A quarterback ranking article that incorporates what has happened on the football field

Whenever there isn’t football being played, quarterback rankings are popular. This offseason has been no different, with Cian Fahey recently raising some eyebrows when he released his quarterback tiers based on the 2017 season, which ranked players like Jacoby Brissett, Jared Goff, and Mitch Trubisky well ahead of Russell Wilson.

In contrast to Fahey’s wholesale rejection of statistics, I set out to make a ranking of quarterbacks that is primarily based on how effective they have been as measured by increasing their team’s chance of scoring to see how it compared.

Criteria

I don’t care about things that happen that a QB has no control over:

• His team’s rushing game (outside of when the QB himself carries the ball)
• His team’s defense

I do take into account factors that affect a QB’s passing efficiency:

• Receiving talent (players like Cam Newton get a boost because of the lack of talent around them)
• Pass protection (players like Russell Wilson get a boost because of how frequently they are under pressure)

Below I will show expected points added (EPA) per play among plays that involved QBs from 2009 through 2017. This includes pass attempts, sacks, scrambles, and designed QB runs (including read option keepers). Because rushing plays very little role in determining who wins NFL games, it is safe to ignore rushes by non-QBs: the effectiveness of a team’s offense will largely be determined by the level of its quarterback play. ESPN’s QBR is built on expected points with some modifications, and predicts out of sample wins better than any other QB metric tested. To the extent that a QB has control over who wins and loses games, it will show up in EPA.

My rankings below are primarily driven by the extent to which a quarterback can consistently be above average over many seasons. If a quarterback is great, his team should expect to have an effective offense year in and year out. QBs with fewer than three seasons starting are not ranked.

Let’s move on to the rankings. I will show QBs in tiers, with QBs ranked within a tier, and commentary below in italics.

The Rankings

Tier 1: Consistently top tier

This is a nine year stretch where the only two years of below average QB efficiency were the Packers in 2015 (when Rodgers was inexplicably bad) and the Packers in 2017 (when Rodgers was injured). As we will see when moving down into lower tiers, being consistently above average — and way above average in most cases -- is very rare.

1. Tom Brady

2. Drew Brees

You could reverse the order here and I wouldn’t argue much, but these two are clearly a cut above the rest.

3. Matt Ryan

4. Aaron Rodgers

If this ranking had been done at the conclusion of the 2014 season, Rodgers would have been ahead of Ryan, but Ryan’s last three seasons have been substantially more efficient. He’s also younger and doesn’t have the injury history that Rodgers does.

Tier 2: Consistently above average, sometimes great

5. Philip Rivers

6. Russell Wilson (starter since 2012)

7. Ben Roethlisberger

Ranking these three is difficult. How much of a boost should Rivers and Roethlisberger get for being incredible in 2009-2011, given how much time has passed since then? When comparing Wilson to Roethlisberger, how much of a boost should Wilson get for his massive disadvantage in supporting cast (not having an offensive line or Antonio Brown)? I’m putting Wilson ahead of Roethlisberger because of his durability and being forced to play the past two seasons without pass protection, but I don’t feel strongly about who should get the #6 spot. These three are clearly above the next two:

8. Cam Newton (starter since 2011)

9. Kirk Cousins (full-time starter since 2015)

Newton goes above Cousins because of a longer track record and playing with less receiving talent. He is underrated by metrics that omit rushing: first downs are valuable, and he has rushed for 332 of them.

Tier 3: Consistently above average

As we move down the list, we see fewer elite seasons and a lot of seasons that are pretty close to average, but these QBs have a track record of being above average.

10. Matt Stafford (full-time starter since 2011)

11. Andrew Luck (starter since 2012)

12. Jameis Winston (starter since 2015)

13. Alex Smith (starter in Kansas City since 2013)

Stafford goes first by a narrow margin because he has two of the best three seasons among this group (2011 and 2016). Luck goes second due to having a couple good seasons without much of a supporting cast. Winston is the biggest surprise for me: I tend to think of him and Mariota in a similar light, but Winston has been consistently above average in efficiency and Mariota has not. Smith is last because he only has one season meaningfully above average and he did so while in a great situation.

Tier 4: Some seasons above average, some seasons below

14. Andy Dalton

15. Eli Manning

16. Tyrod Taylor

I don’t know what to make of any of these QBs other than that they are all heavily reliant on their supporting casts. The trends for Manning and Dalton are concerning and I’d probably drop Manning a tier if he has a third consecutive below average season in 2018. Taylor could arguably be in the tier below but I bumped him up here for being close to average while the team around him was dismantled.

Tier 5: Consistently below average

17. Blake Bortles (starter since 2014)

18. Joe Flacco

19. Derek Carr (starter since 2014)

What a mess. How to rank these players? I’m putting Bortles first because he has the most recent good season and Flacco second because he’s the only one with multiple seasons above average, even though most of them were a long time ago. Carr had one season where he was barely above average that he somehow received MVP votes for, so he has that going for him, which is nice.

20. Marcus Mariota (starter since 2015)

21. Sam Bradford (I’m showing STL since most of his time starting was there. He was below average in PHI in 2015 and above average in MIN in 2016)

22. Ryan Tannehill (starter since 2012)

Mariota’s place is a big surprise. Despite having a good offensive line and rushing a fair amount (which EPA models typically like from QBs), his 2015 and 2017 seasons were well below average. Bradford and Tannehill both have health questions and several seasons well below average. With all of these players, it’s possible that they have been primarily held back by coaching. Hopefully Mariota will see a boost in 2018 as the result of being freed from Mike Mularkey, but until that happens, this is the tier that his efficiency has placed him in.

Tier 6: Incomplete (not enough seasons as a starter)

I am not ranking QBs who have not yet started three seasons. This includes Carson Wentz, Dak Prescott, Jared Goff, Case Keenum, Jimmy Garoppolo, Deshaun Watson, and Mitch Trubisky. Wentz and Prescott could join Tier 3 with good seasons and possibly Tier 2 with great seasons in 2018. Goff would probably join Tier 3 with a repeat of his 2017 season. The remainders need two more seasons to qualify based on my filters.

Conclusion

Fahey and I agree on the four Tier 1 QBs, but after that we diverge strongly. The placement of Newton and Winston is substantially higher than I expected coming in, and Mariota lower. With QBs that have only played three full seasons (e.g., Cousins, Mariota, and Winston), their 2018 seasons have the chance to move them up or down.

The Seattle Seahawks have a good QB that they should not trade. Enjoy it while it lasts. Being a franchise that flounders in the Tier 5 wasteland for a decade is no fun.