The data-driven analysts look down on mouth-breathing jocks engaging in confirmation bias and word salad trying to parse Cover-3 buzz vs. cloud. For the tape grinders, the analytics nerds only believe in the numbers and don’t care about the “how” or “why.” They couldn’t tell you a thing about real football, just what’s scribbled on the papers in their mom’s basement. Of course, they’re both wrong, at least about the the other side, and what’s more, in order to form a full picture of what’s happening, it’s essential to take into account both sides of the equation, no math pun intended.
Too often we define the quality of a player by All-Pros or Pro Football Focus grades or DVOA alone. The more learned analysts and fans will combine one or two markers, but generally we tend to use the means through which our argument can be best made. It’s why Tom Brady stans love quarterback wins and why Vikings fans suddenly became experts on acronyms like EPA and CPOE.
But is there a way to combine these elements? To join the qualitative, the quantitative, and the gray areas in between to tell us who the best players are at the most important positions in football? Well, I tried.
To skip the explanation, scroll down to the QB chart to see the quarterback rankings and explanation below.
How are we actually doing this?
The goal of this project was to determine who had the best premium position player grouping in football: quarterback, receiver, pass rusher, cornerback, offensive tackle. Just take the best guy the team has at each spot and find out who puts together the top group of five. I had a suspicion I knew, but wanted to find out if there was a better way to prove it than relying solely on a single stat or an analyst’s feelings.
For me, the quantitative informs the qualitative. There’s no substitute for tape study, but understanding trends and data challenges preconceived notions and upends unconscious biases. So I looked to a host of measures, some assessing quality on a subjective scale, others purely analytic, while still others attempted to combine the two in some way.
For the subjective piece, I used PFF grades and multiple grades in some cases for non-quarterback positions. For example, coverage grade is particularly important for a cornerback even if their overall grade is slightly lower due to tackling for example.
On the analytic side, we used Pro Football Reference’s Approximate Value for every player, a measure of a player’s value to his team’s success. For quarterbacks, I also used Ben Baldwin’s CPOE+EPA index stat for a purely outcome-based metric. This is the inherent problem with—and where I often butt heads with Ben on—these kinds of tools. They only tells us the outcome of the play. On the other hand, over time these numbers tend to be stable, accounting for the peaks and valleys of inconsistencies around a quarterback.
Sports Info and Solutions created Total Points to blend the subjective and data sides of the equation. They take film breakdowns, assign credit for a given play and use EPA to assign a real value for a player on that play. Not all four-yard runs are equal. If an offensive linemen blows a block that could have led to a bigger gain, that debits his play. If a quarterback makes a good throw but the receiver drops it, the quarterback still gets some value creation for that throw even though you wouldn’t see it in the stat sheet, and thus not in places like CPOE or EPA.
Finally, we add a critical piece, Football Outsiders’ DVOA, which adjusts for schedule. This provides added context. Don’t simply judge the outcomes, judge them based on the opponent. Putting these four metrics (or five, in the case of quarterbacks), we get a fuller picture of who is contributing the most to his team’s success.
In other to be included in the final list, a player needed to be in the top-16 of at least one of these categories, mostly just for my own sanity to not have to run this for every player in the league. Isolate the best guys and parse from there.
Each player is benchmarked relative to the best player in each category to determine how good they were relative to the top, as well as to make it easier to compare players across metrics. They’re then assigned points based on their place in the league and those points are added up so at the end we’re left with the summation of traits for the player based on these metrics.
The 2019 Quarterback Index
For the first part of our series, we’re looking at the quarterbacks. Here’s how the 2019 class fared last season. Each ranking is scaled to a maximum value of 10, so with five metrics analyzed here the highest number of points available is 50. Additionally, we considered only quarterbacks with at least 150 dropbacks last season.
2019 Quarterback Rankings
|Name||Index||PFF Grade||SIS Total Points||AV||EPA+CPOE||DVOA|
|Name||Index||PFF Grade||SIS Total Points||AV||EPA+CPOE||DVOA|
The top of the list looks mostly as we’d expect. Lamar Jackson’s 2019 season provided brilliant and unique quarterback play and provides an ideal example for why quarterback rating so often misses the mark. Jackson’s value stems not only from his ability to create through the air; his running threat would have made him essentially the best running back in the league last year by efficiency. To be sure, the threat of his legs opens up the passing game for him, but that’s point of the game. We see his running value, in part, expressed through his passing success.
Packers fans won’t be surprised to see where Aaron Rodgers lands on this list, coming in at 9th in the league. He produced last season in important moments, but he lacked consistency and the combustibility we’re used to seeing with him. Considering the names above him, it’s hard to argue he should be higher, perhaps with the exception of Derek Carr, the only surprising name at the top of the list.
This shouldn’t be taken to mean Carr or Kirk Cousins is better than Rodgers in a vacuum, or that you’d rather have those guys than Rodgers to win one game (I know I wouldn’t), but rather reflects what we saw last season.
The more surprising names on the list showed up further down the list. For a player like Carson Wentz to come in behind Jacoby Brissett and Jared Goff speaks to a dramatic drop in his play since his near-MVP season. Wentz’ offensive arsenal faced rising injuries as the season went on, but he still boasted a very good offensive line and the best 1-2 tight end punch in football, to go along with a dynamic backfield.
What’s more, stats like Total Points and PFF’s grades account for quality plays Wentz made which went begging because a receiver dropped the ball. At this point, the fantastic season we saw in 2017 looks more and more like an outlier with Wentz and comparisons between him and Dak Prescott vastly understate good excellent Prescott has been to this point in their respective careers. Prescott is an elite quarterback, period.
And we can’t engage with this analysis and not mention Tom Brady. For all the same reasons we can’t assume Wentz will magically be better in 2020 than 2019, the same goes for Brady. His receivers will be much better this year, but the offensive line will be considerably worse and he’ll play without the best brain trust in football. Brady’s ‘19 isn’t an outlier either. He’s been on a downward trajectory for several years now (not unlike Rodgers, though Brady’s is more precipitous and he’s further down the decline scale).
It’s not just the supporting cast. Brady frankly wasn’t particularly good last year, refusing to stand in and face pressure and struggling if his first read wasn’t there as a result. Balls would float and hit the dirt near receivers if Brady couldn’t get the ball out in rhythm.
Next up in the series, we’ll take a look at receivers as we put together the picture around which team has the best group of premier talent in football.