O’s By The Numbers: What if RBIs Were More Accurate?

Note: This week’s post was inspired by Sky Kalkman’s 2008 article on Beyond the Box Score. We offer both Sky and Beyond the Box Score the most humble of blog-o-sphere glove-slaps.

This week in O’s By the Numbers we’ll look at the most revered traditional baseball statistic: Runs Batted In (RBI). Amongst traditional baseball fans, RBIs tend to be favored over any other statistic. In terms of predicting who the Baseball Writers Association of America (BBWAA) will choose as the MVP for a given season, RBIs offer more predictive power than any other statistic. The anti-RBI argument is fairly well-established. It centers around the observation that the ability of a hitter to collect RBIs is more a function of the opportunity to hit with runners in scoring position than it is a function of the hitter’s actual offensive ability. Thus, RBIs will over-rate the contributions of middle of the order clean-up hitters and under-rate the skill set of lead-off hitters with good on-base skills. Here, we’ll examine the question: What if RBIs accurately reflected a player’s offensive production?

To answer this question we’ll use the advanced metric: Weighted Runs Created (wRC). wRC measures the cumulative offensive contribution of a player by using linear weights to compute the run value of plate appearances. wRC accounts for: non-intentional walks, hit-by-pitches, singles, reached on error, doubles, triples, and homeruns. wRC is not adjusted for different stadiums and it is context-neutral, meaning it does not take into account if there were runners on base for a player’s hit or if it was a close game at the time.

Using wRC, we can offer a new metric, deserved RBI (dRBI), which scales wRC to RBI using the best-fit equation pictured above. This enables comparison between the two statistics. The wRC to dRBI transformation is:

dRBI = 11.16 + 0.72*wRC

Using dRBI instead or RBI, the Orioles’ leaders are:

For the most part RBIs are a reasonable statistic to rank the offensive contribution of the Orioles’ players this season. RBIs under-rate Markakis due to his early position in the order and his average walk rate. Reynolds and Jones have had stellar offensive production this year but in each case the production is slightly over-rated by RBIs because of the number of plate appearances with runners on base. Its important to note that wRC, and thus dRBI, is not adjusted for position. As a result, the contributions of Wieters (catcher) and Hardy (shortstop) appear average compared to their teammates. In actuality, one would expect these positions to contribute the least to the team. By contributing at an average (Wieters) and above average rate (Hardy), both players are helping the team more than dRBIs show. Andino suffers the most when measuring offensive production through RBIs. Despite deserving almost as many RBIs as Derrek Lee, he has notched 23 less! Its nice to see that RBIs don’t obfuscate Vlad’s horrible season; he has been poor by all metrics.

I thought it would also be interesting to look around the league and find the 5 most under and over rated players by RBIs. Here are the most over-rated players:

This larger sample shows what a poor metric RBIs can be. Offensively, Ryan Howard has been comparable with Adam Jones this year, yet due to his position in the order and the lineup surrounding him, Howard has 23 more RBIs. Recently, the New York Times looked at Ryan Howard’s RBI prowess in depth. Historically, Howard has shown the ability to accrue more RBIs than he “deserves”. The NY Times piece is worth the time of the By the Numbers enthusiast. Next, we’ll look at the top 5 most under-rated players by RBIs.

Again, the larger sample shows the flaws of RBIs. Despite having 58 less RBIs than Ryan Howard, Reyes deserves 4 more! For the most part this list is composed of lead-off hitters who have less opportunities to to hit with runners in scoring position. In sum, RBIs are hardly the bain of the existence. However, using them as the chief metric to grade offensive production is flawed. Advanced metrics like wRC are more accurate and can be translated to match RBIs to give traditionalists an advanced metric on a familiar scale.