One of the most interesting types of analysis that comes out of gathering multiple big boards is to see which players people disagree on. We already know that there are different approaches that analysts take to ranking players, and the biggest ones are represented by the differences between the evaluator boards and forecaster boards.
In the past, player polarity came from the same sources of uncertainty that caused differences in evaluator vs. forecaster boards: school size, injury concerns and off-field character flags. That holds true today, too.
The fact that there are a different set of players implies that some of the differences in evaluation came from information imbalance—the players who appeared in the previous article were evaluated uniformly but were dropped by forecasters because of information the forecasters had that they have not shared.
In 2014, the most polarizing players (Jeff Matthews, Mike Evans, Ha Ha Clinton-Dix, Marcus Smith, Calvin Pryor, Laurent Duverney-Tardif, Zach Kerr, Yawin Smallwood, Michael Campanaro and Anthony Barr) were generally a good and underdrafted group.
The 2015 group of polarized players (Todd Gurley, Dorial Green-Beckham, Jameis Winston, DeAndre Smelter, Randy Gregory, Eric Rowe, Frank Clark, Ellis McCarthy, Danny Shelton and Shane Ray) produced a more mixed bag, with some clear failures (like Dorial Green-Beckham), clear successes (like Danny Shelton) and players whose talent is still to-be-determined (like Todd Gurley).
And in 2016, we saw a number of players that are impossible to judge yet. Jaylon Smith can now lift his toes, and it will be exciting to see if Corey Coleman can produce a bigger year with a better quarterback. Jonathan Bullard had a down rookie year, but there’s reason to believe he’ll improve. Time will tell with Robert Nkemdiche, but there’s generally a feeling of pessimism surrounding him. Keanu Neal and Chris Jones are quite good, and I think the jury is out on Carson Wentz.
And if people agree that a player should be ranked highly? In 2014, that was a massive sign of success. They included Sammy Watkins, Khalil Mack, Jake Matthews and Jadeveon Clowney. Even the players ranked 20-40 generally worked out; players like Bradley Roby, Brandin Cooks and Allen Robinson.
That same group of agreed-upon players in 2015 didn’t turn out. Devin Smith, Kevin White, Laken Tomlinson and Owa Odighizuwa still have yet to perform for their teams at a high level. The lower-level players in this group generally outperformed their draft status, however, which is the opposite of what you’d expect. Jamison Crowder, Jake Ryan, Tyler Lockett and David Johnson provide a solid corps of overperformers that bring up the average.
I’m not sure we can draw many conclusions at this point for what it means for a player to be polarizing, except that quarterbacks are almost always going to draw big differences in opinion. I think the general understanding of a polarizing player as being high-ceiling, low-floor might be true, however.
The polarizing players that have hit have performed well above expectation. Some of them have varied wildly from season to season (like Anthony Barr or Todd Gurley). A lot of those polarizing players were cut pretty quickly, too—Dorial Green-Beckham, Eric Rowe and DeAndre Smelter (though Smelter made it back onto the team from the practice squad).
Who are the most polarizing players in the draft?
Rank | Player | School | Position | Polarity Score |
4 | Jonathan Allen | Alabama | DL5T | 39.4 |
176 | Blair Brown | Ohio | OLB | 34.6 |
46 | Joe Mixon | Oklahoma | RB | 30.1 |
187 | Ejuan Price | Pittsburgh | ER | 29.0 |
170 | Conor McDermott | UCLA | OT | 27.7 |
95 | ArDarius Stewart | Alabama | WRS | 27.5 |
120 | Jamaal Williams | Brigham Young | RB | 25.7 |
174 | Joe Mathis | Washington | ER | 25.3 |
147 | Shaquill Griffin | UCF | CB | 24.8 |
85 | Jourdan Lewis | Michigan | NCB | 20.5 |
153 | Davon Godchaux | LSU | DL5T | 19.9 |
14 | Derek Barnett | Tennessee | ER | 19.0 |
29 | Jabrill Peppers | Michigan | S | 18.7 |
50 | Derek Rivers | Youngstown | ER | 18.7 |
31 | Patrick Mahomes II | Texas Tech | QB | 18.4 |
195 | Brendan Langley | Lamar (TX) | CB | 18.1 |
44 | Tim Williams | Alabama | ER | 17.4 |
2 | Solomon Thomas | Stanford | ER | 17.3 |
75 | Marcus Maye | Florida | S | 17.3 |
175 | Lorenzo Jerome | Saint Francis (PA) | S | 17.3 |
18 | John Ross | Washington | WRF | 17.2 |
There’s a quarterback (Mahomes), a few small-school players (Brown, Griffin, Rivers, Langley and Jerome), injury worries (Joe Mathis), off-field concerns (Mixon, Lewis and Godchaux), tweeners (Allen, Price, Peppers and Thomas) and unique athletic profiles (McDermott, Barnett and Williams).
That leaves ArDarius Stewart, Jamaal Williams, Marcus Maye and Jon Ross. My intuition is that three of those players (Stewart, Williams and Ross) represent a specific, one-dimensional role of a player to most evaluators, and they value that role differently. I’m not sure about Maye, except to say that this kind of thing sometimes happens without a clear, unifying reason.
Oh, and the players people most agreed on?
Rank | Player | School | Position | Polarity Score |
1 | Myles Garrett | Texas A&M | ER | 0.0 |
134 | Jeremy McNichols | Boise State | RB | 2.5 |
171 | Marquez White | Florida State | CB | 3.1 |
165 | Corey Clement | Wisconsin | RB | 3.1 |
135 | James Conner | Pittsburgh | RB | 3.2 |
57 | Alvin Kamara | Tennessee | RB | 3.2 |
129 | Kendell Beckwith | LSU | ILB | 3.4 |
7 | Reuben Foster | Alabama | ILB | 3.5 |
105 | Jordan Leggett | Clemson | TE | 3.9 |
6 | Marshon Lattimore | Ohio State | CB | 3.9 |
148 | Julie’n Davenport | Bucknell | OT | 4.3 |
156 | Michael Roberts | Toledo | TE | 4.3 |
161 | Tanzel Smart | Tulane | DL3T | 4.5 |
111 | Tanoh Kpassagnon | Villanova | ER | 4.6 |
40 | Quincy Wilson | Florida | CB | 4.6 |
37 | Jarrad Davis | Florida | ILB | 4.7 |
180 | Josh Carraway | TCU | ER | 4.8 |
102 | Chad Hansen | California | WRF | 5.0 |
127 | Corn Elder | Miami (Fla.) | NCB | 5.0 |
That’s an odd group. For the most part, I think agreements need less explanation than disagreements, but it’s worth focusing on one specific player: Reuben Foster.
He was one of the biggest disagreements when comparing the evaluator and forecaster boards but one of the biggest agreements here. Something worth pointing out is that Foster ended up sixth on the evaluator board not because he was ranked sixth by most boards, but that there were only five players consistently ranked above him.
In other words, if he was ranked eighth on one board, only five of the players ranked above him were ranked above him on another board. That means that he can universally be considered the 10th-best player, but with only five players universally ranked above him, the 10 or 15 players that are only sometimes ranked above him drop down.
In this case, Foster was universally ranked eighth or ninth by a good number of boards. Those boards disagreed about which seven or eight players they thought were better, but they all agreed that regardless of those other circumstances, he was just inside of the top 10.
Other than that, there are a surprising number of small-schoolers (Kpassagnon, McNichols, Smart and Davenport) but otherwise, players who are generally considered easy evaluations.
Take these results how you will. Even as a bit of trivia, it’s fascinating.