In mitchell-thomann/faabulousR: Calculates and Visualizes Schedule Free Wins

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Schedule Free Wins (SFW) is a metric that measures fantasy performance by week. The goal of SFW is to provide fantasy players a metric to measure a team's performance against the entire league's schedule instead of just against a single opponent, which can be highly luck driven.

The intuition behind this statistic is simple, how many expected wins given the scores of the entire league. We have all been in the situation where we score the second highest points for a week, yet play the highest scoring team. This results in a loss on our record, which is highly unlucky because if we had played any other opponent, we would have won our game. Fantasy is the perfect setting for this statistic because each team has no effect on the opponents score and the schedules typically are completely random. This is a better statistic than say total points score for measuring team performance because a few outliers can dramatically affect that statistic whereas this statistic is on the wins scale so outliers contribute only to a single week's SFW.

In the situation above, say in a ten team league, the player that scores second most points in a given week is $sfw = \frac{8}{9}$ because there are $9$ opponents and amongst those opponents, the player would have defeated $8$. This is calculated for each league member to determine the SFW for the week. Note that the cumulative total for the week is equal to the total wins that are awarded on a given week - this statistic is on the same scale as wins so a players' SFW can be compared to their actual wins.

To calculate the cumulative SFW for a season, it is simply the total SFW for each player for each week. That is for player $i$:

$$ SFW_i = \frac{1}{K-1}\sum_{j=1}^N{(rank_{ij}-1)} $$

where $K$ is the total number of teams in the league and $rank_{ij}$ is the rank for player $i$ in the $jth$ week.

Package

This package provides a simple method for calculating and visualizing SFW. This is driven by the functions sfw(), plot_cumsfw and plot_scatsfw.

Example:

library(sfwR)

# Calculate SFW
fantasy_sfw <- sfw(fantasy)

#display sorted cumulative sfw
fantasy_cum <- dplyr::filter(fantasy_sfw,week==11)
dplyr::arrange(fantasy_cum, desc(sfw_cum))

#plot cumulative sfw by week
plot_cumsfw(fantasy_sfw)

#plot sfw vs actual wins
plot_scatsfw(fantasy_cum,wins_actual = dplyr::data_frame(team=unique(fantasy_cum$team),
                                                         wins=c(5,9,7,8,6,6,6,5,4,4,3,3)))

The dotted line in the scatterplot represents actual wins equals SFW, the vertical distance from this line represents the difference between a team's SFW and the expected SFW for the team's actual wins. A positive value for this distance represents a "low luck"" team and a negative value represents a "high luck"" team. Values closer to zero for this distance represents the expected outcome with SFW and wins converging. High values for this distance represent SFW and actual wins not in agreement and these teams are experiencing fewer or greater numbers of wins than their scores indicate.

mitchell-thomann/faabulousR documentation built on Sept. 30, 2019, 10:06 a.m.