The Right Stuff+: How Reds rank in new pitching models

For over a century, players, coaches, analysts, fans, and commentators have thrown around somewhat vague terms like “stuff” and “command” to describe pitchers’ effectiveness or lack thereof. But there were only indirect ways to measure these skills, such as ERA or strikeouts and walks. Now, with technology in place to measure nearly every aspect of every pitch — from velocity and spin to movement, release point, and location — data wizards have gone to work trying to measure the previously immeasurable.

Enter the Pitching+ model created by Eno Sarris and Max Bay.

The model has been around for a few years, but only recently was it added to the FanGraphs leaderboards, opening up a new way for more fans to understand the intricacies of pitching. The model is broken down into three components: Stuff+, Location+, and Pitching+.

What exactly does each component measure? And what do these numbers mean? Glad you asked!

A basic explanation of Stuff+, Location+, and Pitching+

In layman’s terms, these metrics are meant to show what makes a pitcher and individual pitches truly “good.”

Here’s how it works: the model seeks to measure how “stuff” and ability to locate pitches in ideal spots compares to other pitchers. It looks at the characteristics of each pitch thrown, compares it to real-life outcomes on similar pitches across the league, and assigns it a run value that is better or worse than average.

These metrics are more process-oriented metric than results-oriented. We know Max Scherzer is a hall-of-fame caliber pitcher based on the results he’s posted over the last decade. But what are the ingredients that have helped him produce those results? And why are those ingredients better than the ones used by other pitchers? The Pitching+ model helps to answer those questions.

The output for each metric should look familiar; like other “plus” stats — wRC+, OPS+, etc. — a score of 100 is average. Anything above that is better than average, and anything below that is less than average. (Note: 100 is not exactly average for individual pitch types, but it’s generally very close.)

Here are the specific components that go into each metric:

  • Stuff+ accounts for the physical makeup of a pitch, including velocity, spin rate, spin axis, vertical and horizontal movement, and release point. The most valuable traits are, unsurprisingly, velocity and movement. For secondary offerings, the model takes into account how the pitch plays off the pitcher’s fastball (e.g., the differences in velocity and movement).
  • Location+ accounts for a pitcher’s ability to throw a certain pitch in the ideal place, depending on the ball-strike count.
  • Pitching+ ties it all into a nice, neat ribbon. It takes the elements of Stuff+ and Location+ — physical pitch characteristics, location, and count — to grade a pitcher’s overall quality. That is, it assesses all the parts of a pitcher’s process.

There’s much more that goes into each metric, and here’s an explainer from FanGraphs that goes deeper into each one.

Let’s take a breath here and use some examples.

Off the top of your head, picture who you think is the nastiest pitcher in the league — one who lights up the radar gun and gets ridiculous movement on their breaking pitches. Did you picture Gerrit Cole? Corbin Burnes? Shohei Ohtani? Dylan Cease? Sandy Alcantara? Those were the top five pitchers in Stuff+ among qualifying starting pitchers in 2022. That certainly checks out. When we throw relievers in the mix, you probably won’t be surprised to learn that Emmanuel Clase was tops in Stuff+ last year, followed by Ryan Helsley, Felix Bautista, and Edwin Diaz.

Now picture a pitcher who consistently paints the corners and throws pitchers where batters simply can’t do anything with them. They’re often pitchers who don’t walk many batters. Aaron Nola might come to mind. Or Justin Verlander. Or Kevin Gausman. Or Miles Mikolas. All ranked in the top five in Location+ in 2022.

Which pitchers excel in both areas? No surprises on the Pitching+ list, either: Cole, Gausman, Verlander, Burnes, and Nola were the top five in baseball last year. If we lower the innings threshold a bit, Braves phenom Spencer Strider was the leader in Pitching+ last season.

In a nutshell, the Pitching+ metrics show how nasty a pitcher’s stuff is, how well his pitches play off each other, and how well he can locate them, relative to the rest of the league. Some other important items to note as you wrap your mind around these new stats and what they mean:

  • Pitching+ has already proven to be very predictive when it comes to projecting future performance.
  • While Stuff+ and Location+ can vary from outing to outing, they also stabilize fairly fast. This allows us to more reliably analyze a pitcher using smaller samples than we could with many other metrics. Per the FanGraphs article linked above, Stuff+ becomes reliable after only 80 pitches, while Location+ stabilizes after about 400 pitches.
  • As the FanGraphs article also notes, Location+ is more important for a pitcher’s success but Stuff+ is more stable from season to season.
How Reds pitchers rank in the Pitching+ model

Just how good is the raw stuff of Hunter Greene, Nick Lodolo, and Graham Ashcraft? What do the Reds see in some of the relievers who aren’t Alexis Diaz (other than their low price tags)? The Pitching+ model can help provide some answers.

Let’s look at the grades for every current Reds pitcher (plus some of the injured ones) since the start of the 2022 season:

For the most part, this data confirms the eye test:

  • Greene’s stuff is off the charts.
  • Ashcraft, Lodolo, Diaz, and Lucas Sims also have nasty stuff. In what might come as a surprise to some, so does Ian Gibaut.
  • Pitchers who’ve been serviceable despite having below-average stuff are often able to locate their pitches well — see Luis Cessa, Reiver Sanmartin, or Connor Overton.
  • Erratic pitchers who tend to walk a lot of batters (Sims, Diaz, Fernando Cruz) are reflected by below-average Location+ grades.

If there are any surprises, it’s that Lodolo ranks as roughly average in Location+ and Pitching+ since last year. But you probably won’t be surprised to see he’s showing signs of getting better. Through two starts in 2023, he has a 109 Location+ — tying him for fourth among all qualified pitchers. Remember that it takes about 400 pitches for Location+ to stabilize, but it’s encouraging that Lodolo is doing a better job of throwing his already-tough-to-hit pitches in spots that make him even harder to square up.

There are some other interesting nuggets in the 2023 data so far. Perhaps most notably, Ashcraft’s Stuff+ is up significantly, jumping to 128. Only one qualified pitcher ranks above Ashcraft through the first two weeks of the season: Shohei Ohtani. When looking at some of the parts under the hood, it certainly checks out. Ashcraft is getting more spin on all of his pitches, he’s throwing his slider harder, his cutter and slider are getting more horizontal movement, and he’s tightened up his release points. Given that Stuff+ stabilizes so quickly, it’s a promising sign that Ashcraft can take a leap forward in his second season.

Another pitcher who stands out is Derek Law, a 32-year-old reliever who has bounced around between five different organizations since 2018. He’s currently flashing as above average in all three categories (107 Stuff+, 107 Location+, 105 Pitching+), although it should be noted that he’s only thrown 64 pitches as of Monday. The Reds are surely seeing this data, too, which may explain Law’s early-season usage in high-leverage situations.

On the flip side of the coin, some notables include Buck Farmer and Overton. Farmer’s velocity is down, and his Stuff+ has plummeted to 87. Meanwhile, Overton has a 97 Location+ through two starts. It may not seem like a huge drop from 102 last year, but the margin of error isn’t large for a pitcher with well below-average stuff.

In our next post, you can take a look at the individual pitches for Reds hurlers and see which ones stand out in the Pitching+ model.

Photo by Peter Joneleit/Icon Sportswire

Matt Wilkes

Matt Wilkes got hooked on Reds baseball after attending his first game in Cinergy Field at 6 years old, and he hasn’t looked back. As a kid, he was often found imitating his favorite players — Ken Griffey Jr., Adam Dunn, Sean Casey, and Austin Kearns — in the backyard. When he finally went inside, he was leading the Reds to 162-0 seasons in MVP Baseball 2005 or keeping stats for whatever game was on TV. He started writing about baseball in 2014 and has become fascinated by analytics and all the new data in the game. Matt is also a graduate of The Ohio State University and currently lives in Chicago. Follow him on Twitter at @_MattWilkes.

4 Responses

  1. citizen54 says:

    Not sure I buy into the Pitching+ model as it is seemingly goes against the data at Baseball Savant and CSW. In theory, it sounds good but for whatever reason hitters are squaring up more than they should against Greene and whiffing more than they should against Lodolo. At Baseball Savant, Greene is a sea of blue in both expected stats and movement of pitches whereas Lodolo is mostly red. if you go buy Pitching + then Greene should be a much better pitcher than Lodolo, who according to the model is an average pitcher. This is over about 2k pitches. As for CSW, Greene is only slightly above average at 30% while Lodolo is at 31.7%.

    I realize the model is predictive and the sample size isn’t the largest but so far the actual results aren’t matching up with the predicted results. Maybe Green is experiencing a bunch of bad luck while Lodolo is getting lucky. I have only looked at Greene and Lodolo so maybe it is accurate for other pitchers and those two are outliers.

    • Matt Wilkes says:

      For me, Lodolo’s lower-than-expected Stuff+ score was the only real surprise among Reds pitchers. Was also a little surprised at Gibaut’s average Location+ score, as I thought it would be below average given his high BB%. But other than that, the scores mostly line up with other data I’m seeing and the old-fashioned eye test.

      The chart in the post is data from the last two seasons, so it skews heavily toward the larger sample from 2022. Early in 2023, Greene still has the higher Stuff+ grade, but Lodolo is well ahead in Location+ (109 to 103), which is making a big impact on the differences in batted-ball data you’re seeing. There’s definitely a bit of bad luck at play for Greene, too, as opponents have a .457 BABIP against him.

      That said, I think you have to look at far more than just two pitchers to determine the efficacy of the Pitching+ model. Like all of the stats you cite, the Pitching+ metrics are just a piece of the puzzle. Important pieces of the puzzle, to be sure, but they don’t always tell the whole story by themselves. Some pitchers had a great Pitching+ score last year but still had subpar results (Tyler Wells, Jose Berrios, Chris Archer, Dylan Bundy, Nick Pivetta, etc.).

      I’d also add that Pitching+ isn’t the only pitching model out there. I didn’t cover it here, but Pitching Bot is another prominent one and it grades Lodolo and Greene much more closely. It grades on the 20-80 scouting scale where 50 is average. Greene is at 63 since last year while Lodolo is at 59; this year, they’re at 65 and 61, respectively.

  2. Anonymous says:

    Great Pitching+ for Dummies writeup; thank you Matt!!

  3. dbp says:

    Great Pitching+ for Dummies writeup; thank you Matt!!