ChiBearsStats/RIP Parkey & the Insignificance of Kickers

## RIP Parkey & the Insignificance of Kickers
library(dplyr)
library(nflscrapR)
library(tidyverse)
library(ggrepel)
library(ggimage)


pbp_all <- readRDS("pbpdata2009to2019.rds") #I just use prescraped data, but if you don't know where to get this, I would recommend reading Ben Baldwin's tutorial:
#(https://gist.github.com/guga31bb/5634562c5a2a7b1e9961ac9b6c568701)

field_goals <-
  pbp_all %>%
  filter(play_type == "field_goal") %>%
  select(play_type, desc, field_goal_result, kick_distance, fg_prob, field_goal_attempt, kicker_player_name, posteam, kicker_player_id, season, wpa, epa) %>%
  mutate(made_kick = ifelse(field_goal_result == "made", 1, 0),
         kick_distance = ifelse(kick_distance < 56, kick_distance, 56))


expected_field_goals <-
  field_goals %>%
  group_by(kick_distance) %>%
  summarize(expected_makes = sum(made_kick), attempts = n()) %>%
  mutate(expected_kick_percentage = expected_makes/attempts) %>%
  select(-c(expected_makes, attempts))

actual_field_goals <-
  field_goals %>%
  group_by(kicker_player_name, season, kick_distance) %>%
  summarize(actual_makes = sum(made_kick), attempts = n(),
            epa = sum(epa, na.rm = TRUE), wpa = sum(wpa, na.rm = TRUE)) %>%
  left_join(expected_field_goals, by = c("kick_distance")) %>%
  mutate(expected_makes = (expected_kick_percentage * attempts)) %>%
  group_by(kicker_player_name, season) %>%
  summarise(attempts = sum(attempts),
            actual_makes = sum(actual_makes),
            expected_makes = sum(expected_makes),
            makes_over_expectation = (actual_makes - expected_makes),
            epa = sum(epa, na.rm = TRUE),
            wpa = sum(wpa, na.rm = TRUE))


#Density Plot of FGoE

  ggplot(actual_field_goals, aes(x=makes_over_expectation, y=..density..)) +
    geom_vline(data = subset(actual_field_goals, kicker_player_name == "C.Parkey" & season == "2018"),
               aes(xintercept = makes_over_expectation, color = "orange")) +
    geom_density() +
    annotate("label", x = -1.5, y= .025,label = "Cody Parkey, 2018 \n -2.76", size = 2.5, color = "black") +
    theme_classic() +
    theme(axis.text.y=element_blank()) +
    guides(color = FALSE) +
    labs(
      x = "Made Field Goals Over Expectation",
      y = "Density",
      title = "How Much Do Kickers Out-Kick or Under-Kick Expectations?",
      subtitle = "2009-2018",
      caption = "Data from nflscrapR \n \n @ChiBearsStats"
    )

  ggsave('fieldgoalsoverexpectation.png', dpi=1000)


#How much EPA do Kickers bring?
  ggplot(actual_field_goals, aes(x=epa, y=..density..)) +
    geom_vline(data = subset(actual_field_goals, kicker_player_name == "C.Parkey" & season == "2018"),
               aes(xintercept = epa, color = "orange")) +
    geom_density() +
    annotate("label", x = -7, y= .01,label = "Cody Parkey, 2018 \n -12.0", size = 2.5, color = "black") +
    theme_classic() +
    theme(axis.text.y=element_blank()) +
    guides(color = FALSE) +
    labs(
      x = "Total EPA",
      y = "Density",
      title = "Kickers EPA",
      subtitle = "2009-2018",
      caption = "Data from nflscrapR \n \n @ChiBearsStats"
    )
  ggsave('kickerepa.png', dpi=1000)


  #How much WPA do Kickers bring?
  ggplot(actual_field_goals, aes(x=wpa, y=..density..)) +
    geom_vline(data = subset(actual_field_goals, kicker_player_name == "C.Parkey" & season == "2018"),
               aes(xintercept = wpa, color = "orange")) +
    geom_density() +
    annotate("label", x = -.2, y= .1,label = "Cody Parkey, 2018 \n -.45", size = 2.5, color = "black") +
    theme_classic() +
    theme(axis.text.y=element_blank()) +
    guides(color = FALSE) +
    labs(
      x = "Total WPA",
      y = "Density",
      title = "Kickers WPA",
      subtitle = "2009-2018",
      caption = "Data from nflscrapR \n \n @ChiBearsStats"
    )
  ggsave('kickerwpa.png', dpi=1000)

  #Let's take a look at accuracy vs attempts

  attempts_lm = lm(makes_over_expectation ~ attempts, data=actual_field_goals)
  rsquared <- summary(attempts_lm)$r.squared
  rsquared <- signif(rsquared, digits = 2)
  rsquaredlabel <- expression(paste(" ",R^2 ,"= .05"))

  ggplot(actual_field_goals, aes(x = attempts, makes_over_expectation)) +
    geom_point() +
    geom_point(data = subset(actual_field_goals, (kicker_player_name == "C.Parkey")),
               color = "orange",
               size = 2) +
    geom_image(
      data = subset(actual_field_goals, (kicker_player_name == "C.Parkey" & season == 2018)),
      aes(image = "https://upload.wikimedia.org/wikipedia/commons/thumb/5/5c/Chicago_Bears_logo.svg/100px-Chicago_Bears_logo.svg.png"), size = .05) +
    annotate("text", x = 50, y= -4,label = rsquaredlabel, size = 4, color = "black") +
    geom_hline(yintercept = 0, lty = 4, color = "black") +
    geom_text_repel(
      data = subset(actual_field_goals, (makes_over_expectation > 5)
                    | attempts > 50
                    | makes_over_expectation < -5),
      aes(label = paste(kicker_player_name, substr(season, 3, 4)))) +
    geom_text_repel(
      data = subset(actual_field_goals, (kicker_player_name == "C.Parkey")),
      aes(label = substr(season, 3, 4)),
      color = "blue",
      nudge_y = 1,
      nudge_x = 1) +
    theme_classic() +
    labs(
      x = "Total Attempts",
      y = "Made Field Goals Over Expectation",
      title = "FGoE vs Attempts (Cody Parkey Seasons Highlighted)",
      subtitle = "2009-2018",
      caption = "Data from nflscrapR \n \n @ChiBearsStats"
    )
  ggsave('madefieldgoalsvsattempts.png', dpi=1000)


  #Careers

  actual_field_goals_career <-
    field_goals %>%
    group_by(kicker_player_name, kick_distance) %>%
    summarize(actual_makes = sum(made_kick), attempts = n(),
              epa = sum(epa, na.rm = TRUE), wpa = sum(wpa, na.rm = TRUE)) %>%
    left_join(expected_field_goals, by = c("kick_distance")) %>%
    mutate(expected_makes = (expected_kick_percentage * attempts)) %>%
    group_by(kicker_player_name) %>%
    summarise(attempts = sum(attempts),
              actual_makes = sum(actual_makes),
              expected_makes = sum(expected_makes),
              makes_over_expectation = ((actual_makes - expected_makes)/expected_makes),
              epa = sum(epa, na.rm = TRUE),
              wpa = sum(wpa, na.rm = TRUE)) %>%
    filter(attempts > 20)

  ggplot(actual_field_goals_career, aes(x = attempts, makes_over_expectation)) +
    geom_point() +
    geom_point(data = subset(actual_field_goals_career, (kicker_player_name == "C.Parkey")),
               color = "orange",
               size = 2) +
    geom_hline(yintercept = 0, lty = 4, color = "black") +
    geom_text_repel(
      data = subset(actual_field_goals_career, (makes_over_expectation > .05)
                    | attempts > 300
                    | makes_over_expectation < -.15),
      aes(label = kicker_player_name)) +
    geom_text_repel(
      data = subset(actual_field_goals_career, (kicker_player_name == "C.Parkey")),
      aes(label = kicker_player_name),
      color = "blue",
      nudge_y = -.005,
      nudge_x = .1) +
    theme_classic() +
    labs(
      x = "Total Attempts",
      y = "Made Field Goals Over Expectation",
      title = "Field Goals Over Expectation (%)",
      subtitle = "2009-2018, minimum of 20 Attempts",
      caption = "Data from nflscrapR \n \n @ChiBearsStats"
    )
  ggsave('careerfgattempts.png', dpi=1000)


  #Year Over Year Kicker Accuracy
  actual_field_goals_yoy <-
    actual_field_goals %>%
    filter(attempts >= 10) %>%
    mutate(following_season = season + 1) %>%
    left_join(select(actual_field_goals_yoy, kicker_player_name, season, makes_over_expectation),
                     by = c("kicker_player_name", "following_season" = "season")) %>%
    filter(!is.na(makes_over_expectation.y))

  yoy_lm = lm(makes_over_expectation.x ~ makes_over_expectation.y, data=actual_field_goals_yoy)
  rsquared_yoy <- summary(yoy_lm)$r.squared
  rsquared_yoy <- signif(rsquared_yoy, digits = 2)
  rsquaredlabel_yoy <- expression(paste(" ",R^2 ,"= .043"))

  #Graph
  ggplot(actual_field_goals_yoy, aes(x = makes_over_expectation.x, makes_over_expectation.y)) +
    geom_point() +
    geom_point(data = subset(actual_field_goals_yoy, (kicker_player_name == "C.Parkey")),
               color = "orange",
               size = 2) +
    annotate("text", x = 5, y= -4,label = rsquaredlabel_yoy, size = 4, color = "black") +
    geom_abline(intercept = 0, slope = 1, lty = 2) +
    geom_vline(xintercept = 0, lty = 2) +
    geom_hline(yintercept = 0, lty = 2) +
    geom_text_repel(
      data = subset(actual_field_goals_yoy, (makes_over_expectation.y > 5)
                    | makes_over_expectation.x > 5
                    | makes_over_expectation.y < -5
                    | makes_over_expectation.x < -5),
      aes(label = paste(kicker_player_name, substr(season, 3, 4), "-", substr(following_season, 3, 4)))) +
    geom_text_repel(
      data = subset(actual_field_goals_yoy, (kicker_player_name == "C.Parkey")),
      aes(label = paste(substr(season, 3, 4), "-", substr(following_season, 3, 4))),
      color = "blue",
      nudge_y = .5,
      nudge_x = .5) +
    theme_classic() +
    labs(
      x = "Year X",
      y = "Year X + 1",
      title = "Field Goals Over Expectation, Year Over Year",
      subtitle = "2009-2018, Minimum 10 Attempts",
      caption = "Data from nflscrapR \n \n @ChiBearsStats"
    )
  ggsave('yearoveryearfieldgoals.png', dpi=1000)

  #Let's do it proportionally to make sure we're not penalizing high attempt seasons
  actual_field_goals_yoy <-
    actual_field_goals %>%
    filter(attempts >= 10) %>%
    mutate(following_season = season + 1,
           difference_percentage = ((actual_makes - expected_makes)/expected_makes)) %>%
    left_join(select(actual_field_goals_yoy, kicker_player_name, season, difference_percentage),
              by = c("kicker_player_name", "following_season" = "season")) %>%
    filter(!is.na(difference_percentage.y))

  yoyper_lm = lm(difference_percentage.x ~ difference_percentage.y, data=actual_field_goals_yoy)
  rsquared_yoyper <- summary(yoyper_lm)$r.squared
  rsquared_yoyper <- signif(rsquared_yoy, digits = 2)
  rsquaredlabel_yoyper <- expression(paste(" ",R^2 ,"= .043"))

  #Graph
  ggplot(actual_field_goals_yoy, aes(x = difference_percentage.x, difference_percentage.y)) +
    geom_point() +
    geom_point(data = subset(actual_field_goals_yoy, (kicker_player_name == "C.Parkey")),
               color = "orange",
               size = 2) +
    annotate("text", x = .2, y= -.2,label = rsquaredlabel_yoy, size = 4, color = "black") +
    geom_abline(intercept = 0, slope = 1, lty = 2) +
    geom_vline(xintercept = 0, lty = 2) +
    geom_hline(yintercept = 0, lty = 2) +
    geom_text_repel(
      data = subset(actual_field_goals_yoy, (difference_percentage.x > .2)
                    | difference_percentage.y > .2
                    | difference_percentage.x < -.2
                    | difference_percentage.y < -.22),
      aes(label = paste(kicker_player_name, substr(season, 3, 4), "-", substr(following_season, 3, 4)))) +
    geom_text_repel(
      data = subset(actual_field_goals_yoy, (kicker_player_name == "C.Parkey")),
      aes(label = paste(substr(season, 3, 4), "-", substr(following_season, 3, 4))),
      color = "blue",
      nudge_y = .05,
      nudge_x = .05) +
    theme_classic() +
    labs(
      x = "Year X",
      y = "Year X + 1",
      title = "Field Goals Over Expectation (%), Year Over Year",
      subtitle = "2009-2018, Minimum 10 Attempts",
      caption = "Data from nflscrapR \n \n @ChiBearsStats"
    )
  ggsave('yearoveryearfieldgoalspercentage.png', dpi=1000)

  #Lastly, let's look at "clutch" kicks
  actual_field_goals_clutch <-
    pbp_all %>%
    filter(play_type == "field_goal") %>%
    select(game_seconds_remaining, game_half, score_differential, play_type, desc, field_goal_result, kick_distance, fg_prob, field_goal_attempt, kicker_player_name, posteam, kicker_player_id, season, wpa, epa) %>%
    mutate(made_kick = ifelse(field_goal_result == "made", 1, 0),
           kick_distance = ifelse(kick_distance < 56, kick_distance, 56)) %>%
    filter((game_seconds_remaining < 300 & score_differential <= 3 & score_differential >= -3) |
             game_half == "Overtime") %>%
    group_by(kicker_player_name, kick_distance) %>%
    summarize(actual_makes = sum(made_kick), attempts = n(),
              epa = sum(epa, na.rm = TRUE), wpa = sum(wpa, na.rm = TRUE)) %>%
    left_join(expected_field_goals, by = c("kick_distance")) %>%
    mutate(expected_makes = (expected_kick_percentage * attempts)) %>%
    group_by(kicker_player_name) %>%
    summarise(attempts = sum(attempts),
              actual_makes = sum(actual_makes),
              expected_makes = sum(expected_makes),
              difference_percentage = ((actual_makes - expected_makes)/expected_makes),
              makes_over_expectation = (actual_makes - expected_makes)) %>%
    filter(attempts > 10) %>%
    left_join(actual_field_goals_career, by = c("kicker_player_name")) %>%
    mutate(difference_percentage_career = ((actual_makes.y - expected_makes.y)/expected_makes.y))

  clutch_lm = lm(difference_percentage ~ difference_percentage_career, data=actual_field_goals_clutch)
  rsquared_clutch <- summary(clutch_lm)$r.squared
  rsquared_clutch <- signif(rsquared_clutch, digits = 2)
  rsquaredlabel_clutch <- expression(paste(" ",R^2 ,"= .17"))

  ggplot(actual_field_goals_clutch, aes(x = difference_percentage,difference_percentage_career)) +
    geom_point() +
    geom_point(data = subset(actual_field_goals_clutch, (kicker_player_name == "C.Parkey")),
               color = "orange",
               size = 2) +
    geom_smooth(method = lm) +
    annotate("text", x = .2, y= -.05,label = rsquaredlabel_clutch, size = 4, color = "black") +
    geom_abline(intercept = 0, slope = 1, lty = 2) +
    geom_vline(xintercept = 0, lty = 2) +
    geom_hline(yintercept = 0, lty = 2) +
    geom_text_repel(
      data = subset(actual_field_goals_clutch, (difference_percentage > .05)
                    | difference_percentage_career > .05
                    | difference_percentage < -.15
                    | difference_percentage < -.22),
      aes(label = kicker_player_name)) +
    geom_text_repel(
      data = subset(actual_field_goals_clutch, (kicker_player_name == "C.Parkey")),
      aes(label = kicker_player_name),
      color = "blue",
      nudge_y = .005,
      nudge_x = -.005) +
    theme_classic() +
    labs(
      x = "Clutch Kicks Above Expectation (%)",
      y = "All Kicks",
      title = "Clutch Kicking",
      subtitle = "2009-2018, Minimum 10 Attempts",
      caption = "Clutch Kick = <5 minutes remaining in Reg. or in OT in a 3 point game. \n Data from nflscrapR \n @ChiBearsStats"
    )
  ggsave('clutchkicks.png', dpi=1000)
	library(dplyr)
	library(nflscrapR)
	library(tidyverse)
	library(ggrepel)
	library(ggimage)


	pbp_all <- readRDS("pbpdata2009to2019.rds") #I just use prescraped data, but if you don't know where to get this, I would recommend reading Ben Baldwin's tutorial:
	#(https://gist.github.com/guga31bb/5634562c5a2a7b1e9961ac9b6c568701)

	field_goals <-
	pbp_all %>%
	filter(play_type == "field_goal") %>%
	select(play_type, desc, field_goal_result, kick_distance, fg_prob, field_goal_attempt, kicker_player_name, posteam, kicker_player_id, season, wpa, epa) %>%
	mutate(made_kick = ifelse(field_goal_result == "made", 1, 0),
	kick_distance = ifelse(kick_distance < 56, kick_distance, 56))


	expected_field_goals <-
	field_goals %>%
	group_by(kick_distance) %>%
	summarize(expected_makes = sum(made_kick), attempts = n()) %>%
	mutate(expected_kick_percentage = expected_makes/attempts) %>%
	select(-c(expected_makes, attempts))

	actual_field_goals <-
	field_goals %>%
	group_by(kicker_player_name, season, kick_distance) %>%
	summarize(actual_makes = sum(made_kick), attempts = n(),
	epa = sum(epa, na.rm = TRUE), wpa = sum(wpa, na.rm = TRUE)) %>%
	left_join(expected_field_goals, by = c("kick_distance")) %>%
	mutate(expected_makes = (expected_kick_percentage * attempts)) %>%
	group_by(kicker_player_name, season) %>%
	summarise(attempts = sum(attempts),
	actual_makes = sum(actual_makes),
	expected_makes = sum(expected_makes),
	makes_over_expectation = (actual_makes - expected_makes),
	epa = sum(epa, na.rm = TRUE),
	wpa = sum(wpa, na.rm = TRUE))


	#Density Plot of FGoE

	ggplot(actual_field_goals, aes(x=makes_over_expectation, y=..density..)) +
	geom_vline(data = subset(actual_field_goals, kicker_player_name == "C.Parkey" & season == "2018"),
	aes(xintercept = makes_over_expectation, color = "orange")) +
	geom_density() +
	annotate("label", x = -1.5, y= .025,label = "Cody Parkey, 2018 \n -2.76", size = 2.5, color = "black") +
	theme_classic() +
	theme(axis.text.y=element_blank()) +
	guides(color = FALSE) +
	labs(
	x = "Made Field Goals Over Expectation",
	y = "Density",
	title = "How Much Do Kickers Out-Kick or Under-Kick Expectations?",
	subtitle = "2009-2018",
	caption = "Data from nflscrapR \n \n @ChiBearsStats"
	)

	ggsave('fieldgoalsoverexpectation.png', dpi=1000)


	#How much EPA do Kickers bring?
	ggplot(actual_field_goals, aes(x=epa, y=..density..)) +
	geom_vline(data = subset(actual_field_goals, kicker_player_name == "C.Parkey" & season == "2018"),
	aes(xintercept = epa, color = "orange")) +
	geom_density() +
	annotate("label", x = -7, y= .01,label = "Cody Parkey, 2018 \n -12.0", size = 2.5, color = "black") +
	theme_classic() +
	theme(axis.text.y=element_blank()) +
	guides(color = FALSE) +
	labs(
	x = "Total EPA",
	y = "Density",
	title = "Kickers EPA",
	subtitle = "2009-2018",
	caption = "Data from nflscrapR \n \n @ChiBearsStats"
	)
	ggsave('kickerepa.png', dpi=1000)


	#How much WPA do Kickers bring?
	ggplot(actual_field_goals, aes(x=wpa, y=..density..)) +
	geom_vline(data = subset(actual_field_goals, kicker_player_name == "C.Parkey" & season == "2018"),
	aes(xintercept = wpa, color = "orange")) +
	geom_density() +
	annotate("label", x = -.2, y= .1,label = "Cody Parkey, 2018 \n -.45", size = 2.5, color = "black") +
	theme_classic() +
	theme(axis.text.y=element_blank()) +
	guides(color = FALSE) +
	labs(
	x = "Total WPA",
	y = "Density",
	title = "Kickers WPA",
	subtitle = "2009-2018",
	caption = "Data from nflscrapR \n \n @ChiBearsStats"
	)
	ggsave('kickerwpa.png', dpi=1000)

	#Let's take a look at accuracy vs attempts

	attempts_lm = lm(makes_over_expectation ~ attempts, data=actual_field_goals)
	rsquared <- summary(attempts_lm)$r.squared
	rsquared <- signif(rsquared, digits = 2)
	rsquaredlabel <- expression(paste(" ",R^2 ,"= .05"))

	ggplot(actual_field_goals, aes(x = attempts, makes_over_expectation)) +
	geom_point() +
	geom_point(data = subset(actual_field_goals, (kicker_player_name == "C.Parkey")),
	color = "orange",
	size = 2) +
	geom_image(
	data = subset(actual_field_goals, (kicker_player_name == "C.Parkey" & season == 2018)),
	aes(image = "https://upload.wikimedia.org/wikipedia/commons/thumb/5/5c/Chicago_Bears_logo.svg/100px-Chicago_Bears_logo.svg.png"), size = .05) +
	annotate("text", x = 50, y= -4,label = rsquaredlabel, size = 4, color = "black") +
	geom_hline(yintercept = 0, lty = 4, color = "black") +
	geom_text_repel(
	data = subset(actual_field_goals, (makes_over_expectation > 5)
	\| attempts > 50
	\| makes_over_expectation < -5),
	aes(label = paste(kicker_player_name, substr(season, 3, 4)))) +
	geom_text_repel(
	data = subset(actual_field_goals, (kicker_player_name == "C.Parkey")),
	aes(label = substr(season, 3, 4)),
	color = "blue",
	nudge_y = 1,
	nudge_x = 1) +
	theme_classic() +
	labs(
	x = "Total Attempts",
	y = "Made Field Goals Over Expectation",
	title = "FGoE vs Attempts (Cody Parkey Seasons Highlighted)",
	subtitle = "2009-2018",
	caption = "Data from nflscrapR \n \n @ChiBearsStats"
	)
	ggsave('madefieldgoalsvsattempts.png', dpi=1000)


	#Careers

	actual_field_goals_career <-
	field_goals %>%
	group_by(kicker_player_name, kick_distance) %>%
	summarize(actual_makes = sum(made_kick), attempts = n(),
	epa = sum(epa, na.rm = TRUE), wpa = sum(wpa, na.rm = TRUE)) %>%
	left_join(expected_field_goals, by = c("kick_distance")) %>%
	mutate(expected_makes = (expected_kick_percentage * attempts)) %>%
	group_by(kicker_player_name) %>%
	summarise(attempts = sum(attempts),
	actual_makes = sum(actual_makes),
	expected_makes = sum(expected_makes),
	makes_over_expectation = ((actual_makes - expected_makes)/expected_makes),
	epa = sum(epa, na.rm = TRUE),
	wpa = sum(wpa, na.rm = TRUE)) %>%
	filter(attempts > 20)

	ggplot(actual_field_goals_career, aes(x = attempts, makes_over_expectation)) +
	geom_point() +
	geom_point(data = subset(actual_field_goals_career, (kicker_player_name == "C.Parkey")),
	color = "orange",
	size = 2) +
	geom_hline(yintercept = 0, lty = 4, color = "black") +
	geom_text_repel(
	data = subset(actual_field_goals_career, (makes_over_expectation > .05)
	\| attempts > 300
	\| makes_over_expectation < -.15),
	aes(label = kicker_player_name)) +
	geom_text_repel(
	data = subset(actual_field_goals_career, (kicker_player_name == "C.Parkey")),
	aes(label = kicker_player_name),
	color = "blue",
	nudge_y = -.005,
	nudge_x = .1) +
	theme_classic() +
	labs(
	x = "Total Attempts",
	y = "Made Field Goals Over Expectation",
	title = "Field Goals Over Expectation (%)",
	subtitle = "2009-2018, minimum of 20 Attempts",
	caption = "Data from nflscrapR \n \n @ChiBearsStats"
	)
	ggsave('careerfgattempts.png', dpi=1000)


	#Year Over Year Kicker Accuracy
	actual_field_goals_yoy <-
	actual_field_goals %>%
	filter(attempts >= 10) %>%
	mutate(following_season = season + 1) %>%
	left_join(select(actual_field_goals_yoy, kicker_player_name, season, makes_over_expectation),
	by = c("kicker_player_name", "following_season" = "season")) %>%
	filter(!is.na(makes_over_expectation.y))

	yoy_lm = lm(makes_over_expectation.x ~ makes_over_expectation.y, data=actual_field_goals_yoy)
	rsquared_yoy <- summary(yoy_lm)$r.squared
	rsquared_yoy <- signif(rsquared_yoy, digits = 2)
	rsquaredlabel_yoy <- expression(paste(" ",R^2 ,"= .043"))

	#Graph
	ggplot(actual_field_goals_yoy, aes(x = makes_over_expectation.x, makes_over_expectation.y)) +
	geom_point() +
	geom_point(data = subset(actual_field_goals_yoy, (kicker_player_name == "C.Parkey")),
	color = "orange",
	size = 2) +
	annotate("text", x = 5, y= -4,label = rsquaredlabel_yoy, size = 4, color = "black") +
	geom_abline(intercept = 0, slope = 1, lty = 2) +
	geom_vline(xintercept = 0, lty = 2) +
	geom_hline(yintercept = 0, lty = 2) +
	geom_text_repel(
	data = subset(actual_field_goals_yoy, (makes_over_expectation.y > 5)
	\| makes_over_expectation.x > 5
	\| makes_over_expectation.y < -5
	\| makes_over_expectation.x < -5),
	aes(label = paste(kicker_player_name, substr(season, 3, 4), "-", substr(following_season, 3, 4)))) +
	geom_text_repel(
	data = subset(actual_field_goals_yoy, (kicker_player_name == "C.Parkey")),
	aes(label = paste(substr(season, 3, 4), "-", substr(following_season, 3, 4))),
	color = "blue",
	nudge_y = .5,
	nudge_x = .5) +
	theme_classic() +
	labs(
	x = "Year X",
	y = "Year X + 1",
	title = "Field Goals Over Expectation, Year Over Year",
	subtitle = "2009-2018, Minimum 10 Attempts",
	caption = "Data from nflscrapR \n \n @ChiBearsStats"
	)
	ggsave('yearoveryearfieldgoals.png', dpi=1000)

	#Let's do it proportionally to make sure we're not penalizing high attempt seasons
	actual_field_goals_yoy <-
	actual_field_goals %>%
	filter(attempts >= 10) %>%
	mutate(following_season = season + 1,
	difference_percentage = ((actual_makes - expected_makes)/expected_makes)) %>%
	left_join(select(actual_field_goals_yoy, kicker_player_name, season, difference_percentage),
	by = c("kicker_player_name", "following_season" = "season")) %>%
	filter(!is.na(difference_percentage.y))

	yoyper_lm = lm(difference_percentage.x ~ difference_percentage.y, data=actual_field_goals_yoy)
	rsquared_yoyper <- summary(yoyper_lm)$r.squared
	rsquared_yoyper <- signif(rsquared_yoy, digits = 2)
	rsquaredlabel_yoyper <- expression(paste(" ",R^2 ,"= .043"))

	#Graph
	ggplot(actual_field_goals_yoy, aes(x = difference_percentage.x, difference_percentage.y)) +
	geom_point() +
	geom_point(data = subset(actual_field_goals_yoy, (kicker_player_name == "C.Parkey")),
	color = "orange",
	size = 2) +
	annotate("text", x = .2, y= -.2,label = rsquaredlabel_yoy, size = 4, color = "black") +
	geom_abline(intercept = 0, slope = 1, lty = 2) +
	geom_vline(xintercept = 0, lty = 2) +
	geom_hline(yintercept = 0, lty = 2) +
	geom_text_repel(
	data = subset(actual_field_goals_yoy, (difference_percentage.x > .2)
	\| difference_percentage.y > .2
	\| difference_percentage.x < -.2
	\| difference_percentage.y < -.22),
	aes(label = paste(kicker_player_name, substr(season, 3, 4), "-", substr(following_season, 3, 4)))) +
	geom_text_repel(
	data = subset(actual_field_goals_yoy, (kicker_player_name == "C.Parkey")),
	aes(label = paste(substr(season, 3, 4), "-", substr(following_season, 3, 4))),
	color = "blue",
	nudge_y = .05,
	nudge_x = .05) +
	theme_classic() +
	labs(
	x = "Year X",
	y = "Year X + 1",
	title = "Field Goals Over Expectation (%), Year Over Year",
	subtitle = "2009-2018, Minimum 10 Attempts",
	caption = "Data from nflscrapR \n \n @ChiBearsStats"
	)
	ggsave('yearoveryearfieldgoalspercentage.png', dpi=1000)

	#Lastly, let's look at "clutch" kicks
	actual_field_goals_clutch <-
	pbp_all %>%
	filter(play_type == "field_goal") %>%
	select(game_seconds_remaining, game_half, score_differential, play_type, desc, field_goal_result, kick_distance, fg_prob, field_goal_attempt, kicker_player_name, posteam, kicker_player_id, season, wpa, epa) %>%
	mutate(made_kick = ifelse(field_goal_result == "made", 1, 0),
	kick_distance = ifelse(kick_distance < 56, kick_distance, 56)) %>%
	filter((game_seconds_remaining < 300 & score_differential <= 3 & score_differential >= -3) \|
	game_half == "Overtime") %>%
	group_by(kicker_player_name, kick_distance) %>%
	summarize(actual_makes = sum(made_kick), attempts = n(),
	epa = sum(epa, na.rm = TRUE), wpa = sum(wpa, na.rm = TRUE)) %>%
	left_join(expected_field_goals, by = c("kick_distance")) %>%
	mutate(expected_makes = (expected_kick_percentage * attempts)) %>%
	group_by(kicker_player_name) %>%
	summarise(attempts = sum(attempts),
	actual_makes = sum(actual_makes),
	expected_makes = sum(expected_makes),
	difference_percentage = ((actual_makes - expected_makes)/expected_makes),
	makes_over_expectation = (actual_makes - expected_makes)) %>%
	filter(attempts > 10) %>%
	left_join(actual_field_goals_career, by = c("kicker_player_name")) %>%
	mutate(difference_percentage_career = ((actual_makes.y - expected_makes.y)/expected_makes.y))

	clutch_lm = lm(difference_percentage ~ difference_percentage_career, data=actual_field_goals_clutch)
	rsquared_clutch <- summary(clutch_lm)$r.squared
	rsquared_clutch <- signif(rsquared_clutch, digits = 2)
	rsquaredlabel_clutch <- expression(paste(" ",R^2 ,"= .17"))

	ggplot(actual_field_goals_clutch, aes(x = difference_percentage,difference_percentage_career)) +
	geom_point() +
	geom_point(data = subset(actual_field_goals_clutch, (kicker_player_name == "C.Parkey")),
	color = "orange",
	size = 2) +
	geom_smooth(method = lm) +
	annotate("text", x = .2, y= -.05,label = rsquaredlabel_clutch, size = 4, color = "black") +
	geom_abline(intercept = 0, slope = 1, lty = 2) +
	geom_vline(xintercept = 0, lty = 2) +
	geom_hline(yintercept = 0, lty = 2) +
	geom_text_repel(
	data = subset(actual_field_goals_clutch, (difference_percentage > .05)
	\| difference_percentage_career > .05
	\| difference_percentage < -.15
	\| difference_percentage < -.22),
	aes(label = kicker_player_name)) +
	geom_text_repel(
	data = subset(actual_field_goals_clutch, (kicker_player_name == "C.Parkey")),
	aes(label = kicker_player_name),
	color = "blue",
	nudge_y = .005,
	nudge_x = -.005) +
	theme_classic() +
	labs(
	x = "Clutch Kicks Above Expectation (%)",
	y = "All Kicks",
	title = "Clutch Kicking",
	subtitle = "2009-2018, Minimum 10 Attempts",
	caption = "Clutch Kick = <5 minutes remaining in Reg. or in OT in a 3 point game. \n Data from nflscrapR \n @ChiBearsStats"
	)
	ggsave('clutchkicks.png', dpi=1000)