host of other things haven’t changed from last year to this, that framing difference has mattered. That’s worth knowing, and it’s worth considering as we watch him pitch.

Out of respect for your time, I won’t go any further into the data here. This is just the beginning—an introduction to the concept that first bedeviled George Plimpton, and a little bit of data relevant to the Cubs. Over the next few months, I look forward to diving deeper into these numbers and seeing what it can tell us. I know it won’t tell us that it’s possible for me to be an adequate major-league pitcher, if only I faced the Phillies in 35-degree weather every day. But it might tell us a little bit about how pitchers differ from one another, and that’s interesting enough for me.

* Because these numbers are not adjusted on a rate basis—not divided by total batters faced—a pitcher who’s pitched to 700 batters will almost certainly have a higher ‘Value’ than one who’s faced only ten batters. Just something to keep in mind. Also, this: there’s a few more—very minor—adjustments that DRA incorporates, so ‘Value’ minus ‘ADJ-R’ will not equal Runs Allowed, in most cases. It doesn’t make much of a difference, but it’s worth noting. 

Lead photo courtesy Jason Getz—USA Today Sports.