The Baseball Economist: The Real Game Exposed (22 page)

OBP does not discriminate between the walks and hits or singles and home runs. All it counts is how many times a player reaches base without making an out. It is measured as a rate, in terms of the number of opportunities a player has to make an out. Sacrifice flies are added to the denominator, because although the player may give himself up to advance a base runner, it’s not clear that was his intention. Sacrifice bunts are ignored, because the hitter clearly is not trying to reach base in this situation.

On-base percentage is one of the most underrated statistics in the game. It’s often ignored, because it has no flash. A guy who gets on base by walking is not nearly as interesting as someone who reaches base by hitting the ball on the field or over the fence for a home run. There seems to be a mentality that there is something cheap about getting on base via a base on balls. In fact, a walk does not have as much of an impact on scoring runs as a single, because a single can advance existing runners beyond a single base. However, not making outs is an important skill, because a team’s opportunities to score runs are constrained by its outs. In a nine-inning game a team only gets twenty-seven outs, three outs per inning, to put up more runs than the opposing team. Every out brings the end of the inning and game closer, thereby lowering the team’s opportunity to score runs. Also, getting on base, at the minimum, puts your team one base closer to scoring a run.

The batting average is the gold standard of baseball hitting statistics. Almost all fans accept AVG as the benchmark for comparing hitters, and it’s normally the first statistic you see next to a player’s name. In fact, the “batting title” for the best hitter in the game is awarded to the player with the highest AVG.

AVG measures the rate at which a batter reaches base by hitting the ball relative to the opportunities the player has to hit the ball. At-bats are the total number of times a player steps to the plate when he does not walk, get hit by a pitch, or sacrifice himself to advance a runner via a bunt or fly ball. Unlike with OBP, batters are not punished for sacrifice flies. This exclusion creates opportunity for a player to have an AVG higher than his OBP, which seems to indicate the mathematical impossibility of a batter reaching base from hitting more than he reached in total. In fact, it’s just the result of the different denominators of the two statistics. Having an AVG greater than one’s OBP is still a rare feat, yet Billy Beane—the OBP-loving protagonist of
Moneyball
—did just this in 1989. AVG rewards players for hitting their way on base, it is neutral toward walks and being hit by a pitch, and it punishes nonsacrificial outmaking. Like OBP, it weights all hit types the same.

The slugging percentage (SLG) is a batting average that accounts for hitting power. Each hit type is weighted by the number of bases the hitter advances.

SLG quantifies the power punch per at-bat. Certainly, a player who hits lots of doubles and home runs is more valuable to the team than a singles hitter with an identical AVG. And at some point, the team might even prefer a player with a lower AVG if his SLG is high enough, because he can push himself and his teammates farther around the bases when he does hit the ball. While SLG certainly tells us more about a player’s value to the team than AVG, it’s not perfect. First, the arbitrary weighting of bases from one through four, though intuitively appealing, is wrong. Using some sophisticated techniques to estimate the run values of the different types of hits, we know that doubles, triples, and homers are not two, three, and four times more valuable than singles in producing runs. But for a rough approximation of hitting power, SLG has the benefit of being very simple to calculate.

A second deficiency of SLG is that it captures more than just hitting power. Because its denominator is at-bats, it also captures hit frequency. For example, take two hitters with one hundred at-bats. Steve Single hits forty singles, and Donny Double his twenty doubles.

Clearly, Donny has more power than Steve, but their slugging percentages are identical. One way to remedy this problem is to look at only the extra-base hits. The metric known as “isolated power” (or iso-power) weights only extra-base hits, and does not count singles.

Using isolated power, it’s easy to see Donny is the player with greater hitting power, with an iso-power of .200 compared to Steve’s .000. Again, no metric is perfect, but it’s always good to know the strengths and weaknesses.

Runs batted in (RBI) is the second category in the “triple crown” of batting competitions—the AVG and home runs are the first and third. It measures the number of times a player drives in runners, except when he hits into a double-play, because even a run isn’t worth two outs. The RBI is a stat that we often judge a player by, but probably should not. Why? Well, RBI totals depend on not just the quality of the player, but several outside factors. A player who bats clean-up, behind three excellent players with high OBPs, will generate a lot more RBI than a lead-off man, who often bats with no players on base. For reasons other than hitting ability, we would expect very different RBI totals from these players. It’s possible that the lead-off hitter does many things that if he were given the opportunity, would generate more RBIs than the man in the four-hole. It’s just very hard to know.

For this reason, the RBI is a nearly useless statistic to judge baseball players. Certainly, to have a high RBI total, you have to be a good baseball player. But there are lots of good baseball players, so it’s not a good idea to compare players on this statistic. If you want to judge a player’s ability to knock players in, certainly a worthy goal of a team, SLG is a far better measure of this ability. SLG is not dependent on the base-runner configurations that a batter normally faces over a season.

Baseball is about scoring runs, right? Thus, a player who scores more runs is preferred to a player that does not. However, runs (RS), like RBI, are largely dependent on the play of teammates. There is only one case in which a player does all of the work to score a run, when he hits a home run. However, rarely do home run hitters lead the league in runs. Don’t get me wrong, producing runs is valuable, but the run total of any player is largely determined by the players who bat behind him in the lineup. Lead-off hitters typically lead the RS category, because they precede good hitters in the lineup. Certainly, a player with a lot of runs has to be good at reaching base and can enhance his scoring ability with speed. But the run totals largely reflect things that other players do.

A batter’s batting average with runners in scoring position (RISP AVG) is a commentator classic: “You know, Johnny’s had a tough time keeping his average above the Mendoza line, but when it matters most, he’s been a real clutch hitter, batting .425 with runners in scoring position.” RISP AVG is simply a batting average calculated only when a player bats with at least one runner on second base or beyond. There is no doubt that hits in these situations are more valuable than hits with no men on base. Certainly, a manager loves for his players to get hits in these situations. But I wouldn’t advise any manager to bench a .300 hitter with a .200 RISP AVG for a .200-hitter with a .400 RISP AVG. The problem is that hitting with RISP is not a skill, or at least not much of one that we can identify, but a statistical anomaly.
⁶⁷

Just because we can observe a number does not mean it contains useful information regarding a skill that a player has. Take early season batting averages. Everyone understands that regular batting averages for players can be absurdly high or low early in the season, simply because bad or good luck has not had time to even out. With few opportunities and a couple of odd bounces, in early May a player can have an AVG that is far from his norm. This is also what is going on with RISP AVGs, as well as other situational statsistics. Players don’t have as many RISP opportunities over the season, for some of these statistics to fall back or rise to a player’s normal AVG.

Is it really such a good thing for a player to have a RISP AVG higher than his regular AVG? If I were the manager, my question to the player in question would be, “What the heck are you doing not hitting when there’s no one on base? We need base runners, and it’s awfully selfish of you to hold back unless there’s some potential RBI out there!” If hitting with RISP is something a hitter can purposely alter, I have a hard time believing he is holding something back in non-RISP situations. It might be possible that some hitters hit the ball in ways that are more likely to succeed with runners on due to a common defensive configuration, but there has been little evidence of this.

Now, the next obvious step is to evaluate what these statistics tell us about run production. The goal of the offense is to score more runs. Which statistics tell us the most about a team’s ability to produce runs? To do this, let’s analyze team-level statistics using a simple statistical method to predict team run scoring.
⁶⁸
We observe teams instead of players, because individuals can only contribute to runs scoring. Only in the case of a home run does the offense of a single player cause a run to occur. Players need the help of their teammates to score runs; therefore, looking at teams will tell us something about the value of these stats in predicting runs. We’ll exclude a few of the statistics discussed above for obvious reasons. Runs are out, because, at the team level, the sum of runs scored will obviously be equal to the runs scored. RBI suffers from almost the same problem, although some runs can score without an RBI being awarded. Let’s ignore RISP AVG since it contains a lot of luck. A team with a good RISP AVG will obviously score more runs than a team with a poor average. Getting hits in key spots is a good thing, but not largely controlled by players or teams.