Advanced Metrics and How to Use Them: Batting
By Chris Greene
Oct 3, 2015; New York City, NY, USA; Washington Nationals right fielder Bryce Harper (34) watches his home run ball during the eighth inning against the New York Mets at Citi Field. Washington Nationals won 3-1. Mandatory Credit: Anthony Gruppuso-USA TODAY Sports
I am a proponent of advanced stats like WAR, wRC+, FIP, wOBA, and all the other acronyms. When I write, though, I sometimes worry that fans of the game who haven’t spent countless hours delving into the minutia of advanced stats simply don’t know what I’m talking about. Since the point of writing is to communicate – hopefully clearly, this poses a problem.
This series (Advanced Stats and How to Use Them) is my attempt to fix said problem. I want to pull back the curtain and show some of the inner workings behind these modern metrics, focusing on the ones that I use most commonly. I’ll try to keep this pretty basic and straightforward, so much of it you may have already encountered. Just remember that this isn’t supposed to be a course on sabermetrics. This isn’t even designed to be the introduction to a course on sabermetrics. It’s just a glimpse at the very basics of how these stats work.
More from St Louis Cardinals News
- Cardinals: Jose Quintana’s Market Price is Overvalued
- Cardinals: Paul Goldschmidt builds home for single mother and her family
- Cardinals: Is Lars Nootbaar underrated or overrated?
- 3 reasons why the Cardinals should wait to address rotation until 2024
- 5 Cardinals who must improve this off-season
In each of these posts, I’m going to look at some traditional statistics and the problems with them, and then dive into some of the more advanced metrics that I like to look at. So let’s get started!
Before we dive into these stats, let’s figure out what we’re trying to get at. Sabermetrics and advanced stats are best used in an attempt to get at “true talent.” True talent is the constantly fluctuating, ever changing, actual ability of a player. It is, at least in theory, an objective skill level that exists within every baseball player, devoid of luck, chance, and circumstance. For instance, you could pretty easily say that Pete Kozma is a true talent zero home run player. Throw him out there season after season and he might occasionally hit one, but really that’s more fluke than skill.
Now why would you want to get at this true talent? I mean, who cares if Pete Kozma is a true talent zero home run guy if he gets lucky 30 times in 2016, right? Well, kind of. But it’s very difficult to guess at what’s going to happen with luck or chance, and we’re interested in projecting the future. Luckily, this true talent doesn’t change as much, and we want to know who we should play at shortstop in 2016. Everyone betting that Kozma will hit 30 home runs and be an All-Star raise your hand. Ok. Now if you think that Jhonny Peralta will hit .270 and be the better choice, raise your hand. Presumably, most of you voited for Peralta. That’s because you already understand true talent and know that he’s simply a better player than Kozma.
So, as we get into these statistics, think of using them to determine a player’s true talent level in an effort to better predict how well he will perform in the future. Now let’s look at some stats.
The Ugly: Runs Batted In (RBIs)
Let’s start with the RBI. As most of you probably know, RBIs count how many runs score on hits from a particular player. What could possibly be bad about that?
The Problem: RBIs are fun to look at, and seeing the leader every year is always entertaining, but it’s a really bad measure of a good player, especially with regards to future performance. Don’t believe me? Well let’s take a look at an example. Player A has 200 hits, 50 home runs, 10 triples, and 40 doubles. Player B has 150 hits, all singles. Who would you rather have on your team? Player A, obviously. But I left out RBIs. Player A only had 50, because every time he came to the plate the bases were empty because his lineup stinks. Player B collected 300 because the bases were loaded every time he got a hit and two runs always scored.
Drastic and extreme example? Absolutely. But, do you still see how RBIs provide (at best) a warped view of things? As a better example, Mike Trout only managed to collect 90 RBIs in 2015, while Matt Kemp gathered 100. That’s not a huge difference, but who would you rather have on your team? If you’re looking at RBIs, it’s Kemp. If you’d pick Kemp as the better offensive player, you might want to see a psychiatrist. The difference is that Trout is all of the offense for the Angels while Kemp is a just a piece of the puzzle in San Diego.
How to Use It: Look at RBIs because it’s fun, but not because it’s terribly useful in a meaningful way. If you’re arguing that a player is better than another player based on RBIs, you probably need to reevaluate your stance.
The Bad: Batting Average (BA)
Batting average? This guy wants to kill off batting average? How can you think such a simple stat is bad? It’s just hits divided by at bats! Easy to calculate, easy to understand, and insightful. So why can’t we use it?
The Problem: Batting average isn’t a bad statistic, it’s just incomplete. The complaints about it stretch back over 100 years, but it persists, probably because it is so simple. And I like simple as much as the next guy. But let’s forget baseball for a minute and talk about my friend. I want to know how much money he has, and so I ask him (I’m rude that way). He tells me he has seven bills. That’s not useless, right? I have some notion of how much money he has, and I can even guess that they aren’t all hundreds, because that would be rare, and no one carries fifties.
But, even after removing the two highest possible denominations, he could still have anywhere from seven to one hundred and forty dollars. That’s not very precise. In much the same way, batting average pretends that singles, doubles, triples, and home runs are all worth the same amount. We know that they aren’t all worth the same, but we try to pretend that they are. Looking at it mathematically, team batting average only has about a 65% correlation with runs scored. That’s not terrible, but it’s not great either.
How to Use It: Pair it with other statistics! Batting average does tell us how frequently a player puts the ball into play for a hit, which is useful. When paired with on-base percentage and slugging (as in the slashline I favor), you can start to get a better picture of how well a player is performing. If you need a single number to use, however, OPS and wRC+ are much better options.
The Good: On-Base Plus Slugging (OPS)
This is the advanced stat for the guy who doesn’t like advanced stats. It’s simple enough for anyone to calculate: all you have to do is add together on-base percentage and slugging percentage, and you’ve got it. But why would you want to do that?
Why It’s Good: So, team batting average correlates about 65% with runs scored, and we agreed that’s not very good. Team OPS, on the other hand, comes closer to having a 90% correlation with runs scored. Basically, if you’re good at OPS, you’re probably good at scoring runs. There are better metrics (like wRC+), but if you want to argue with your sabermetric-savvy friends, but don’t want to deal with anything you can’t calculate yourself, then this is your statistic.
It includes value for walking and each different hit type, giving you a rather complete view of what a batter does at the plate. In terms of range, .600 and below is terrible, .700-.800 is pretty average, .800-.900 is good, and anything above .900 is great. If you’re above 1.000, you’re probably the league’s MVP.
How to Use It: OPS is great for using all on it’s own to evaluate how good a player is offensively. Feel free to whip it out as a trump card without needing to back it up with much else. High OPS = lots of runs scored, and that’s good.
The Great: Weighted Runs Created Plus (wRC+)
And now we get to the really fun stuff: weighted runs created plus. What the heck does all of that mean? Why isn’t OPS good enough?
Why It’s Great: This is actually simpler than it seems, and it adds a lot to what OPS starts for us. At the center of the stat is the idea of “runs created.” In short, it attempts to figure out how many runs a player is worth offensively and quantify that. Weighted runs created plus factors in park effects and adjusts for different leagues.
Finally, it puts the whole thing on a nice easy scale where 100 is average, and each point above or below is a percentage point above or below average. This past season, Bryce Harper lead all of baseball with 197 wRC+, which means that he was 97% better than an average offensive player, adjusting for park effects and league. In his career Babe Ruth was also 97% better than average, so Harper was as good this past season as Ruth was for his career.
Compared to OPS, wRC+ is a little bit better at correlating to runs scored, but it has another great advantage: consistency across eras. For example, in 1969, the Baltimore led the league with a 112 wRC+ and had a .756 OPS in the process. Thirty years later, the Cleveland Indians led baseball with an identical 112 wRC+, but .840 OPS. So what’s the difference?
In 1969, the offensive environment was a lot less favorable than it was in 1999. Thus, a statistic like wRC+ which compares offense to league average gives us a better idea of how good a player’s performance actually was. The same is true with other leagues, including the minor leagues where run scoring environments vary.
How to Use It: If you’re looking to compare players across eras or leagues, wRC+ is your tool. It can also be very helpful when looking at the minor leagues and trying to see past different league or park effects.
Other Useful Metrics: BABIP and ISO
BABIP: After studying decades of data, baseball analysts have found something really interesting. If you put a ball in play (between the foul lines, in the field), it lands for a hit almost exactly 30% of the time. Like, almost always for everyone. Sometimes individual players are especially good at getting balls in play to land as hits, but not frequently.
Over short periods of time, however, this number can fluctuate due to good or bad luck, and raise or lower this batting average on balls in play from the standard .300. So if you see a generally bad player having an outstanding stretch, or a great player having an awful one, look at BABIP. A high BABIP means great luck, and a low BABIP means bad luck. That can help you filter out the lucky results and get at the true talent ones.
ISO: Sometimes, you just want a home run. Isolated power – or ISO – tracks only a player’s extra base hits while ignoring singles, thereby giving you a good look at how much power he’s generating. The simplest way to calculate it is to subtract batting average from slugging percentage. Average is about .150, and anything above .200 is great, while anything below .100 is bad. So when you’re checking out free agents trying to find the power hitter you think the Cardinals need, ISO is a good place to start.
If you’re interested in looking into even more advanced metrics and learning more about what goes into these statistics, I’d encourage you to look into Fangraphs. In addition to wonderful articles and lots of player data, they have a phenomenal glossary that offers both simple definitions and in depth looks at a whole host of useful offensive statistics.
So there’s the primer on a couple of my favorite offensive stats. Next week, we’ll take a look at pitching and some new ways to measure the value that hurlers provide to their teams.