Oldies But Not Goodies: Discarding RBI, Runs Scored, and Pitching Wins

While most traditionalist stats are all but completely gone from most Major League Baseball front offices, the casual fan and the older generations of die-hard fans have stuck with the statistics. Classic and easy to remember, a century of celebrating records has some people holding on to these numbers for good. Sometimes, classic theories and ideas can still be relevant, but in this case, that isn’t true.

baseball glove benchI have nothing against people who like these stats. Many of them have spent their entire lives having RBI records, batting average records, win-loss records, and other statistical feats drilled into their heads. These were the only options, so people didn’t question them.

There are three stats in particular, however, that need to go. This may be obvious to many fans but they can’t hold a candle to some of the new metrics that have popped up in recent decades. RBI, runs scored, and pitching wins and losses are arguably the most irrelevant and useless popular statistics.

Before diving in to each one, consider one thing. What is the goal of using statistics for individual players? The answer is simple: to isolate production. Simply, to tell how talented a player is. In theory, the best statistics are affected only minimally by other players, otherwise a statistic can be as much a measure of a team’s ability as the individual player’s. If you were a general manager signing a player from the Red Sox for $200 million, given only statistics that are significantly affected by other Red Sox players, would you consider making the deal.

Runs Batted In (RBI)

The Goal: Runs Batted in has two joint goals (or at least it is perceived to have two): to evaluate how good a player is in important situations and how productive he is overall.

The flaws: This statistic rose to mass use in the 1920s. This is not a problem, but it shows that the statistic was created before baseball was understood as much as it is today.

RBI is as much of a team statistic as it is an individual one, it’s main problem. There is only one method of getting an RBI without any baserunners and that’s hitting a home run. Say two hitters each hit home runs, except the second hitter did it with a runner on first base. Does the second batter deserve twice as much “credit” as the first? No, not at all. The second home run hitter probably had little to no effect at all on whether that runner reached base, but he still gets credit for “driving in” that person. Not only that, but players on better teams tend to get more opportunities to drive runners in. Two players may be driving in the same percentage of baserunners but one may have far fewer RBI than the other.

One more thing to consider about RBI is that it treats every situation equally. What good is a second inning RBI single when your team is down 9-0? Which leads me to…

Alternatives: There are a number of different alternatives for RBI but the most popular is probably Win Probability Added. Remember how every RBI is treated the same regardless of situation? That is where WPA comes in. Baseball has been played for over a century and almost every situation imaginable has repeated itself over and over again. One thing is certain: there are always calculable odds of who is more likely to win. Every action affects a team’s odds of winning a game, whether it is small or large. A walk-off home run obviously has a bigger impact than a one out single in the third inning with nobody on. WPA uses linear weights, a complex way of saying the odds of winning added (or lost) from each action. Players with a higher WPA tend to have had bigger impacts on games (although it is not predictive), specifically in high-pressure situations, which statisticians have debated the effects of with no real consensus yet. This is still a stat where a team must put a player into position to have a bigger impact, but it certainly quantifies that impact far better than RBI. (To read my article from last summer going in depth on WPA, click here.)

Runs Scored (Individual)

The Goal: This stat is rather murky in its presumed goal. Really, it is likely meant to measure both production overall and baserunning.

The flaws: Again, this is a stat in which it depends so much on the surrounding team, probably even more than RBI. A player can bat 1.000 and still never score a run. Of course, these theoretical situations aren’t relevant to the real world of baseball, but the idea holds true: teams set you up to score a run. Sure, the player may have successfully made it to home plate without falling flat on his face, or he could have even dove into home plate well. However, the hitting  team still had to do something to allow him to cross home plate and even the team in the field often time chooses not to throw to home, instead opting to hit the cutoff man and settle at that.

Alternatives: There are a ton of alternatives to Runs Scored, satisfying both purposes. Getting into them could take another thousand words, but there are plenty of viable alternatives. For overall production, OPS, OPS+, all the way down to wOBA and wRC+ do the job better than runs scored as they isolate that particular player more. For baserunning, there are complex metrics like UBR out there, as well as some of Baseball-Reference’s statistics that even include a player’s ability to avoid getting thrown out at first on a double play. There is some very interesting stuff out there that can even break down the type of baserunning a hitter is good at.

Wins and Losses (For Pitchers)

The Goal: To evaluate the performance of an individual pitcher

The flaws: Where to begin? There are so many flaws with wins and losses. As a general rule, I say that wins and losses are half affected by the offense and some by defense as well. A pitcher can be on his game striking batters and getting weak ground balls and still get the loss. In order to get a win, the offense of the pitcher’s team must score more runs than the other. Say what you will about pitching to the score, but that’s what it comes down to.

Additionally, a pitcher’s defense behind him can let him down, whether measured in errors or not. Even if the pitcher allows only unearned runs, the loss is still given.

There are also plenty of situations where the pitcher throws a great game but leaves tied, giving a reliever an opportunity to get credit, even if he comes in only to pick off a baserunner. It has happened before, and it is so often the pitcher with the best night that gets cheated.

Alternatives: Rate stats are the way to go here. Looking at game logs is fine as well, but not all wins and losses are created equal, but even an undeserved win will show up in rate stats. Specifically, FIP and xFIP are great alternatives as they take out the fielding aspect as well as the hitting aspect, which ERA does not completely do.

*   *   *

There are a number of statistics like these that are very flawed and should be essentially discarded from use by the average fan. As someone who’s skeptical of almost anything, I noticed early on that there were flaws. With baseball especially, it’s important to not let tradition get in the way of realizing the flaws of the different ways people analyze the game.


About Connor O'Brien 334 Articles
Connor O'Brien is a third-year economics student at Rutgers University, a longtime writer here at MetsMerized Online, and an aspiring economist. He embraces sabermetrics but also highly values scouting. Follow him on Twitter at cojobrien.