2020 NFL Draft — An Analytical Overview
“There is nothing more dangerous than the right answer to the wrong question” — Russell Carleton, ‘The Shift’
The NFL held its first-ever virtual draft between Thursday, April 23rd, and Saturday, April 25th. Beyond the obvious, this NFL draft was different than most others for many reasons. One of which is that more people are paying attention now more than ever and draft content is leading most sports websites that don’t have much else to discuss right now.
Typically, during the weeks immediately proceeding the NFL draft, many analysts are busy posting their draft grades. Draft grades are usually dependent on where a team was able to draft a player versus where that player was expected to be drafted. The problem that can arise from this type of analysis is it is inherently susceptible to groupthink.
Throughout the pre-draft season, NFL fans are inundated with mock drafts and it becomes easy to believe that we already know how the draft should go, before it even starts. After the draft takes place, mock drafts can often be weaponized as an argument about where a player should, or should not have been taken.
For example, many mock drafts had CeeDee Lamb being drafted within the first ten picks. When he fell to the Cowboys, it was easy to believe that they had gotten great value with that pick. However, that is the right answer to the wrong question.
The wrong question: Overall, did our team get players that were expected to come off the board earlier in many mock drafts?
The right question: Did our team add draft capital in order to draft highly ranked players at highly valued positions?
The second question is obviously more nuanced, and admittedly less fun. But it is quantifiable and unbiased.
I broke down each step of this question, and looked at which teams can answer “yes” and which teams must answer “no”.
Before I go through the process, I want to reinforce that being able to answer “yes” or “no” is not a predictor for future success. Many teams that answer “no” have the luxury of not needing to draft many highly ranked players at highly valued positions. Moreover, where this analysis and the public opinion deviate is where some really interesting analysis begins.
Part 1: Draft Pick Value
During the late 1980s and early 1990s, Jimmie Johnson created the “draft pick value chart” that was used as a reference guide to make it easier to trade draft picks. What the chart does is assigns a value to each draft pick starting with the highest value assigned to pick 1.01.
To see this chart in practice, you can look at the trade between the Patriots and the Chargers on the first night of the draft. The Patriots gave up pick 1.23 (draft value of 760) for picks 2.05 (draft value of 530) and 3.07 (draft value of 235). The Patriots come out ahead in this scenario by 5 “points”, which is the equivalent of an early 7th round pick.
This chart not only gives us the ability to analyze trades, but it also gives us a quantifiable value for each pick and allows us to compare picks to one another.
For this analysis, I used the Chase Stuart model developed in 2012 that uses the “approximate value” a player drafted at each position returned to their team five years after they were drafted. For all intents and purposes, its similar to the Jimmie Johnson chart, but with an analytical foundation to back up the evaluations. Each pick has an assigned value, that starts with the highest value at pick 1.01 and descends exponentially from there.
Question 1: Did you add or lose draft capital?
Inspired by a similar Warren Sharp analysis, the first part of this analysis is looking at whether teams add or lost draft capital during the draft using the draft pick value model. Below is a chart of every team’s draft capital before the draft.
And here’s a chart of how that changed throughout the draft.
Throughout the course of the draft, ten teams added draft capital, nine teams did not add or lose draft capital, and 13 teams lost draft capital. Notable draft capital “losers” are teams that traded up in the first round like the Packers, Buccaneers, and Chargers.
This in itself is not enough to answer whether any given team had a good or bad draft, but it can give us direction in their approach. While Miami lost draft capital, they had the most draft capital going into the draft and still led the league after the draft. The Texans added a relatively significant amount of draft capital (equivalent to a mid-5th round pick), but they still ranked only 30th in overall draft capital after the draft.
Part 2: Positional Values
Quantifying the value of a draft pick is easy (mostly because other people have done it for us). Quantifying the value of a position is more difficult.
Intuitively we know that quarterbacks are the most valuable, and positions like long-snapper, punter, and fullback are among the least valuable. However, these values are all relative and team-dependent. We can’t simply rank the positions because the difference in the value of a quarterback and the second most valuable position may be much larger (or smaller) than the difference in value between the second and third most valuable positions.
There are metrics that analysts have used to define positional value (like WAR) but for this analysis, I let the draft “market” define that value for me.
I looked at the results of the last five drafts and, using the draft pick value chart, was able to come up with an average value for each position. This positional value basically translates to the average price teams have paid for that position in the draft.
Below is a chart that showcases the positional values teams gave to each position in this year's draft versus the results from the previous five drafts.
And this chart more clearly illustrates the change in value between this year’s draft and the previous five drafts.
While the wide receivers have been getting all the attention (and rightfully so), it was the tackle position that received the largest premium in this year’s draft.
Question 2: Did you use high draft capital on highly valued positions?
I wanted to combine the findings from the past two parts of the analysis in order to get a more holistic picture of how each team used their draft capital. To do that, I found the best way was to create a positional value “multiplier” and combined that with the draft pick value.
The “multiplier” was based on the positional values of the draft results from the last five years, translated to a non-linear scale. This was then multiplied to the draft pick value in order to gain a new variable that I am calling “positional pick value”. The positional pick values for each team were then added together.
The below chart illustrates the total positional pick values for each team.
And the average optimized pick value for each team.
Clearly teams that spent high draft capital on quarterbacks fair better in this analysis, but it also illustrates the relationship between draft capital and positional value. While the Packers spent their first-round pick on a quarterback, they traded up to do so and therefore gave up draft capital later in the draft. They ranked 20th in total positional pick values and 24th in average positional pick values.
Part 3: Player Values
The final piece of this analysis is also the most subjective and relies heavily on the 2020 NFL Draft Consensus Big Board that can be found at theathletic.com.
Apart from draft capital and positional value, I wanted to look at the player that team’s actually selected and their hypothetical value based on their rankings in the consensus big board.
To break this down, I looked specifically at the overall rankings, and then at the distribution of the players within the rankings at each position. For instance, the gap between quarterbacks was fairly narrow as they have Joe Burrow listed #2 overall and Tua Tagovailoa listed at #5 overall. However, center is a different story. Cesar Ruiz is listed at #31 overall and the next center, Matt Hennessy, is listed at #73 overall.
I took those rankings, and their distributions, and created a percentage scale for each position. Then I assumed every #1 ranked player at a given position was “100%” and created a linear scale off of that assumption using the standard deviation. To simplify, the #1 center (Cesar Ruiz) is assumed to return 100% of the positional value, whereas the #2 center (Matt Hennessy) is assumed to return only 93%, and so on. Similarly, Joe Burrow is assumed to return 100% of the QB positional value, whereas Tua is assumed to return 96%.
Question 3: Did you draft highly ranked players at highly valued positions with high draft capital?
A player’s position rank is a function of the position, and as such the equation to answer the question is above is this:
Absolute Pick Value = Pick value * (Positional Value (%) * Player Value(%))
This new variable, absolute pick value, helps us find the final answer to our question. Below are the total absolute pick values by each team.
And the average absolute pick values by each team.
It is obvious that Cincinnati benefited from selecting at the top of each round and selecting the #1 ranked QB first overall. Looking at their strategy helps us identify what other teams did wrong.
The Bengals didn’t make any trades and selected a player at the top of each round. They picked highly ranked players at highly valued positions.
Conversely, if you look at the progression of the Minnesota Vikings through the three parts of this analysis you can see a very different strategy and outcome.
Minnesota added the third most draft capital and used it on highly valued positions, like WR, CB, Tackle, and DE. However, 11 of their 15 total picks came in the fourth round or later (four in the seventh round) and they were forced to select low ranked players at those positions. They finished the draft 11th in total absolute pick value and 27th in average absolute pick value.
Step 4: What we know, and what we don’t
The large glaring hole of this analysis is “team needs”. However, team needs are difficult, if not impossible, to identify. I find that team needs can be used to explain away discrepancies in the data included in this analysis.
This analysis gives us a clear comparison of how every team utilized their draft picks. Many teams attempted to optimize their pick values by selecting highly ranked players, others targeted highly valued positions. Some refused any offer to trade down, others took every opportunity to add draft capital.
As teams and their fans ask themselves, “Did we draft highly ranked players at highly valued positions with the capital we added?”, some teams can point to the chart above and say yes, while others must face the reality that the answer is no.
