In bayesball/CalledStrike: Functions for Graphing Statcast Data

knitr::opts_chunk$set(echo = TRUE,
                      warning = FALSE,
                      message = FALSE)

Outline of work on pitch values

This illustrates new functions and data from the CalledStrike package to compute and graph pitch values.

library(dplyr)
library(readr)
library(ggplot2)
library(CalledStrike)
library(mgcv)
library(metR)

Read in Statcast data

Read in Statcast data from 2019 season:

sc <- read_csv("~/Dropbox/2016 WORK/BLOG Baseball R/OTHER/StatcastData/statcast2019.csv")

Runs values

The runs value of each count and the runs value for each end-of-PA outcome are contained in the dataset count_values contained in the CalledStrike package.

count_values

Compute the pitch values

The function compute_pitch_values() will create count variables and compute pitch values for all pitches in a Statcast dataset.

sc <- compute_pitch_values(sc)

Remove some plays

Remove plays like stolen bases and pickoff plays (with pitch value of 0) that don't affect pitch value.

sc %>%
  filter(abs(Runs.y) > 0) -> sc

Graph of Smoothed Pitch Values

This gives graph for a specific pitch type. Note that I have changed the sign on the pitch values so we get perspective from the pitcher.

Pitch values of changeups to right-handed hitters.

pitch_value_contour(filter(sc, 
                           pitch_type == "CH",
                           stand == "R"), 
    title = "Pitch Values of Changeups, Stand = R")

Compare pitch values of changeups to left and right-handed hitters.

changeups <- filter(sc, pitch_type == "CH")
d <- split(changeups, changeups$stand)
pitch_value_contour(d,
      title = "Pitch Value of Changeups")

Compare for left and right sided pitchers.

sc_L <- filter(sc, p_throws == "L")
sc_R <- filter(sc, p_throws == "R")
d <- c(split(sc_L, sc_L$stand),
      split(sc_R, sc_R$stand))
names(d) <- c("Pitch L, Stand L",
              "Pitch L, Stand R",
              "Pitch R, Stand L",
              "Pitch R, Stand R")
pitch_value_contour(d,
      title = "Pitch Value of Changeups")

Tile Plots

The function pitch_value_plot() constructs heat maps using the geom_tile() geometric object from the ggplot2 package. They don't require an input of values for the contour lines.

Pitch values of changeups to right-handed hitters.

pitch_value_plot(filter(sc, 
                           pitch_type == "CH",
                           stand == "R"), 
    title = "Pitch Values of Changeups, Stand = R")