Skip to content

ffb-optimal-draft/ffb-draft-strategy

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

93 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

"Winning Draft Strategy" : Fantasy Football and Cheat sheet for PPR Leagues

Project Description

Data Scientists using; pro-football-reference.com, webscraping data from football.fantasysports.yahoo.com, and yahoofantasy API, datasets to develop a machine learning model that will help to predict the players total season projections.

Project Goals

  • Create a fantasy football drafting strategy cheat sheet based off of historical NFL and fantasy league data.
  • Create a machine learning model that predicts the upcoming NFL FFL 2023 total season PPR points for Quarter Backs, Running Backs, Wide Recievers, Tight Ends.
  • Gather findings, draw conclusions and recommended next steps for Fantasy Football team owners.

Questions to answer

  • What position should I draft this round?
  • Is it better to draft a RB, WR, TE or should draft a QB next?

Initial Thoughts and Hypothesis

Initially, the data records the past 7 years of Real-Word NFL Statistics and captured over 200 fantasy league. This large dataset will help draw accurate predictions for season projections.

Planning

  • Perform Data acquisition and Preparation
  • Analyze Data features and analyze sports trends and patterns.
  • Establish baseline of model using baseline methods.
    • We took the mean of the number of ppr pts scored for a baseline of 113.
  • Create Machine Learning model to predict and forecast season projections.
    • We used XGBRegressor and RidgeRegressor functions for our two primary models.
    • We split the data frame up into 4 different dataframes. One for each position.
    • We added new columns such as reception percentage, interception percentage, etc.
    • Afterwards, we trained our models on how many ppr points a player scored did from 2018-2020.
    • We validated our model on the years 2021 and 2022.
    • Finally our test/projection set is projections for the upcoming 2023 season.
  • Draw and record conclusions
    • Our conclusions are the model is predictings as expected.
    • The number of ppr points scored are within reason for each position.
    • The players in the top 25-50 for each position make sense and are within reason.

Data Dictionary

Target Variable Definition
Target The number of ppr points a player scored in the following year.
Feature Definition
Player The name of each player in the dataset.
Rk The rank of each player.
Team The NFL Team Each player is rostered on.
Pos The position each player is listed as.
Age The age number of each player.
G The amount of games each player has played in the (regular) season.
GS The amount of games started each player has played in the (regular) season.
Cmp The number of completed passes for each player.
Pass_att The number of attempted passes for each player.
Pass_yds The number of passing yards for each player.
Pass_tds The number of passing touchdowns for each player.
Int The number of interceptions each player.
Rush_att The number of rushing attempts for each player.
Rush_yard The number of rushing yards for each player.
Y/A The average number of yards per attempts for each player.
Rush_tds The number of rushing touchdowns for each player.
Rec The number of completed recieving catches for each player.
Y/R The average number of recieving yards per reception catches for each player.
Rec_tds The number of recieving touchdowns for each player.
Fmb The number of fumbles recorded for each player.
Fl The number of fumbles lost to opposing teams recorded for each player.
Rush_rec_tds The total TDs(Rushing and Recieving) recorded for each player.
Ppr_pts The total number of PPR season points for each player.
Vbd The variance between each player's total fantasy point and the point total of a "baseline" player at the same position gives us a relative number, which can then be compared across positions.
Pos_rank The number of seasonal rank standings for each player in their position.
Year The year of the regular season for each player.
ADP The number of the average draft position for each player in Fantasy Football.
Adp_by_pos The number of the average draft position for each player in Fantasy Football where the player falls in their respecitive position.
Success Whether each player was a sucessfull pick in the regular season based on actual season statistics.
Round The number of the average draft position for each player in Fantasy Football Leagues by Team Owners.

Conclusions and Recommendations

  • Acquired 13 years of data including Average Draft Position.
  • Explored position by position and round by round for all 13 years.
  • Modeled player performance using RidgeRegressor and XGBRegressor.
  • Created Ultimate Draft Guidebook using the information discovered in exploration.
  • Created Cheatsheet of players predicted PPR points scored next season.

Next Steps

  • Draft Simulation for all 15 rounds with a 12 man league.
  • Test how fantasy teams do with players predicted points.
  • Test predictions against the upcoming season to determine how accurate our model is at predicting.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •