Sabermetrics Hitting 202: Intermediate Hitting Stats

Last week I went over the basic hitting stats and gave some more information about them. Today, we go to the next level. Per comments from last week, I will add a bit more information on how to use these stats. Here we go!

---

Intermediate Hitting Stats and Metrics

Strikeout Rate (K%) and Walk Rate (BB%)

Formula: K% = Strikeouts per plate appearance = K/PA

BB% = Walks per plate appearance = BB/PA

What does it actually mean: These two stats measure how often a hitter strikes out or walks, respectively, per plate appearance.

Is this useful? If so, how? These stats are both very useful because they provide a strong measure of a player’s plate discipline and contact skills. As we discussed last week, K% is a strong factor in a player’s batting average, a vital component to both OBP and SLG, while BB% is the other key component in OBP. Together, K% and BB% have a fairly significant part in determining a player’s overall hitting ability (mostly on the “avoiding outs” portion). Both stats are easy to use and are remarkably reliable.

How can I use it? Utilizing K% and BB% can be very straight forward - they are a simple rate stat and can quickly give you an idea of a player’s approach with a quick glance. These stats also tend to “stabilize” very quickly - ~60 PA for K% and ~120 PA for BB% - so you don’t need a lot of data before you have a good estimate for the player’s ability. This can also help in determining if a player’s skill in either aspect has changed.

K% is meaningful because not only does it measure a player’s strike out frequency, it also gives an idea of their ability to make contact. Players with extremely low K% tend to have very high contact rates (See 2016 Jean Segura) while players with high K% tend to have very poor contact rates (Peter O’Brien). Also, a high K% is generally correlated with a batter that has poor pitch recognition or selection skills. It is very difficult for a player with a very high K% to be a successful MLB hitter unless they have either very high walk rates or power (or both).

BB% is a very important skill because it can help a batter avoid a lot of outs and therefore, raise his “floor” as a batter. Generally speaking, a batter with a high walk rate means they are great at distinguishing between balls and strikes and often means they make better contact when they do swing. The best hitters almost always have strong walk rates for three reasons:

1. More outs avoided via walks adds more value to a batter
2. Strong pitch recognition leads to better contact
3. A better hitter is more likely to be pitched around

Note that reasons 1 and 2 are inherent skills to the batter while the third reason is a result of the batter being good.

Deeper Dive: Let’s look at the general ranges for K% and BB% as of 2016/2017.

Note that this table is a bit subjective (these descriptors are a general idea) and will vary from year to year. To determine a batter’s ability, it is best to compare the player to that year’s league averages. Batters in 2017 may have substantially higher K% than hitters of the past, but they are not actually worse hitters, as the game has changed.

In a perfect world, the best hitters would have low strikeout and high walk rates. However, there is also a very strong correlation between strikeouts and power - that is, players with good power also tend to have higher strikeout rates. Usually these players don’t stray into the “bad” range for K%, but they’ll often be above average (Paul Goldschmidt is a perfect example).

Weighted On-Base Average (wOBA)

Formula: [(0.69 x uBB) +(0.72 x HBP) + (0.89 x 1B) + (1.27 x 2B) + (1.62 x 3B) + (2.10 x HR)] / (AB + BB - IBB + SF + HBP)

(note: these values will change slightly year-by-year)

What does it actually mean: wOBA is a rate statistic that attempts to properly credit a hitter for each outcome by the type of outcome (walk vs single vs homer, etc.) rather than treating all hits or times on base the same.

Is this useful? If so, how? Created by Tom Tango, wOBA is an incredibly useful and powerful metric to measure a player’s overall hitting value. Remember how I mentioned last week how OPS had flaws because it treated OBP and SLG as equals while also misrepresenting the run values of the types of hits in slugging (single = 1, double = 2, etc.)? Well, wOBA is the result of what OPS was trying to achieve. wOBA combines all outcome for a hitter and combines them into a single metric that is weighted by the actual run value of each outcome.

Don’t let the formula for wOBA intimidate you - you should never need to calculate it by hand. Instead, look at what it is trying to tell you: the run values for the various types of hits. For instance, SLG tells you that a double (weighted value of 2) is worth twice as much a single (weighted value of 1), while wOBA states that a double is worth about 43% more than a single (1.27/0.89). By using the “weights” (aka the number in front of each stat) you can compare the overall run value of each outcome. Essentially, wOBA is a more accurate form of OPS.

Do note that wOBA is context neutral, meaning that a batter is not given more/less credit based upon the number of runs actually scored by that single outcome. For example, the possible run outcomes for a home run can vary from 1 (solo home run) to 4 (grand slam), depending on the number of baserunners. wOBA does not want to include the runs attributed via these baserunners as the batter has no control over how many batters are on base when he comes up to bat. That’s why you see a fixed linear rate for home runs at 2.10: essentially, the “average” home run contributed about 2.10 runs per HR when you factor in all of the baserunners.

Lastly, wOBA is scaled to closely align with the league average OBP. So, the league average wOBA should be very close to the league average OBP.

How can I use it? wOBA is very simple, really. All outcomes are scaled appropriately such that you can accurately compare and evaluate players without the issues faced when using OPS. An estimated scale of wOBA from 2016:

Bad: .300 or lower

Average: .320

Elite: .370 or higher

wOBA only measures a player’s output with the bat. Stolen bases and baserunning values are captured via other metrics.

Deeper Dive: wOBA is also useful because it can be easily converted to weighted runs above average (wRAA) which is a good baseline for determining how far above/below average a hitter is on a per run basis. The formula:

wRAA = [(wOBA - League_wOBA)/wOBA_scale) x PA)

League_wOBA is the league average wOBA for the given year and wOBA_scale is an empirical scalar to scale to OBP (more here). Simple take a player’s wOBA, subtract the league average, divide by wOBA_scale, then multiple that value by the total number of PA and you get the number of runs above/below average for that player. It’s not exact, but the general rule of thumb is that .020 points of wOBA is approximately 10 runs above average per 600 PA. So a 2016 player with a .320 wOBA would have 0 wRAA (0 means average, not above or below) and a player with a .340 wOBA would have 10 wRAA.

Weighted Runs Created (wRC) and Weighted Runs Created Plus (wRC+)

Formula: wRC = [((wOBA-League_wOBA)/wOBA_scale) + (League_R/PA)] x PA

wRC+ = (wRAA/PA + League_R/PA) + (League_R/PA - (Park_Factor x League_R/PA)) / (League_wRC_excluding pitchers/PA) x 100

What does it actually mean: wRC+ is a rate statistic that measures a hitter’s total value while adjusting for park effects and the current league run environment. Furthermore, wRC+ is scaled such that league average is set to 100 every year and every point above or below 100 is equal to one percentage point better or worse than league average.

Is this useful? If so, how? Again, don’t let the math and formulas confuse you as these will generally be automatically calculated by spreadsheets. wRC+ is effectively the premier stat for measuring a player’s overall value as a hitter. wRC+ is the final step after wOBA because it takes into account park factors (e.g. hitters shouldn’t be valued as “better” hitters because they play in Coors Field where it is much easier to hit homers) and the league run environment. Converting it to a linear scale also makes comparisons very simple: for example, a player with a wRC+ of 150 created 50% (150 - 100 = 50%) more runs than average while a player with a wRC+ of 50 created 50% less runs than average.

wRC is very rarely used on its own but is included as the general basis for wRC+. wOBA and wRAA are generally the stats seen on a player’s statbook.

How can I use it? wRC+ is the go-to stat for overall player valuation both within a year and from year-to-year. The park factor scaling is a big factor as parks can vary from about -5% to +20% of average, so a player that plays all their games in Colorado would gain a roughly 20% boost to their overall stats compared to the same player playing at an average ballpark. wRC+ accounts for this factor so that a player is valued appropriately regardless of where they play.

A similar mindset applies from year-to-year. The run environments can vary dramatically from year-to-year. For example, the league average wOBA in 2016 was .318 while in 2000 it was .341. Mike Trout lead baseball with a 171 wRC+ last year with a .418 wOBA. In 2000, Gary Sheffield had a 173 wRC+ but a much higher wOBA of .451. If you looked at just wOBA, one would assume that Gary Sheffield was a much better hitter than Mike Trout. However, when accounting for league run environments, both hitters ended up about 70 percent better than average. Generally speaking, the best players are never more than 60-100% better than average (or wRC+ of 160-200). Since 1990, there have only been 7 season with a wRC+ better than 200, 4 of which belong to Barry Bonds.

Deeper Dive: wRC+ is essentially a more accurate form of OPS+, though they do have a strong correlation with each other. An important thing to note is that the number of PA is also important, though not specifically calculated: a wRC+ of 150 over 600 PA is much, much harder to achive than a wRC+ of 150 over 60 PA.

---

So, there we have it. Between last week and this week, we have the majority of the tools that are used to measure and value hitter performance. A key factor here is that these measures don’t necessarily indicate future performance - this is a whole other ballpark.

Predicting future performance is the main goal of Sabermetrics but in order to do that, we need to know how to properly value and measure offensive value, first. And now we have the tools to do that. Predicting future performance requires deeper analysis and more knowledge of statistical methods - something we will cover in the future.

Next week, I think I’ll take a break from the stats and do some sort of analysis on the team. After that we’ll dive into the basics of pitching stats.

Now, go finish up your taxes!