Count-Dependent Pitch Quality
In the first part of the series, we took a stab at the what (which pitches), when (which counts), why (quality-command tradeoff), and how (velocity, movement) behind pitchers changing their pitch profiles depending on the count of an at-bat. There were notable velocity and movement differences across the league depending on if the pitcher was ahead, even, or behind in the count. This piece will utilize Stuff+ to investigate how pitchers manipulate the quality of their pitches depending on the count, count-specific pitch quality behavior, as well as list pitchers who stood out amongst the rest of the league.
To review, the motivation behind manipulating pitch profiles with respect to the count can be explained by a quality-command tradeoff. While ahead, pitchers are often looking to exploit their leverage in the count by upping their pitch quality via some combination of velocity and movement to achieve a called strike or swing and miss while the penalty for a ball is low.
Simply put, they throw their nastiest stuff when the reward for a whiff is high and the consequence of a ball is low. On the contrary, when behind in the count, the penalty for throwing a ball is higher, therefore, command normally takes a slight precedence over pitch quality. In this instance, some pitchers may sacrifice pitch quality, within reason, in favor of locating the baseball to get back in the count.
Pitcher Behavior
For reference: “Ahead-Behind Stuff+ Diff” is the average stuff+ in ahead counts minus the average stuff+ in behind counts. A positive value indicates a higher stuff+ in ahead counts, a negative value indicates a lower stuff+ in ahead counts. In this instance, a difference of 5 equals a 5% increase in pitch quality in ahead counts
By bucketing counts into ahead, even, and behind by pitcher on a league-wide scale, we can see pitcher behavior of pitch types in each count. Amongst all pitch types, pitchers displayed a 6.5 and 4.8 bump in stuff+ while ahead compared to behind and even, respectively. This backs the initial theory that pitchers get nastier while they’re ahead, specifically with two strikes, on purpose. On the flipside, there is little difference in pitch quality between even count and behind count pitches.
We can also see the influence that the count weighs on the pitch quality of individual pitch types. The quality of curveballs and sliders were most improved while ahead, while changeups were largely unaffected by count. In theory, the “get-me-over” and “put-away” labels normally apply to breaking balls and, as we saw in the previous blog, breaking balls tend to have more intentional variation in velocity and movement than other pitch types. As a result, it makes sense that the quality of sliders and curveballs varies more drastically than other pitches.
Count Specific Behavior
Stuff+ is our internal metric for pitch quality, where league average is a score of 100.
For example: a stuff+ of 109 means the pitch is 9% better than league average.
Digging deeper, when we break down each pitch type by their respective counts, we observe that there is a very clear trend of increased pitch quality transitioning from hitter friendly to pitcher friendly counts.
For example, we see a major bump in pitch quality in two-strike counts when pitchers are looking to punch tickets, likely for two reasons. First, as shown in Table 1, pitchers consciously improve the quality of their offerings when they get ahead. Second, pitchers also throw their nastiest pitches with two strikes, tossing their best pitch type when they need a whiff.
We can see this effect quite clearly when looking at 0-0 CBs, where hitters see more “get-me-over” pitches with an average stuff+ for curveballs of 89.1, about 11 percent below league average in terms of pitch quality. However, in 0-2 counts, hitters see CBs with an average stuff+ of 116, 16 percent above league average. The difference in stuff+ between the 0-2 and 0-0 curveball is just about 27%, a very substantial change in the quality of pitch, which represents improvements in velocity and movement.
Across all pitch types, pitchers will throw roughly league average pitches in all counts excluding two-strike counts and 3-0. The 3-0 obviously sees a massive drop in pitch quality, as the sole purpose of the 3-0 count is to throw a strike knowing most of the time (not all the time) the batter isn’t going to swing. In two-strike counts, all pitch types are thrown with an average stuff+ of 107-108.6, the highest of any other counts, and for good reason.
This gives us a general scope of league wide pitch quality behavior. Due to the volatility of pitch profiles, some pitchers stood out from the league by how they manipulated their pitches. The next section will identify outliers in the league, as well as in-depth examples of certain pitchers.
League Leaders and Outliers
Listed below are the MLB outliers in changing the velocity, movement, and quality of their pitches. For simplicity, only ahead-behind count differences are included.
This table represents the pitches that have the largest differences in profile or pitch quality in ahead counts compared to behind. Within the outliers, some of them were rather interesting. Seth Lugo’s ahead-count fastball (150 stuff+) is nearly 50 percent better in terms of raw stuff compared to his behind-count fastball (104 stuff+), mostly due to a velocity difference of 2.3 mph. Yu Darvish throws his cutter almost 5mph harder with a gain of 4.7in of vertical in ahead counts versus behind. Cam Hill’s slider gains an additional 4.8 inches of sweep.
Tyler Clippard’s changeup sticks out because he throws a traditional changeup in even/behind counts and a split-grip variation in ahead counts. As a result, Clippard’s changeups have the greatest velocity and movement differences in the league due to a massive difference in spin direction and spin rate.
More on them in detail below.
Patrick Corbin, Slider
73 MPH SLIDER: https://baseballsavant.mlb.com/sporty-videos?playId=8a43c6dd-2829-4980-bf82-da4a9108a262
83 MPH SLIDER: https://baseballsavant.mlb.com/sporty-videos?playId=890af37f-3f38-4e50-8a73-2a2685f5a520
Patrick Corbin’s slider has been one of the most successful and consistently effective sliders in recent years and has played a major role in his ability to be a successful big league starter. Looking at his splits by count, his slider movement from 2020 doesn’t change depending on the count, however it is possible it changes due to other factors (batter handedness, scouting reports, hitter swing feedback). However, in the table above, we can see that he throws the pitch substantially harder depending on the count. It is rare for a pitcher to add velocity to a breaking ball without sacrificing some spin efficiency in the process.
On average, Corbin throws his slider 2.6 mph harder in ahead counts than behind. For context, sliders averaged a bump of 0.57 mph in ahead counts compared to behind in the 2020 season. Corbin threw his slider 4.5-times harder in ahead counts compared to behind relative to the behavior of the league, which results in an increase of 12.7 stuff+ while ahead and a 9.9 bump in stuff+ even counts compared to his behind count profile.
Our model has had difficulty reflecting the effectiveness of Corbin’s slider, as it grades well-below MLB average across all counts in terms of pitch quality despite his continued success with the pitch. It is likely Corbin is able to achieve impressive results from a rather pedestrian profile by exceptional command, pitch tunneling, and his ability to manipulate the speed of the pitch to disrupt hitter timing in various counts.
Yu Darvish, Cutter
83 MPH CUTTER: https://baseballsavant.mlb.com/sporty-videos?playId=163b6d19-d478-41ae-80d8-1a3061a01e8b
91 MPH CUTTER: https://baseballsavant.mlb.com/sporty-videos?playId=51e57568-5112-454d-857f-62931278120c
With a confirmed 10 different pitches, there’s little room on a break chart that Darvish doesn’t occupy and that’s clearly evident by watching him pitch. Darvish has stated he possesses a “slow cutter” and a “cut fastball,” which are both labeled as cutters, but not the same pitch in terms of profile and how he uses them. His cutter is one of the largest count-dependent profiles in all of baseball. Where some pitchers may alter some aspect of the pitch, Darvish significantly changes this pitch across the board for velocity, shape, and movement. For velocity, his cutter comes in, on average, 4.3 mph harder in ahead counts compared to behind. That manipulation of intent also brings a change in shape. His profile has a massive shift in spin axis from 98 degrees/9:16 (slider shape) to 145 degrees/10:50 (cutter shape), trading sweep and depth for lift and velocity in ahead counts.
It can be assumed the slower version is primarily for getting back in counts, stealing strikes, and locating with an easier profile to control. The ahead count cut-fastball is thrown with maximum intent to produce a swing and miss. While both types are above league average, his ahead count, harder cutter has a stuff+ of 135, while the behind-count, slower cutter is 120, so his ahead count cutter is 15 percent better in terms of raw stuff compared to the cutter he throws behind in the count. This is likely a compromise he makes in order to improve accuracy, as the importance of a behind count pitch is throwing a strike, rather than getting a hitter to chase.
In a recent interview with Rob Freidman (@PitchingNinja on twitter), Darvish explains his cutter grip, as well as how he uses the placement of his thumb to manipulate the speed and movement of the pitch.
Tyler Glasnow, Curveball
78 MPH CURVEBALL: https://baseballsavant.mlb.com/sporty-videos?playId=40a11d01-d811-4eec-b126-4dfebfa26016
85 MPH CURVEBALL: https://baseballsavant.mlb.com/sporty-videos?playId=c35d8225-9bcd-45c3-b1ca-e259fd84e0e6
Tyler Glasnow’s curveball is one of the best pitches in all of baseball. His 12-6 has an average spin rate just beneath 3,000 rpms, is thrown very hard, and has plus movement. When we break down his curveball profile by count, he saw one of the biggest pitch quality differences in the league, but was just beneath the minimum sample required for behind-count pitches to be included amongst the league outliers. The theory here is simple: he throws his curveball harder when he is ahead in the count, specifically, 2.1 mph harder, which results in his stuff+ skyrocketing from 140.59 while behind, to 205.70 while ahead, a difference of 65.1.
Sergio Romo, Slider
HARD/SMALLER SLIDER: https://baseballsavant.mlb.com/sporty-videos?playId=6f735a6e-9dcf-4609-8104-f4670bf4c8ff
SLOW/BIGGER SLIDER: https://baseballsavant.mlb.com/sporty-videos?playId=be7f2dba-b950-408a-bdcb-d0ee0fe5296b
Mentioned in the previous blog, Sergio Romo owes the vast majority of his career success to his ability to throw a slider. Contrary to the norm, Romo throws his slider more than any other pitch in his arsenal, using it 64.6 percent of the time in 2020, good for the highest mark of his career. Since 2008, Sergio Romo has a .225 wOBA against his slider.
As you can see from the break chart above, his slider takes on many different shapes, due to Romo intentionally manipulating the shape and speed of the pitch. It is likely that Romo manipulates his slider based on a variety of conditions such as scouting reports, batter swing feedback, or batter handedness, but we also see the count of the at-bat plays a role as well.
Romo’s slider behavior is actually counterintuitive in that it decreased in quality last season when he was in advantageous counts, largely due to an increase in velocity while behind. In ahead counts, Romo shows an increase of both spin and spin efficiency, which results in a 4-5 inch increase in sweep in ahead counts. Where the trend from the rest of the league has been increasing velocity when ahead, it looks like Romo is going for more sweep when he is looking for a punchout.
In the interview below for Jomboy Media, Sergio talks about his mindset behind the slider when he struck out Miguel Cabrera for the last out in the 2012 World Series. While discussing his sequencing, he talks about his 0-0 slider he likes to throw “to get a strike if I’m trying to get ahead or if a guy hasn’t faced me, different look.” When he gets ahead on Cabrera, he says “I gotta throw him my best strikeout one right here.” While his pitch quality technically decreased when he got ahead, the bigger movement profile may be better suited for whiffs rather than raw run values, giving Romo a very specific slider for a very specific circumstance.
Cam Hill, Slider
SMALLER SLIDER: https://baseballsavant.mlb.com/sporty-videos?playId=1b342fc2-5210-4daa-a7e1-5c21b2984447
BIGGER SLIDER: https://baseballsavant.mlb.com/sporty-videos?playId=1b342fc2-5210-4daa-a7e1-5c21b2984447
Hill’s slider led the league with the largest difference in horizontal break in ahead counts compared to behind. This is evident qualitatively by the width of slider movement on his break chart. When we look at his profiles, his slider gained an extra 4.8 inches of sweep in ahead counts due to a significant increase in spin efficiency from 44.6 percent while behind to 64.8 perfect while ahead. This change resulted in a jump of 20 stuff+ for his ahead count slider. During his debut season in 2020, Hill’s slider proved to be a major weapon for him with a wOBA of .168 and a whiff% of 43.3 percent. While it may not be directly tied to the success of Hill’s slider, it is possible his ability to produce greater gloveside movement contributed to the effectiveness of the pitch in ahead counts, especially with two strikes, as well as his ability to flash it for strikes when necessary.
Tyler Clippard, Changeup
CIRCLE CHANGEUP: https://baseballsavant.mlb.com/sporty-videos?playId=55608537-2fee-4383-84ea-8752bc80877b
SPLIT CHANGE: https://baseballsavant.mlb.com/sporty-videos?playId=4e0a5ffb-f73e-4d81-a519-647abba24475
The profile differences for Tyler Clippard’s changeup nearly jump off the page as he led the league in multiple categories. Across the board, the pitch is considerably different in velocity, spin, shape movement, and pitch quality. His ahead count changeup is 2.6mph harder while losing ~500 rpms, 10.6 inches of run, and 12 inches of lift. On paper, you would assume they are two entirely different pitches… and that’s because they are.
Upon viewing in-game video of his changeup during ahead counts and behind, it is clear he throws a split-finger variation strictly while ahead, and a circle changeup during even and behind counts. This explains the increase in velocity and massive drop in movement seen, as the split-finger grip kills spin and can result in a drop in spin efficiency.
The mindset of the split-finger variation while ahead for Clippard is likely to achieve whiffs with something harder with a more north-south movement profile than the standard changeup. Normally, pitchers make these changes to improve their offerings in two strike counts, while Clippard counterintuitively decreases his stuff+ by 30-40 compared to his changeup in other counts. Although the quality of the pitch decreases, it is possible the north-south, split-grip changeup is better suited to miss barrels in two-strike counts than the horizontal fashion of the standard changeup.
Player Development
Understanding how pitchers change their pitches relative to count states is potentially important and can be a strategy in player development. While it may not be wise to recommend every pitcher should have two shapes of every pitch or throw pitches differently depending on the count, there are certain scenarios where manipulating the velocity or shape of the pitch can help improve an athlete’s effectiveness on the bump. However, this likely takes a significant period of trial and error with finding something that works and actually accomplishes what we want. A few points to consider:
- What is the risk/reward of the change?
- In what situation are we changing it?
- What is the current quality of the pitch we’re changing?
- Is it repeatable?
In a perfect world, a pitcher locates their best stuff and maximizes pitch quality in any count. In reality, there’s a reason these trends exist at the highest level of the game. The game of baseball evolves over time and for good reason. Successful trends and strategies stick, negative ones fade away. If there were no underlying benefits of manipulating pitch types, it is likely it would not have started occurring or continue to occur in the game today at the highest level.
By Connor Hinchliffe
Comment section