A python program to calculate player comparison fantasy football stats.
The mean for a position group is calculated using only the middle 80% of player values (removing the top 10% and bottom 10% per standard population statistic calculations). The program exports CSV files for the position groups QB, RB, WR, TE with the fields noted below.
- Player name
- Total points (weeks 1-16)
- Avg points per game (weeks 1-16)
- Auction value (relative positional median value times relative positional top qtr value then squared and scaled by availability)
- Availability (percentage of number of 2014 games played)
- +/- above the position mean (weeks 1-16)
- +/- above the position median (weeks 1-16)
- +/- above the top quarter (weeks 1-16)
- +/- above the position mean (weeks 1-8)
- +/- above the position median (weeks 1-8)
- +/- above the top quarter (weeks 1-8)
- +/- above the position mean (weeks 9-16)
- +/- above the position median (weeks 9-16)
- +/- above the top quarter (weeks 9-16)
-
relative positional median value - a player's average points scored per game divided by the median of average points scored per game cumulative for a position
-
relative positional top quarter value - a player's average points scored per game divided by the top quarter median of average points scored per game cumulative for a position
-
multiplied value of relative positional median & top quarter value - relative positional median value multiplied by relative positional top quarter value
-
squared value of the multiplied value - squared value of the multiple of relative positional median & top quarter value
-
squared value scaled by availability - squared value of the multiple of relative positional median & top quarter value, then scaled by the percentage of games a player played in weeks 1 through 16
Note: Power Ranking has an applied floor value of 0 for a 0-100 rating system.
--
Note 1: Execute in the terminal via python init.py 2016 in the src directory (requires python 2.7 and nfldb module)
Note 2: Calculations are based on weeks 1-16 with week 17 discarded, as the last week is commonly not apart of the fantasy season or playoffs.
Note 3: The CSV file may be imported into a spreadsheet program (i.e. Google Sheets) to be converted into a user-friendly format for analysis. Find the CSV files without executing the program inside the csv directory.
Full credit is given to the author, BurntSushi, of nfldb located at https://github.com/BurntSushi/nfldb. The nfldb python module provides a database of NFL statistics gathered via NFL gamecenter json data and an API query to interface with the data.