Me and my coworkers recently had the draft for our fantasy football league. Although I’ve done many fantasy leagues in the past, this was my first time doing an auction draft. I left the draft room feeling pretty good about my team (the B-Sze Bees), but also excited about the possibilities for analysis. Given the finite bank roll of each player, the auction draft leaves a lot more room for someone to have a horrible (or amazing) draft, depending on how they allocate their resources.
Although its hard to decide who had a good or bad draft before the season is over, I did think it was worthwhile to look at the draft results and use them to determine what our league valued, based on spending habits. The first, and most obvious piece of analysis, was to look at how spending broke down by position, across all teams.
While it initially seems that our league spent more on running backs and wide receivers, you have to remember that each team needs to draft two and a half starters at each position (since the flex can go to a WR or a RB). Splitting the 39% budget allocation given to RB and the 34% budget allocation given to WR, we find that our draft valued quarterbacks about equally to running backs and wide receivers. This goes against conventional drafting logic, but is unsurprising given a) the rise of the pass-heavy offense in the NFL b) the relative lack of depth at running back this year and c) the high variability of points receivers accumulate on a week-by-week basis.
Drilling down further, the next step was to ask, “based on spending at a particular position, who did our league indicate were the ‘elite’ players?” I took a stab at this using some rough, back-of-the-envelope style computation. For each position, I calculated the average amount paid over the entirety of the draft, as well as the standard deviation of those prices. The results are found below:
Quarterback Average: $14.30
Quarterback Standard Deviation:$18.29
Running Back Average: $17.12
Running Back Standard Deviation: $17.52
Wide Receiver Average: $15.37
Wide Receiver Standard Deviation: $13.51
Tight End Average: $9.80
Tight End Standard Deviation: $12.30
Kicker Average: $1.59
Kicker Standard Deviation: $0.94
Defense / Special Teams Average: $2.46
Defense / Special Teams Standard Deviation: $1.71
Given these numbers, who were the players at each position for whom someone paid at least one standard deviation more than the average price at that position? Here are the results:
QB: Aaron Rodgers, Drew Brees, Tom Brady
RB: Adrian Peterson, Arian Foster, Chris Johnson, Darren McFadden, LeSean McCoy, Marshawn Lynch, Matt Forte, Maurice Jones Drew, Ray Rice, Ryan Mathews
WR: A.J. Green, Andre Johnson, Antonio Brown, Brandon Marshall, Calvin Johnson, Greg Jennings, Julio Jones, Larry Fitzgerald, Marques Colston, Mike Wallace, Percy Harvin, Roddy White, Wes Welker
TE: Antonio Gates, Jermichael Finley, Jimmy Graham, Rob Gronkowski
K: David Akers, Mason Crosby, Stephen Gostkowski
D/ST: San Francisco 49ers
It turns out this metric is a great proxy for players who are considered ‘top players’ at their positions. At each position, however, I’d say there are a few surprising omissions (e.g., Eli Manning or Matt Stafford at QB) and unexpected inclusions (e.g., Jermichael Finley at TE).
At this point, I wondered: What informs our appraisal of who will be an ‘elite’ player in the 2012-2013 fantasy year? The obvious answer is the Yahoo! or ESPN rankings, but I was curious as to whether our drafting was more informed by these fantasy sites’ 2012 projections, or the actual 2011 results. I plotted auction draft price against 2011 point totals and 2012 project point totals for the ‘elite’ players at all six positions, according to the previously used critera.
While the trend is present, but not especially strong when compared with last season’s results, we see a very strong linear correlation between 2012 projected point totals and auction price paid. This is, in some sense, expected — the 2012 projections should be a synthesis of information that we know from last year in addition to offseason developments (e.g., Larry Fitzgerald’s 2012 projection takes into account his excellent 2011 season, but also takes into account he’ll be receiving passes this year from John Skelton, Kevin Kolb, or some unknown 3rd-stringer). I thought it was particularly interesting that when looking at the auction price versus 2012 projections, TEs fall into the same linear regression as WRs and RBs, despite the common knowledge that TEs are less valuable. This reinforces the idea that when you draft an elite TE like Rob Gronkowski or Jimmy Graham, you are expecting WR- or RB-like performance.
Finally, I asked, where did people overspend on a player who had a great 2011 season, but is projected to regress in 2012? Similarly, who were the ‘great pick-ups’, with respect to players that are expected to have a breakout year in 2012?
I looked here at auction price vs. expected improvement from 2012 to 2011. A data point in the bottom right would be a steal (a player expected to do much better this year, who the bidder paid very little for), while a data point in the top left would be very bad (a player for whom heavy regression is expected and a large price was paid). It is worth noting that the strong outlier to the right for WRs is Andre Johnson, who was injured a large portion of 2011 (hence the huge expected increase in output). While most of the points in this plot fall in the +/- 20 range (pretty mild fluctuations for a player when integrated over an entire season), it seems that people tended to dump a considerable amount of money into running backs who were projected to perform much better in 2012 than they did in 2011. Given the number of under-performers in 2011 at the position (e.g., Chris Johnson, Darren McFadden), and the consequent perceived lack of depth at the position, it’ll be interesting to see how the money spent on running backs pans out this year.