Welcome to Sabermetrics, Scouting, and the Science of Baseball, our eighth annual charity baseball research conference. This year, we are proud to support the Angioma Alliance, an organization by and for those affected by cavernous angiomas and their loved ones, health professionals, and researchers. We hope you have a fun time, learn some new things, think differently about a problem, and meet some great people.
We used a Diamond Kinetics SwingTracker sensor in conjunction with a custom circuit wired to a pitching machine in order to experimentally determine the commit point, that is, the last moment a hitter can check his or her swing. This is interesting from a hitting and a pitching perspective. From the hitter's perspective, given the short amount of time it takes for a pitch to reach the front of the plate, it is advantageous to be able to begin swinging and stop short of a swinging strike after recognizing the pitch type and location. From the pitcher's perspective, knowing how long it takes a swing to reach the point of no return helps inform pitch tunneling conversations. In January of this year, Tom Tango (of MLBAM) worked with one of us (Joe Petrich) to estimate the commit point based on Diamond Kinetics' trigger-to-impact time data from elite hitters. This estimate is used by the Statcast 3d pitch visualization tool, that shows multiple overlayed pitches with release point, recognition point, and point of no return highlighted. The study discussed in this presentation refines that estimate based on experimentally collected data. We used an Arduino microcontroller, light gate, and LED to detect pitches exiting a pitching machine and then flash a light indicating to a hitter that they should not swing. We decided whether or not the light would turn on randomly, and introduced randomness in the timing of turning on the light in order to determine the latest time we could tell the hitter to check his swing and have him be successful. We judged a successful check swing to be one in which no part of the bat broke a plane formed by the front edge of home plate, perpendicular to the ground. The hitter was instructed to swing with maximum effort with the goal of hitting the pitch as hard as possible. Balls were fed into the pitching machine at an anticipated interval so the hitter was prepared for the pitch, but the hitter was not informed before the experiment how often the light would come on indicating he should not swing, and the hitter was not told to attempt to check his swing at the last moment. We used the acceleration data from the SwingTracker sensor to ensure the swings and reaction times we collected were at or near maximum effort. In addition to the experimentally determined commit point, theoretical calculations will be presented showing the force required to start and stop a swing, and variations in commit point based on an individual hitter's strength and swing path will be discussed.
Pitch tunneling is an exciting new concept in pitching, and is being measured with increasing accuracy. This presentation will explore pitch tunneling from the pitchers' perspective, specifically addressing the following concepts:
- The factors pitchers consider when choosing each individual pitch, including count, arsenal, confidence, a hitters' scouting report, swing mechanics, and more.
- Common pitch pairings used by pitchers with different arsenals
- When tunneling works, and when it might not make as much sense
- How pitchers anticipate the outcomes of the pitches they choose
- How visualization factors into pitch calling and how pitcher's use tunneling to target their moving pitches
- Can pitch tunneling data be integrated into a pitcher's overall strategy?
In summary, pitchers use as much information as they can gather to select the pitch that they expect will produce their desired outcome, either immediately or later in an at-bat. This presentation will share a pitcher's perspective on how pitch-calling has traditionally been done, and how new data can potentially improve pitch-calling and outcomes.
Conventional wisdom says Japanese pitchers rely heavily on breaking and off-speed pitches, compared to their stateside counterparts. Is it true? If so, to what degree? Some of the best analytical minds in Japan put the matter under the microscope: the differences between NPB and MLB in general breaking and off-speed percentage, in usage and in specific ball-strike counts. We'll also take a look at why Japanese pitchers tend to have deeper arsenal of pitches, and the cultural and developmental backgrounds behind all that.
Michael Fishman, current Yankees AGM and former Director of Quantitative Analysis, will answer questions about baseball analysis in the big apple.
Conventional batting metrics report outcomes, not player contributions. With BP's Deserved Runs framework, we introduce a new statistic for batters, Deserved Runs Created (DRC+), which focuses on a batter's most likely contribution to run-scoring. By combining multilevel modeling with a novel multinomial structure, DRC+ provides substantially improved accuracy over existing publicly-available metrics, particularly in volatile run environments.
How can we leverage tracking data to improve performance? Here, we report the approach that we have undertaken in Japan, with the goal of using tracking data to improve pitching ability. We also introduce our new system, "BACS", that is currently being used in Japanese baseball.
Current Red Sox Vice President Pitching Development/Assistant Pitching Coach and former Collegiate Second Baseman Brian Bannister will answer your questions.
With baseball pitching, a large number of injuries and lost time is attributed to upper extremity injuries such as impingement, rotator cuff injury, shoulder instability, labral tears, and elbow ulnar collateral ligament (UCL) injury. During pitching, repetitive throwing and large amounts of torque on the shoulder cause microtrauma. There has been tremendous focus on evaluating and addressing intrinsic risk factors of shoulder and elbow injury such as range of motion, muscle imbalances, and scapular kinematics; however, extrinsic risk factors are often not addressed by healthcare professionals. Extrinsic risk factors such as cumulative throwing load, rest and recovery, pitching mechanics, modality use, nutrition, and sleep may play a larger role on development of injury risk than the intrinsic risk factors that healthcare professionals have traditionally been focusing on. Therefore, the purpose of this session is to evaluate the influence of extrinsic risk factors on shoulder and elbow injury risk and evaluate current evidence on effectively managing these risk factors to decrease shoulder and elbow injuries in baseball pitchers.
I will start with a brief presentation exploring whether the strike zone is measurably different at various minor league levels. I used a sample of 20,000 minor league pitches from an organization that wishes to remain anonymous. I built regression and classification models for each level to predict both new pitches as well as test sets from within the data. The results give a window into strike-calling tendencies at different rungs of the ladder. This will be followed with open Q&A about professional and international baseball.
Eye-Tracking - the Brain-Eye connection. How attention and fatigue can be objectively measured through the eyes. How can it predict level of play? Can we train eye-tracking to be a better player? We know that eye-tracking is critical for any ball sport and especially baseball. However, the impact is far greater than just being able to hit the ball. Based on $30M dollars and years of D.O.D. research and testing scientists have established that eye-tracking can effectively and objectively measure attention and fatigue as well as help injury. Is a player ready to perform at their best? Eye-tracking can tell us. The eyes truly are the window into the brain and perhaps the best way to be able to predict performance in sports as well as life.
The ulnar collateral ligament (UCL) is the most commonly injured ligament in overhand throwing athletes. UCL injuries have been, historically, treated reactively. Over 14 years, while coaching catchers, infielders, and outfielders, I have developed a simple way to correct throwing mechanics that can both allow for healing of the UCL, as well as prevent UCL injuries in overhand throwing athletes. Every athlete that has incorporated these mechanical adjustments experienced reduced pain. Some athletes adopted the modifications permanently, and were pain-free for the remainder of their career. By teaching athletes these modifications, unsafe throwing mechanics can be corrected before the cumulative stress and damage to the ligament is diagnosed with an overuse injury. My innovative method for addressing UCL injuries can be taught to any coach or medical professional working with overhand throwing athletes.
TrackMan's phased-array Doppler radar has recently replaced PITCHf/x as MLB's primary pitch-tracking system. Beyond recovering estimates of 3-D pitch trajectories, the TrackMan radar also provides a measurement of the magnitude of a pitch's total spin. This additional information provides an unprecedented opportunity to characterize pitchers and their pitches. Using TrackMan data extracted from BaseballSavant and MLB Gameday along with weather data from wunderground, we follow analysis from Alan Nathan to estimate the transverse spin vector, the gyro spin vector, and the 3-D spin axis for each MLB pitch tracked in 2017. These parameters provide a new set of dimensions along which we characterize and compare pitchers and their pitches after accounting for the correlations and variances among the various variables. Since the large majority of the measured total spin for fastballs is due to transverse spin, we also use this data to study the relationship between the lift force which determines pitch movement and transverse spin. This relationship has been previously modeled using smaller data sets obtained using laboratory measurements. The new parameters can also be used to improve pitch classification accuracy and to analyze the effect of altitude and weather on pitch characteristics.
This study seeks to find the true value of what is termed as a 'secondary pitch' when a pitcher possesses a 65 to 70+ grade fastball. The belief is that this will lead to a more successful and efficient pitch repertoire for these pitchers. We have always been taught that we need a secondary pitch but the question is how dispensable is the pitch if it does not attain a level comparable or equal to the primary pitch. There has been distinct interest in the baseball community for the development of methods by which acquire a consistent competitive advantage this may be an option to be explored for teams that have compiled power arms that may have control issues.
Major League Baseball recently released its official report investigating the source of last season’s Home Run Surge. The independent committee concluded that the cause was a decrease in the ball’s drag coefficient, but were unable find the reason for this change. After taking apart and examining baseballs from 2014 and 2016/2017, I found a single statistically-significant difference—specifically, that the laces on the more recent balls were 9% thicker. Since thicker laces have greater tensile strength, they may be keeping the ball more spherically symmetric, thus leading to lower drag coefficients. Increased lace thickness is also a likely candidate for causing the recent spike in pitcher blisters. Thicker laces will produce prouder stitches, creating a “bumpier” seam with a rougher texture. Since blisters are often associated with tightly gripping or rubbing the seams, this change could be a strong factor in higher injury rates. I discuss the potential impact of this small and unexpected change on the game going forward, with suggestions for follow-up studies and ways to mitigate current and future problems.
Analysis of ejections during the Statcast era (2015-7) shows that non-white players are ejected at rates disproportionate to their representation in the league. Major League Baseball has become increasingly non-white over the past decade, with non-white players comprising about 42.5% of MLB players in 2017. Using data from the Umpire Ejection Fantasy League, analysis found that non-white players have been ejected at rates disproportionate to league representation over the past three seasons, particularly 2015 and 2016. 2017 ejections of non-white players were roughly proportionate to league representation; it's unclear whether this decrease in disproportionate ejection represents a change in umpire practices, a chance effect, or something else. Controlling for ejection of players for 'nonviolent' offenses - removing fighting, unsportsmanlike conduct, and throwing at as reasons for ejection - did not alter this pattern of disproportionate ejection. This pattern holds when considering ejections by the home plate umpire of position players, with the majority of ejections for both white and non-white players as ejections for arguing ball vs. strike calls.
Annually, players from across the country whether they play NCAA, NAIA or Junior College baseball are shipped to various locations around the United States to play in the summer. For many, this is a chance to gain exposure against top talent in summer collegiate leagues such as The Cape Cod League, The Northwoods League and The Valley League just to name a few. The tension for many coaches, who are responsible for players from different colleges, is how to blend winning baseball games and managing player health. Pitching staffs for summer leagues are often met with the hurdle of strategizing how to utilize their best pitchers while being under the constraints of pitch counts and throwing limits. This study specifically will look at the 2018 Charlottesville TomSox pitching staff ,in the Valley Baseball League, and how they will use sabermetrics, technology such as Motus, Rapsodo, and Diamond Kinetics as well as a multi-camera system to better manage their players. Specifically, the TomSox will gather workload data on each individual pitcher to quantify the amount of stress each individual pitcher endures over the course of a season to better manage each pitchers' arm health. This presentation will look to present on the success of combining workload information from Motus Baseball and sabermetrics over the course of a college summer baseball season to minimize arm injuries while competing for a championship and enhancing player development.
This research seeks to track and analyze how arm slot, spin axis, release point (hand/fingers), among other things correlate with baseball movement. Data was collected during a Pitcher's Development and Design Program for college-level pitchers. Pitchers wore a Motus Sleeve and throwing in front of a Rapsodo camera while the following pitches were evaluated: 4-seam fastball, 2-seam fastball, change-up, slider, slurve and curveballs. Arm slots were measured with a Motus sleeve. Spin axis and horizontal and vertical movement data were measured using a Rapsodo camera in order to find the relationship that exists between the amount of movement a pitcher has in correlation to their arm slot, release point and finger grip.
The purpose of this paper is to analyze the the economic impact of the 2017 Major League Baseball Collective Bargaining Agreement (CBA) on the market for international amateurs. The Previous CBA covered years 2012-2016 and instituted slot values for the domestic draft and international draft pools. The 2012 CBA allotted teams between $2 million to $5.5 million to spend on international amateurs per year before incurring a penalty. These industry allotments ranged from $78 to $87 million. Penalties include paying 75% up to 100% on every dollar spent above a team's bonus pool and additional maximum deals for following years. The richest contract was given to Yoan Moncada in 2015 by the Boston Red Sox at $31.5 million; the total acquisition cost was $63 million for the team. The 2017 CBA increases the total industry allotment to $153 million, but also instituted a hard-cap on international spending. In this way the industry ensures smaller signing bonuses for players of increasing talent levels during the next five years. The model looks at the top thirty richest international amateur contracts signed per period from 2012-2017, including the first year under the new labor agreement. In this way we can compare the signing bonuses earned under the 2012 CBA and how those changed in the 2017-2018 signing period with the hard cap on bonus pools.
Every at bat is a battle between the batter and pitcher, and every ball or strike has the power to swing an at bat in one or the other?s favor. In this study, we follow up on Dan Meyer's 2015 The Hardball Times article, 'Dynamic Run Value of Throwing a Strike (Instead of a Ball),' and look to find if the dynamics of a count have changed in the 'Juiced Ball Era.' With drastic changes in how the game is played from higher home run and strikeout rates to shifting being more common, it is fair to wonder how run values by count have been affected. This will be our primary analysis. Using Statcast data from Baseball Savant, we will find the wOBAs of each count in the juiced ball seasons (the second half of 2015 through the first half of 2018). We will then calculate run values based on those wOBAs to see how a ball or a strike affects the batter's success in at bats containing that count. Additionally, with the introduction of Statcast, we have new data with which to see how the game is changing on a pitch-by-pitch level.
We live in an increasingly data-driven world where evidence-based decision making is being used in medicine, public policy, and other areas. This talk tells the story of one case of evidence-based medicine. We will examine the evidence, see how it was presented, and consider how well the evidence was understood by those who presented it and those who had to use it to make an important decision.