As draft season approaches, fantasy football owners will be inundated with player projections and rankings. Median projections are an excellent tool for organizing players into tiers (you can find Mike Braude‘s for free here), but when applying projections to actual drafts, owners should be taking it a step further.
A player is a draft day value if they are likely to meet or exceed their draft position. In order to estimate the likelihood of a player reaching value we have to consider their range of outcomes, which I’ll attempt to lay out for you here.
When I started this project, my initial plan for projecting expected scoring rates was to look at historical stats from each position as a whole, and calculate averages and standard deviations from each group. I did end up using these calculations, but some discussion with RotoViz‘s Justin Winn convinced me that scoring rates need to be broken down by individual, not just by position.
The Sim Score App gives a list of 25 player comps from one season, along with a multiplier that is used to adjust for any struggles with finding close comps, and tells us what those players did the following year. Using those player comparables and the multiplier, I calculated an adjusted average for receptions and PPR fantasy points per reception (FP/Rec) for the player in question for the upcoming season. I then took the standard deviation from each set of comps to give us a range of their most likely outcomes in 2014.
MFL ADP is currently the best gauge of a player’s true stock, because MFL10s are the only real drafts being done in high quantity right now. ADP was filtered for only real money MFL10s, which are PPR leagues, starting after June 15.
The last piece to the puzzle is how many fantasy points a player needs to score to meet value. This is obviously going to vary wildly from year to year, but in an effort to set some parameters, I took the average end of season PPR scores for wide receivers finishing 1 -65 from the last 3 years. A comparison of the range of outcomes to those averages will at least give us some idea of where a player might finish in fantasy scoring.
Range of Outcome Analysis
Antonio Brown has seemingly been the punching bag for regression to the mean arguments this season, and he’s the main reason I started this project; every player can’t have the same mean, and I wanted to know what his was.
Here are the adjusted averages and standard deviations for what Antonio Brown’s closest comps did their following year:
The following matrix shows us how Antonio Brown, whose current ADP is WR8, will perform should his receptions and scoring rate fall within 1 standard deviation of his expected mean.
|FP/Rec||72.9 Receptions||83.4 Receptions||93.8 Receptions||104.2 Receptions||114.7 Receptions|
|3.2||229.3 (WR17)||262.1 (WR11)||294.9 (WR5)||327.7 (WR3)||360.5 (WR1)|
|2.9||208.2 (WR21)||238.0 (WR16)||267.8 (WR8)||297.6 (WR4)||327.3 (WR3)|
|2.6||187.1 (WR27)||213.9 (WR20)||240.6 (WR14)||267.4 (WR9)||294.2 (WR5)|
The bold cells show when Antonio Brown would meet or exceed his current ADP, which, according to these scenarios, is about 47% of the time, roughly the same bust rate as 1st round running backs. In 2 other scenarios, Brown finishes at least as a WR1, so there is probably about a 60% chance that owners are going to be happy with his production.
What This Means for 2014
This process is an inexact science but it gives us a very good idea of best and worse case scenarios. Even if Antonio Brown does regress some, he’s clearly the best wide receiver in Pittsburgh, and it’s highly unlikely that we see him drop to the low end of the spectrum for receptions.
Brown’s roughly 50/50 chance of meeting value tells us that he’s being drafted right around where he should be, but you definitely don’t want to reach for him at this point. As we project more players’ ranges and gain more reference points, we will get a better idea of how often we expect certain groups of players to meet their value and how safe they are in relation to their peers.