adjustment.R

###############################################
# working with complete set of 2015 play-by-play data
# collected using the getData function in the openWAR
# package (retrieves MLBAM GameDay files)
# currently have this saved as a Rdata file
###############################################
load("alldata2015.Rdata")

# computes the run values of all plate appearances

library(dplyr)
d2015 <- mutate(d2015,
                R1B=!is.na(start1B),
                R2B=!is.na(start2B),
                R3B=!is.na(start3B),
                E1B=!is.na(end1B),
                E2B=!is.na(end2B),
                E3B=!is.na(end3B))
(runs_expectancy <- summarize(group_by(d2015, 
                                       startOuts, R1B, R2B, R3B),
                              Runs=mean(runsFuture)))
d2015 <- inner_join(d2015, runs_expectancy,
                    by=c("startOuts", "R1B", "R2B", "R3B"))
d2015 <- inner_join(d2015, runs_expectancy,
                    by=c("endOuts"="startOuts",
                         "E1B"="R1B",
                         "E2B"="R2B",
                         "E3B"="R3B"))
d2015 <- mutate(d2015,
                Runs=Runs.y - Runs.x + runsOnPlay)

###############################
# adjustment for ballpark
###############################

# look at batter means

S <- summarize(group_by(d2015, batterName),
               R=mean(Runs),  N=n())

# get team identifier for each player

d2015 <- mutate(d2015,
                bat_team=ifelse(half=="top", 
                                as.character(away_team), 
                                as.character(home_team)))

TeamData <- summarize(group_by(d2015, batterName),
    Team=names(sort(table(bat_team), decreasing=TRUE))[1])

S1 <- inner_join(S, TeamData,
                 by="batterName")

# plots mean run values against PA for all players
# with Rockie players identified

library(ggplot2)
ggplot(S1, aes(N, R)) + geom_point(alpha=.2) +
  ylim(c(-.15, .15)) + geom_smooth() +
  geom_point(data=filter(S1, Team=="col"), 
             aes(N, R), color="red") +
  ggtitle("Mean Run Values for all Players in 2015 Season\nRockies Players in Red")

###### regress run values on bat team ids and computes residuals

fit <- lm(Runs ~ 0 + bat_team, data=d2015)
d2015$Residual <- fit$residuals

R <- summarize(group_by(d2015, batterName),
               Residual=mean(Residual),  N=n())
R1 <- inner_join(R, TeamData,
                 by="batterName")

# plots mean residual values against PA for all players
# with Rockie players identified

ggplot(R1, aes(N, Residual)) + geom_point(alpha=.2) +
  ylim(c(-.15, .15)) + geom_smooth() +
  geom_point(data=filter(R1, Team=="col"), 
             aes(N, Residual), color="red") +
  ggtitle("Mean Residual Values Adjusted for Ballpark\nRockies Players in Red")

# look at arm and batter side effects

A <- summarize(group_by(d2015, batterName, stand, throws),
               Mean=mean(Runs), N=n())  
A <- mutate(A, Platoon=paste(stand, 'hitter against', 
                             throws, 'pitcher')) 

A.mean <- summarize(group_by(A, Platoon),
                    Mean=sum(Mean * N) / sum(N))

# graphs mean runs and PA for all players in four platoon 
# situations

ggplot(A, aes(N, Mean)) + geom_point() +
  facet_wrap(~ Platoon, ncol=2)+ ylim(c(-.15, .15)) +
  geom_hline(yintercept=0) +
  geom_hline(aes(yintercept=Mean), data=A.mean, color="red") +
  ggtitle("Mean Run Values for all Players in 2015 Season\nPlatoon Effects")

d2015 <- mutate(d2015, platoon=factor(paste(stand, throws)))
fit2 <- lm(Runs ~ 0 + platoon, data=d2015)
d2015$Residual2 <- fit2$residuals

A1 <- summarize(group_by(d2015, batterName, stand, throws),
               Mean=mean(Residual2), N=n())  
A1 <- mutate(A1, Platoon=paste(stand, 'hitter against', throws, 'pitcher')) 
A1.mean <- summarize(group_by(A1, Platoon),
                     Mean=sum(Mean * N) / sum(N))

# graphs mean residuals and PA for all players in four platoon 
# situations

ggplot(A1, aes(N, Mean)) + geom_point() +
  facet_wrap(~ Platoon, ncol=2)+ ylim(c(-.15, .15)) +
  geom_hline(yintercept=0) +
  geom_hline(aes(yintercept=Mean), data=A1.mean, color="red") +
  ggtitle("Mean Residual Values for all Players in 2015 Season\nPlatoon Effects")