Monthly Archives: April, 2020

Increase in Home Runs Due to Batter Behavior

Introduction to the Home Run Story

If you live on this planet, you are probably familiar with the home run story in Major League Baseball. There has been a remarkable increase in home run in the last few seasons topped by a record number (6776) of home runs hit in the 2019 season. Although the increase has been pretty obvious, the reasons for this increase are less obvious. In particular, many prominent reporters are stating that the sole reason for the increase in home run is due to changes in the composition of the ball. I understand this is a great story, since it seems to suggest that MLB and the Rawlings company are in a conspiracy to change the nature of baseball.

I was part of a scientific committee that looked at this problem extensively — our latest report was released to the public in December 2019. Although we agree that changes to the baseball, notably the drag coefficient, appears to account for much of the increase in HR hitting, we also stated that part of the increase in home runs is due to batter behavior. Since this reason seems to be neglected or forgotten, I thought it would be worthwhile to revisit our explanation about the batter behavior effect. In particular, I will focus on the changes between the 2018 and 2019 seasons and look at changes in batter behavior for some of the top home run hitters.

By the way, this change in batter behavior makes a good story and Jered Diamond just released Swing Kings that writes about these home run hitters who worked on their swing with great success.

The Change in Home Run Hitting Has Two Components

From 2018 to 2019, there was an increase of 6676 – 5585 = 1191 home runs. We want to figure out how much of this increase is attributable to changes in the baseball and how much is due to change in launch conditions (batter behavior variables including launch angle (LA), exit velocity (EV), and spray angle (SA)). This is what we did:

  1. First we fit a statistical model on the 2018 balls in play to understand how the probability of a HR depends on LA, EV, and SA. (By the way, other variables such as temperature and ball park play a role, but these effects are relatively small and tend to average out over a whole season.) This model helps us to understand the effect of the ball used in the 2018 season, so we can call this the “2018 ball” model.
  2. Using this 2018 ball model, we predict the number of 2019 home runs using the 2019 launch variables — we obtain 5930. This is an increase of 5930 – 5585 = 345 home runs from 2018 which represents about 30% of the total increase in home runs.
  3. The remaining 70% of the home run increase from 2018 to 2019 is due to other variables, such as changes to the baseball.

In our report, we also looked at home run changes from 2015 to 2016, 2016 to 2017, and 2017 to 2018. The effect of the ball was variable (up in 2017, down in 2018, and up again in 2019), but there has been a steady increase in home runs due to launch conditions.

Look at this at the Player Level

Since the launch condition effect in home run production is not negligible, it is worthwhile to examine how changes in launch conditions have affected home run hitting for specific players. Let’s illustrate this analysis for two extreme players, Joey Gallo and the “247” Khris Davis.

Joey Gallo

  1. Joey Gallo hit 40 home runs in 292 balls in play (BIP) in 2018. Using Gallo’s launch conditions for the 2019 season, I would predict him to hit 0.200 HR per BIP if the 2018 ball was being used.
  2. I would predict him to hit 292 (0.200) = 58.4 HR in 292 BIP in the 2019 season using the 2018 ball.
  3. The change 58.4 – 40 = 18.4 represents a big increase in home run hitting (in 292 BIP) based on his improved batting behavior in 2019.
  4. Actually, Gallo only hit 22 home runs in 2019 since he only had 129 BIP this season.

Khris Davis

  1. Davis hit 48 HR in 407 BIP in 2018. Using 2019 launch conditions, I predict his to hit 0.06 HR per BIP with the 2018 ball.
  2. If Davis had 407 BIP, I would predict him to hit 407 (0.6) = 24.6 home runs in 2019.
  3. The change (predicted minus 2018 HR) is 24.6 – 48 = -23.4. In contrast to Gallo, there was a big dropoff in home run hitting due to his batting behavior.
  4. Actually, Davis only hit 23 HR in 337 BIP in 2019 — the 2019 ball did not appear to help him much this season.

The Top 20 Home Run Hitters in 2018

The following graph shows the launch condition effects for the top 20 home run hitters in the 2108 season. Some of the hitters such as Joey Gallo benefited with improved batting behavior, but actually a majority of these batters (such as Khris Davis) lost home runs due to changing launch conditions. It would be interesting to learn what was happening in 2019 for these particular hitters.


I wrote this post to clarify one issue in our December 2019 report — here are the main points.

  • Hitters are changing their batting behavior and it is making a difference in terms of home runs hit. Batter behavior is a large effect and is not overwhelmed by the ball effect.
  • The ball effect affects all hitters in a somewhat uniform way which helps to explain the substantial home run hitting among a large group of players. In contrast, the launch condition effect varies among hitters. We are able to measure the effect of batting behavior and the size of this effect can be significant.
  • Any paper or article that talks about the home run hitting should recognize that it is more than a ball effect.
  • It is possible that the 2020 (or 2021) baseball has more drag which would negatively impact home runs. In contrast, I would expect that more hitters would be aware of the significant role of launch conditions and home runs due to batting behavior would continue to rise.
  • By the way, all of the R code for this exercise is available on request. Given the Statcast data, it is easy for the interested reader to reproduce my work.