Sabermetrics And You: The Big Three, Part 1 - Batting Average

You can read all of the articles in the "Sabermetrics And You" series here.

Sabermetrics Reading Level (1-5): 1 - Beginner. This is an introductory-level article designed to give you a better understanding of basic sabermetric principles and/or stats. It should be well-trodden ground for Amazin' Avenue veterans.

It used to be that if you wanted the measure of a hitter you needed to know only three things about him: his batting average, his RBIs, and his home runs. Maybe you could throw stolen bases in there, but really those were the big three of traditional batter evaluation.

• Batting average told you how good a hitter your guy was. The old axiom was that a good hitter, a .300 hitter, still failed 70% of the time. If a guy hit .300 you knew he could swing the bat and was definitely the kind of player you'd want in the first half of the lineup. Heck, fill your squad with eight or nine .300 hitters and you'd be set.
• RBI told you pretty clearly whether or not your guy was a run producer. It didn't so much matter what his batting average was. If he knocked in 100 runs a season that guy was a middle-of-the-order threat and an All-Star. He knew how to drive in runs; that was his job.
• Home runs told you if your hitter was a slugger or not. He probably pulled the ball a lot and hit 25+ dingers a season. He likely also had a lot of RBIs, but some guys could pile up RBIs with a bunch of doubles, sacrifice flies, and clutch singles. Those were fine, but our heroes hit home runs.

Today we're going to take a look at batting average, what it tells us, what it doesn't, and what better alternatives we have to address its shortcomings.

WHAT'S WRONG WITH BATTING AVERAGE?

Batting average (AVG or BA) tells you one very specific thing: how often a player got a hit. The formula is strikingly simple: just divide hits by at-bats. Almost anyone can understand it which is probably what has made it so appealing. However, it has two crippling problems that almost entirely drain it of its usefulness.

1. Batting average completely ignores walks, which are often just as good as singles. On balance are they quite as good as singles? No, not really, but they're about two-thirds as valuable as singles, and batting average doesn't count them at all. Nor does it count hit batsmen, which are roughly as valuable as walks.

   Let's take a very simple example.

   Batter A comes to the plate 100 times, collecting 30 singles and grounding out 70 times.
   Batter B comes to the plate 100 times, collecting 15 singles and 50 walks, grounding out 35 times.

   Batter A has a .300 batting average and reached base safely 30% of the time.
   Batter B has a .300 batting average and reached base safely 65% of the time.

   Batter B is clearly the better hitter, and by a significant margin. Batting average wouldn't tell us that, though. Instead, batting average would see these as two identical hitters, which is a massive lapse of reason.

   Here's a real example from 2010.

   The Rangers' Vladimir Guerrero hit .300 (178 hits in 593 at-bats). He walked 19 times and was struck by a pitch four times.
   The Indians' Shin-Soo Choo hit .300 (165 hits in 550 at-bats). He walked 83 times and was struck by a pitch 11 times.

   The currency of a baseball game is the out. A team has a finite number of outs per game (usually 27), and the team that maximizes its opportunities while conserving its outs is most often the winner. In plain terms, while the fundamental objective of an offense is to score runs, the surest way to score runs is to avoid making outs. In roughly the same number of plate appearances, Choo reached base (i.e. avoided making an out) 259 times while Guerrero did so just 197 times. That's a huge difference in value that batting average sweeps under the rug.

2. Batting average treats all hits as if they were equal. Singles are not as good as doubles, nor are doubles as good as triples, nor are triples as good as home runs. Singles, then, are clearly not as good as triples, and they're laughably inferior to home runs. Batting average is blind to these distinctions that even a tee-ball player would instantly recognize. If batting average is the first thing you reach for when determining a player's offensive production, this fact should trouble you.

   Here's another simple example.

   Batter A comes to the plate 100 times, collecting 30 singles and grounding out 70 times.
   Batter B comes to the plate 100 times, mashing 30 home runs and grounding out 70 times.

   Again, Batter A looks like a scrawny hobo compared to the muscly-armed Batter B, but batting average tells you they're exactly the same player.

   Here's a real example from 2010.

   The Mariners' Chone Figgins hit .259. He had 132 singles, 21 doubles, two triples, and one home run.
   The Blue Jays' Jose Bautista hit .260. He had 56 singles, 35 doubles, three triples, and 54 home runs.

   These two hitters came to the plate approximately as often as one another and had virtually the same batting average. The vast majority of Figgins's hits were singles; just 24 of his hits went for extra bases. Meanwhile, Bautista, the AL home run champ, collected 92 extra-base hits, four times as many as Figgins. Bautista hit for prodigious power, while Figgins slapped the ball around, yet batting average tells you nothing about that difference in potency.

WHAT'S BETTER?

At the very least, batting average should be relegated to the "marginally useful" drawer in your baseball research toolbox. Without even getting into more advanced baseball metrics, replacing batting average in your go-to repertoire with two other stats -- on-base percentage (OBP) and slugging percentage (SLG) -- would leave you far better equipped to evaluate and compare individual hitters. Each stat addresses one of the two areas that batting average neglects.

On-base percentage tells you how often a hitter avoids making an out, or, in more obvious terms, how often he reaches base safely. Its inclusion of walks and hit-by-pitches as vital components of offensive value give you a far better accounting of a player's contributions than batting average ever could.

Going back to our earlier example, while Vladimir Guerrero and Shin-Soo Choo each hit .300 this past season, on-base percentage reveals the stark difference between the two hitters. Equals in batting average, Choo's .401 OBP was considerably better than Guerrero's .345 mark. On-base percentage tells us quite distinctly what batting average could not -- that Choo was a much better hitter than Guerrero despite identical batting averages.

Here's the math, for those who are interested:

Slugging percentage addresses problem #2 above, in that it does a far better job than batting average at distinguishing between different types of hits. Home runs count four times as much as singles and twice as much as doubles, doubles are twice as valuable as singles, and so on.

In our earlier real-world example, Chone Figgins and Jose Bautista each hit around .260 in 2010. However, while Figgins slugged just .306, Bautista slugged .617! These two players could hardly be less similar, but their batting averages suggest they're essentially the same player.

Again, the math if you're interested:

THIS ARTICLE IS LONG. CAN YOU SUM IT UP FOR ME?

Batting average neglects important offensive attributes like walks and hitting for power. On-base and slugging percentages account for these shortcomings and help provide a much more comprehensive profile of a batter's contributions at the plate.

Even if you don't read another article in this series, simply replacing batting average with on-base percentage and slugging percentage will dramatically improve your baseball IQ. Equipped with these useful but alarmingly simple tools, you'll be prepared the next time someone tells you that so-and-so is a good hitter because he hit .300. "No," you'll retort. "He barely walked and he hit nothing but singles. That guy isn't really a good hitter at all."

SO, IS BATTING AVERAGE COMPLETELY USELESS?

Not at all, it's just not nearly as useful as it should be given its ubiquitous use in traditional baseball writing and reporting. Batting average generally does a good job of describing how well a player hits the ball and reaches base safely as a result. That's certainly a valuable skill, and while it is included in OBP and SLG (as a portion of a larger, more comprehensive measurement of batting ability), if you need to know specifically how often a player reached base via a hit, batting average will do exactly that (and little else). In a later column we will discuss how batting average (and to a lesser extent on-base and slugging percentages) is a function of not just a player's hitting ability, but also of things that he has very little control over: defense and luck.

WHAT ELSE?

If two stats are too many for you, you can add on-base and slugging percentages together to arrive at a single stat abbreviated OPS (On-base Plus Slugging). Doing so will give you one easy number that can be used for quick-and-dirty comparisons. OPS is not without its problems. For one, it assumes on-base and slugging percentages are of equal value when they're not. OBP is almost twice as valuable as slugging, so on-base percentage is shortchanged a bit by OPS. Nevertheless, adding OBP and SLG together, clumsy as it may be, gives you a pretty useful single stat to work with.

If you want to take AVG, OBP, and SLG analysis even further, check out run correlations, which illustrate the impact of individual metrics on overall run scoring. Fascinating stuff, but perhaps more math than some are comfortable with.

We'll cover ways of adjusting OPS -- and more advanced alternatives to it -- in future columns.

QUESTIONS & ANSWERS

I know that .300 is a good batting average. What's a good on-base percentage or slugging percentage?

Since different seasons (and different eras) have varying degrees of offensive dominance, your first step in evaluating a player's on-base or slugging percentage is to compare it to the rest of the league. You can find average league values on Baseball-Reference.com's league pages. For instance, in 2010 the average NL on-base percentage was .324 and the average slugging was .399. Generally, for recent seasons a .350 on-base percentage is pretty good and a .400 on-base percentage is outstanding; a .450 slugging percentage is pretty good and a .550 slugging percentage is outstanding.

Have a question about this article? Send us an e-mail and we might answer it here.