Sabermetric Series, Part 2: Applying Metrics To MLB Splits

Welcome to Part 2 of the Sabermetric Series! In Part 1 (which you can find here), we began with quality of contact and batted balls for hitters. Here in Part 2, we are going to round out the batted ball portion of that and move on to understanding some of the commonly used metrics such as wOBA and OPS+. We are then going to take what we’ve learned and apply them to MLB splits data for hitters. Let’s roll!

Previous Installments

Sabermetric Series, Part 1: Quality of Contact and Batted Balls

HR/FB%

Last week we left off at HR/FB% (home run per fly ball rate), so we’ll dive in right there. The stat itself is basic, simply calculating the percentage of a batter’s fly balls that go for home runs. In 2018, the league average HR/FB% was 12.7%. For the most part, if a player is substantially higher or lower than that 12.7% mark, we can project regression to the mean.

To make sure that is the case, we have to take into account a number of things: FB%, Hard%, Pull%, and the player’s career HR/FB%. FB% alone doesn’t dictate anything about HR/FB%, but obviously the higher your FB%, the more chances you have to put one in the seats. A high Hard% and Pull% can feed into HR/FB% quite a bit, though, depending on a player’s approach.

If a player suddenly has a spike in HR/FB% from his career rate, we look at his batted-ball profile. Unless there has been a major change there, that player’s HR/FB% can be expected to normalize back toward his career rate over a larger sample size.

In 2018, Christian Yelich posted a 35% HR/FB%. That’s more than double the MLB average. His 47.6% hard contact rate was 7th best among qualified batters, but he hit just 23.5% fly balls with a mere 34.9% pull rate. That fly ball rate ranked 133rd out of 140 qualified batters, ranking just a touch higher than Dee Gordon’s 22.4% FB%. Batters who hit fewer fly balls tend to carry higher HR/FB rates. Still, while Yelich is a tremendous hitter in a strong hitter’s ballpark, he will be facing serious home run regression heading into 2019.

Someone like Brian Dozier, who doesn’t have the easy all-fields power of an Aaron Judge or Joey Gallo, has shaped his career by pulling the ball for power. As a Twin, Dozier knew that Target Field was deep in center field and had a high wall in right field, so he used a 50%-plus pull-rate to yoke balls out of the park to left field.

Conversely, a slap hitter like Dee Gordon will never approach a 13.7% HR/FB% because of his contact and batted ball distribution — he has a career 16.7% Hard%, 20.6% FB%, and 30.8% Pull%. His game is speed and batting average, so he hits the ball on the ground to all fields. If ever you see a player with that sort of profile hit five homers over a couple of weeks, you know to chalk it up as an anomaly instead of a trend.

On To The Advanced Metrics!

Now let’s move on to some of the commonly used metrics. There are far more in existence than I’ll profile here, but these are the main ones that are commonly used in baseball analysis.

Let’s begin with my personal favorite: wOBA (weighted on-base average). wOBA is based on a similar scale as OBP (on-base percentage), where the average is .315 (league average OBP in 2017 was .318). wOBA takes into account a hitter’s outcomes and weighs them based on the value of the event. A triple is worth more than a double, a double is worth more than a single, etc.

It’s a very handy tool that allows you to just look at a number and get a good idea of how effective a batter has been. 2018’s wOBA leader was Mookie Betts at .449, but if you see any player higher than .370, that means they are performing as one of the best hitters in the league.

If for some reason you don’t like wOBA, you can also use OPS (on-base plus slugging) and come to a similar conclusion about a player.

To calculate OPS, all you have to do is add a player’s Slugging% to his OBP and bingo bango, you’ve got OPS.

The league average OPS in 2018 was .728. To use the same example as wOBA, Mike Trout led the league in OPS with a 1.088 mark, and any player with a .900 or better mark is playing at an elite level.

However, wOBA is a better tool than OPS, to put it simply. While OPS just mashes OBP and SLG together, wOBA properly weights every outcome and more properly values the non-intentional base on balls. Plus, “whoa-bah” is fun to say. Whoa-bah. Say it again, it’s a soothing de-stresser. Whoa-bah. Ahhh.

Quickly, a few more advanced stats to consider: OPS+ is an offshoot of OPS that is adjusted for league and park effects and is based on a totally different scale where 100 is average. A 110 OPS+ is a player 10% better than league average, a 95 OPS+ is 5% below average, etc.

Since everything is normalized, OPS+ is a good tool when comparing players to each other across different leagues and ballparks, and over different years.

However, just like wOBA is better than OPS, wRC+ is better than OPS+. wRC+ (weighted runs created plus) uses the linear weights that wOBA uses and adjusts them for league and park, also basing it on a scale where 100 is league average. For example, Nolan Arenado’s 2018 wRC+ of 132 tells us that playing in Coors Field wasn’t the only reason he was amazing since he was still 32% better than the league average player even when adjusted for ballpark.

Do The MLB Splits

Now let’s apply what we know to MLB splits for hitters.

In 2018, Matt Chapman enjoyed a breakout year. That said, his wOBA at home in pitcher-friendly Oakland was a mediocre .326. On the road is where he buttered his bread, hitting for an obscene .414 wOBA.

MLB splits for home/away aren’t generally that telling, but it can be beneficial to at least be aware of whether a player typically performs better at home or on the road. Lefty/righty MLB splits are much more applicable in player analysis. However, often times a sample size — even over the course of a full season — can be too small to take much away when simply looking at home runs or batting average. So we will use quality of contact and wOBA to help determine a player’s true skill over a shorter sample of at-bats.

Let’s take Justin Turner’s 2016 MLB splits as an example. He raked against righties with a .305 average and .385 wOBA. Against lefties, though, he hit just .209 over 172 at-bats for an ugly .283 wOBA.

However, if we look at his quality of contact, he actually hit lefties well. His 38.6% Hard% was actually better than he hit against righties (37.2%). Additionally, his BB/K (walks per strikeout; MLB average is 0.4) was 0.68, twice as good as he was against righties (0.34). That tells us Turner still has good command of the strike zone against lefties.

His BABIP against lefties was just .230, so if we put that all together, we can easily determine that Turner was incredibly unlucky against lefties in 2016, and over time that should correct. Flash forward to 2017, and what do we have but a .484 wOBA against lefties over 142 at-bats. Turner didn’t have some breakthrough against lefties; he just got the proper balls to fall.

An interesting player to watch for 2019 is Max Kepler. Left-handed hitters often struggle against same-handed pitching, and that was much the case with Kepler in 2017. While he hit for a .350 wOBA and 36% hard contact against righties, he posted an awful .203 wOBA and 22.8% hard contact rate against lefties.

However, in 2018, he improved against lefties to a .323 wOBA and 39.8% hard contact rate. Despite holding his 36% hard contact rate against righties though, his wOBA vs RHP dipped to .314. A .219 BABIP against righties is the culprit here, masking the huge gains Kepler made against lefties. A lot of that BABIP was deserved due to a surge in fly ball rate and the lowest HR/FB of his career (9.9%). Still, with a little better batted ball fortune and more balls leaving the park, Kepler could enjoy a breakout in 2019.

That’s gonna do it for this week. Next time we are going to dive into plate discipline for hitters, looking at contact rate, reach rate (O-zone%), and much more. Until then, go dive into some player splits!