Launch Angles, Still Rising

The concept of launch angle has successfully embedded itself in the minds and lexicons of baseball players, coaches, and fans over the past decade. Several high-profile players have acted as ambassadors to the so-called fly ball revolution along the way: from J.D. Martinez to Justin Turner to Mookie Betts, brand-name players have popularized the metric. There are more recently some late adopters too, as Jake Malihot of FanGraphs recently pointed out in the case of Eric Hosmer. Given this shift, what follows glimpses the season over season transitions in qualifying player launch angles over the first five seasons (2015-2019) that such metrics have been available via Statcast. It also takes a brief look at how things are shaping up to this point in 2020.

First though, it makes sense to provide some context for why launch angle adjustments have swept through baseball. In part at least, launch angles garnered attention only as soon as they could be reliably tracked. In their book “The MVP Machine,” Ben Lindbergh and Travis Sawchik highlight both the newfound availability of this metric along with the compelling initial results (italics added here):

“The fly-ball revolution also spread thanks to Statcast, which in 2015 began to measure the launch angle and exit velocity of almost every ball in every MLB game. Teams started to outfit their batting cages at all levels with swing- and ball-tracking technology. The data confirmed that air balls were better than grounders: with every 10 degree increment from minus 30 degrees to 30 degrees (where zero is a line drive), the league-wide wOBA on contact in 2018 increased, easily surpassing the all-angles average of .315. (The average home run left the bat at a vertical angle of 28.2 degrees.)“

To visualize season over season launch angle developments, I have made use of mirror histograms. These histograms illustrate the distribution of qualifying hitter average launch angles for one season in the positive (above the x-axis) space on the histogram and the following season’s distribution in the negative space (it should be noted that those percentages of course are not negative, although they occupy space below 0 on the y-axis). The mirror histogram below juxtaposes the 2015 and 2016 seasons, respectively.

There is a clear offset across seasonal histograms above. The 2016 season, in orange, saw a higher maximum, minimum, average, median, and modal average launch angles. In 2015, the first season of Statcast, nearly one-fifth (19.84%) of qualifying hitters had average launch angles below 7.5 degrees. One season later, just 10.27% did.

The next three visualizations feature the same views but with 2016 vs. 2017 data, 2017 vs. 2018 data, and 2018 vs. 2019 data, respectively.

This final mirror histogram juxtaposes the initial 2015 season data to the most recent full season, 2019. The shift is stark. Additionally, the spread of launch angles is more recently less normal and features less dispersion. Anecdotally, the difference between those two seasons seems to be a difference in information. What was a hitherto untracked statistic became mainstream across this five-year stretch. Launch angles, once more or less unknown, were correspondingly randomly and normally distributed; now, that distribution has been distorted by concerted intent. The limits of effective launch angles are being pushed, ground-ball hitters have either adjusted or disappeared, and players have flocked to the (new) center of the distribution.

Finally, below is a quick chart of summary statistics of Statcast average launch angle data (for qualifying hitters) since its creation.

2020, to this point, represents more of the same trend. The samples are small, the confidence intervals are huge, and regression to the mean will surely begin to pinch those numbers, but average launch angles are up again despite those factors. Boundaries are being pushed too: Matt Chapman, Max Kepler, and Joey Gallo each own average launch angles over 26 degrees while Rhys Hoskins is heading toward consecutive seasons with an average launch angle greater than 23 degrees. Lindbergh and Sawchik cited the average home run launch angle of 28.2 degrees; in this era of the home run, it doesn’t take much effort to speculate where some of these players are headed.