Skip to contents

Get started with nflfastR

Ben Baldwin & Sebastian Carl

Source: vignettes/nflfastR.Rmd
nflfastR.Rmd

If you are new to R or are having trouble understanding the code in the below sections we highly recommend the nflfastR beginner’s guide in vignette("beginners_guide").

The Main Functions

nflfastR comes with a set of functions to access NFL play-by-play data and team rosters. This section provides a brief introduction to the essential functions.

nflfastR processes and cleans up play-by-play data and adds variables through it’s models. Since some of these tasks are performed by separate functions, the easiest way to compute the complete nflfastR dataset is build_nflfastR_pbp(). The main input for that function is a set of game ids which can be accessed with fast_scraper_schedules(). The following code demonstrates how to build the nflfastR dataset for the Super Bowls of the 2017 - 2019 seasons.

library(nflfastR)
library(dplyr, warn.conflicts = FALSE)
ids <- nflfastR::fast_scraper_schedules(2017:2019) %>%
  dplyr::filter(game_type == "SB") %>%
  dplyr::pull(game_id)
pbp <- nflfastR::build_nflfastR_pbp(ids)
#> ── Build nflfastR Play-by-Play Data ───────────── nflfastR version 4.6.1.9008 ──
#> • 19:16:40 | Start download of 3 games...
#> ✔ 19:16:45 | Download finished. Adding variables...
#> ✔ 19:16:45 | added game variables
#> ✔ 19:16:46 | added nflscrapR variables
#> [19:16:46] WARNING: src/learner.cc:553: 
#>   If you are loading a serialized model (like pickle in Python, RDS in R) generated by
#>   older XGBoost, please export the model by calling `Booster.save_model` from that version
#>   first, then load it back in current version. See:
#> 
#>     https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html
#> 
#>   for more details about differences between saving model and serializing.
#> ✔ 19:16:46 | added ep variables
#> ✔ 19:16:46 | added air_yac_ep variables
#> [19:16:46] WARNING: src/learner.cc:553: 
#>   If you are loading a serialized model (like pickle in Python, RDS in R) generated by
#>   older XGBoost, please export the model by calling `Booster.save_model` from that version
#>   first, then load it back in current version. See:
#> 
#>     https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html
#> 
#>   for more details about differences between saving model and serializing.
#> 
#> [19:16:46] WARNING: src/learner.cc:553: 
#>   If you are loading a serialized model (like pickle in Python, RDS in R) generated by
#>   older XGBoost, please export the model by calling `Booster.save_model` from that version
#>   first, then load it back in current version. See:
#> 
#>     https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html
#> 
#>   for more details about differences between saving model and serializing.
#> ✔ 19:16:47 | added wp variables
#> ✔ 19:16:47 | added air_yac_wp variables
#> [19:16:47] WARNING: src/learner.cc:553: 
#>   If you are loading a serialized model (like pickle in Python, RDS in R) generated by
#>   older XGBoost, please export the model by calling `Booster.save_model` from that version
#>   first, then load it back in current version. See:
#> 
#>     https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html
#> 
#>   for more details about differences between saving model and serializing.
#> ✔ 19:16:47 | added cp and cpoe
#> ✔ 19:16:47 | added fixed drive variables
#> ✔ 19:16:47 | added series variables
#> • 19:16:47 | Cleaning up play-by-play...
#> ✔ 19:16:47 | Cleaning completed
#> ✔ 19:16:47 | added qb_epa
#> • 19:16:47 | Computing xyac...
#> [19:16:48] WARNING: src/learner.cc:553: 
#>   If you are loading a serialized model (like pickle in Python, RDS in R) generated by
#>   older XGBoost, please export the model by calling `Booster.save_model` from that version
#>   first, then load it back in current version. See:
#> 
#>     https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html
#> 
#>   for more details about differences between saving model and serializing.
#> ✔ 19:16:49 | added xyac variables
#> • 19:16:50 | Computing xpass...
#> [19:16:50] WARNING: src/learner.cc:553: 
#>   If you are loading a serialized model (like pickle in Python, RDS in R) generated by
#>   older XGBoost, please export the model by calling `Booster.save_model` from that version
#>   first, then load it back in current version. See:
#> 
#>     https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html
#> 
#>   for more details about differences between saving model and serializing.
#> ✔ 19:16:50 | added xpass and pass_oe
#> • 19:16:50 | Decode player ids...
#> ✔ 19:16:51 | Decoding of player ids completed
#> ── DONE ────────────────────────────────────────────────────────────────────────

In most cases, however, it is not necessary to use this function for individual games, because nflfastR provides both a data repository and two main play-by-play functions: load_pbp() and update_db(). We cover load_pbp() below, and please see Example 8: Using the built-in database function for how to work with the database function update_db().

The easiest way to access the data in the data repository is the new function load_pbp(). It can load multiple seasons directly into memory and supports multiple data formats. Loading all play-by-play data of the 2018-2020 seasons is as easy as

pbp <- nflfastR::load_pbp(2018:2020)

Joining roster data to the play-by-play data set is possible as well. The data can be accessed with the function fast_scraper_roster() and its application is demonstrated in Example 10: Working with roster and position data.

Application Examples

All examples listed below assume that the following two libraries are installed and loaded.

Example 1: replicate nflscrapR with fast_scraper

The functionality of nflscrapR can be duplicated by using fast_scraper(). This obtains the same information contained in nflscrapR (plus some extra) but much more quickly. To compare to nflscrapR, we use their data repository as the program no longer functions now that the NFL has taken down the old Gamecenter feed. Note that EP differs from nflscrapR as we use a newer era-adjusted model (more on this in this post on Open Source Football).

This example also uses the built-in function clean_pbp() to create a ‘name’ column for the primary player involved (the QB on pass play or ball-carrier on run play).

readr::read_csv("https://github.com/ryurko/nflscrapR-data/blob/master/play_by_play_data/regular_season/reg_pbp_2019.csv?raw=true") %>%
  dplyr::filter(home_team == "SF" & away_team == "SEA") %>%
  dplyr::select(desc, play_type, ep, epa, home_wp) %>%
  utils::head(6) %>%
  knitr::kable(digits = 3)
desc play_type ep epa home_wp
J.Myers kicks 65 yards from SEA 35 to end zone, Touchback. kickoff 0.815 0.000 NA
(15:00) T.Coleman left guard to SF 26 for 1 yard (J.Clowney). run 0.815 -0.606 0.500
(14:19) T.Coleman right tackle to SF 25 for -1 yards (P.Ford). run 0.209 -1.146 0.485
(13:45) (Shotgun) J.Garoppolo pass short middle to K.Bourne to SF 41 for 16 yards (J.Taylor). Caught at SF39. 2-yac pass -0.937 3.223 0.453
(12:58) PENALTY on SEA-J.Reed, Encroachment, 5 yards, enforced at SF 41 - No Play. no_play 2.286 0.774 0.551
(12:38) J.Garoppolo pass incomplete short left to M.Goodwin [J.Clowney]. pass 3.060 -0.698 0.577 
nflfastR::fast_scraper("2019_10_SEA_SF") %>%
  nflfastR::clean_pbp() %>%
  dplyr::select(desc, play_type, ep, epa, home_wp, name) %>%
  utils::head(6) %>%
  knitr::kable(digits = 3)
desc play_type ep epa home_wp name
GAME NA 1.474 0.000 0.546 NA
5-J.Myers kicks 65 yards from SEA 35 to end zone, Touchback. kickoff 1.474 0.000 0.546 NA
(15:00) 26-T.Coleman left guard to SF 26 for 1 yard (90-J.Clowney). run 1.474 -0.554 0.546 T.Coleman
(14:19) 26-T.Coleman right tackle to SF 25 for -1 yards (97-P.Ford). run 0.920 -0.814 0.528 T.Coleman
(13:45) (Shotgun) 10-J.Garoppolo pass short middle to 84-K.Bourne to SF 41 for 16 yards (24-J.Taylor). Caught at SF39. 2-yac pass 0.107 2.427 0.498 J.Garoppolo
(12:58) PENALTY on SEA-91-J.Reed, Encroachment, 5 yards, enforced at SF 41 - No Play. no_play 2.534 0.600 0.573 NA

Example 2: scrape a batch of games very quickly with fast_scraper

This is a demonstration of nflfastR’s capabilities. While nflfastR can scrape a batch of games very quickly, please be respectful of Github’s servers and use the data repository which hosts all the scraped and cleaned data whenever possible. The only reason to ever actually use the scraper is if it’s in the middle of the season and we haven’t updated the repository with recent games (but it is automatically updated overnight every day).

# get list of some games from 2019
games_2019 <- nflfastR::fast_scraper_schedules(2019) %>%
  utils::head(10) %>%
  dplyr::pull(game_id)
tictoc::tic(glue::glue("{length(games_2019)} games with nflfastR:"))
f <- nflfastR::fast_scraper(games_2019)
tictoc::toc()
#> 10 games with nflfastR:: 7.604 sec elapsed

Example 3: Completion Percentage Over Expected (CPOE)

Let’s look at CPOE leaders from the 2009 regular season.

As discussed above, nflfastR has a data repository for old seasons, so there’s no need to actually scrape them. Let’s use that here with the convenience function load_pbp() which fetches data from the repository (for non-R users, .csv and .parquet are also available in the data repository).

tictoc::tic("loading all games from 2009")
games_2009 <- nflfastR::load_pbp(2009) %>% dplyr::filter(season_type == "REG")
tictoc::toc()
#> loading all games from 2009: 2.128 sec elapsed
games_2009 %>%
  dplyr::filter(!is.na(cpoe)) %>%
  dplyr::group_by(passer_player_name) %>%
  dplyr::summarize(cpoe = mean(cpoe), Atts = n()) %>%
  dplyr::filter(Atts > 200) %>%
  dplyr::arrange(-cpoe) %>%
  utils::head(5) %>%
  knitr::kable(digits = 1)
passer_player_name cpoe Atts
D.Brees 7.5 509
P.Rivers 6.6 474
P.Manning 6.5 569
B.Favre 6.1 527
B.Roethlisberger 5.4 503

Example 4: Using Drive Information

When working with nflfastR, drive results are automatically included. We use fixed_drive and fixed_drive_result since the NFL-provided information is a bit wonky. Let’s look at how much more likely teams were to score starting from 1st & 10 at their own 20 yard line in 2015 (the last year before touchbacks on kickoffs changed to the 25) than in 2000.

pbp <- nflfastR::load_pbp(c(2003, 2015))

out <- pbp %>%
  dplyr::filter(season_type == "REG" & down == 1 & ydstogo == 10 & yardline_100 == 80) %>%
  dplyr::mutate(drive_score = dplyr::if_else(fixed_drive_result %in% c("Touchdown", "Field goal"), 1, 0)) %>%
  dplyr::group_by(season) %>%
  dplyr::summarize(drive_score = mean(drive_score))

out %>% 
  knitr::kable(digits = 3)
season drive_score
2003 0.206
2015 0.305

So 20.6% of 1st & 10 plays from teams’ own 20 would see the drive end up in a score in 2003, compared to 30.5% in 2015. This has implications for Expected Points models (see this article).

Example 5: Plot offensive and defensive EPA per play for a given season

Let’s build the NFL team tiers using offensive and defensive expected points added per play for the 2005 regular season. Creating data viz including NFL team logos (or wordmarks, or headshots), we recommend the nflverse R package nflplotR.

When using load_pbp(), the helper function clean_pbp() has already been run, which creates “rush” and “pass” columns that (a) properly count sacks and scrambles as pass plays and (b) properly include plays with penalties. Using this, we can keep only rush or pass plays.

library(nflplotR)
pbp <- nflfastR::load_pbp(2005) %>%
  dplyr::filter(season_type == "REG") %>%
  dplyr::filter(!is.na(posteam) & (rush == 1 | pass == 1))
offense <- pbp %>%
  dplyr::group_by(team = posteam) %>%
  dplyr::summarise(off_epa = mean(epa, na.rm = TRUE))
defense <- pbp %>%
  dplyr::group_by(team = defteam) %>%
  dplyr::summarise(def_epa = mean(epa, na.rm = TRUE))
offense %>%
  dplyr::inner_join(defense, by = "team") %>%
  ggplot2::ggplot(aes(x = off_epa, y = def_epa)) +
  ggplot2::geom_abline(slope = -1.5, intercept = c(.4, .3, .2, .1, 0, -.1, -.2, -.3), alpha = .2) +
  nflplotR::geom_mean_lines(aes(y0 = off_epa, x0 = def_epa)) +
  nflplotR::geom_nfl_logos(aes(team_abbr = team), width = 0.07, alpha = 0.7) +
  ggplot2::labs(
    x = "Offense EPA/play",
    y = "Defense EPA/play",
    caption = "Data: @nflfastR",
    title = "2005 NFL Offensive and Defensive EPA per Play"
  ) +
  ggplot2::theme_bw() +
  ggplot2::theme(
    plot.title = ggplot2::element_text(size = 12, hjust = 0.5, face = "bold")
  ) +
  ggplot2::scale_y_reverse()

Example 6: Expected Points calculator

We have provided a calculator for working with the Expected Points model. Here is an example of how to use it, looking for how the Expected Points on a drive beginning following a touchback has changed over time.

While I have put in 'SEA' for home_team and posteam, this only matters for figuring out whether the team with the ball is the home team (there’s no actual effect for given team; it would be the same no matter what team is supplied).

data <- tibble::tibble(
  "season" = 1999:2019,
  "home_team" = "SEA",
  "posteam" = "SEA",
  "roof" = "outdoors",
  "half_seconds_remaining" = 1800,
  "yardline_100" = c(rep(80, 17), rep(75, 4)),
  "down" = 1,
  "ydstogo" = 10,
  "posteam_timeouts_remaining" = 3,
  "defteam_timeouts_remaining" = 3
)
nflfastR::calculate_expected_points(data) %>%
  dplyr::select(season, yardline_100, td_prob, ep) %>%
  knitr::kable(digits = 2)
season yardline_100 td_prob ep
1999 80 0.33 0.64
2000 80 0.33 0.64
2001 80 0.33 0.64
2002 80 0.34 0.82
2003 80 0.34 0.82
2004 80 0.34 0.82
2005 80 0.34 0.82
2006 80 0.34 0.81
2007 80 0.34 0.81
2008 80 0.34 0.81
2009 80 0.34 0.81
2010 80 0.34 0.81
2011 80 0.34 0.81
2012 80 0.34 0.81
2013 80 0.34 0.81
2014 80 0.35 0.98
2015 80 0.35 0.98
2016 75 0.38 1.46
2017 75 0.38 1.46
2018 75 0.41 1.47
2019 75 0.41 1.47

Not surprisingly, offenses have become much more successful over time, with the kickoff touchback moving from the 20 to the 25 in 2016 providing an additional boost. Note that the td_prob in this example is the probability that the next score within the same half will be a touchdown scored by team with the ball, not the probability that the current drive will end in a touchdown (this is why the numbers are different from Example 4 above).

We could compare the most recent four years to the expectation for playing in a dome by inputting all the same things and changing the roof input:

data <- tibble::tibble(
  "season" = 2016:2019,
  "week" = 5,
  "home_team" = "SEA",
  "posteam" = "SEA",
  "roof" = "dome",
  "half_seconds_remaining" = 1800,
  "yardline_100" = c(rep(75, 4)),
  "down" = 1,
  "ydstogo" = 10,
  "posteam_timeouts_remaining" = 3,
  "defteam_timeouts_remaining" = 3
)
nflfastR::calculate_expected_points(data) %>%
  dplyr::select(season, yardline_100, td_prob, ep) %>%
  knitr::kable(digits = 2)
season yardline_100 td_prob ep
2016 75 0.41 1.81
2017 75 0.41 1.81
2018 75 0.44 1.84
2019 75 0.44 1.84

So for 2018 and 2019, 1st & 10 from a home team’s own 25 yard line had higher EP in domes than at home, which is to be expected.

Example 7: Win probability calculator

We have also provided a calculator for working with the win probability models. Here is an example of how to use it, looking for how the win probability to begin the game depends on the pre-game spread.

While I have put in 'SEA' for home_team and posteam, this only matters for figuring out whether the team with the ball is the home team (there’s no actual effect for given team; it would be the same no matter what team is supplied).

data <- tibble::tibble(
  "receive_2h_ko" = 0,
  "home_team" = "SEA",
  "posteam" = "SEA",
  "score_differential" = 0,
  "half_seconds_remaining" = 1800,
  "game_seconds_remaining" = 3600,
  "spread_line" = c(1, 3, 4, 7, 14),
  "down" = 1,
  "ydstogo" = 10,
  "yardline_100" = 75,
  "posteam_timeouts_remaining" = 3,
  "defteam_timeouts_remaining" = 3
)
nflfastR::calculate_win_probability(data) %>%
  dplyr::select(spread_line, wp, vegas_wp) %>%
  knitr::kable(digits = 2)
spread_line wp vegas_wp
1 0.55 0.51
3 0.55 0.60
4 0.55 0.64
7 0.55 0.74
14 0.55 0.87

Not surprisingly, vegas_wp increases with the amount a team was coming into the game favored by.

Example 8: Using the built-in database function

If you’re comfortable using dplyr functions to manipulate and tidy data, you’re ready to use a database. Why should you use a database?

  • The provided function in nflfastR makes it extremely easy to build a database and keep it updated
  • Play-by-play data over 20+ seasons takes up a lot of memory: working with a database allows you to only bring into memory what you actually need
  • R makes it extremely easy to work with databases.

Start: install and load packages

To start, we need to install the two packages required for this that aren’t installed automatically when nflfastR installs: DBI and RSQLite (advanced users can use other types of databases, but this example will use SQLite). The if statements make sure the packages won’t be updated if they are already installed:

As with always, you only need to install these once. They don’t need to be loaded to build the database because nflfastR knows how to use them, but we do need them later on when working with the database.

Build database

There’s exactly one function in nflfastR that works with databases: update_db(). Some notes:

  • If you use update_db() with no arguments, it will build a SQLite database called pbp_db in your current working directory, with play-by-play data in a table called nflfastR_pbp.
  • You can specify a different directory with dbdir.
  • You can specify a different filename with dbname.
  • You can specify a different table name with tblname.
  • If you want to rebuild the database from scratch for whatever reason, supply force_rebuild = TRUE. This is primarily intended for the case when we update the play-by-play data in the data repo due to fixing a bug and you want to force the database to be wiped and updated.
  • If you want to rebuild specified seasons, this can also be supplied to force_rebuild (e.g. force_rebuild = c(2019, 2020)).
  • The parameter db_connection is intended for advanced users who want to use other DBI drivers, such as MariaDB, Postgres or odbc. Please note that dbdir and dbname are dropped when a db_connection is provided but the argument tblname will still be used to write the data table into the database.

Let’s say I just want to dump a database into the current working directory. Here we go!

nflfastR::update_db()
#> ── Update nflfastR Play-by-Play Database ──────── nflfastR version 4.6.1.9008 ──
#> ℹ 19:17:34 | Can't find the data table "nflfastR_pbp"
#> in your database. Will load the play by play data from
#> scratch.
#> • 19:17:34 | Starting download of 25 seasons between 1999 and 2023...
#> • 19:18:52 | Checking for missing completed games...
#> ℹ 19:18:54 | You have 6703 games and are missing 0.
#> ✔ 19:18:54 | Database update completed
#> ℹ 19:18:54 | Path to your db: ./pbp_db
#> ── DONE ────────────────────────────────────────────────────────────────────────

This created a database in the current directory called pbp_db.

Wait, that’s it? That’s it! What if it’s partway through the season and you want to make sure all the new games are added to the database? What do you run? update_db()! (just make sure you’re in the directory the database is saved in or you supply the right file path)

nflfastR::update_db()
#> ── Update nflfastR Play-by-Play Database ──────── nflfastR version 4.6.1.9008 ──
#> • 19:18:54 | Checking for missing completed games...
#> ℹ 19:18:55 | You have 6703 games and are missing 0.
#> ✔ 19:18:56 | Database update completed
#> ℹ 19:18:56 | Path to your db: /home/runner/work/nflfastR/nflfastR/vignettes/pbp_db
#> ── DONE ────────────────────────────────────────────────────────────────────────

If it’s partway through a season and you want to re-build a season to allow for data corrections from the NFL to propagate into your database, you can specify one season to be rebuilt:

nflfastR::update_db(force_rebuild = 2020)
#> ── Update nflfastR Play-by-Play Database ──────── nflfastR version 4.6.1.9008 ──
#> • 19:18:56 | Purging season 2020 from the data table "nflfastR_pbp" in your
#> connected database...
#> • 19:18:57 | Starting download of the 1 season 2020
#> • 19:19:00 | Checking for missing completed games...
#> ℹ 19:19:00 | You have 6703 games and are missing 0.
#> ✔ 19:19:01 | Database update completed
#> ℹ 19:19:01 | Path to your db: /home/runner/work/nflfastR/nflfastR/vignettes/pbp_db
#> ── DONE ────────────────────────────────────────────────────────────────────────

Connect to database

Now we can make a connection to the database. This is the only part that will look a little bit foreign, but all you need to know is where your database is located. If it’s in your current working directory, this will work:

connection <- DBI::dbConnect(RSQLite::SQLite(), "./pbp_db")
connection
#> <SQLiteConnection>
#>   Path: /home/runner/work/nflfastR/nflfastR/vignettes/pbp_db
#>   Extensions: TRUE

It looks like nothing happened, but we now have a connection to the database. Now we’re ready to do stuff. If you aren’t familiar with databases, they’re organized around tables. Here’s how to see which tables are present in our database:

DBI::dbListTables(connection)
#> [1] "nflfastR_pbp"

Since we went with the defaults, there’s a table called nflfastR_pbp. Another useful function is to see the fields (i.e., columns) in a table:

DBI::dbListFields(connection, "nflfastR_pbp") %>%
  utils::head(10)
#>  [1] "play_id"      "game_id"      "old_game_id"  "home_team"    "away_team"   
#>  [6] "season_type"  "week"         "posteam"      "posteam_type" "defteam"

This is the same list as the list of columns in nflfastR play-by-play. Notice we had to supply the name of the table above ("nflfastR_pbp").

With all that out of the way, there’s only a couple more things to learn. The main driver here is tbl, which helps get output with a specific table in a database:

pbp_db <- dplyr::tbl(connection, "nflfastR_pbp")

And now, everything will magically just “work”: you can forget you’re even working with a database!

pbp_db %>%
  dplyr::group_by(season) %>%
  dplyr::summarize(n = dplyr::n())
#> # Source:   SQL [?? x 2]
#> # Database: sqlite 3.45.0 [/home/runner/work/nflfastR/nflfastR/vignettes/pbp_db]
#>    season     n
#>     <int> <int>
#>  1   1999 46136
#>  2   2000 45491
#>  3   2001 44969
#>  4   2002 47355
#>  5   2003 46810
#>  6   2004 46706
#>  7   2005 46823
#>  8   2006 46299
#>  9   2007 46266
#> 10   2008 45917
#> # ℹ more rows
pbp_db %>%
  dplyr::filter(rush == 1 | pass == 1, down <= 2, !is.na(epa), !is.na(posteam)) %>%
  dplyr::group_by(pass) %>%
  dplyr::summarize(mean_epa = mean(epa, na.rm = TRUE))
#> # Source:   SQL [2 x 2]
#> # Database: sqlite 3.45.0 [/home/runner/work/nflfastR/nflfastR/vignettes/pbp_db]
#>    pass mean_epa
#>   <dbl>    <dbl>
#> 1     0  -0.0977
#> 2     1   0.0733

So far, everything has stayed in the database. If you want to bring a query into memory, just use collect() at the end:

russ <- pbp_db %>%
  dplyr::filter(name == "R.Wilson" & posteam == "SEA") %>%
  dplyr::select(desc, epa) %>%
  dplyr::collect()
russ
#> # A tibble: 6,946 × 2
#>    desc                                                                      epa
#>    <chr>                                                                   <dbl>
#>  1 (14:12) 3-R.Wilson pass short right to 18-S.Rice to SEA 34 for 9 yar…  1.13  
#>  2 (12:53) 3-R.Wilson pass incomplete deep left to 18-S.Rice. PENALTY o…  2.68  
#>  3 (11:25) (Shotgun) 3-R.Wilson pass incomplete short right to 18-S.Ric… -1.31  
#>  4 (10:24) (Shotgun) 3-R.Wilson pass short left to 18-S.Rice to ARI 31 …  0.928 
#>  5 (9:47) 3-R.Wilson scrambles right end ran ob at ARI 27 for 4 yards (… -0.0194
#>  6 (8:35) 3-R.Wilson pass incomplete short right to 18-S.Rice.           -0.426 
#>  7 (7:54) (Shotgun) 3-R.Wilson left end pushed ob at ARI 9 for 4 yards … -1.17  
#>  8 (:27) 3-R.Wilson sacked at SEA 17 for -5 yards (51-P.Lenon). Penalty… -1.13  
#>  9 (14:28) (Shotgun) 3-R.Wilson pass short right to 17-B.Edwards to SEA…  1.94  
#> 10 (13:59) 3-R.Wilson pass incomplete deep left to 87-B.Obomanu.         -0.453 
#> # ℹ 6,936 more rows

So we’ve searched through about 1 million rows of data across 300+ columns and only brought about 6950 rows and two columns into memory. Pretty neat! This is how I supply the data to the shiny apps on rbsdm.com without running out of memory on the server. Now there’s only one more thing to remember. When you’re finished doing what you need with the database:

DBI::dbDisconnect(connection)

For more details on using a database with nflfastR, see Thomas Mock’s life-changing post here. More detailed information on dbplyr (the dplyr database back-end) are given in the second edition of Hadley Wickham’s R for Data Science (2e).

Example 9: working with the expected yards after catch model

The variables in xyac are as follows:

  • xyac_epa: The expected value of EPA gained after the catch, starting from where the catch was made.
  • xyac_success: The probability the play earns positive EPA (relative to where play started) based on where ball was caught.
  • xyac_fd: Probability play earns a first down based on where the ball was caught.
  • xyac_mean_yardage and xyac_median_yardage: Average and median expected yards after the catch based on where the ball was caught.

Some other notes:

  • epa = air_epa + yac_epa, where air_epa is the EPA associated with a catch at the target location. If a receiver loses a fumble, it is removed from his yac_epa
  • Expected value of EPA at catch point = air_epa + xyac_epa
  • So if we want to get YAC EPA over expected, we need to compare yac_epa to xyac_epa, as in the example below
  • To get first downs over expected, we could compare first_down to xyac_fd
  • These fields are populated for all pass attempts, whether caught or not, but restrict to completed passes when measuring, for example, YAC EPA over expected
  • The expected YAC EPA model doesn’t take receiver fumbles into account, so actual minus expected YAC is slightly negative due to fumbles happening

Let’s create measures for EPA and first downs over expected in 2015:

nflfastR::load_pbp(2015) %>%
  dplyr::group_by(receiver, receiver_id, posteam) %>%
  dplyr::mutate(tgt = sum(complete_pass + incomplete_pass)) %>%
  dplyr::filter(tgt >= 50) %>%
  dplyr::filter(complete_pass == 1, air_yards < yardline_100, !is.na(xyac_epa)) %>%
  dplyr::summarize(
    epa_oe = mean(yac_epa - xyac_epa),
    actual_fd = mean(first_down),
    expected_fd = mean(xyac_fd),
    fd_oe = mean(first_down - xyac_fd),
    rec = dplyr::n()
  ) %>%
  dplyr::ungroup() %>%
  dplyr::select(receiver, posteam, actual_fd, expected_fd, fd_oe, epa_oe, rec) %>%
  dplyr::arrange(-epa_oe) %>%
  utils::head(10) %>%
  knitr::kable(digits = 3)
receiver posteam actual_fd expected_fd fd_oe epa_oe rec
D.Johnson ARI 0.500 0.391 0.109 0.334 50
J.White NE 0.489 0.434 0.055 0.264 47
T.Ginn CAR 0.800 0.734 0.066 0.249 45
D.Lewis NE 0.472 0.309 0.163 0.238 36
L.Green LAC 0.629 0.526 0.103 0.216 35
O.Beckham Jr. NYG 0.692 0.706 -0.014 0.207 91
G.Bernard CIN 0.373 0.289 0.083 0.204 51
T.Riddick DET 0.400 0.304 0.096 0.203 80
D.Woodhead LAC 0.468 0.354 0.114 0.172 77
T.Lockett SEA 0.588 0.548 0.040 0.163 51

The presence of so many running backs on this list suggests that even though it takes into account target depth and pass direction, the model doesn’t do a great job capturing space. Alternatively, running backs might be better at generating yards after the catch since running with the football is their primary role.

Example 10: Working with roster and position data

At long last, there’s a way to merge the new play-by-play data with roster information. Use the function to get the rosters:

roster <- nflfastR::fast_scraper_roster(2019)

Now let’s load play-by-play data from 2019:

games_2019 <- nflfastR::load_pbp(2019)

Here is what the player IDs look like because nflfastR now automatically decodes IDs to look like the old format with GSIS IDs:

games_2019 %>%
  dplyr::filter(rush == 1 | pass == 1, posteam == "SEA") %>%
  dplyr::select(name, id)
#> ── nflverse play by play data ──────────────────────────────────────────────────
#> ℹ Data updated: 2024-03-07 14:39:28 UTC
#> # A tibble: 1,207 × 2
#>    name     id        
#>    <chr>    <chr>     
#>  1 C.Carson 00-0033594
#>  2 R.Wilson 00-0029263
#>  3 R.Wilson 00-0029263
#>  4 C.Carson 00-0033594
#>  5 R.Wilson 00-0029263
#>  6 C.Carson 00-0033594
#>  7 R.Wilson 00-0029263
#>  8 C.Carson 00-0033594
#>  9 R.Wilson 00-0029263
#> 10 R.Wilson 00-0029263
#> # ℹ 1,197 more rows

Now we’re ready to join to the roster data using these IDs:

joined <- games_2019 %>%
  dplyr::filter(!is.na(receiver_id)) %>%
  dplyr::select(posteam, season, desc, receiver, receiver_id, epa) %>%
  dplyr::left_join(roster, by = c("receiver_id" = "gsis_id"))
# the real work is done, this just makes a table and has it look nice
joined %>%
  dplyr::filter(position %in% c("WR", "TE", "RB")) %>%
  dplyr::group_by(receiver_id, receiver, position) %>%
  dplyr::summarize(tot_epa = sum(epa), n = n()) %>%
  dplyr::arrange(-tot_epa) %>%
  dplyr::ungroup() %>%
  dplyr::group_by(position) %>%
  dplyr::mutate(position_rank = 1:n()) %>%
  dplyr::filter(position_rank <= 5) %>%
  dplyr::rename(Pos_Rank = position_rank, Player = receiver, Pos = position, Tgt = n, EPA = tot_epa) %>%
  dplyr::select(Player, Pos, Pos_Rank, Tgt, EPA) %>%
  knitr::kable(digits = 0)
Player Pos Pos_Rank Tgt EPA
M.Thomas WR 1 199 105
T.Kelce TE 1 179 100
C.Godwin WR 2 123 87
D.Adams WR 3 161 77
T.Lockett WR 4 140 76
J.Jones WR 5 164 72
G.Kittle TE 2 129 56
C.McCaffrey RB 1 148 52
D.Waller TE 3 124 45
A.Ekeler RB 2 113 44
J.Cook TE 4 75 43
Z.Ertz TE 5 147 42
J.White RB 3 105 27
D.Cook RB 4 77 26
K.Juszczyk RB 5 28 23

Not surprisingly, all 5 of the top 5 WRs in terms of EPA added come in ahead of the top RB. Note that the number of targets won’t match official stats because we’re including plays with penalties.

Example 11: Replicating official stats

The columns like name, passer, fantasy etc are nflfastR-created columns that mimic “real” football: i.e., excluding plays with spikes, counting scrambles and sacks as pass plays, etc. But if you’re trying to replicate official statistics – perhaps for fantasy purposes – use the *_player_name and *_player_id columns.

Leaderboards

Let’s try to replicate this page of passing leaders.

nflfastR::load_pbp(2020) %>%
  dplyr::filter(season_type == "REG", complete_pass == 1 | incomplete_pass == 1 | interception == 1, !is.na(down)) %>%
  dplyr::group_by(passer_player_name, posteam) %>%
  dplyr::summarize(
    yards = sum(passing_yards, na.rm = T),
    tds = sum(touchdown == 1 & td_team == posteam),
    ints = sum(interception),
    att = dplyr::n()
  ) %>%
  dplyr::arrange(-yards) %>%
  utils::head(10) %>%
  knitr::kable(digits = 0)
passer_player_name posteam yards tds ints att
D.Watson HOU 4823 33 7 544
P.Mahomes KC 4740 38 6 588
T.Brady TB 4633 40 12 610
M.Ryan ATL 4581 26 11 626
J.Allen BUF 4544 37 10 572
J.Herbert LAC 4336 31 10 595
A.Rodgers GB 4299 48 5 526
K.Cousins MIN 4265 35 13 516
R.Wilson SEA 4212 40 13 558
P.Rivers IND 4169 24 11 543

These match the official stats on NFL.com (note the filter for season_type == "REG" since official stats only count regular season games). Note that we’re using passing_yards here because yards_gained is not equal to passing yards on plays with laterals.

While the above works, we’ve also provided a function that does this all for you: calculate_player_stats(). This function takes an nflfastR play-by-play dataframe as an input along with one other argument, weekly, which defaults to FALSE. When weekly is true, a week-by-week dataframe is returned (rather than an aggregate over the whole provided dataframe). Let’s again replicate the top 10 players in passing yards:

nflfastR::load_pbp(2020) %>%
  dplyr::filter(season_type == "REG") %>%
  nflfastR::calculate_player_stats() %>%
  dplyr::arrange(-passing_yards) %>%
  dplyr::select(player_name, recent_team, completions, attempts, passing_yards, passing_tds, interceptions) %>%
  utils::head(10) %>%
  knitr::kable(digits = 0)
player_name recent_team completions attempts passing_yards passing_tds interceptions
D.Watson HOU 382 544 4823 33 7
P.Mahomes KC 390 588 4740 38 6
T.Brady TB 401 610 4633 40 12
M.Ryan ATL 407 626 4581 26 11
J.Allen BUF 396 572 4544 37 10
J.Herbert LAC 396 595 4336 31 10
A.Rodgers GB 372 526 4299 48 5
K.Cousins MIN 349 516 4265 35 13
R.Wilson SEA 384 558 4212 40 13
P.Rivers IND 369 543 4169 24 11

We can do the same for rush attempts to replicate the NFL leaderboard:

nflfastR::load_pbp(2020) %>%
  dplyr::filter(season_type == "REG") %>%
  nflfastR::calculate_player_stats() %>%
  dplyr::arrange(-rushing_yards) %>%
  dplyr::select(player_name, recent_team, carries, rushing_yards, rushing_tds, rushing_fumbles_lost) %>%
  utils::head(10) %>%
  knitr::kable(digits = 0)
player_name recent_team carries rushing_yards rushing_tds rushing_fumbles_lost
D.Henry TEN 378 2027 17 2
D.Cook MIN 312 1557 16 2
J.Taylor IND 232 1169 11 1
A.Jones GB 201 1104 9 0
D.Montgomery CHI 247 1070 8 0
J.Robinson JAX 240 1070 7 0
N.Chubb CLE 190 1067 12 1
J.Jacobs LV 273 1065 12 2
L.Jackson BAL 159 1005 7 2
M.Gordon DEN 215 986 9 2

Again, this matches up exactly.

Yards from scrimmage

What if we want total yards from scrimmage? We’ll demonstrate three methods here. The hardest way is to use the fantasy_player_name column, which is the rusher on rush plays and receiver on receiving plays:

nflfastR::load_pbp(2020) %>%
  dplyr::filter(season_type == "REG", !is.na(down)) %>%
  dplyr::group_by(fantasy_player_name, posteam) %>%
  dplyr::summarize(
    carries = sum(rush_attempt),
    receptions = sum(complete_pass),
    touches = sum(rush_attempt + complete_pass),
    yards = sum(yards_gained),
    tds = sum(touchdown == 1 & td_team == posteam)
  ) %>%
  dplyr::arrange(-yards) %>%
  utils::head(10) %>%
  knitr::kable(digits = 0)
fantasy_player_name posteam carries receptions touches yards tds
D.Henry TEN 378 19 397 2141 17
D.Cook MIN 312 44 356 1918 17
A.Kamara NO 187 83 270 1688 21
S.Diggs BUF 1 127 128 1536 8
D.Montgomery CHI 247 54 301 1508 10
J.Taylor IND 232 36 268 1468 12
A.Jones GB 201 47 248 1459 11
T.Kelce KC 0 105 105 1416 11
J.Robinson JAX 240 49 289 1414 10
D.Hopkins ARI 1 115 116 1408 6

Looking at the PFR scrimmage stats, these columns are an exact match.

But we could also just use calculate_player_stats() again:

nflfastR::load_pbp(2020) %>%
  dplyr::filter(season_type == "REG") %>%
  nflfastR::calculate_player_stats() %>%
  dplyr::mutate(
    yards = rushing_yards + receiving_yards, 
    touches = carries + receptions, 
    tds = rushing_tds + receiving_tds
  ) %>%
  dplyr::arrange(-yards) %>%
  dplyr::select(player_name, recent_team, carries, receptions, touches, yards, tds) %>%
  utils::head(10) %>%
  knitr::kable(digits = 0)
player_name recent_team carries receptions touches yards tds
D.Henry TEN 378 19 397 2141 17
D.Cook MIN 312 44 356 1918 17
A.Kamara NO 187 83 270 1688 21
S.Diggs BUF 1 127 128 1536 8
D.Montgomery CHI 247 54 301 1508 10
J.Taylor IND 232 36 268 1468 12
A.Jones GB 201 47 248 1459 11
T.Kelce KC 0 105 105 1416 11
J.Robinson JAX 240 49 289 1414 10
D.Hopkins ARI 1 115 116 1408 6

And we get the same thing.

The third way is to use the load_player_stats() function, which can load a data frame of player-level stats for every week since 1999.

nflfastR::load_player_stats(seasons = 2020) %>%
  dplyr::filter(season_type == "REG") %>%
  dplyr::group_by(player_id) %>%
  dplyr::summarize(
    player_name = dplyr::first(player_name),
    recent_team = dplyr::first(recent_team),
    yards = sum(rushing_yards + receiving_yards),
    touches = sum(carries + receptions),
    carries = sum(carries),
    receptions = sum(receptions),
    tds = sum(rushing_tds + receiving_tds)
  ) %>%
  dplyr::ungroup() %>%
  dplyr::arrange(-yards) %>%
  dplyr::select(player_name, recent_team, carries, receptions, touches, yards, tds) %>%
  utils::head(10) %>%
  knitr::kable(digits = 0)
player_name recent_team carries receptions touches yards tds
D.Henry TEN 378 19 397 2141 17
D.Cook MIN 312 44 356 1918 17
A.Kamara NO 187 83 270 1688 21
S.Diggs BUF 1 127 128 1536 8
D.Montgomery CHI 247 54 301 1508 10
J.Taylor IND 232 36 268 1468 12
A.Jones GB 201 47 248 1459 11
T.Kelce KC 0 105 105 1416 11
J.Robinson JAX 240 49 289 1414 10
D.Hopkins ARI 1 115 116 1408 6

And again the output is identical.

Fantasy points

Let’s calculate PPR fantasy points per game in the first 16 weeks of the season among wide receivers who appeared in more than 5 games.

nflfastR::load_pbp(2020) %>%
  dplyr::filter(week <= 16) %>%
  nflfastR::calculate_player_stats() %>%
  dplyr::mutate(
    ppg = fantasy_points_ppr / games
  ) %>%
  dplyr::filter(games > 5) %>%
  # only keep the WRs
  dplyr::inner_join(
    nflfastR::fast_scraper_roster(2020) %>% 
      dplyr::filter(position == "WR") %>% 
      dplyr::select(player_id = gsis_id),
    by = "player_id"
  ) %>%
  dplyr::arrange(-ppg) %>%
  dplyr::select(player_name, recent_team, games, fantasy_points_ppr, ppg) %>%
  utils::head(10) %>%
  knitr::kable(digits = 1)
player_name recent_team games fantasy_points_ppr ppg
D.Adams GB 13 341.8 26.3
T.Hill KC 15 328.9 21.9
S.Diggs BUF 15 314.0 20.9
C.Ridley ATL 14 270.9 19.3
D.Hopkins ARI 15 280.3 18.7
D.Metcalf SEA 15 268.2 17.9
K.Allen LAC 14 245.1 17.5
A.Thielen MIN 14 244.3 17.4
W.Fuller HOU 11 188.9 17.2
A.Robinson CHI 15 257.2 17.1

Comparing to the FantasyPros website, this is an exact match.

Frequent issues

The drive column looks wacky

Use fixed_drive and fixed_drive_result instead. See Example 4: Using Drive Information.

Why are there so many win probability columns?

vegas_wp and vegas_home_wp incorporate the pregame spread and are much better models.

I’m trying to do X. Help!

Please ask in the Discord channel.

This section is a helper that holds the hyperlinks to the above chapters. It’s a workaround for the missing sections anchor bug in pkgdown which will hopefully be fixed with this pull request at some point in the future.