Sabermetrics: Why They Matter

Ignorance more frequently begets confidence than does knowledge: it is those who know little, not those who know much, who so positively assert that this or that problem will never be solved by science.
– Charles Darwin

The implementation of advanced metrics into baseball has created a rift between old and new school ways of thought. On one end of the spectrum, batting average, home runs, and RBI are still used as the ultimate ways to judge a hitter, while xwOBA, wRC+, and WAR dominate the other side. A reason for not adopting new metrics that gets frequently used is that these stats are all “made up” or “fake”. But what advanced stats actually do is analyze the process of a hitter or pitcher, rather than results. I’ll be doing my best to explain most advanced metrics used in the game today, and hopefully I’ll have convinced a few more people that these “made up stats” aren’t actually so bad for the game.

To start, I should probably explain why most traditional stats aren’t good for measuring player performance. RBI, one of the most commonly used measure for evaluating hitters among traditionalists, is never indicative of how a player is hitting in a given year. The #1 and #2 vote getters in MVP voting last year, Mookie Betts and Mike Trout, had 82 and 79 RBI last year, respectively. Neither of those totals even came close to the top 10, finally showing a willingness to go away from RBI as a measure of value. The problem with RBI is that it’s entirely based on opportunity, and better teams will afford their players more opportunities to hit with runners on base. For example, Mike Trout only had 130 plate appearances with runners in scoring position last year, with 44 of those ending in walks. He only had 78 at bats with RISP last year and still turned in 42 RBI in those limited opportunities. In contrast, J.D. Martinez had 200 plate appearances with RISP last season, and ended up with 158 at bats with RISP, a number that surpassed Trout’s plate appearances in those situations. With 80 more at bats in run scoring chances, there was no reason for J.D. to not finish with more RBI than Trout had simply because of the extra opportunities he was afforded. Now, RBI aren’t entirely worthless, as most of the league leaders in any given year will represent some of the top hitters in the game. However, most of the advanced stats I’m going to introduce will as well, and unlike RBI, will not be based on situational opportunities afforded by playing with a good offense.

Exit Velocity/Launch Angle

These two are truthfully why I wrote this article in the first place. I happened upon a Twitter thread that claimed that Aaron Judge should “focus less on exit velocity and launch angle and worry about making good contact”, or something to that effect. That guy, and several others, were essentially saying that Aaron Judge should stop worry about hitting the ball hard in the air, and, worry about, making good contact instead? Exit velocity isn’t a new metric. It’s literally just how fast the ball leaves the bat at the point of contact, yet far too many people treat it as a new metric. Launch angle is also not new, as people now just have a way of tracking the angle at which the ball leaves the bat. I can pretty much guarantee that all of Hank Aaron’s 755 home runs had a good exit velocity and launch angle, otherwise they probably wouldn’t have been home runs. Bad launch angle = Pop Flies and Grounders (check out the Pop Flies and Grounders podcast hosted by our boy Denis Ackerman #ad). Good launch angle = line drives and fly balls. Can a player mess up their swing by trying to create an ideal launch angle? Of course they can! That doesn’t mean the concept is new. The only thing new is the name.

wOBA

wOBA, or weighted on base average, attempts to do with one stat what on base percentage (OBP) and slugging percentage (SLG) were meant to do. It takes into account the amount of times a player reaches base, but it also rewards them for the way they reached base. OBP treats all methods of getting on base as the same, while SLG can be inherently flawed because of how weighted doubles and triples can be. wOBA accounts for the run environment in a given season, so different methods of reaching base will be assigned different values based on how likely that scenario will lead to a run. For example, Fangraphs gives the example of 2014, where a home run was worth 2.101 times on base, and a walk was worth 0.69. So a player who had a 1-4 game with a home run and a walk would have a wOBA of .558, calculated by adding 2.101 + 0.69 and then dividing by the 5 plate appearances in a game. wOBA can still be used on a similar scale as OBP; the 2019 leader in wOBA, Christian Yelich, has a mark of .441, while the OBP leader, Mike Trout, has a .437. The two stats typically have similar marks to each other, but the slight fluctuations in calculations provide a fuller picture to what a hitter is actually doing.

Why it’s useful:

wOBA is a more accurate measure of how valuable a player is offensively. OBP assumes that all times on base are equal, and SLG can look inflated when presented without context. Since wOBA uses calculations that take into account how valuable a given outcome is in comparison to how often it happens in the league that year, it gives a far more accurate valuation of the value a player is bringing to the plate in a given season.

xwOBA

Have you ever seen a player go 4-4 in a game, but all 4 hits were bloop hits or infield singles? How about a player go 0-4 where every ball was a line drive right at an outfielder, or a deep fly ball caught at the wall? This is where xwOBA comes in. In scenario 1, the batter’s xwOBA would be significantly lower than in scenario 2. xwOBA stands for “expected weighted on base average”. The “expected” part comes from the exit velocity and launch angle of a baseball, which is compared to similarly hit baseballs since the introduction of Statcast tracking back in 2015. Using the data stored from Statcast tracking, batted balls are able to be assigned a “hit probability” based on the results of similarly hit balls in the past. Batters are then assigned an xwOBA from the hit percentage of the ball they hit. xwOBA is used as a predictive tool for sustainability of hitters, as it tracks the quality of contact a hitter makes rather than the results produced from the at bats they take. It takes out the luck involved of placing a baseball well enough to get a hit, and instead rewards batters consistently hitting hard line drives and fly balls. If you’re curious to know how accurate xwOBA is, here are the top 7 in xwOBA for 2019:

1) Mike Trout
2) Cody Bellinger
3) Nelson Cruz
4) Yordan Alvarez
5) Christian Yelich
6) Anthony Rendon
7) J.D. Martinez

All 7 of those guys are having monster years, so it’s safe to say this is a pretty reliable metric to use.

Why it’s useful:

While xwOBA can vary extremely from the results a player is actually putting up, it is usually a good measure of how sustainable current performance is, and thus is used as a predictive tool to predict future success. For example, Ryan O’Hearn had a wOBA of .398 in a breakout 2018 campaign, but had the 3rd highest disparity between wOBA and xwOBA with his xwOBA checking in at .345. Looking at the drop off there, it would have been a good bet to assume that O’Hearn was going to regress going into this year. On the other side, J.D. Davis had the highest change from wOBA to xwOBA in the other direction, with his wOBA being .218 and his xwOBA being .299. While it was unreasonable to expect this kind of breakout from him, everything pointed to Davis being far better than what the results showed in 2018.

wRC+

wRC+, or weighted runs created +, is currently regarded as one of the best tools to measure a player’s overall offensive production. In short, wRC+ is used similarly to wOBA in that all outcomes are valued properly instead of valuing every time on base as the same. wRC+ also accounts for the offensive environment of the current season and offensive factors of a team’s ballpark and sets numbers on a scale to 100, with 100 being league average and every number above or below representing a 1% difference from league average. For example, a wRC+ of 125 would represent 25% above league average, and a wRC+ of 90 would represent 10% below league average.

Why it’s useful:

Unlike wOBA and xwOBA, wRC+ gives a direct comparison to the rest of the league in order to show how much better a player is than the players he is currently playing against. wRC+ takes all measures of hitting in, so a player who hits 50 home runs with a .310 OBP won’t rank as high as a player with a .420 OBP and 35 homers. It also does a good job of measuring park factors too, so the wRC+ leaders in a given year will never be a bunch of Rockies. wOBA and xwOBA are hard to use by themselves without context, but wRC+ is measured in way that doesn’t need context because it already makes comparisons for you. As a stat, it most accurately measures the overall contributions a player makes on an offensive side, without putting too much emphasis on team based stats like RBI or R.

The top 10 in wRC+ this year are:

1. Mike Trout (180)
2. Christian Yelich (169)
3. Cody Bellinger (166)
4. Nelson Cruz (162)
5. Anthony Rendon (161)
6. Alex Bregman (159)
7. George Springer (153)
8. Carlos Santana (148)
9. J.D. Martinez (147)
10. Michael Brantley (147)

Again, that list is full of players who are all mashing in 2019, which backs it up as a reliable stat to use for offensive performance. One of the biggest obstacles to using these new stats in daily practice is understanding them. What most of these stats do, is take the important parts of traditional stats and use them as an overall measure for performance instead of trying to use a lot of stats to compare with each other. Traditional counting stats are still fun to track, but hopefully we are moving towards a day where advanced stats become the norm.