aptmac/dfs-helper-nfl

Python scripts for optimizing Daily Fantasy Football line-ups. In the 2021 season it got ranked Diamond (99th percentile) in Yahoo DFS quickmatches, and placed 504/223186 in the season-long 2021 Yahoo cup.

Hi there 👋

dfs-helper-nfl's Issues

output results to file (json & csv)

It would be nice to have the results printed to file.

Yahoo DFS allows for lineups to be submitted as CSV, so this would make it even quicker. Just need to figure out exactly what the csv format looks like, and give it a try.

upon running, seems Kamara is printed twice.

--- (4/4) Formatting the results ---
name position team salary points
6 Lamar Jackson QB BAL 36.0 25.8
14 Alvin Kamara RB NO 33.0 19.0
17 Alvin Kamara RB NO 32.0 19.0
66 Stefon Diggs WR BUF 24.0 16.1
79 Darren Waller TE LV 23.0 14.2
258 Courtland Sutton WR DEN 14.0 12.1
267 Emmanuel Sanders WR BUF 14.0 12.0
275 Michael Carter RB NYJ 14.0 12.5
425 San Francisco 49ers DEF SF 10.0 6.8
Total used amount of salary cap: 200.0
Projected points: 137.5
--- Completed ---

place the projected scores data directly into the dataframe after fetching

Currently the fetch script iterates over the list of positions, and saves each individual page to a file. It would be nice if instead we just toss it into the dataframe, and don't need to save as many files.

fetch script doesn't work on new FP tables

FP changed the URL for their weekly projected ranking tables, but more importantly they have been updated to be similar to the new draft rankings table. The problem here is that the table is populated via JavaScript a short time after the website initially loads, so scraping the html becomes a bit more difficult.

This can use the same approach as draftaid-react, and use Selenium on a timer to wait until the data populates and then scrape the data.

allow for custom formats to be passed via the command line

It would be nice if configuration options could be passed to the solver program, here via the command line.

Ideal attributes would be:
--budget to specify a different budget
e.g., --budget=125

--positions to specify which positions we'd like to use
e.g., --positions=1qb,2rb,3wr,1te,1flex,1dst or --positions=1superflex,4flex

--teams to specify which teams to include
e.g., --teams=gb,sf for tonight's Packers/49ers SNF

--time to specify which game time to use
e.g., --time=mnf
options would include 1 (for 1:00PM EST starts), 4, 8/snf, thursday/tnf, monday/mnf

adjust the solver script to read json format

At the moment the solver script accepts the information in csv format, but it would be nice if this could be adjusted to read json as well

Use OwnersBox CSV edit lineup feature

OwnersBox has a nifty feature that creates a custom CSV file for you to use to create your own rosters. It contains the players of the contest you're playing, and once you arrange your players into the appropriate columns you can re-upload it and it should take care of the rest.

This would be huge because I tried manually adjusting my roster this past weekend and the UI is not very intuitive.. it took way longer than it should have.

The CSV is also in a nasty looking format because I'm thinking it was designed with Excel in mind.. this is going to be a bit more annoying than it should be but hopefully isn't too bad overall.

set up format for Yahoo Single Game Faceroll contests

The single game faceroll contests use $125 budget instead of the regular $200, and runs only five positions; 1 MVP "Superflex", and 4 flex positions.

The "Superflex" gets 1.5x the points, and AFAIK runs the same positions as flex.
Flex in this case is anything (QB/WR/RB/TE/DST)

Rotowire no longer provides OwnersBox data for free

Wanted to give this a go again now that regular fantasy is over. Turns out OwnersBox projections are no longer provided by Rotowire without a subscription. Will have to find a way around this ..