Different plays with same play_id in 2000 pbp data

[ajreinhard]

It looks like there might be multiple plays in the 2000 play-by-play data that have been given the same play_id. No other seasons have a similar issue.

library(nflfastR)
#> Warning: replacing previous import 'vctrs::data_frame' by 'tibble::data_frame'
#> when loading 'dplyr'
library(tidyverse)
#> Warning: package 'tidyverse' was built under R version 3.6.3
#> Warning: package 'ggplot2' was built under R version 3.6.3
#> Warning: package 'tibble' was built under R version 3.6.3
#> Warning: package 'dplyr' was built under R version 3.6.3
  
pbp_df <- load_pbp(2000)

pbp_df %>% 
  group_by(game_id, play_id) %>% 
  filter(n() > 1) %>% 
  ungroup %>% 
  select(game_id, play_id, down, ydstogo, yardline_100, game_seconds_remaining, desc) %>% 
  arrange(game_id, play_id)
#> # A tibble: 12 x 7
#>    game_id   play_id  down ydstogo yardline_100 game_seconds_re~ desc           
#>    <chr>       <dbl> <dbl>   <dbl>        <dbl>            <dbl> <chr>          
#>  1 2000_03_~    2767     1      10           16              900 (15:00) T.Couc~
#>  2 2000_03_~    2767     2      10           16              900 (15:00) T.Couc~
#>  3 2000_03_~    2768    NA       0           16               NA Timeout #1 by ~
#>  4 2000_03_~    2768     3       4           10              900 (15:00) T.Couc~
#>  5 2000_06_~    1825    NA       0           48               NA Timeout #3 by ~
#>  6 2000_06_~    1825     4       9           69             1846 (:46) T.Barnha~
#>  7 2000_06_~    1825     1      10           52             1833 (:33) D.McNabb~
#>  8 2000_06_~    1825     2       8           50             1816 (:16) D.McNabb~
#>  9 2000_11_~    2323     2       3           94             1429 (8:49) B.Gries~
#> 10 2000_11_~    2323     3       1           96             1386 (8:06) T.Davis~
#> 11 2000_11_~    2323     4       2           95             1349 (7:29) J.Elam ~
#> 12 2000_11_~    2323    NA       0           30             1345 J.Elam kicks 5~

[guga31bb]

Sadly nothing to be done about this one #33 (comment)

Note that there are fewer games affected by this since we switched to a different data source for 2001-2010, but there's no way to fix the games from 2000, I don't think

[mrcaseb]

Since we are talking about 12 plays in 3 games I think we could hard code the correct order of the problematic plays using the game books. Maybe we can choose a new play_id? Is there any system in the play ids?

Example

[guga31bb]

I think there's a system in play IDs because they can be linked to external sources (ESPN, Big Data Bowl) but for games from 2000 this isn't that relevant and hard coding some play IDs is probably fine

[guga31bb]

Closing since it's the same as #33