Quantifying Pitcher Offense: Introducing oERA

I have a theory: I think that a pitcher who can hit effectively is more valuable than another pitcher of equal pitching ability who can’t contribute offensively at all. Sure, that seems fairly intuitive and, I think, uncontroversial. The issue is that no one has ever tried to quantify exactly what value an offensively competent pitcher brings to a team (or if they have, the stat hasn’t stuck).

Pitcher offense has consistently and repeatedly been overlooked and cast aside as a potential tool to help win a baseball game. Pitchers are instructed to focus on pitching mechanics and drills and told not to worry about hitting. Additionally, in the minor leagues (and increasingly in college and high school) pitchers aren’t asked to hit at all. As a result, the vast majority of pitchers at the Major League level are abysmal Major League hitters. In a sport where seemingly minuscule strategic adjustments–think defensive positioning–are agonized over, it comes as a bit of a surprise that something as large as 1/9 of a teams’ offensive potential is chalked up as a lost cause without much of a second thought.

Of course, not all pitchers are bad hitters. Some, like Tim Hudson and Zach Greinke, are viewed as novelties and not much thought is given to their offensive production outside of a surprised cheer from the crowd and the announcers every time they manage a hit. Others, like Micah Owings, prove so adept at hitting that they are moved off the mound to another position in order to get more consistent at bats.

I think this is a huge mistake. Pitcher offense should be a prized commodity. The potential value is immense. If one team is starting a pitcher who can hit and the opponent’s starter doesn’t know which end of the bat to hold, the former team is essentially playing a 9 against 8 game, a huge potential advantage.

If National League teams are looking for Moneyball-style market inefficiencies, a good place to start would be trying to find pitchers who can also offer some offensive production.

It’s a balance, however. While offensive production may be a nice bonus (a game-changing bonus in some cases, as we’ll see), it is understandable that teams would be unwilling to sacrifice much, if any, pitching value for the sake of bonus offense.  Thus, in a cumulative statistic purporting to quantify a pitcher’s offensive production, pitching must be weighted much more heavily than hitting.

So, with the above as my working theory, here’s my premise: every run created on offense is equivalent to -1 run surrendered on defense, and vice versa. This is obviously not a new idea when applied to position players, but may be novel when talking about pitchers.

In order to produce a stat that succeeds in quantifying pitcher offensive value, I wanted to be sure to satisfy a few criteria. The stat had to be simple, so it could be easily compared between players, and it had to be a per-game rate stat so that it could work for both starters and relievers–pitchers with wildly unequal innings pitched (Chris Davis v Kershaw?). Additionally, it was important that this stat not rely on any relative stats as components, because often the sample size with pitchers on offense is so small that tiny statistical outliers can wreak havoc on accuracy. (For instance, wRC+, is a great stat for hitters that ranks them by relation to the league average at their position. The highest wRC+ in the MLB this year among position players was posted by Miguel Cabrera and was awfully close to 200. The highest wRC+ for a pitcher? Zach Duke’s 497 over exactly 2 plate appearances. So wRC+ isn’t going to be very helpful.)

By contrast, wRC, an excellent statistic that estimates a players “runs created” seemed like the perfect component to measure a pitcher’s offensive output. In my estimation, one offensive run created (a wRC of 1) should be equal to one earned run allowed for a pitcher. Basically, every time a pitcher makes a run on offense, its like not giving up an earned run (a run the pitcher is responsible for) on the mound.

So, if we factor wRC into ERA (Earned Run Average, obviously), we can account for offensive production in the total number of runs, on a per game basis, that a pitcher is responsible for. I call this stat offensively weighted ERA or oERA.

In order to calculate oERA, all we have to do is subtract the player’s wRC from their ER and then recalculate ERA based on these new values. This takes into account the idea that one offensive run created negates one defensive run allowed

Let’s take a look at how this stat works on some real life examples.

Some things that immediately jump out based on this random sampling:

• Zach Greinke is even better than he’s given credit for. While it’s true that he posted a stellar 2.63 ERA, when you take into account his league leading 10 runs created, his oERA is really an entire half a point better. So, in truth, the 8 Dodgers not named Greinke only had to come up with 2.14 runs– not 2.64–every time he pitched.
• Mat Latos is a great pitcher but his offense is bad enough that it costs his team runs. In fact, he’s responsible for spotting the opponent .14 more runs than his ERA would suggest simply because the opposing pitcher gets to face him a couple of times a game.  (Whether Dusty Baker’s usage of the pitcher in the batter’s box contributes to Latos’ offensive inefficiencies is the subject of a different study.)
• Tim Lincecum’s recent struggles on the mound are exacerbated by futility with the bat.
• While Bronson Arroyo may seem like a quality veteran starter who, at 36, is still capable of posting sub-4 ERA’s, in fact, his unrivaled ineptitude at the plate takes a lot of the shine off his otherwise decent pitching stats.

In the National League, a pitcher is more than just a pitcher. And positive offensive contributions are more than just some amusing novelty. They are an integral piece of the a pitchers complete value, and oERA is is a great first step in deciphering that riddle. It can help augment more traditional evaluations of pitching ability.

-Max Frankel

Deeper Curiosity: The next step, which I hope to get to in the next few days, is to assign a dollar value to oERA differences. I’d like to be able to say that one pitcher, identical in every way  to another pitcher except that his oERA is one point lower, should be paid $X more than the latter pitcher. It would be a simple task ordinarily, but the inaccessibility of the formulas that make up WAR makes it a bit more of a challenge.

Please share any questions, comments, or concerns about oERA or my methodology below, or at [email protected].