Breaking down Brian Gutekunst’s Packers draft classes by RAS

As any Green Bay Packers fan who follows the NFL Draft closely knows, the Packers value overall athleticism as much as or more than any other team in the NFL. The Packers routinely draft players with elite size and athleticism, and have done so in every draft since Brian Gutekunst took over as general manager in 2018.

One of the best tools for measuring a player’s overall athleticism in a simple number is Relative Athletic Score, or RAS, which is compiled and published by Kent Lee Platte (aka @MathBomb). Kent’s work has been invaluable for giving a quick snapshot of a player’s innate ability, putting that onto an easy-to-understand 0-10 scale.

Do NFL teams use Kent’s RAS values? Probably not exactly, but teams unquestionably have some methodology for assessing players’ abilities based on athletic testing results and those internal tools probably end up taking a similar form as RAS. Therefore, RAS becomes the best proxy that is widely (and freely) available for assessing a player’s athleticism.

So just how athletic have the Packers’ draft classes been over Gutekunst’s five drafts? And have there been any trends or shifts in his approach to getting athletes over the years? I set out to examine those questions by taking a deep dive into the RAS numbers for the Packers’ 48 draft picks since 2018, including their 11 selections this past weekend.

Draft RAS by Year

As we know, Gutekunst (and his predecessor and mentor, Ted Thompson) has always put a high priority on getting great athletes in the Draft. This is no surprise, as his first draft in 2018 was notable for having just one player below an 8.0 RAS. That was fifth-round guard Cole Madison, who was the only player below a 5.0 in Gutey’s first two years at the helm.

What I have done here is compile a summary of each of Gutekunst’s five draft classes, broken down with a few key pieces of information. First, I took the simple average of all of the RAS values for each year’s draft class (omitting any players who did not test) and called this “aRAS.” Then, I computed weighted aRAS values using two different methodologies.

The first of these I call “PaRAS” for “Pick-weighted average RAS.” This assigns a “pick weight” value to every selection in the NFL Draft, ranging from 1 through the maximum number of selections in that year’s draft, and going in reverse order of selections. For example, in 2022 there were 262 selections, so the first overall pick is assigned 262 points, the second pick gets 261, and all the way down to one point for the 262nd overall pick. I then multiplied each of the Packers’ picks’ weights by the RAS of the player they selected with that pick, summed the total of those values, and divided that sum by the total of all of the pick weights from the Packers’ class (again, omitting players who did not have an RAS number). This method gives picks values that decrease linearly as you move later in the draft.

The second average value I computed I will call “VaRAS” for “Value-weighted average RAS.” In this case, instead of the pick weighting as defined above, I used the draft slot trade values for each pick as pulled from the 2022 Rich Hill draft chart, then computed the average the same way as above. This value function is roughly exponential, putting much higher weight on earlier picks and tailing off considerably as the draft goes along, but leaving less of a difference between consecutive picks later in the draft compared to early on. In the table below, I also will include the total value of all of the picks the Packers used in each of these classes for reference (i.e. the total “draft capital,” though for that column I did not omit players who did not test).

While the aRAS is itself somewhat useful, it can be significantly skewed by one player with a particularly poor testing regimen, even if that player is drafted very late, as they will have the same impact on aRAS as a first-round selection. The weighted averages look to adjust that while giving somewhat different weighting values. Personally, my inclination is to use VaRAS as the most useful measure because of how it defines the weighting factors.

Here are the results dating back to 2018:

What can we learn from this? One observation that jumps out is the difference between aRAS and PaRAS/VaRAS for almost every draft. Only in 2018 and 2019 was aRAS higher than either PaRAS or VaRAS, and the VaRAS numbers are higher in every draft than aRAS. This means that in general, the Packers are valuing elite athleticism more early than they are late in the draft, a trend we’ll examine shortly.

Additionally, the differences between aRAS and PaRAS/VaRAS values for 2022 are extreme, ranging from just under 1.0 (PaRAS) to over 1.5 (VaRAS). There are two key takeaways here. First is that the Packers had more high-value picks in 2022 than usual, with three selections in the top 34 picks, and that all three of those players had RAS values above 9.6. The second observation is that the team drafted one player with a RAS below 6, Jonathan Ford with a 3.54, but because his selection came in round 7, it is minimized in the weighted averages. Fifth-rounder Kingsley Enagbare (6.26) and seventh-rounder Samari Toure (6.16) also bring down the aRAS of this class without overly dropping PaRAS or VaRAS.

We can also see that in all three averages, 2021 was the Packers’ least-athletic draft class of the Gutekunst area. Part of this is driven by the fact that there was no testing information for the team’s second-round pick, center Josh Myers; had he come in with a high number, that could have skewed the overall class back up. However, another big factor was the poor testing performance of third-round pick Amari Rodgers, whose 5.37 RAS is one of the lowest for any Gutekunst draft pick in the first four rounds of the draft.

Meanwhile, the Packers’ 2018 class had an exceptionally high aRAS of 8.9, with only one player — guard Cole Madison at a 4.58 — coming in below 8 individually. However, the fact that the high picks were in the 9-9.5 range rather than above 9.6 is what helps push the 2022 class up higher on PaRAS/VaRAS numbers.

RAS by Round

We know that athleticism is a must if you’re going to hear the Packers call your name in the early rounds. But that is only specifically true for the first two rounds of the draft. Indeed, some of the Packers’ worse RAS values in the Gutekunst era have come in the third. Jace Sternberger (5.20), Josiah Deguara (6.69 at tight end), and Rodgers (5.37) were all sub-7 players in round three, a surprising departure from the team’s trends through the draft as a whole.

In fact, it seems that the end of round two is a cutoff for the Packers begin to be willing to reach on players with less-than-ideal athleticism. Take a look at this scatterplot of the Packers’ picks and note where you first see a data point below 8 on the vertical axis:

This shows that while the Packers still are much more likely to draft good athletes at all times in the draft, they begin getting more comfortable with lesser athleticism starting in the third round. Furthermore, they are increasingly willing to reach for lesser and lesser athletes as the draft gets later and later.

The widening spread between athletic reaches and athletic marvels in later rounds has the impact of pushing the trend line downwards as you get farther into the draft, even though many of the late picks still remain clustered above 8.0 RAS numbers.

---

Addenda

There are a few minor notes to mention regarding this analysis. I re-ran the RAS values for some players who switched positions from their assumed default position entering the draft (and therefore their default position for RAS values). For example, Josiah Deguara is listed as a fullback and Jon Runyan as an offensive tackle, instead of tight end and guard, respectively, but I used their TE and G numbers instead.

Additionally, I used a RAS value of 7.47 for 2022 draft pick Romeo Doubs, who ran very late in the draft process. That value used the 4.53 40 time that he mentioned in his first conference call with Packers media, though reports indicated that the Packers timed him around 4.50.

Finally, for the PaRAS, the 2018 through 2020 classes had weighting values from 256 to 1, while 2021 had values 259 to 1 and 2022 was 262 to 1 because of the addition of more compensatory draft picks in those two years. The weighting for VaRAS simply extends out the value of pick 256 (0.9 points) to later picks beyond 256.