PHOENIX—The pitcher initiates action in a baseball game, and the consequence of his act—how he throw the ball, the predictive nature of his biometrics—became the focal point for two research presentations during Day Two of the SABR Analytics Conference.
An Introduction To In-Game Biomechanics Data
Sportvision furnishes the public with Pitch f/x data that captures the velocity and movement of every pitch thrown in the major leagues since 2007. The company’s Hit f/x, Command f/x and Field f/x technologies paint a picture—for major league organizations who pay for the service, anyway—of batted balls, pitcher precision and a defender’s range.
In his research presentation, Graham Goldbeck, the manager of data analytics and operations at Sportvision, focused on the company’s data-gathering capabilities as diagnostic tools for pitchers, with similar potential for batters.
Sportvision’s cameras measure four possibly instructive pieces of biomechanical information every time a pitcher throws the ball to his catcher, the location of his: plant foot, shoulder at release, elbow at release and hand at release. The company measures similar data for batters, including hands at contact point, tip of the bat in the hitting plane and location of the back and front feet.
Owing to the limited scope of the data presented for this study—all pitchers who pitched at Oakland’s O.co Coliseum as well as Arizona Fall League games at Mesa in 2013—only narrow conclusions could be drawn. The data gathering process is ongoing, but Goldbeck shared Sportvision’s initial findings for the 100-plus pitchers and few dozen batters on record.
Sportvision camera systems determined that the typical stride range for pitchers—the distance between the rubber and plant foot—was between five and seven feet, with a mean of 6 feet. Goldbeck said that no clear relation existed between a pitcher’s height and his stride length. In other words, some tall pitchers take a short stride, and some short pitchers take a long stride.
At release point, the average pitcher’s shoulder measures at 30 degrees, using the sidearm (zero degrees) and straight overhand (90 degrees) release points as axes. In scouting parlance, 30 degrees would be equivalent to the natural three-quarters arm slot used by most major league pitchers. Goldbeck pointed out that many pitchers raise their arm slot in order to impart maximum downward break on their curveballs, so the notion is not just a stray observation from pitching coaches.
Sportvision determined that the values for the elbow at release were somewhat similar to the corresponding shoulder at release values. This means that the elbow says little more than the shoulder can say, and that the elbow at release is not a telling measure by itself.
The typical big league pitcher extends his hand, at the point of release, about 5 feet, 8 inches from the rubber. For most pitchers, this amounts to a few inches behind the location of his plant foot.
Goldbeck would not reveal how Sportvision cameras isolate the lead foot, the shoulder, the elbow or the hand of major league pitchers during their deliveries, but he indicated that the data could have implications for front offices or pitching coaches interested in tracking biomechanical data for pitchers in game situations, as opposed to the artificial, laboratory setting that requires pitchers to wear an array of sensors or even an entire sensor suit. In the future, a pitching coach armed with this technology might be able to consult his iPad from the dugout to confirm what his eyes (or the catcher’s) see: that the pitcher is dropping his elbow, varying his stride length or missing his ideal plant-foot location.
Furthermore, clubs that value pitchers with a particular stride length, plant-foot location, arm angle or hand extension could limit the pool of candidates quickly with access to the data. This could take a lot of the guesswork out of acquiring pitchers, while also having potential as a diagnostic tool for their own pitchers.
Goldbeck expanded upon last year’s Analytics Conference presentation about contact point for batters, showing off the capabilities of Hit f/x technology to pinpoint the contact point along the barrel. In other words, the Sportvision cameras may be able to accurately measure the frequency with which a player finds the bat’s sweet spot.
Similarly, a batter’s path through the hitting zone could reveal whether he’s consistently being tricked by pitch types or whether he’s swinging late and simply not recognizing the ball out of the pitcher’s hand. Ultimately, Goldbeck foresees a time when bat path, bat speed and the bat’s time in the hitting zone can be quantified and expressed with similar precision as Pitch f/x data today.
Until that day arrives, Goldbeck said, Sportvision plans to partner with Major League Baseball Advanced Media on the overarching tracking technology it plans to implement beginning at three major league parks this season. He said that Sportvision’s ultimate goal is to have tracking technology at all levels of the game, from amateur ball to pro ball down to traveling teams of 6-year-olds.
Using Pitch f/x Data To Help Determine Pitcher Injury Risk
Finding an island in the choppy seas of big data, Jeff Zimmerman of FanGraphs.com isolated three trends in Pitch f/x readings that tend to indicate when a pitcher is injured or will soon break down. As with all things baseball, exceptions to the rule exist, but his Pitcher Abuse Index, or PAIN, identified numerous pitchers in 2013 who would spend time on the disabled list or, as it turns out, would pitch through pain.
The two attributes most predictive of injury in the long term, Zimmerman found, were changes in velocity or in percentage of pitches thrown in the strike zone (Zone%), with the latter carrying more weight over a stretch of time. In fact, any pitcher throwing fewer than 52 percent of his pitches for strikes stands a good chance of going on the DL. Sustained velocity loss, likewise, not only tends to result in decreased performance but also might signify depleted health.
Zimmerman singled out Rangers righthander Yu Darvish and Reds rookies lefty Tony Cingrani as two pitchers who threw fewer strikes and threw with less velocity last September. Only after the season did they reveal they were pitching through injuries: nerve inflammation in the lower back for Darvish and a lower back injury for Cingrani. Both pitchers veered dangerously into the red on the PAIN index during the final month of the season.
Short-term indicators for potential pitcher injury are familiar to pitching coaches everywhere: decreased in-game velocity and inconsistent in-game release point (save for those who intentionally throw from different angles, such as Diamondbacks righthander Bronson Arroyo or Royals lefty Bruce Chen). Zimmerman cited Rangers righty Alexi Ogando (who made three DL trips in 2013) and White Sox righty Felipe Paulino (who made more rehab starts than big league ones in 2012-13) as two pitchers who showed fluctuating velocities and release points just prior to an injury stint.
One notable example of a pitcher who evaded the PAIN index: Mets righty Matt Harvey, who showed no velocity loss, no lapse in control and no difference in release point prior to his elbow injury last summer, one that required offseason Tommy John surgery.
Zimmerman spelled out a few attributes that apply to many injured pitchers: 1) In general, the hardest throwers tend to get hurt most often, 2) Pitchers who throw breaking balls more than 30 percent of the time tend to get hurt most often, and 3) Scariest of all, Zimmerman found that about 40 percent of regular starters and 20 percent of regular relievers in any one year will spend time on the DL the following season.