Baseball
Add news
News

Massive Expansion of Play-by-Play and Stats on Pro Football Reference

0 0

Ever wondered how effective Jerry Rice was with Joe Montana throwing him the ball? How John Elway performed in the 4th quarter? Who the best short-yardage rusher was in the 1980s? What defense could never get off the field on 3rd down in the late '70s? No? Well, we're happy to announce that we've got answers for you anyways! We are thrilled to announce that Pro Football Reference now has play-by-play and related statistics/splits for all regular-season and playoffs games (with the exception of four minutes and 47 seconds of a single game) all the way back to 1978, which was the first season of the NFL's 16-game schedule era. This sort of comprehensive, digitized play-by-play is not available anywhere else in the public realm. We are very proud to be able to democratize this data and present it to our users.

First of all, we would like to credit Aaron Schatz of FTN Fantasy and Jeremy Snyder of Quirky Research for partnering with us on this initiative. The play-by-play was all generated by Schatz and Snyder, who have also used it to generate the sort of stats Schatz has been famous for since his Football Outsiders days (like DVOA), which can now be found at FTN.

Previous to this expansion, our play-by-play coverage only went back to 1994, so this update represents 16 new seasons of coverage and means we have nearly 50 seasons worth of play-by-play coverage and related data on Pro Football Reference now (in addition to play-by-play of every Super Bowl, which we have had for many years). We'd like to run through some of the places where this new data is apparent. Previously these statistics were stocked back to 1994 (or 1992 in the case of targets), but they now all go back to 1978:

Some important things to note about this data:

  • The numbers you can derive from this PxP line up incredibly well with official yardage and statistical totals, but it is important to remember that play-by-play is unofficial and did not always get the same corrections that game-level totals did. This is especially apparent with sacks, which receive a good deal of scrutiny in post-game review and often get corrected. Furthermore, we have elected to not use this play-by-play to generate tackle or passes defended totals. This is because official scorers at different stadiums were wildly inconsistent about handling these. Some stadiums would mark the nearest defender on any incompletion in the play-by-play, while others never did. Additionally, many stadiums only ever listed the primary tackler in the play-by-play, omitting any assists. We do have plans for expanding our tackles coverage in the future via other means, though.
  • Many official scorers neglected to list intended receivers on interceptions and DPIs. Schatz and Snyder have, through video review, added these in many instances that were previously missing.
  • Additionally, many scorers neglected to indicate 'forced fumbles' in the play-by-play. Schatz/Snyder have used video review to fill in many holes in the historical record here, as well.
  • Speaking of fumbles, the standards for what constituted a fumble were quite different in the earlier days of this dataset. Fumbles where the runner would clearly be marked down these days are not uncommon. Same for fumbles on 'catches' that would be ruled incomplete these days. Furthermore, Schatz and Snyder discovered that official scorers used to routinely charge fumbles to the wrong player on botched handoffs. By rule, it should be charged to the player who last possessed the ball (which is almost always the QB), but many of these were charged to running backs. These days they are accurately credited to the QB.
  • Schatz/Snyder have also filled in a ton of instances where the official scorer never marked a tackler. Another thing to note about the tackles is that in modern NFL play-by-play on plays with multiple tacklers, a semicolon between their names indicates that the first player gets a solo tackle and the second player gets an assist; if a comma appears between their names, then both players get an assist. In the older gamebooks no such convention existed, so it is impossible to ascertain the proper breakdown of solo/assisted tackles just from the play-by-play.
  • This new dataset will allow us to calculate and display previously unreleased game and season totals for player forced fumbles and game totals for fumble recoveries.
  • Convention in gamebooks during this era was to list the time on the clock at the start of a drive or for plays in the final two minutes of a half. So most plays do not have a time left on the clock associated with them.
  • The final 4:47 of the September 9, 1979 Cardinals/Giants game is missing. If anyone has a gamebook that does not have these final moments of the game cut off, or has video of the game, please let us know.

We'd also like to share some of our favorite stats that we encountered in testing that this material was showing up in the right places. We hope you're able to do the same with Stathead!

In addition to thanking Schatz and Snyder, we would also like to thank the Professional Football Researchers Association and the late Rupert Patrick for putting together the best collection of NFL gamebooks publicly accessible (for a very cheap membership rate!) anywhere, as well as Jon Kendle at the Pro Football Hall of Fame, Chris Willis at the NFL Films library and John Turney for sharing data to help get this project over the finish line. Please keep your eyes peeled for some further updates stemming from this work in the near future, as well.

Comments

Комментарии для сайта Cackle
Загрузка...

More news:

World Baseball Softball Confederation
Mets Merized Online
World Baseball Softball Confederation

Read on Sportsweek.org:

Other sports

Sponsored