What hard–hit percentage and average launch angle ranges are common among MLB hitters with high barrel percentages?

I have always enjoyed baseball, and I like looking up players’ statistics. While exploring stats, I learned that the barrel metric is very important because a barrel is defined as a batted ball that typically produces at least a .500 batting average and a 1.500 slugging percentage based on its exit velocity and launch angle.

Now it is the offseason, I think the barrel metric would also play an important role in evaluating hitters. Instead of looking only at barrel percentage, I wanted to analyze the two components that barrels are based on: hard–hit percentage and average launch angle. I wanted to find out the typical hard–hit percentage and average launch angle ranges for hitters who generate a lot of barrels, and that is why I chose this research question.

The data was collected from baseballsavant then cleaned for data analysis by Kaggle uploader. The dataset contains MLB batting exit velocity metrics and the players from 2015 to 2022. Minor league players and hitters with very small sample sizes are not included. The limitations are the small sample hit count players because they may have unstable statistics. Also, MLB Statcast accuracy varies slightly by stadium and angle. Barrels definition is standardized but not a perfect representation of quality, so this data can also be biased. So, I will not say that one thing causes another. I will only say that they are related or show a pattern.

First, I removed players who had fewer than 100 batted–ball events to avoid unreliable results. Then, I created a new dataframe using only the variables I needed.

To answer my research question, I created a heatmap using all three variables: hard–hit percentage, average launch angle, and barrel percentage. I wanted to see, at a glance, how average launch angle and hard–hit percentage each relate to barrel percentage, so I create a 2D heatmap.

And I faced a problem here. Although a typical barrel is defined as a batted ball with an launch angle between about 26–30 degrees, my heatmap did not show the highest values in that range. To understand why, I also created scatter plots that showed the relationship between barrel percentage and each variable (hard–hit percentage and average launch angle) separately.

In the hard–hit percentage vs. barrel% scatter plot, the points formed a diagonal line. This means that hitters with higher hard–hit percentages generally produce more barrels.

However, the average launch angle vs. barrel% scatter plot did not show a clear straight-line pattern. Average launch angles ranged widely, starting from around –5 degrees and going much higher, but most of the data points were grouped between about 5 and 20 degrees. The barrel percentages in this range were also very similar.

From this, I learned that hitters have many different launch angles, and their average launch angle usually falls between 10 and 20 degrees because a barrel happens only at the ideal moment. Hitters who average around 10–20 degrees tend to have a balanced mix of different batted-ball angles, which helps them generate more barrels overall.

As a result, I think that when teams evaluate hitters, it may be useful not only to look at barrel percentage but also to consider hitters with an average hard–hit percentage above 50% and an average launch angle of around 10–15 degrees.