Introduction
As I am writing this post, I have seen 17 days of baseball and 487 home runs. Since MLB has changed the composition of the baseball this season, one wonders how this change will impact home run hitting. Since temperature affects home run running, it does not make sense to compare this season with the 2020 season when games were played in warmer weather. So I will focus on comparing 2021 with home run hitting during the same time frame of the 2019 season. I created a Shiny app that facilitates this comparison. After some initial comments, I’ll use the app to compare home run rates between 2019 and 2021 over specific regions of values of the launch variables. Based on what I learn in this exploration, I’m pretty confident the total 2021 home run will not exceed the 6776 home runs hit during the 2019 season.
Comparing 2021 with 2020 Home Runs
What have we learned about home run hitting in the first 17 days in April? Let’s make a quick comparison of 2021 with the previous season (2020).
- We have observed 487 home runs and the rate per plate appearance is 3.04%. As a comparison, the 2020 home run rate was 3.46%, but 2020 games were played in warmer weather.
- It is important to note that the strikeout rate in 2021 is currently 24.5% which is more than a full percentage point higher than the 2020 strikeout rate of 23.4%. The 2021 in-play rate is 63.8% compared to the 2020 in-play rate of 65.3%. So there are fewer opportunities to hit home runs in 2021.
- The home rate in 2021 among balls put into play is currently 4.77%. Looking at previous Aprils, this rate lies between the 2018 April and the 2019 April home run rates.
The Shiny App
The Shiny app allows for “brushed” home run comparison on in-play events across specified time intervals for the 2019 and 2021 seasons. Here I will compare home run hitting for 2019 and 2021 for the period April 1 through April 17. After I choose the time intervals, the app graphs the values of launch angle and exit velocity for all batted balls for both seasons where the color of the point corresponds to a home run. Note that I am focusing on the region where the launch angle is between 15 and 50 degrees and the exit velocity is between 90 and 115 mph. In the 2019 season, 99.7% of the home runs had launch variables in this region.
Brushing
On the surface, the 2019 and 2021 plots look very similar — for example, the regions where home runs are likely to occur look the same in the two plots. To make finer comparisons, I choose a rectangle by brushing over the plot. Here I am brushing a region where the launch angle is between 23 and 28 degrees and the exit velocity is between 95.2 and 103.2 mph.
The app then displays two relevant rates:
- the count (BIP) and rate of batted balls (Rate_Pct) in that (launch angle, exit velocity) region for each season
- the count (HR) and rate of home runs (HR_Rate_Pct) among the batted balls in that launch variable region for each season
By viewing these two rates, Rate_Pct and HR_Rate_Pct, for choices of different launch variable regions, we’ll get some good insights comparing 2019 and 2021 home run hitting.
Entire region
I begin by selecting the entire region of the display. For the entire region, we see below that we see a slightly higher number of batted balls in this region in 2019 than 2021 (respective rates of 23.8% and 23.4%). Of the batted balls in this region, 21.7% were home runs in 2019 compared to 20.0% in 2021. (Remember we are focusing on this 17-day period in April.)
Balls around 100 mph, good launch angles
An interesting borderline region are launch angle values in (27, 33) degrees and exit velocity values in (95.4, 104.6) mph. I call this “borderline” since the chance of a home run in this region is close to 50%.
Here we see similar counts of batted balls in this region for the two seasons (323 and 329), but the HR rate in this region is substantially higher in 2019 (46.4% compared to 37.4%) in this region leading to 150 – 123 = 27 additional home runs in this region in 2019.
Balls over 100 mph, good launch angles
Next we select the region of batted balls where the exit velocity exceeds 100 mph and the launch angle is in a “HR-friendly” interval of (27, 33) degrees. This is a region where it is likely that the batted ball is a home run. Interestingly, for both seasons, exactly 248 batted balls were hit in this region for this 17-day period in the two seasons. But 74.6% of these batted balls were home runs in 2019 compared with 67.7% in 2021.
Hard-hit balls over 105 mpg
Next I focus on “hard-hit balls” that are hit over 105 mpg for any angle in (14, 50) degrees. There were a higher rate of hard-hit balls in 2021 (4.3% compared to 4.0%) but the HR rate of balls in this region was higher in 2019. In this particular case, the two factors cancel out, resulting in 233 “hard-hit” home runs from this region in each season.
Balls hit over 100 mpg, low angles
Here I am looking at balls hit over 100 mph for low launch angles (between 15 and 24 degrees). There was a higher rate of these batted balls in 2019 and there was also a higher HR rate, resulting in 56 more home runs in 2019 in this region.
Balls hit over 100 mpg, all angles
Next I look at balls over 100 mpg for all angles between 15 and 51 degrees. Here we have similar rates of batted balls in this region, but there was a higher HR rate in 2019, resulting in 59 more home runs.
Take-Aways
- Home run production in 2021 is “low”, but that could be just an impact of the low temperatures in April. So I wanted to make a fairer comparison by looking at home run hitting in 2021 and 2019 during the same time period in April. The differences we are observing are not due to temperature.
- The Shiny app is helpful for comparing home rates for different time intervals of the two seasons, and for comparing home run rates for specific regions of launch variables.
- There is some evidence that batted balls are hit harder in 2021. But by comparing 2019 and 2021 rates across different regions of the (launch angle, exit velocity) space, the rates of hard-hit balls seem pretty similar in the two seasons.
- Our comparison shows that the home run rates in balls in different regions are not the same. The 2019 balls appear to be carrying further, leading to higher home run rates. In other words, the drag coefficients appear to be higher in 2021.
- We will see higher in-play home run rates in 2021 as we move towards warmer weather. I’ll continue to compare 2019 and 2021 home run production over similar time periods using the app. Assuming the 2021 ball does not go through more changes, it is likely we’ll see a lower home count in 2021 than the 6776 home runs we saw in the 2019 season.
- I haven’t provided the Shiny code since it is pretty similar to other brushing apps such as the BrushingZone() function available through my ShinyBaseball package.
Exploring Baseball Data with R 