Last year, Pitcher List’s own Chaz Steinberg started a weekly column dedicated to ranking and analyzing pitchers based on their plate discipline metrics. The column was a fascinating way to look at pitchers from another angle and helped form the foundation for understanding these crucial statistics that we now use every day. Although Chaz was unable to continue the column this year, we didn’t want the rankings to die, so we’re bringing them back! Every week my incredible colleague Alex Drennan and I (I’ll try not to bog things down too much I promise) will take what Chaz started and help update them for the 2019 season. As the season progresses, we hope to bring a pair of new voices to the project while staying to true to the heart of the column and we also plan to add to the rankings system as we discover new things and grow in our knowledge. Before we get into the initial 2019 Pitcher Plate Discipline rankings, we’d like to introduce ourselves individually and explain what drew us to the project and what we hope to accomplish throughout the season, so bear with us for a few minutes before we get to the good stuff.
Daniel Port: By its nature, fantasy baseball is a results-based exercise in statistics. We have a habit (for obvious reasons) of focusing on the outcome of a situation rather than how we got there. The standings don’t care how a pitcher got to a 6 IP, 3 ER, 8 K outing. But for those of us who analyze baseball, the results are often only part of what care about. Any given event in baseball is usually the result of several different forces, often independent, all coming together to affect the outcome of every single play. I often feel like it’s easier to explain how the Three Mile Island accident happened than it is to explain all the factors that went into any event in baseball. That’s the beauty of baseball. It’s essentially a group of highly skilled athletes coming together to play a game in a grass-covered chaos dome. It’s fantastic. If we grant the premise that there are a ton of factors that go into the outcome of most at-bats, then how do we properly evaluate a pitcher’s true skills and, perhaps indirectly, whether or not they are properly using those skills? That, my friends, is exactly what Alex and I are hoping to find out.
Now many of you will say we already have lots of statistics that do this. You’ll point to FIP, xFIP and SIERA (my personal favorite). I use them all the time. These are a great start to figuring out what results are affected by a pitcher’s skills and what is affected by the myriad of factors outside of the pitcher’s control. The thing is though, these measurements are what we call index statistics. In other words, they’re a combination of several connected statistics with the goal of creating a neat tidy number that, at a glance, tells a broad view of what we need to know. Anyone who has talked baseball with me before or has read my work knows I have a real love/hate relationship with index statistics. They’re awesome for drawing big picture, all-encompassing conclusions. The thing is they rarely tell you why those conclusions came about.
In college, I got to spend a month living in Spain. Between the sangrias on the beach and getting my butt kicked in soccer by the locals, I fell in love with Spanish art, especially paintings done by Spanish Renaissance painter El Greco. One of his pieces, a painting titled The Burial Of Count Orgaz, is still my favorite painting. I’ve tried linking to it, but any image on the internet does not do justice to the sheer size of this thing. Sitting at roughly 15 by 11 ft, it literally covers one entire wall of the church where it is housed in Toledo, Spain (no, the Mudhens don’t play there). When you walk into the church’s single, open atrium it’s honestly the only thing you can look at; it’s that enormous and stunning and beautiful. The real genius of the painting though is when you finally catch your breath and take the time to get right up to it. Here you take in the excruciatingly fine details of the painting, how every little brush stroke comes together to create the whole picture. Then, when you step back again to take in the whole painting, it is that much more stunning and beautiful because now you’ve seen how it was created. This is how I look at stats like FIP, xFIP, and SIERA. They are the painting when you first walk into that church in Toledo. My goal with this column is to show you their brushstrokes. I do firmly believe that plate discipline stats lie at the heart of demonstrating a pitcher’s skill. Over the rest of this article, Alex and I hope to take the foundation that Chaz laid out last year when he created this column and show how these plate discipline numbers can help give context to a pitcher’s skills, and hopefully then be able to better understand those big picture index statistics and the art of pitching as a whole.
Alex Drennan: I had a falling out with baseball when I was entering high school. I don’t have a definitive answer for why it happened, but September 2007 certainly played a part in it. September 2008 was a less gut-wrenching let down on its own, but taken in conjunction with 2007 it was so painful. Maybe it was the teenage angst that made me give up on the Mets, or maybe I was just too afraid to be let down again. Whatever the cause, I was pretty removed from baseball until 2012. 5.1 IP, 0 ER, 11 K was all it took to draw me back in. That was the line from the MLB debut for Flushing’s white knight, who went on to become the Dark Knight in 2013. Sure, the Mets were still terrible, but Matt Harvey was dominating batters in a way I’d never seen before and it was captivating. My relationship with Harvey has soured since then (I was more than happy to get Jacob deGrom‘s personal catcher in return), but my relationship with the art pitching has aged like a fine wine.
Since’s Daniel already explained how the indexes should be viewed in the landscape of baseball statistics and what the plate discipline metrics are trying to shed more light on, I’ll try to explain how we’re planning to build on the foundation Chaz laid out last year. We fully acknowledge that this is an imperfect measure of a pitcher’s worth, just as every other metric attempting to is. Our goal is to make this piece as encompassing and understandable as possible. Some things currently missing in the metric, at least explicitly, are:
- Strike Zone Plot Evaluation
- Individual Component Grades
- Impact of Velocity
At this point in time, I’m not sure how easily we’ll be able to incorporate these things into our piece or how important they even are, but that’s exactly why they’re worth investigating. In addition to the new stuff, we’ll be reviewing what’s already in place to see if it needs tweaking. The weighting system for each PD component Chaz has outlined makes sense on the surface, but wOBA and wRC don’t weight a double as two times the worth of a single like SLG does, so maybe SwStr% should be worth three-and-a-half times what FS% is worth instead of just three times. This is a learning effort for us, so we’re open to any suggestions or feedback that would help us improve the piece. Please leave feedback in the comments and we’ll be more than happy to address it.
Part One: Which stats are we using, and what are we hoping they will tell us?
We’ll be starting with the foundation and formula that Chaz laid down and using much of his findings and his data as a reference point. I waffled back and forth on whether I should simply regurgitate this information here in its entirety or summarize as best I can while directing you to his original piece in order to make sure that Chaz gets the credit. After much consultation, I’ve decided on the latter. What follows is summary on each stat included in this exercise, but if you want to see the initial research and findings that Chaz put together check out his original primer here. It’s well worth the read and has tons of useful information.
For the purposes of these rankings, we’re going to be looking at two groupings of stats:
Group #1 – K% and BB%
Group #2 – O-Swing%, Contact%, SwStrik%, and F-Strike%
We will mostly look at Group 2 within the context of how those stats correlate with those of Group 1. This is playing off the idea that the two factors a pitcher has the most control over are when he walks a batter and when he strikes them out, hence K% and BB%. You’ll see as we break down each stat in Group 2 that we can actually predict K% to some degree based off these numbers and therefore hopefully have some idea of the pitcher’s skill. First though, some brief context on K% and BB%.
K% – This is a pretty straightforward stat. It’s simple of measure of what percent of batters faced a pitcher strikes out. We typically consider anything above 20% to be above average, but really we like to see a starting pitcher getting a K% somewhere north of 23% before getting excited about them. As Chaz pointed out in his primer from last year, while K% is a really great indicator for both pitcher skill and pitcher success, it does have some flaws built into it. First off, it can have its share of luck-skewed outliers. Like Chaz says, a strikeout can be dependent on an umpire’s strike zone and whether or not a well hit ball falls foul instead of becoming a hit. The same thing goes for a caught foul tip on strike three. It also doesn’t account in any way for the quality of the opponents that the pitcher has faced, nor how strikeout-prone they may be. With that being said, it is still widely considered a darn good barometer for how well a pitcher is throwing and how much success he should be having.
BB% – We’re not going to talk too much about BB% other than to note that the lower it is, the more likely the pitcher is going to be successful. It’s essential that a pitcher doesn’t give up baserunners and it’s even more important that he doesn’t give up the ones he has the most control over. The same umpire-related caveat applies here as well, but this stabilizes pretty quickly, so we should get an idea right away if it’s bad luck or skill-related. It’s also worth noting a BB is BB here: There’s no way of noting if it was a strategic walk; they all count equally.
Now that we’ve touched on the stats in Group 1, let’s take a look at the stats in Group 2 that help lend context to those numbers.
O-Swing% – This is simply a measure of how often the pitcher gets a batter to swing at a pitch outside the zone. Being able to get a hitter to chase pitches is a quintessential skill if a pitcher is going to be successful. As Chaz showed in his primer, a higher O-Swing% has a positive effect on both K% and BB%. This makes sense. Getting hitters to swing at bad pitches would certainly lead to strikeouts and throwing throwing fewer balls leads to fewer walks. It doesn’t seem to have quite as strong of an effect on the two metrics as some others, but there’s a reason we include it in the Money Pitch standards. Note the 40% or higher is considered elite and we’re looking for at least 35%.
SwStr% – This is the sexy stat. Everyone loves a swinging strike and for good reason. It’s the single outcome that the pitcher controls entirely on his own. The umpire can’t affect it, nor can the pitcher’s defense, nor his park. It doesn’t take a genius to figure out that a high SwStr% correlates pretty heavily with a high K%. I would love as the season goes on to find a useful way to incorporate our CSW% statistic, which incorporates called strikes as well, but we’ll see if this is best used as a separate stat or as a replacement to this one. Anything above 10% is above average, but we start getting excited between 11%-12%. The fascinating thing that Chaz found regarding SwStr% is that due to its high correlation with K%, you can actually use it to calculate an Expected K% by doubling the SwStr% and adding one. This will be instrumental throughout the year in helping us get an idea of how much of a pitcher’s strikeout performance is luck-based or skill-based.
Contact% – This measures how often a hitter makes contact when he swings. Including this metric makes perfect sense. Outside of walks, the only major way a hitter can get on base is by making contact with the ball. As Chaz states, you can draw major conclusions about a pitcher’s ability to strike players out by its inverse relationship to Contact%. If they swing and don’t hit it, that’s the desired result. When put into this perspective it becomes pretty clear that there is a strong relationship between Contact% and K%. It is worth noting that Contact% doesn’t include foul balls. We can also use this stat to predict K%, according to Chaz, by taking Contact% and simply subtracting it from 100%. The remaining percent is your Estimated K%. We will actually combine this number with the Expected K% above and use that to calculate our final K% to make sure we’ve covered all our bases.
F-Strike% – Last One. Here we are told how often a pitcher throws a first-pitch strike to start off an at-bat. It doesn’t have a ton of impact on K% (I have a sneaking suspicion that what you do with the second pitch of an at-bat has a larger impact on K%), but it does have a decent effect on reducing BB%, which is key to a pitchers success, so we’re going to include it. Following Chaz’s initial formula, we won’t give it nearly as much value as the other three metrics we just went over.
Part Two: The Grading Scale
We wanted to continue Chaz’s method of creating an overall score for a pitcher’s plate discipline metrics and then ranking them using this score. As Alex and I tweak our methodology throughout the year this may change a bit, but for the first few weeks we are going to roll with the old weighted system:
Contact%: Multiply by 3
SwStr%: Multiply by 3
O-Swing%: Multiply 3
F-Strike%: Leave as is
Then we add them all together and divide by four. This should leave us a score somewhere between 60 and 100. Then we’ll assign the player a letter grade that follows along the lines of the old tried and true grading scale we used in school, that way you have a way to get a quick reference point for how well a pitcher is performing with regards to their plate discipline numbers. Each week we’ll dive into a few players to either further explain their score or perhaps lay out the reason for a gap between their results and what their score tells us about their skills. Note that this early in the season we may see some extreme grades, but it will all settle out as the sample size gets larger. In case you would like to see what a full season sample size looks like, we’ve run the full 2018 results through the system and have included them as an appendix in this article. This can give you a good sense of what scores and stats might look out of place or could potentially be early season outliers.
Part Two and a Half: How could we forget Statcast data?
In addition to the four metrics we used for the Part Two grading, there’s a Statcast component to the grading as well. We’ll be using Baseball Savant’s xSLG for two reasons:
1) Strikeout and walk rate are captured in xwOBA and we’re trying to add a different flavor rather than adding too much of the same thing to the mix.
2) We’re trying to capture true skill without any luck-based factors, which is why we went with xSLG rather than SLG.
For the grading, we’re subtracting xSLG from 100%, since a higher xSLG is worse for pitchers, and then multiplying that by 1.5 to get roughly the same grading scale we used for the other four metrics. The xSLG grade is weighted the same as Part Two’s, so the final Plate Discipline grades are an average of Part Two and Two and a Half.
Part Three: The Rankings
[table id=60 /]
There you go! There are the first round of rankings and grades for the 2019 season! If you’re skeptical of this measure right now it’s understandable, but if you look to the 2018 leaderboard below, that should help remove some of the skepticism. Either Alex or I will be back next week with some overall takes and players we would like to highlight. Until then, have fun with the table and let us know in the comments if any pitchers, in particular, catch your eye!
Appendix A: The 2018 PD Rankings and Grades
[table id=61 /]
(Photo by Jimmy Simmons/Icon Sportswire)