Illustrates finding spacings and implementing permutation test for Mike Trout's 2015 hitting data

trout3.R

# loads in data frame d2015
# Retrosheet play-by-play data for 2015 season

load(url("http://www-math.bgsu.edu/~albert/retrosheet/pbp.2015.Rdata"))

# function will extract 1/0 data for a specific batter and a particular outcome

get_streak_data <- function(retrodata, batter, outcome){
  require(dplyr)
  require(Lahman)
  names <- unlist(strsplit(batter, " "))
  retroid <- filter(Master, nameFirst==names[1], 
                nameLast==names[2])$retroID
  data <- filter(retrodata, BAT_ID==retroid, 
                 AB_FL==TRUE)
  mutate(data, Date=substr(GAME_ID, 8, 12),
          H=as.numeric(H_FL > 0),
          HR=as.numeric(H_FL == 4),
          SO=as.numeric(EVENT_CD == 3)) %>% 
            arrange(Date) -> data
  if (outcome=="H") 
    select(data, H) -> Outcomes
  if (outcome=="HR")
    select(data, HR) -> Outcomes
  if (outcome=="SO")
    select(data, SO) -> Outcomes
  Outcomes[, 1]
}

# HITS
H2015 <- get_streak_data(d2015, "Mike Trout", "H")

# HOME RUNS
HR2015 <- get_streak_data(d2015, "Mike Trout", "HR")

# Strikeouts
SO2015 <- get_streak_data(d2015, "Mike Trout", "SO")

# load in BayesTestStreak package
# available from https://github.com/bayesball/BayesTestStreak

library(BayesTestStreak)

# this function implements a permutation test for randomness

permutation.test(H2015)
permutation.test(SO2015)
permutation.test(HR2015)

# this function finds all of the spacings for each -- gaps between 1's

find.spacings(H2015)
find.spacings(SO2015)
find.spacings(HR2015)