How to Win in the NFL: Don’t Overpay Your Quarterback

The end of the NFL regular season is finally here, and 12 teams remain. As a loyal New York Jets fan, this season was just another chapter in an increasingly discouraging fandom, and yet I remain a relatively frequent football watcher (although science around brain trauma and CTE have definitely decreased my love of the sport).

Over the past two days, I have read two NFL-related stories on ESPN that I found interesting. The first (read yesterday) details Jay Gruden’s “lukewarm” review of Kirk Cousins’ play this year, while the second (read today) reviews the best and worst quarterbacks in the league based on QBR, a metric used to evaluate quarterback performance. While the inspiration for this post comes from material from both yesterday and today, this is an entry that I’ve known from the start that I wanted to post, so I’m going to cheat a little bit and write about my quarterback theory today.

For a few years now, I have had a theory about quarterbacks, and more broadly, what it takes to win a super bowl in the NFL. The theory posits that there are three tiers of quarterbacks in the NFL:

• Tier 1: Quarterbacks who, regardless of how much a team pays them, can carry their teams to super bowls. Example: Tom Brady
• Tier 2: Quarterbacks who have the potential to carry their teams to super bowls if their supporting cast is good enough. However, they are not good enough to win super bowls when they are “highly paid.” When this happens, there is not enough money to go around for other players, and the quarterback is not good enough to carry the team. Example: Joe Flacco
• Tier 3: Quarterbacks who, under no circumstances and regardless of how little they are paid, will not quarterback a super bowl winning team. Example: Trevor Siemian

The underlying assumption behind this theory is that quarterback performance is inextricably tied to team success in the NFL. I don’t think this claim is very controversial, but it is notable that this theory posits that teams’ chances of winning super bowls are entirely dependent on (a) how good their quarterbacks are and (b) how much they pay them.

For a long time, I’ve wanted to test this theory by seeing if we can predict the success of Team X simply by using the following two variables:

• QBR of the quarterback who played the most snaps for Team X
• Annual salary of the highest paid quarterback on Team X

Before getting started, these variables deserve a bit of explanation. We use the QBR of the quarterback on a team who played most throughout the season as a proxy for overall quarterback performance. There are some other advanced metrics around, but QBR is pretty widely accepted, so that’s what we’ll use.

We use annual salary of the highest paid quarterback on a roster, rather than the salary of the one who played the most. We do this because, at the end of the day, this theory tries to measure the “bang for your buck” that teams get at the quarterback position. If a team’s star quarterback gets injured (e.g. Aaron Rodgers), tough luck! That was still money spent on quarterback production, rather than on strengthening the rest of the roster.

So, let’s look at the data. First, the salary cap hits of the highest-paid quarterbacks on every team, from highest to lowest:

I have organized quarterbacks into two groups, one on the left and one on the right. On the left, we have the tier of “highly paid” quarterbacks. I set the cutoff at $15m per year, which almost cuts the population of 32 quarterbacks exactly in half. Moreover, the average contract for the quarterbacks above is ~$13.5m, meaning the “highly paid” group makes above the average of the salaries of the highest paid quarterbacks on all 32 NFL teams.

This part of the analysis is definitely inexact; we could set the cutoff somewhere else, but based on the salaries above, this seems like the most natural cutoff in terms of salary. Notably, some teams (like the Dolphins, who paid Jay Cutler $10m this year) have multiple very expensive quarterbacks, which messes with the actual quarterback spend per team. However, because I’m unwilling at the moment to calculate those exact figures and I think they probably don’t impact the analysis that much, we’ll use the salaries of the most expensive quarterbacks as average spend per team.

Next, QBR of the quarterback who played the most snaps on each team. Importantly, this quarterback may differ from the one who was paid the most. According to ESPN, the data is as follows:

I have divided up the 32 quarterbacks among three performance tiers: tier 1, tier 2, and tier 3. Tier 1 quarterbacks had a QBR of over 60 for the season, tier 2 quarterbacks had a QBR between 45 and 60, and tier 3 quarterbacks had a QBR below 45. These divisions work rather nicely — we have 8 quarterbacks in tier 1, 13 in tier 2, and 11 in tier 3. Importantly, the Texans and 49ers did not have any quarterback play enough to qualify for the QBR rankings. The teams were near mirror images of each other at the quarterback position this year, with the Texans receiving outstanding play at the beginning of the season from Deshaun Watson and the 49ers receiving similarly high-level play from Jimmy Garoppolo at the end of the season.

Combining the above two tables nets an analysis as follows:

To remind ourselves, the original thesis was: teams can only win super bowls with either tier 1 quarterbacks or tier 2 quarterbacks who are not “highly paid.” So, let’s look at what the data says— overall, pretty good results! I have highlighted the 3 teams of the 32 total that seemed to have performed contrary to the thesis:

• Cowboys: Dak Prescott was a tier 1 quarterback who is not highly paid. However, the team missed the playoffs.
• Panthers: Cam Newton was a tier 2 quarterback who is highly paid. However, the team made the playoffs.
• Saints: Drew Brees was a tier 2 quarterback who is highly paid. However, the team made the playoffs.

First of all, making the playoffs is very different than winning the super bowl, so if the Panthers or Saints win the super bowl this year, the thesis will be more in jeopardy. However, it is important to note that not all quarterbacks in the same tier are created equal; Drew Brees was the highest rated quarterback in tier 2, and Cam Newton’s running ability makes QBR a less effective performance metric for him.

Overall, the results are satisfactory — all playoff teams either have excellent quarterbacks or decent quarterbacks who are underpaid. On the other hand, teams with low-performing but highly paid quarterbacks (such as the Bengals, Colts, Dolphins, and Giants) all missed the playoffs.