third and long

library(tidyverse)

passes <- read_csv('data/third_and_long_pass_plays.csv')

grouped_passes <- passes %>% 
  group_by(air_yards) %>%
  summarize(catch_rate = mean(passing_cmp)) %>% 
  filter(air_yards !=0) %>% 
  filter(air_yards >= -4)

lg_avg_completion_percentage <- function(p) {
  0.00028 * p ^ 2 - 0.0239 * p + 0.787
}

ggplot(grouped_passes, aes(x = air_yards, y = catch_rate)) +
  theme_minimal() +
  theme(plot.title=element_text(size=16, hjust=0, vjust=-1),
        plot.subtitle=element_text(size=12, hjust=0)) +
  stat_function(fun = lg_avg_completion_percentage, color = '#FFB704') +
  geom_point(alpha=.4, color='#018417') +
  geom_hline(yintercept = 0, alpha = .4) +
  geom_smooth(
    fill = '#CCCCCC',
    alpha = .2,
    color = '#018417',
    size = .6,
    span = .3
  ) +
  scale_x_continuous(
    limit = c(-5, 50),
    breaks = c(seq(
      from = -5,
      to = 50,
      by = 5
    )),
    minor_breaks = NULL
  ) +
  scale_y_continuous(
    limit = c(0, 1),
    breaks = c(seq(
      from = 0,
      to = 1,
      by = .1
    )),
    minor_breaks = NULL,
    labels = scales::percent
  ) +
  labs(fill = "Player name", x = "Depth of Target", y = "Completion Percentage",
       title = "Completion Percentage by Depth on 3rd Down and Long", subtitle = "(Note: 2009-2016 league average completion percentage in orange)")