This is the third part of a series on how to evaluate baseball pitchers. Jordan Barhorst will pick up the series from here, connecting this stuff to the Reds.
In the first two parts, we examined metrics based on the outcome of pitched balls. Today, we’ll look at whether we can judge pitchers based on the innate qualities of individual pitches.
The pitcher’s “arsenal” consists of the attributes of the pitches he throws, such as the pitch type, pitch velocity, vertical and horizontal movement of the pitch, the ball’s spin rate and the pitch’s location in the strike zone.
The less formal word we’ve used for this is “stuff.”
In addition to judging a pitcher based on the outcome of his pitches – were they hit for home runs, did they strike out the batter – we can evaluate a pitcher by looking at his stuff.
One virtue of this approach is that a pitcher’s arsenal remains pretty constant from year to year. Spin rate and vertical movement don’t depend on fielding, bullpen or luck. It’s technology and physics.
Measuring Arsenal with Technology
Analyzing a pitcher’s arsenal isn’t new. It’s the way scouts have talked about prospects for decades. Beyond that, major league pitchers are known for pitch attributes. Nolan Ryan and Roger Clemens were linked to blazing fastballs. Pedro Martinez had a devastating change-up. Sandy Koufax and Bert Blyleven were known for curveballs. Mo Rivera had the cutter. And, of course, we obsessed over every one of Aroldis Chapman’s radar gun readings.
What is new is the precision with which we can measure stuff and then use fancy math to figure out what components of stuff matter.
PITCHf/x cameras debuted in 2006 with its revolutionary real-time data on velocity and movement. In 2017, every major league stadium transitioned to Statcast, a more advanced system that tracks the full trajectory of every pitch (and players, too) using TrackMan radar and cameras. Many minor league ballparks have TrackMan, as well.
This spring, we’ve read about Rapsodo and Edgertronic.
Rapsodo is a similar technology to TrackMan except it comes in a small, portable black-and-red box. The latest version sits between the pitcher and catcher. The software allows immediate interface with users. At one-tenth the cost of TrackMan, Rapsodo is perfect for Spring Training.
The Edgertronic camera is another technology that helps analyze pitch arsenal and has made a splash this spring. It’s a high-speed camera that comes in a little blue box. It shows in super slow motion (1,000 frames per second) how the baseball comes out of a pitcher’s hand. Edgertronic allows pitchers and coaches to analyze the pitcher’s grip.
The Reds have used high-speed cameras and portable pitch tracking devices in the past, but 2019 marks the first time they have an official, organization-wide commitment to it. The Dodgers under hitting coach Turner Ward, who is now the Reds coach, were among early adopters.
Analysts can use the data to find correlations between arsenal variables and the best results for individual pitchers. Technology helps identify optimal movement and spin combinations. Pitchers can adjust grips to get more tilt and spin.
Pitch velocity is the most widely collected and used arsenal data. Fans see MPH readings in stadiums and on TV broadcasts for every pitch. Pitchers throw harder today than ever before. It hasn’t been that long since 95-mph fastballs were rare.
Average pitch velocity has been increasing for a combination of reasons. Pitchers are bigger, stronger and use longer strides to the plate. They are trained to throw harder when they are younger. As more innings are given to hard-throwing relievers, average velocity increases.
Velocity is highly correlated with strikeout rates. Higher velocity has been a major contributing factor in record-breaking numbers of strikeout. That makes sense. A 95-mph fastball reaches home plate in 400 milliseconds. The bat swing takes 150 milliseconds. So hitters have a quarter-second to see the ball and decide whether and where to swing. As velocity increases, decision time shrinks further. Higher pitch velocity leads to more swings-and-misses and weak contact.
Strikeouts are a huge factor in pitcher success. Strikeouts prevent the ball from going into play. It’s the one outcome where hitters won’t get on base and runners don’t advance. Fly balls and ground balls can produce hits. Small increases in pitch velocity can make a big difference in a pitcher’s success. It’s a great way to evaluate pitchers.
A pitched baseball moves downward (vertical) and to the side (horizontal). The new technologies measure both. Pitchers can be graded based on their pitch movement.
Movement disrupts timing and anticipation. In terms of having an effective arsenal, vertical movement has been shown to be more important than horizontal. Vertical movement is correlated with strikeouts and weak contact. When the ball moves horizontally, it often stays on the same swing plane for the hitter.
Research has shown that movement on a fastball is substantially less important than velocity. A curveball’s effectiveness depends on both velocity and movement, depending on whether the goal is whiffs or ground balls. Velocity and drop (vertical movement) are important for sliders.
Spin Rate and Efficiency
The spin rate of a baseball affects how much it moves in the air. The pitcher’s grip, pressure on the ball and release affect the spin.
If physics makes your eyes glaze ignore this paragraph. As a baseball rotates (spins), air moves around it. The axis of tilt on the ball and the rate of spin determine the force of the air pressure. The baseball’s seams push the air and change direction of the ball. That is called the Magnus force.
Beyond that, there are types of spin. Some spin (transverse spin) makes the ball move, but a different type of spin (gyrospin) doesn’t. Gyrospin is the kind of spin a quarterback uses when throwing a spiral, or a bullet has when fired from a gun. The goal with gyrospin is to prevent vertical or horizontal movement.
Trackman measures the “spin efficiency” of every pitch, which separates transverse spin from gyrospin. In short, it measures how much “good” spin the pitch has. Matt Wilkes wrote about pitch spin rate here a couple years ago. He analyzed how spin rates affect each kind of pitch differently. All types of fastballs induce more swings-and-misses as the spin rate increases.
Spin efficiency is an important way to evaluate pitchers.
The Problem with Measuring Command
The trickiest aspect of developing an Arsenal Score is adding a component for pitcher command. Command is distinguished from control. The latter is about balls and strikes while the former is putting the ball where the pitcher wants it. Command is where in the strike zone or how close to the edge of the zone a pitch goes.
The inherent problem with measuring command is the analyst or camera has no knowledge of the pitcher’s intent. To put it simply, you can’t measure command because you don’t know where the pitcher wanted to throw the ball. There are circumstances, particularly with off-speed pitches or breaking balls, when a pitcher is deliberately throwing the ball outside the strike zone. How do you judge a pitcher when he’s trying to throw it off the edge but the ball ends up Right Down Broadway?
The challenge in estimating pitcher intent hasn’t stopped researchers from trying. Maybe the catcher’s target can serve as a proxy. Studying an individual pitcher’s pattern over a period of time might give insight into where he’s trying to throw it.
Arsenal Scores are tools that use composite measurements for an individual pitcher’s pitch portfolio. They grade each pitch on several criteria and weight the frequency at which the pitch was thrown.
Just like with ERA and FIP, there are different versions of Arsenal Scores based on what they measure. Early arsenal models (Pavlidis, 2013; Sarris 2014) looked at defense-independent outcomes of pitches – swings and misses, ground balls, popups, etc. They didn’t use box score results like singles, home runs and groundouts. Other writers (Edwards, 2017) tweaked those models by including measures of called strikes and using different weightings (Chamberlain, 2017).
Sauceda finds that ACES (Arsenal Combination Estimate Scores) produces a tight evaluation of pitchers. The correlation of an ACES score from one year to the next is 76% or seven times higher than that for ERA. The ACES score doesn’t include any batted ball outcome or run production, just actual pitch characteristics, and it better predicts ERA than ERA itself.
A couple other examples of this type of research: Mailhot (2019) used his Arsenal Score (Stuff+) this week to evaluate pitchers based on pitch types. Combining (Sonne, 2017) the two types of Arsenal Scores – outcomes plus pitch qualities – may lead to higher level of prediction.
One way for pitchers to keep hitters guessing is to make the delivery of two pitches look the same for a longer amount of time. Pitchers strive for identical release points and a common initial trajectory out of the hand. This skill requires a pitcher to control his body and repeat his delivery.
This simple concept is called pitch tunneling.
Pitch tunneling delays the hitter from recognizing the pitch. Batters use cues to help identify pitches and decide whether and where to swing. Many of those signs come from body mechanics of the pitcher, including release point. The later the pitcher can postpone the hitter’s decision, the better the outcome for the pitcher.
Studies show that successful pitch tunneling is correlated to worse plate discipline by the batter, in getting him to swing at pitches out of the zone and take pitches in the zone. Swings-and-misses, strikeouts and weak contact are also positively associated with tunneling.
Pitch tunneling statistics are another way to evaluate a pitcher based on physics, not the outcome of a batted ball.
We’ve reached the end of this part of the series. Getting from Pitcher Wins to Spin Efficiency has been an adventure. If you read all three posts, thanks for sticking with it.
Jordan takes over from here, applying the theory to Reds pitchers. But I wanted to make a few concluding remarks related to my sections.
1. There is no one way to appreciate baseball. You can love the sport without looking at a single statistic.
2. If you do check out various statistics, understanding the strengths and weaknesses of each one will help you form the most accurate opinion of the player.
3. None of the metrics are perfect. Some are considerably better than others. They’re all pretty lousy with small sample sizes.
4. Don’t rely on a single measure to evaluate a player.
5. Understand what each statistic is telling you about the player. If you know the difference between FIP and xFIP, or ERA and SIERA, then a gap between them can help you evaluate the player. But don’t average similarly scaled metrics. Mixing a bad pot of soup with a good one doesn’t taste the best.