The Cardinals have their work cut out for them this off-season with regards to the pitching staff. The bullpen’s performance was dreadful all season long. The rotation performed very well, but is bursting at the seams with options entering 2019. Adam Wainwright will be back with an undefined role, while Alex Reyes is a wild card. Tyson Ross and Bud Norris, each mostly effective out of the bullpen in 2018, are not likely to return. Tyler Lyons was very quietly granted free agency a few weeks ago. Thirty players took the mound for the Cardinals last season, and 21 of them are still on the table as 2019 options. That assumes Ross and Norris are gone. There’s a lot to work through.
There will certainly be some new additions, and some of the 21 remaining will be stashed in AAA, traded, or granted free agency. However, it’s a good idea to know what you have before making those types of moves. You need to know if a pitcher got better or worse results than you can reasonably expect. I want to take a different approach today in making that determination. Let’s play around with velocity and spin rate to see how Cardinal pitchers fared compared to their league-wide counterparts.
I’ve collected the average velocity, average spin rate, and wOBA (weighted on-base average) allowed against each specific pitch for each Cardinal pitcher in 2018. These pieces are readily available on Baseball Savant.
From there, I determined the wOBA allowed by the average pitcher from 2016 to 2018 on similar pitches. Then, that number was compared to each respective Cardinal, who was given a league-relative wOBA (a wOBA+, as it were) comparing their performance to the rest of the league. Note that I did not perform a ballpark adjustment, so keep that in mind as you view the results.
Here’s a real world example. Luke Weaver’s average four-seam fastball this season was 93.7 miles per hour, with a spin rate of 2,324 RPM (revolutions per minute). He allowed a .342 wOBA on fastballs this season. Since 2016, when all pitchers have thrown four-seam fastballs between 92 and 93.9 mph with a spin rate between 2,300 and 2,399, they’ve allowed a .349 wOBA. We then take Weaver’s .342, divide it by the league-wide .349, and multiply it by 100. The resulting figure- 98.0- means that Weaver’s four-seam fastball was 2% more effective than league average, with 100 representing league average. Put another way, in the absence of other factors beyond velocity and spin, you would have expected Weaver to allow a wOBA that was 2% worse against his fastball.
I performed this exercise for four-seam fastballs, sinkers, sliders, and curveballs. I also combined two-seam fastballs and cutters into their own category. I did not research change-ups because so much of the effectiveness of a change-up is inherently wrapped up in other factors. The effectiveness of a change-up says a lot more about the rest of a pitcher’s approach and repertoire compared to the quality of the change-up itself. As such, I excluded it.
I’ve also excluded pitchers who are already gone. You won’t see Sam Tuivailala, Greg Holland, Preston Guilmet, or Ryan Sherriff in here. You also won’t see Jedd Gyorko or Greg Garcia because, come on, why would I do that? Lastly, I’ve excluded Alex Reyes and Giovany Gallegos on the basis of their minuscule sample sizes. However, I’m leaving Lyons, Norris, and Ross in place because there may be something instructive regarding their performances in the underlying data.
We’ll start with four-seam fastballs. I’ve opted to leave the number of results- the number of plate appearances that ended on this particular pitch- in these tables in the interest of transparency. You’ll see why as you dig through the data.
4-Seam Fastballs vs. Lg. Average
It’s easy to dismiss Bowman, Hicks, Hudson, Gregerson, Lyons, and Wainwright, whose results are built on 20 plate appearances or less. Leone, Cecil, and Webb are also teetering on the brink. The biggest outliers here are Norris in the negative and Ross in the positive. Pitchers with Norris’ four-seamer typically see much better results, whereas pitchers with Ross’ four-seamer typically fare much worse. Something in their make-up led to an inversion in results.
Michael Wacha, Miles Mikolas, John Gant, Carlos Martinez, and Jack Flaherty- a group that combined for 112 of the team’s 162 this season- all well exceeded the league averages for similar fastballs.
Two-Seam Fastballs and Cutters
Moving on to the other kinds of fastballs, here are the results for two-seamers and cutters.
Two-Seamers and Cutters vs. League Average
|Luke Weaver||2278||87.1 MPH||24||0.249||0.344||72.4|
|John Gant||2424||93.2 MPH||192||0.283||0.355||79.7|
|Michael Wacha||2129||89.4 MPH||59||0.269||0.337||79.8|
|Daniel Poncedeleon||2354||89.5 MPH||31||0.273||0.331||82.5|
|Bud Norris||2508||91.9 MPH||141||0.262||0.305||85.9|
|Jack Flaherty||1989||91.1 MPH||97||0.329||0.368||89.4|
|Carlos Martinez||2142||91.8 MPH||208||0.355||0.365||97.3|
|Dominic Leone||2424||90.8 MPH||62||0.330||0.333||99.1|
|Miles Mikolas||2215||93.5 MPH||173||0.347||0.345||100.6|
|Brett Cecil||2196||86.1 MPH||20||0.344||0.337||102.1|
|Adam Wainwright||2326||83.9 MPH||33||0.344||0.331||103.9|
|Matt Bowman||2011||91.2 MPH||61||0.389||0.370||105.1|
|John Brebbia||2158||93.5 MPH||5||0.407||0.367||110.9|
|Tyson Ross||2496||90.8 MPH||9||0.400||0.333||120.1|
|Luke Gregerson||2320||87.8 MPH||23||0.452||0.331||136.6|
Small samples more or less eliminate Weaver, Brebbia, Ross, Gregerson, Poncedeleon, Cecil, and Wainwright here. As we’ll see, this is the only pitch where Mikolas had worse results than league average. Gant, Wacha, Norris, and Flaherty stand out as overperformers on two-seamers and cutters. For the first three in that list, they amassed impressive results on these pitches. Flaherty, on the other hand, was still better than the overall league average on two-seamers and cutters, but his velocity and spin typically yield worse results.
It’s a limited group throwing enough sinkers to register, but there are some doozies in there.
Sinkers vs. League Average
It’s tough to deduce much from Lyons and Cecil, though it’s worth noting that they’re showing up with worse results compared to the league average on almost every single pitch. Something beyond velocity and spin prevented them from having better seasons.
Wainwright exceeded expectations, but it has more to do with the league-wide ineffectiveness of sinkers thrown with his spin and velocity. The real beasts here are Hicks and Hudson. The average wOBA against all sinkers, regardless of velocity and spin, is .350. With that in mind, even if Hudson regressed to league average for his spin and velocity class, it would still be an effective pitch. It’s possible that Hicks is due for some regression but I suspect other factors pushed his sinker wOBA down below league average.
Now we start seeing the limitations of sample size for all of our relievers.
Sliders vs. League Average
Seven of our 17 pitchers fail to cross the 30 result threshold. If it seems like these are impressive wOBAs allowed by Cardinal pitchers, bear in mind that the league average on all sliders- regardless of velocity and spin- is .266 over the last three years. Still, we can see some encouraging signs here. Gregorson’s mix of velocity and spin is typically lethal. The same is true for Brebbia, who underperformed on the pitch. If velocity and spin is any indication, he probably has better times coming his way. Mikolas was absolutely devastating with his slider, and the spin and velocity combo suggests he should have been above average. The same is true for Hudson.
Finally, here are curveballs, which seem to be an all or nothing pitch. Bowman, Martinez, Poncedeleon, and Hudson rarely end at-bats with the pitch, rending their results meaningless.
Curveballs vs. League Average
For context, the league-wide wOBA against all curveballs is .266. Flaherty and Wainwright performed exceptionally well on the pitch, and Gomber, Mikolas, and Wacha did fairly well. Cecil continued his trend of underperforming expectations. For both Gant and Weaver, their worst pitch was the curveball. If not for how much damage he yielded on curveballs, Weaver would have had a fairly solid season.
Putting it all together
I’d like to do one more thing here, if you’ll bear with me through one last table. I’ve taken each pitcher’s results and weighted it by the percentage of results they received. In doing so, I’ve accumulated a total expected wOBA for each pitcher based on spin and velocity. The name “xwOBA” is taken, so I’m calling it “spin-velOBA.” I know it’s a weird, cumbersome name, but this is far from official. I compared their spinvelOBA to the actual wOBA they allowed and created yet another league-relative number. The lower the numbers here, the more a pitcher overperformed compared to his expected level based on spin and velocity, spread across all pitches. The higher, the more they underperformed.
The positive is that this should scrub out some of the small sample issues we had with individual pitches, and get us more overall results for some of the outliers.
wOBA vs. Expectation
It’s almost comical how sharply it divides between relievers and starters. Gant, Mikolas, Wainwright, Wacha, Carlos Martinez, Flaherty, and Daniel Poncedeleon accounted for 124 of the Cardinals’ 162 starts, and each allowed a wOBA that was 8% or lower than their expected level based on spin and velocity. Weaver and Gomber are the biggest outliers in the rotation, though Weaver was fairly close to what you would expect within our parameters.
On the other hand, Cecil, Lyons, Mayers, and Gregerson have velocity and spin that suggests they should have received much better results. Given how far Cecil, Lyons, and Gregerson fell from their 2017 results, it’s easy to see how the season spun out of control in high leverage situations. For the three that are returning, it offers a ray of hope that 2019 won’t be as disastrous.
I’d love to draw some firm conclusions here, but it would be intellectually dishonest. This has been a fun exercise and it might shed some light on these pitchers. However, there are several big puzzle pieces missing when evaluating the data above.
Location/Command: We have no idea how well these pitchers are hitting their spots with these pitches.
Life (horizontal and vertical movement): Throwing a high RPM pitch at a good velocity sound great in a vacuum, but not if the pitch is straight. A lot of ground can be made up with the life on a pitch. Granted, velocity and spin can generate horizontal and vertical movement in many cases, but it’s not guaranteed.
Tunneling and/or sequencing: Throwing different back to back pitches through the same tunnel, even if the pitch is lacking in velocity or spin, can very effectively disrupt a hitter’s timing. That’s only indirectly accounted for in the results above.
Similarly, sequencing pitches can disrupt a hitter- changing their eye level, coaxing them to back off the plate or dive out over the plate, or simply speeding up and/or slowing down their bat with off-speed pitches.
Quality of opponent: Austin Gomber’s fastball was a tick worse than league average for similar pitchers, registering a 103 wOBA+. However, we don’t know if he faced an inordinate amount of good hitters, or had to swim upstream against the platoon advantage more than normal. Ultimately, that’s true for all of these pitchers. That piece of information would help fill in some gaps for us.
Defense: There’s nothing in the data above that accounts for the quality, or lack thereof, of the defense behind each pitcher during these events.
Dumb luck: Even if we account for every other piece listed above, there will always be dumb luck that can completely wreck a pitcher’s statistic in these type of categories. It happens all the time in baseball, and this type of research is no different. Hitters get two degrees more loft on a flyball, they swing a millisecond sooner, outfielders get a bad jump and suddenly a flyball becomes a double, pitchers don’t quite get the grip they want on a curveball, a sharp slider turns into a cement mixer, and on and on. Not all of those things are “luck” in the purest sense of the word but they’re random events that can add noise to the data above.
This all leaves us with a fun thought experiment about velocity and spin, and an educational tool for which types of spin and velocity combinations yield better results. They’re building blocks to understanding where these pitchers over or underperformed. What this data doesn’t do is answer why they over or underperformed their expected wOBA. We can save that for another day.
If you’re looking for similar, much more detailed work, I recommend QOP (quality of pitch).