4.4 Correlation

In short, we may want correlation values close to 1.0 or -1.0.

4.4.1 Offense

graph code

cor_value = cor(ch4_data$R, ch4_data$Wpct,
                use = "pairwise.complete.obs")

subset_for_labels <- ch4_data |>
  select(R, Wpct, playoff_bool, yearID, teamID) |>
  group_by(playoff_bool) |>
  mutate(highlight = ifelse(
    R == max(R) | R == min(R),
    1, 0
  )) |>
  ungroup() |>
  filter(highlight == 1) |>
  mutate(year_team = paste(yearID, teamID))

ch4_data |>
  ggplot(aes(x = R, y = Wpct)) +
  geom_point(aes(color = playoff_bool)) +
  geom_label(aes(x = R, y = Wpct, color = playoff_bool, label = year_team),
             data = subset_for_labels) +
  labs(title = "Win Percentage vs Runs Scored",
       subtitle = paste0("correlation: r = ", round(cor_value, 4)),
       caption = "1998 to 2023 seasons",
       x = "runs scored",
       y = "win percentage") +
  scale_color_manual(values = c("#2905A1", "#AAAAAA")) +
  theme_minimal() +
  theme(legend.position = "bottom",
        legend.title=element_blank())

4.4.2 Defense

graph code

cor_value = cor(ch4_data$RA, ch4_data$Wpct,
                use = "pairwise.complete.obs")

subset_for_labels <- ch4_data |>
  select(RA, Wpct, playoff_bool, yearID, teamID) |>
  group_by(playoff_bool) |>
  mutate(highlight = ifelse(
    RA == max(RA) | RA == min(RA),
    1, 0
  )) |>
  ungroup() |>
  filter(highlight == 1) |>
  mutate(year_team = paste(yearID, teamID))

ch4_data |>
  ggplot(aes(x = RA, y = Wpct)) +
  geom_point(aes(color = playoff_bool)) +
  geom_label(aes(x = RA, y = Wpct, color = playoff_bool, label = year_team),
             data = subset_for_labels) +
  labs(title = "Win Percentage vs Runs Allowed",
       subtitle = paste0("correlation: r = ", round(cor_value, 4)),
       caption = "1998 to 2023 seasons",
       x = "runs allowed",
       y = "win percentage") +
  scale_color_manual(values = c("#2905A1", "#AAAAAA")) +
  theme_minimal() +
  theme(legend.position = "bottom",
        legend.title=element_blank())

4.4.3 Run Differential

graph code

cor_value = cor(ch4_data$RD, ch4_data$Wpct,
                use = "pairwise.complete.obs")

subset_for_labels <- ch4_data |>
  select(RD, Wpct, playoff_bool, yearID, teamID) |>
  group_by(playoff_bool) |>
  mutate(highlight = ifelse(
    RD == max(RD) | RD == min(RD),
    1, 0
  )) |>
  ungroup() |>
  filter(highlight == 1) |>
  mutate(year_team = paste(yearID, teamID))

ch4_data |>
  ggplot(aes(x = RD, y = Wpct)) +
  geom_point(aes(color = playoff_bool)) +
  geom_label(aes(x = RD, y = Wpct, color = playoff_bool, label = year_team),
             data = subset_for_labels) +
  labs(title = "Win Percentage vs Run Differential",
       subtitle = paste0("correlation: r = ", round(cor_value, 4)),
       caption = "1998 to 2023 seasons",
       x = "run differential",
       y = "win percentage") +
  scale_color_manual(values = c("#2905A1", "#AAAAAA")) +
  theme_minimal() +
  theme(legend.position = "bottom",
        legend.title=element_blank())