If you follow me on Twitter, you’ve probably seen me discuss my belief that the most efficient method of any kind of player draft is understanding and using opportunity cost in the selection process. Opportunity cost in a draft can be quantified as the drop-off from the current pick to the next within the same position. The most efficient choice given that methodology would be to choose the player who has the largest difference between his grade and the grade of the best player expected to be available with the next pick within each position.
There are a lot of ways to determine which players are likely to be selected between picks. One great example is this great Bayesian Draft Analysis Tool created by Brian Burke at Advanced NFL Analytics.
To keep things a little more simple, I pulled all of the SCOUTS, Inc. grades from the past 10 years and compared them to the draft position of each draft prospect. Using that dataset, I was able to estimate the probability that a player would be chosen over a range of picks based on his grade range. (I had to keep it to ranges, since the sample size of my data was fairly small.)
Because I wasn’t fully satisfied with the grades provided by SCOUTS, I decided to blend them with three other grade sources: those posted at the National Football Post, NFL.com’s draft page, and Draft Ace. These grades were chosen because they graded (almost) every prospect in the draft and were easily available grades, not rankings (because rankings don’t really help in this case). The grades were also normalized to a scale of 100 (National Football Post and NFL.com were in the traditional scouting format, explained here). The grades were then averaged together to get a composite grade per player.
Now that I have the expected draft positions established using the SCOUTS grades and the composite grades per player, I can evaluate the opportunity cost at each pick by position. (A quick note on the positions: I merged all linebackers into one category since these players could move between the outside and inside and there weren’t enough to keep separate. I did the same with interior linemen, guards and centers, with the same thought in mind.) The table below shows Houston’s decision at the first overall pick, along with the expected number of players in each position to be selected before their next choice and sorted by opportunity cost.
Based on this framework, the Texans would have the highest cost of not selecting Jadeveon Clowney at first overall. It also happens that he has the best composite grade in the draft. The best pick may not always be the "best available" player though, since that methodology doesn’t take the value over the player you could select with your next pick into consideration. The players available the next time around are just as important as the ones currently at the top of the board.
Another small wrinkle in the opportunity cost model I’m working with is positional needs. I cherry picked the team needs from Burke’s model (which comes from NFL.com) to get the table below.
The application of this would lead to the following question for Houston: is the positional need for an offensive tackle more important than the higher opportunity cost at defensive end? That answer may vary from team to team (even person to person within a team, I’d imagine), so let’s play out both scenarios.
Both first round outcomes are in the table below. The "Need Pick" column represents the pick based on the player with the highest opportunity cost within the positions of need (also represented in the table). The "Overall OC Pick" column uses just opportunity cost to select a player and needs aren’t taken into consideration (click image to enlarge).
The need-based pick seems to play out really well for the Buffalo Bills (although the guard, Su’a-Filo, getting picked fourth overall doesn’t seem likely), given Evans' fall to them. Evans represents great value at the ninth spot, as his grade (89.4) and opportunity cost (10.7) are both relatively high compared to the surrounding picks.
The true opportunity cost picks aren’t as exciting for Buffalo fans. It probably best represents the "doomsday scenarios" discussed during the WGR 550 morning show, as all three tackles and both top receivers are gone by the ninth pick. It’s likely that the Bills would do their best to trade out of their pick if this scenario plays out.
Neither scenario is as accurate as it could be, however. Teams don’t blindly select players based on opportunity cost and disregard needs. Furthermore, the likelihood of Clowney dropping to the No. 16 pick because no team has a "need" isn’t going to happen (besides, what team doesn't "need" a game-changing defensive end?). I combined the needs qualifier with the overall opportunity cost to come up with what I believe to be a good approximation of how the first round will play out.
The first round estimation is below, along with the pick based on the team needs, the pick with the highest opportunity cost, and the reason ("Note") for the choice of one player over the other.
Once again, the three tackles and top two receivers are gone. The Bills have been rumored to be interested in Beckham, but who really knows at this point in the process. Once again, they likely move out of their pick if this happens.
The "OC Pick" column really highlights how a player (like Aaron Donald, and Anthony Barr) can slip in a draft despite being the best player or the one with the highest opportunity cost. What’s also interesting is that Cleveland gets a quarterback with their second pick in all three simulations (although it’s a different one each time). That shows how closely the three are ranked at this stage and that they shouldn’t draft that quarterback until their second pick.
Using opportunity cost is a great way to assess the most efficient pick. It does have some shortcomings, however, since it is entirely dependent on the composite grades. To fix that, I’d be thrilled if anyone with prospect grades put a link to their work in the comments below. I’ll add them to my current grades and analyze the picks in the draft as it occurs on Thursday night.
Each team has slightly different grades on each prospect. Some teams might draft in a really inefficient or irrational manner. Predicting those irrational picks is impossible, but shouldn’t change a team’s overall strategy of selecting the best player with the highest value over what’s available the next time around. That's opportunity cost in the draft.