mrcaseb

Code for Plots by @mrcaseb

• Preamble
• How to get started
• Completion Percentage Over Expectation (CPOE) along Depth of Target (DOT)
• Effect of Interceptions on Win Probability
• Next Gen Stats Receiving

mrcaseb

library(dplyr)
get_espn_wp_college <- function(espn_game_id) {
  espn_wp <- data.frame()
  tryCatch(
    expr = {
      espn_wp <-
        httr::GET(url = glue::glue("http://site.api.espn.com/apis/site/v2/sports/football/college-football/summary?event={espn_game_id}")) %>%
        httr::content(as = "text", encoding = "UTF-8") %>%
        jsonlite::fromJSON(flatten = TRUE) %>%
        purrr::pluck("winprobability") %>%

mrcaseb

library(tidyverse)

# Lee Sharpe's brightness function ----------------------------------------

brightness <- function(hex) {
  result <- rep(0, length(hex))
  for (i in 2:7)
  {
    ch <- substr(hex, i, i)
    result <- result + ifelse(i %% 2 == 0, 16, 1) * case_when(

mrcaseb

library(magrittr)

df <- tibble::tibble(
  var_a = c("a", "b", "c", "d"),
  var_b = c(1.234, 12.34, 123.4, 1234)
)

num_to_html <- function(num, separator = ".") {
  stringr::str_split_fixed(num, stringr::fixed(separator), n = 2) %>%
    as.data.frame() %>%

mrcaseb

library(dplyr)

st_plays <- nflreadr::load_pbp(2021) |>
  filter(!is.na(epa) & touchback != 1 & special == 1)

epa_kicking <- st_plays |>
  group_by(team = posteam) |>
  mutate(epa = dplyr::if_else(play_type == "kickoff", -epa, epa)) |>
  summarize(epa_kick = mean(epa), plays_kick = n())

mrcaseb

library(tidyverse)
library(nflverse)
#> ── Attaching packages ─────────────────────────────────── nflverse 1.0.1.9000 ──
#> ✓ nflfastR 4.3.0.9008     ✓ nflreadr 1.1.3     
#> ✓ nflseedR 1.0.2.9001     ✓ nflplotR 1.0.0.9000
#> ✓ nfl4th   1.0.1.9000
#> ──────────────────────────────────────────────────────────────── Ready to go! ──
options(dplyr.summarise.inform = FALSE)
options(nflreadr.verbose = FALSE)
s <- nflreadr::load_schedules(TRUE)

mrcaseb

library(dplyr, warn.conflicts = FALSE)
library(ggplot2)

preds <- nflreadr::csv_from_url("https://raw.githubusercontent.com/nflverse/nfldata/master/data/predictions.csv")

g <- nflreadr::load_schedules(2021)

points <- preds |> 
  filter(prediction != 50) |> 
  left_join(g |> select(game_id, week, result), by = "game_id") |> 

mrcaseb

library(tidyverse)

drafts <- nflreadr::load_draft_picks()

draft_order <- rvest::read_html("https://en.wikipedia.org/wiki/2022_NFL_Draft") |> 
  rvest::html_table() |> 
  purrr::pluck(5) |> 
  janitor::clean_names() |> 
  select(round = rnd, pick = pick_no, team = nfl_team, notes) |> 
  mutate(across(c(round, pick), as.numeric)) |> 

mrcaseb

#' Remove Bookmaker’s Vig from a Vector of Probablities

#' @description This function implements the iterative power method to adjust
#' implied probabilities suggested by Vovk and Zhadanov (2009) and Clarke (2016),
#' where bookmakers’ implied probabilities are raised to a fixed power, which
#' never produces bookmaker or fair probabilities outside the 0-1 range
#' (upper limit can be changed) and allows for the favorite long-shot bias.

#' @param probs A vector of probabilities from which to remove the bookmaker’s
#' vig. The probabilities will be adjusted to sum up to `sum_to`.

mrcaseb

load_538_games <- function(seasons = "SB_ERA"){
  s <- nflreadr::csv_from_url("https://projects.fivethirtyeight.com/nfl-api/nfl_elo.csv") |>
    tibble::as_tibble() |>
    dplyr::na_if("") |>
    dplyr::select(
      season,
      season_type = playoff,
      away_team = team2,
      away_score = score2,
      home_score = score1,