Contents

Introduction

For my project, I have chosen to focus on two of the main areas of concentration for the 2021-2022 season: firstly, our two-strike (and non-two-strike) approach, and secondly, some advance scouting.

The data used for this project is a collection of data from the 2021 season, filtered to be only returning players. I have also elected to treat switch hitters as two different hitters. I have also filtered the data to only include data for the top ten hitters in the data with regard to pitches faced, as there were a handful of players with some data available but not enough for any analysis.

To keep things short and sweet, I will only include player-specific plots for a few players, however, all plots have been written as functions that require only a player’s name as the input and will create a plot for that player specifically.

Two-Strike Approach

With our two-strike-approach, one thing I have done is attempted to quantify how often our players are chasing pitches, or swinging at pitches we’d prefer them to take. To do this, I have created three different levels of “chases” to quantify plate discipline.

Here is a visual that shows how these look in practice.

This plot can be easily extended to individual players or subsets of players to identify areas where they may be frequently chasing pitches. Going back to the values of each level of chasing, we can compute for each player their chase rate, which is not an exact rate per se considering we have quantified a semi-chase as 0.25, but the chase rate statistic is still informative as we can see how the players compare to each other with plate disciple. The following plot shows the chase rate for each player in the data, separated into one side for less than two strikes and one side for two strikes.

Unsurprisingly, we observe that chase rates are generally higher with two strikes. We also observe that some players are more consistent with their plate discipline (like Clem, Mac, Johnny), while some have big differences in their chase rates when separated by strike count (Danny as RHH, Frick). This is an instance where the previously shown chasing plot can be effectively applied to an individual player to be more informative. I will here apply it to Frick since his chase rate jumps a lot with two strikes and is the highest on the team.

We can see that most of Frick’s chases are low and/or outside, but we also observe that he doesn’t seem to get many pitches that are inside. This will be true for probably all right-handed hitters, but it reinforces that pitches on the outer half of the plate should be expected. We can also see that with Chases specifically, there is a higher volume of them with two strikes than there is with either 0 or 1 strikes.

Advance Scouting

With the advance scouting section of this project, I will have four different plots, all of which can be reproduced for any player on any team with sufficient Trackman data. Each plot can additionally be split into two plots for two strikes or less than two strikes, but it is often infeasible on a player level because of the data needed to do so (also dependent on which plot).

I will show examples of each of these plots for Angel Zarate, since he has the most data available in the data I am using.

Heat Plot

Hot/Cold Zones

Swing Percentage Zones

Swing and Miss Zones

All of these plots are relatively easily modifiable and are all reproducible for any player we have Trackman data for. I believe that these ideas can be applied to both evaluation of our own team along with use as advanced scouting material. One shortcoming that I think we could address is with the hot/cold zones and other plots that are specific to different parts of the strike zone, player height and batting stance will change the dimensions of the strike zone and this is not something that is currently accounted for. With the hot/cold zones specifically, they are close to what we might get from Synergy, but not exactly the same. Below is the image I have been using to define the strike zone in this project.

Finally, here is a link to all code used for this project.