ffverse / ffopportunity Goto Github PK

for xgb model files. requires xgboost 1.6 but is sig smaller file size to download

Please select sacks in this function to calculate fantasy points with sacks.

For the ultimate cool stuff include "sack yards" ;)

Can you please add the ability to generate the xyac distribution from the model as described in https://opensourcefootball.com/posts/2020-08-30-calculating-expected-fantasy-points-for-receivers/. This would enable users to generate a distribution (range of outcomes) of expected passing and receiving fantasy points in addition to median/mean outcomes.

His numeric stats seem to be doubled and I saw some one else say same thing about Chase Claypool.

The yards_after_catch field in pbp_pass is incorrect. It is currently a decimal between 0 and 1. Based on the data dictionary, it hould be an integer equal to the difference between receiving_yards and air_yards. Reviewing the desc fields confirms this. In the weekly dataframe, the rec_air_yards and rec_yards_gained fields seem accurate, so this error seems limited to the pbp_pass.

library(tidyverse)
library(ffopportunity)

ep_load(
  season = nflreadr::most_recent_season(),
  type = "pbp_pass",
  version = "latest"
) %>% 
  select(yards_after_catch, air_yards, receiving_yards, desc, game_id, play_id) %>% 
  head(4)
#> → <ffopportunity predictions>
#> → Generated 2024-01-15 01:41:41.296844 with ep model version "latest"
#> # A tibble: 4 × 6
#>   yards_after_catch air_yards receiving_yards desc               game_id play_id
#>               <dbl>     <dbl>           <dbl> <chr>              <chr>     <dbl>
#> 1             0.661         6               6 (14:30) (Shotgun)… 2023_0…      77
#> 2             0.496        10              NA (13:16) (Shotgun)… 2023_0…     124
#> 3             0.563        12              12 (13:12) 14-S.Howe… 2023_0…     147
#> 4             0.484        -4 

Attaching the file I pulled on the evening of 12/23/2022 from the following link.

ep_weekly_2022.csv

Focusing on Tom Brady to highlight the issue, it looks like the total_fantasy_points column is being derived from rush_fantasy_points and is ignoring the other point scoring categories pass_* and rec_*

Will see if I can ID the issue in the code but in the mean time I thought I'd raise the alarm in the mean time

It'd be nice to get the outputs for individual scoring options (as discussed via Discord).

Ideally connected to ffscrapr to get the leagues scoring (which may will be hard for some complex scorings requiring minus points for targets/attempts and other stuff like the analytics scoring).

Best regards,
Christian

PS: Atm I'm building a predicted_pbp based on your functions and then summarize it with my leagues points.
Quite cool package, I already love it ❤️

Pre-release:

• Determine scope of modelling documentation to include
• Use nflreadr::load_ff_opportunity backend for ep_load?
• Backlink apps/webpages to main pkgdown?
• Move ep static page to ffverse domain instead?
• Deploy ep dynamic app?

CRAN first-release stuff:

• usethis::use_cran_comments()
• Update (aspirational) install instructions in README
• Proofread Title: and Description:
• Check that all exported functions have @return and @examples
• Check that Authors@R: includes a copyright holder (role 'cph')
• Check licensing of included files
• Review https://github.com/DavisVaughan/extrachecks

Prepare for release:

• usethis::use_version('major')
• devtools::build_readme()
• urlchecker::url_check()
• devtools::check(remote = TRUE, manual = TRUE)
• devtools::check_win_devel()
• rhub::check_for_cran()
• Review pkgdown reference index for, e.g., missing topics
• Draft blog post

Submit to CRAN:

• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted 🎉
• usethis::use_github_release()
• usethis::use_dev_version()
• Finish blog post
• Tweet
• Add link to blog post in pkgdown news menu