Value-Grading the 2014 NFL Draft

When talking about the NFL Draft, value is a funny, imprecise thing.  Mock draft experts, big board builders, prospect rankers, pundits, and preachers are all quick to label any given selection a "reach" or a "steal" based on their own subjective evaluations.  And then we listen to these labels and integrate them into our own personal, subjective evaluations.  In other words, we believe them.  But how do we (or they) know?  How do we know that a team’s selection was actually a reach or a steal?  How do we know the value placed on that particular player by that particular team?  Well, we don’t.  And here is an attempt to objectively explain why.

The average selection in the 2014 NFL Draft was a "reach" by nearly one full round (29 picks).  This is determined by comparing the average draft position in the TPR-based Chaos Theory simulation (hat tip to Brent Cohen) to the actual draft positions in the NFL Draft.  Here is the average difference in draft position by round (negative values are "reaches"; positive values are "steals").

It looks like teams draft much differently in round three, doesn’t it?  Under ideal circumstances, if the draft were conducted in a manner that minimized reaches and maximized steals, the draft position differences in each round would be zero.  In other words, we would have accurately projected the results of the draft.  But, obviously, the real world operates a bit differently.  In reality, the draft offers an alternative definition of normal.  Normal may not always mean taking the best player available and it may not always mean taking the best player at a position of need.  So maybe our definition has been wrong… in today’s NFL Draft, "normal" may simply mean taking the best value.

The rankings above represent the average draft position differences for each team’s 2014 draft class.  The Buffalo Bills seem to have had an excellent draft, stealing players eleven picks later than initially valued.  The Philadelphia Eagles, on the other hand, rank 19^th.  On average, their draft picks could have been available 31 picks later (roughly the NFL average), which is somewhat comforting if a goal of the draft is to ensure "getting your guy" before that guy is taken by the team’s next pick in the following round (apparently something at which Seattle is quite adept).  By most (popular) accounts, the Eagles’ Marcus Smith pick skews things a bit.  Or does it?  Indeed the TPR-inspired Chaos Theory ADP for Smith is 146, so on the surface it seems as though his selection was a HUGE reach.  But, then again, perhaps it wasn’t.

It all depends on how players are valued; on how we view "normal".  Using Expected Utility (Value) Theory as a guide, consider this hypothetical (sorry Chip, I know you hate these!): the Eagles just traded back from the 22^nd to the 26^th position in the draft and are weighing their options.  The Gamble-Kelly-Roseman triumvirate (GKR) is torn between drafting two players: Marcus Smith and Jordan Matthews.  It’s an interesting problem because Marcus Smith is the 146^th player on their board and Jordan Matthews is the 65^th.  You may be saying to yourself,  there must be better options available.  Maybe not.  If looking through the lens of the new "normal"… that is… when trying to determine which selections have more value, things are not as they appear.

Because Marcus Smith plays a position of real interest for several NFL teams, and a position that was not particularly deep in this draft, GKR could have determined that Smith has a 25% chance of being available when the Eagles chose again at pick 54.  If we multiply Smith’s initial, expected value (146) by .25, his adjusted value becomes 36.5.  Compare this to Matthews.  Since he plays a position (WR) that was relatively deep in the draft, GKR determined there was a 75% chance he would be available at pick 54.  Thus, Matthews’ adjusted value becomes 39 (54 x .75), two-and-a-half picks after Smith’s.  Since Smith’s value was better than Matthews, the Eagles selected Smith with the 26^th pick in the draft (and of course traded up to nab Matthews at pick 41, just 2 picks later than he was "valued").

This type of methodology assumes various approaches to risk.  A less risk-averse team (or a risk-seeking team like the Eagles) may have selected Smith, whereas a more risk-averse team (like the Bills) may have selected Matthews, the pick that represented the "surer thing".  This does not mean that mock drafts, big boards, and prospect rankings are irrelevant.  Quite the contrary, these tools (and better yet, the sum total of these tools) represent the draft’s potential when devoid of risk.

And this does not mean teams don’t make mistakes.  Undoubtedly some do.  But when millions of dollars, hundreds of hours, and countless numbers of human resources are spent to prepare for the draft, those mistakes are becoming less and less obvious… and less and less subjective.  Risk is one reason why we’ll never know precisely how a team values a particular player.  And risk is one reason why teams will always get their guy.