What NFL teams had the best drafting operations from 2017 to 2019?

It’s often been said that NFL general managers earn their money on days 2 and 3 of the draft, and that sentiment has a certain intuitive appeal. In looking at the selection of first round picks in the 2017-2019 drafts, there’s actually not a ton of divergence in opinion between the commentating “draftniks” and the professional scouts and GMs working for NFL teams. The pros tend to pick a little better than the commentators, but using a consensus commentator big board would absolutely get you into the same ballpark as the pros in the first round.

After that, things tend to get more fuzzy, however, as the talent pool expands and the depth of evaluations contracts. MattinBrisVegas has written a number of articles discussing how the likelihood of player success drops off fairly dramatically after Day 1 and early Day 2, but we’ve both wondered if, in the midst of a league-wide “fog” of mostly misses, we might be able to devise a means of picking up a “signal” of team-specific success.

Tasked with the specific mission of seeing which teams have drafted best in the recent past, we each went at the task in different - though slightly similar - ways. Here are my results. Matt’s will be available in a subsequent piece.

The Approach

Building on the methodology I used to evaluate the accuracy of consensus big boards (and NFL experts generally) in picking first round talents - based on looking at actual NFL performance several years after the draft - I examined how all 32 NFL teams performed over the draft period between 2017 to 2019. By design, at a minimum, all of the players evaluated had at least 3 actual seasons of NFL performance (or at least the possibility of accruing 3 actual seasons) to examine.

The key pieces of data used in this evaluation were “actual draft position,” “weighted career approximate value (wAV),” and “wAV rank.” “wAV” is simply a global measure of actual performance in the league. It doesn’t necessarily allow for easy comparison between draft classes - because as players accrue years in the league their wAV tends to increase - but it is one of the best publicly available tools to compare relative performance within a draft class, which is how it is used here.

“wAV rank” is the just the re-ordering of drafted players based on their actual wAV scores several years after the draft. So, for instance, Lamar Jackson was the 32nd player taken in the 2018 draft, but his wAV rank, after 4 years in the league, is number 1 in his draft class. Conversely, Baker Mayfield, taken number 1 overall in that class has a current wAV rank of “5.”

Subtracting the wAV rank from the actual pick location gives a variance score for each draftee, which I’ve labeled “wAV rank minus pick (WRMP)” for simplicity’s sake. This allows us to examine which players have outplayed (or played below) their draft position, relative to the talent around them, and by how much.

Again, as an example: In the 2017 draft class, Commanders’ center Chase Roullier was picked in the 6th round (#199). Since that time, he’s been the 51st best player in his class in terms of wAV, essentially performing like one would expect a mid-second round pick to play. As a result, his WRMP is 148, which is one of the best deltas in his entire draft class. A true “steal,” if ever there was one.

The flip side of this approach is that it dramatically dampens the value of what teams do in the first round - unless they completely bomb that pick, like the Eagles did with Andre Dillard at the #22 spot in the 2019 draft (his WRMP is currently -83). This approach assumes that teams will come pretty close to hitting on their first round pick, and penalizes them seriously if they miss badly, which, I think, is appropriate in this context.

I’ve summed the WRMPs for each team’s class, and then divided those “wSums” by the actual number of players taken by each team in the draft to create a “WRMP average (wAVG)” for each franchise in each of the draft years. A positive wAVG indicates that the team is - across the entire draft - picking players who are performing above their draft position.

In a fairly dramatic example, in 2019, the Dolphins were tops on this metric, with a wAVG of 58, meaning that for the 6 players they drafted, they averaged play that was about 2 rounds above expected. This was anchored by nailing the picks of Myles Gaskin (7th), Andrew van Ginkel (5th), and Isaiah Prince (6th), and not missing badly on any of their picks, including first rounder Christian Wilkins.

At the other end of the spectrum in 2019 were the Eagles, who had a disastrous draft, notching a wAVG of -61, meaning that the players they took were overdrafted by about two rounds, on average. This includes the aforementioned Dillard, JJ Arcega-Whiteside, Shareef Miller, and Clayton Thorson. Miles Sanders was the only bright spot from that draft for Philly.

In short, this analysis favors those teams that don’t make big mistakes in the earliest rounds, and that hit on gems in the later rounds.

Enough of the background. So how did each of the 32 teams fare during this period?

As you can see, there’s actually a pretty dramatic spread in average performance across the league during this period. The Bills, Rams, and Dolphins were the cream of the crop from 2017-2019. It probably shouldn’t be too much of a surprise that one of those teams won the Super Bowl last year, and another made it to the AFC championship game. The Dolphins haven’t yet seen the fruits of their labor, but this suggests things may be looking up for them as well.

Here I’ll reiterate an important point: Though the Bills took stud QB Josh Allen in the 2019 draft, they get almost no “credit” for him in this exercise. His WRMP is a measly “4,” basically because you expect the guy drafted #7 overall in his class to be one of the best performers. His fellow classmate, Tremaine Edmunds, also taken in the first, was only a “7,” but the Bills’ draft class basically had no “misses,” and 5th rounder Wyatt Teller, was Roullierian-level steal.

I’ll also remind readers the Rams had ZERO first round picks during this period, and only two second rounds picks. In addition to their creative salary cap management, it’s looking more and more like very solid mid/late round drafting has been a core component of their recent success.

At the other end of the spectrum are the Broncos, Jets, and Giants, which should likely come as surprises to no one. Each of those teams has floundered miserably for the past several years, and had general manager turnover since the beginning of this period.

The Broncos added George Paton in 2021, the Jets added Joe Douglas in 2019 (after this miserable 2019 draft), and the Giants recently jettisoned Dave Gettleman <sheds tear> for the Bills’ Joe Schoen, protégé of arguably the best drafting GM in the business, Brandon Beane (see photo above).

At this point, I’ll take a moment to remind the audience that Schoen was high in the running for the Washington GM job before Martin Mayhew was hired <weeps silently>.

Another Approach

In processing the results from this initial exercise, one potential criticism is that it lumps all the picks in a particular draft together and averages them. Hit on a grand slam steal, or bust badly on a high pick, and it risks elevating the whole draft or bringing it down, depending on the circumstances.

So, I thought I’d also look at the number of “impact steals” each team had in each of these drafts. For the purposes of this exercise, an “impact steal” is a player whose WRMP times their wAV exceeds “1000” in the 2019 draft (and is adjusted upward with each year further back in time). That is, a player whose actual performance is higher than would otherwise be expected, and whose performance itself has been exemplary. So, this sorts out those like 7th round picks who simply enjoyed a modicum of success relative to their wash-out peers, but otherwise were unremarkable in the scheme of things.

Conversely, it still grabs a 7th rounder, like Miami RB Myles Gaskin who has played incredibly well by any standard. The statistic is a bit contrived, but based on the results, it seems to pick up the right group of above average starters one would ideally acquire in a draft, particularly after the 3rd round.

The table below is the total number of “impacts steals” for teams over the three drafts examined. The Bills, who finished 1st in the prior exercise, finish third here, with 4 impact steals from 2017 to 2019. Those players include: Dion Dawkins (2nd round), Matt Milano (5th round), Wyatt Teller (5th round), and Devin Singletary (3rd round). And, these are in addition to their first and second round picks who all performed about as expected, or better: Tre’Davious White, Josh Allen, Tremaine Edmunds, Cody Ford, and Ed Oliver.

The Ravens, who paced the field on this metric, acquired 6 impact steals, including: Chuck Clark (6th round), Lamar Jackson (1st round), Orlando Brown & Mark Andrews (3rd round), Brad Bozman (6th round), and Zach Zieler (7th round).

Washington finished top 10 in both metrics, with 3 impact steals during this window: Chase Roullier, Terry McLaurin, and Cole Holcomb.

In looking at general trends, the Rams, Dolphins, and Chargers finished at the top end of both evaluations, and the Jets, Giants, and Broncos pulled up the rear in each case.

I’m certain there’s much more to glean from this information, and I look forward to discussing it with you in the comments below, but, before I go, I would recommend keeping an eye on this group of teams for (near term) success. In some cases, it will be continued, in other cases, it would be a change of pace:

• Buffalo Bills
• LA Rams
• SF 49ers
• LA Chargers
• Miami Dolphins

And expect these “apparent top drafters” to struggle:

• Minnesota Vikings
• New Orleans Saints
• Tennessee Titans

The Commanders are at the outer edge of first group, and should be playoffs contenders based on this era of drafting. Whether that will persist or not, depends tremendously on the success of the last couple draft classes, which are simply too early to judge.