Finds estimates of true batting averages for all batters against a specific pitcher.

batter.matchup.ggplot.R

# loading in Retrosheet data for the seasons 1960 through 2013 from my website 
load(url("http://bayes.bgsu.edu/baseball/pbp.1960.1979.Rdata"))
load(url("http://bayes.bgsu.edu/baseball/pbp.1980.1999.Rdata"))
load(url("http://bayes.bgsu.edu/baseball/pbp.2000.2013.Rdata"))

batter.matchup.ggplot <- function(Name, graph=TRUE, retroid=FALSE){
  # this function assumes the data frames pbp.60.79, pbp.80.99, and pbp.00.13 are
  # in the workspace
  fit.model <- function(y, n){
    require(LearnBayes)
    fit <- laplace(betabinexch, c(1, 1), cbind(y, n))$mode
    eta <- exp(fit[1]) / (1 + exp(fit[1]))
    K <- exp(fit[2])
    Estimate <- (y + K * eta) / (n + K)
    list(K=K, eta=eta, Estimate=Estimate)
  }
  require(dplyr)
  require(Lahman)
  
  if(retroid==FALSE){
    name <- unlist(strsplit(Name, split=" "))
    retro.id <- filter(Master, nameLast==name[2],
                       nameFirst==name[1])$retroID[1]} else
                         retro.id <- Name                     
  d1 <- filter(pbp.60.79, PIT_ID==retro.id, BAT_EVENT_FL==TRUE)
  d2 <- filter(pbp.80.99, PIT_ID==retro.id, BAT_EVENT_FL==TRUE)
  d3 <- filter(pbp.00.13, PIT_ID==retro.id, BAT_EVENT_FL==TRUE)
  d <- rbind(d1, d2, d3)
  S <- summarize(group_by(d, BAT_ID),
                 AB=sum(AB_FL), H=sum(H_FL >= 1))
  S <- filter(S, AB > 0)
  get.name <- function(j){
    F <- filter(Master, retroID==S$BAT_ID[j])
    paste(F$nameFirst, F$nameLast)}
  S$Batter.Name <- sapply(1:dim(S)[1], get.name)
  
  M <- fit.model(S$H, S$AB)
  
  if(graph==TRUE){
    print(c(eta=M$eta, K=M$K))
    S$Estimate <- M$Estimate
    S <- S[order(S$Estimate, decreasing=TRUE), ]
    S$Num <- 1:dim(S)[1]
    require(ggplot2)
    pt <- ggplot(S, aes(AB, Estimate)) + geom_point(size=3) +
      geom_hline(aes(yintercept=sum(H) / sum(AB)), 
                 color="red", size=2) +
      ggtitle(paste(Name, "Smoothed Averages")) +
      theme(plot.title = element_text(size = rel(3),
            color = "blue")) +
      theme(axis.title = element_text(size = rel(2))) +
      geom_text(data=subset(S, Num < 3 | Num > dim(S)[1] - 2),
                aes(AB, Estimate, label=Batter.Name),
                color="red")
      
    print(pt)
    print(head(S))
    print(tail(S))
    S} else c(M$K, M$eta)
}

This function assumes that the Retrosheet data frames pbp.60.79, pbp.80.99, and pbp.00.13 are in the current workspace. The data frame pbp.60.79 contains play-by-play data for the seasons 1960 through 1979, pbp.80.99 contains data for seasons 1980 through 1999, and pbp.00.13 contains data for seasons 2000 through 2013.