Is Investing a Top Fantasy Football Draft Pick on Unproven Players a Mistake?

2014 was not a great year for taking chances in fantasy football land.

The allure of the possible led the fantasy community to prop up unproven commodities such as Toby Gerhart, Cordarrelle Patterson, Montee Ball and Andre Ellington to heights they fell incredibly short of. While none of those players had produced as a top 24 player at their positions, they were drafted due to their supposed potential to do just that.

In fact, there were eight players selected in the top 60 in 2014 that had yet to post a season in the top 24 of their position group, and only one returned the value on his average draft position. And that one player, Emmanuel Sanders, just so happened to upgrade to Peyton Manning as his quarterback.

With that failure as a backdrop, I wanted to research the return on investment (or ROI) on players in the top 60 of average draft position (ADP) to see if there were any predictive statistics or warning signs in the numbers. Specifically, I wanted to isolate unproven yet highly hyped players to see if the failure of 2014 was just an outlier or if we should dutifully avoid the next wave of overhyped prospects.

For those keeping score, that would be Melvin Gordon, Brandin Cooks, Joseph Randle, Latavius Murray, Carlos Hyde, Martavis Bryant, Todd Gurley, Amari Cooper, Sammy Watkins, T.J. Yeldon, and probably Allen Robinson soon if people aren't careful.

Math, You See

While you can manipulate statistics to prove just about any point you want, the goal of fantasy football research should be to inform your draft decisions and how you approach draft strategy. The question on the table today is whether unproven players with a tremendous amount of pre-draft hype are worth the risk or whether avoiding them for more proven prospects is a justified approach to the first five rounds of your standard draft.

For my research, I charted every player from 2010 to 2014 with an ADP of 60 or above (the first 5 Rounds of a standard 12-team draft) based on their prior season's finish, positional ADP, and final positional ranking. If a player being drafted in the top 60 had never before had a top 24 season at the wide receiver or running back position, they were marked as a "hype" player. While quarterbacks and tight ends were not excluded from this analysis, the sample sizes proved too small to make any definitive judgments.

The results were fascinating.

Just 14 out of the 44 hype players taken in those five years gave you a positive return on your investment, and the average ROI based on ADP was -15.51. That stands in contrast to the -12.92 ROI for all other players. While that may seem like a minuscule difference, when used in reference to such a large sample size, it does demonstrate the greater risk associated with drafting unproven players.

But the more important data point is not necessarily ROI, which is a very broad measuring stick, but instead evaluating if a player you drafted to be a top player at their position actually gave you that value.

To analyze this, I gave a negative grade to a player who landed outside the top 24 at their position group but who were drafted as a top-24 prospect. This allowed me to zoom in on specific position groups as well as if individual players performed as expected.

Wide Receivers

From 2010 to 2014 there were five hits and four misses at wide receiver for hype players, amounting to a hit percentage of 55.56%. That's not a bad score at all, especially when you consider that injuries are not factored out of the equation. But how did that compare to non-hype players drafted in the top 60?

It actually factored very favorably, with a 59% hit rate for all other wide receivers drafted in the top 60.

With the margin for error so similar, the next part of the research was to figure out if we could utilize our Net Expected Points (NEP) metric to help further isolate which hype players might be heading towards a disappointing season. NEP is our in-house metric that quantifies a player's production relative to league expectation level based on historical performance.

Specifically, I looked at Reception NEP per target for each player, which evaluates efficiency on a per-target basis. For context, know that the top 10 receivers in this category in 2014 averaged a score of 0.88.

Here is how each player's Reception NEP per target looked in the year prior to his ADP spike and his final finish at the position in the year when he was drafted highly.

Name	Year	Rec NEP/Target	Final Pos Rank
Percy Harvin	2010	0.90	20
Hakeem Nicks	2010	0.90	8
Dez Bryant	2011	0.73	19
Pierre Garcon	2010	0.69	31
Eric Decker	2012	0.55	7
Danny Amendola	2013	0.52	60
Michael Crabtree	2010	0.52	35
Emmanuel Sanders	2014	0.48	7
Cordarrelle Patterson	2014	0.39	76

While situations play a huge role in determining future success, it is clear based on this research that that above average NEP scores are somewhat predictive of future success. That doesn't always manifest itself as an immediate top 24 finish, but the top four players on that chart posted great scores and have proven to be great fantasy contributors. Sanders broke that trend, but it certainly didn't hurt to move into a Peyton Manning-led offense that was historically dominant.

So with that history in mind, let's look at the projections for the next wave of hype receivers.

Name	Year	Receptions	Rec NEP/T	Current ADP
Brandin Cooks	2014	69	0.60	36
Martavis Bryant	2014	26	0.95	46
Sammy Watkins	2014	65	0.70	51
Allen Robinson	2014	48	0.51	68

While none of the NEP scores scream "bust," Bryant's sample size is so small and deep-ball dependent that it is difficult to make any definitive evaluation. What is true is that he was incredibly effective at creating the big play, which is certainly a good trait to have.

Watkins and Cooks appear capable of breakout potential. Robinson is priced appropriately outside the top 60 right now considering that Jacksonville was the worst passing offense in the league (by far) according to our Adjusted Passing NEP per play metric and that his Reception NEP per target score was average. But no one would be surprised if these players gave you a positive return on your investment.

Running Backs

Here is where things got interesting. Over a five-year period, there were 26 hype running backs drafted in the top 60 overall. 16 of them failed to crack the top 24, meaning the hit rate was just 38.5%.

So for every rookie-year Doug Martin there were almost two David Wilsons.

For all other players, the hit rate was 69.23%! That difference is remarkable and likely speaks to the lack of available talent at the position, causing people to reach for the supposed next big thing instead of taking the solid player standing right in front of them. By season's end, Chris Ivory and Steven Jackson are in the top 24, and Montee Ball is on the bench.

Another interesting data point was that only two out of the 23 non-rookie hype backs managed to finish in the top 10 at their position the next year (LeSean McCoy and Ryan Mathews), and neither finished in the top five. While these players are presumed to have high upside, and thus the extreme hype, more often than not they failed to deliver on that upside.

The next step was to look at Rushing NEP per rush to see if it can help us make more informed decisions. After evaluating the data, it was difficult to draw any strong correlations between Rushing NEP and future success.

For example, of the hype running backs evaluated during a five-year period, our NEP data would indicate that Toby Gerhart would have the most success in 2014, and Rashad Jennings would have the least success that same year. Here are the non-rookie backs from 2013 to 2014, their final positional ranking and their Rushing NEP per rush from the prior season.

Name	Year	Rush NEP/R	Final Pos Rank
Toby Gerhart	2014	0.27	55
Shane Vereen	2014	0.11	28
DeMarco Murray	2012	0.1	26
Lamar Miller	2013	0.08	38
Andre Ellington	2014	0.06	20
Stevan Ridley	2012	0.04	10
Montee Ball	2014	0.03	89
BenJarvus Green-Ellis	2012	0.03	19
Daryl Richardson	2013	-0.02	78
David Wilson	2013	-0.03	93
Ben Tate	2014	-0.06	51
Rashad Jennings	2014	-0.17	29

Gerhart aside, I guess one could postulate that the a Rushing NEP per rush higher than 0.05 for identifying talent. Keep in mind that the average score of the top 10 running backs in 2014 in this metric was 0.07. But this definitely isn't a "rule," with McCoy scoring a -0.07 the year before his breakout season.

What We Can Learn

The goal of research like this isn't to avoid mistakes by being overly cautious as a manger -- missing on players will occur no matter how much data you accumulate. However, by investigating our mistakes and successes we can begin to find trends that provide us data points that we can use to inform our decisions.

So while data would indicate that drafting a quarterback who is under six feet tall is a recipe for failure, there is always room for a Drew Brees or Russell Wilson to prove us wrong.

What the research indicates is that we should be inherently skeptical of these hype players -- specifically at the running back position -- not that we should avoid them altogether.

For 2015, it means that the Carlos Hyde hype should be very concerning, while we can feel more comfortable taking a risk with Randle or Murray. Comfortable, but not confident.

Name	Year	Carries	Rush NEP/R	Current ADP
Joseph Randle	2015	51	0.11	39
Latavius Murray	2015	82	0.05	41
Carlos Hyde	2015	83	-0.09	40

The numbers would imply that two of them are likely to finish outside the 24, with Hyde being the most likely candidate to disappoint. Allowing the statistics to inform your draft tendencies should help steer you away from the common mistakes that occur on draft day.

That being said, each year, player, and situation is different, and there are enough exceptions to every rule that if you feel a player is worth the hype, then you shouldn't let negative statistics stop you. But know that the numbers do demonstrate a need to be more cautious with unproven commodities in the early rounds of your draft, especially running backs.

Know the risk, and proceed with caution.