FO 10th Anniversary: Best Quarterbacks
by Danny Tuccitto
1. Which quarterback owns the record for highest pass DVOA in a single season? What season was it?
2. Which quarterback owns the record for highest pass DYAR in a single season? What season was it?
3. Which quarterback owns the record for highest rush DYAR in a single season? What season was it?
4. Which quarterback owns the record for highest pass DYAR in a single game? What game was it?
5. Which quarterback owns the record for highest rush DYAR in a single game? What game was it?
6. Among quarterbacks who debuted after 1990, who owns the record for highest pass DYAR over the course of his career?
7. Among quarterbacks who debuted after 1990, who owns the record for highest rush DYAR over the course of his career?
8. Among quarterbacks who debuted after 1990, who owns the record for highest total DYAR over the course of his career?
Before reading the rest of this article, write down your answers. And for any cheaters out there, today's special guest proctor is Mr. Selig. His punishments may not be swift, but they can be ruthless enough to make a grown man feign emotion. Consider yourself warned.
Done yet? OK, good.
Moving on, you may be aware that FO is currently celebrating its 10th anniversary. Today's festivities involve a look at the best quarterback performances in Football Outsiders stats, which currently date back to 1991. (1989-1990 are hopefully coming before we start the 2013 season, but we don't have them ready yet.) All of the season stats you'll see are freely available on various pages of the site, and Vince Verhei's weekly "Quick Reads" column details our stats for each game during the season. In short, you could have put most of these lists together yourself.* But seeing as how you probably have a life while I clearly do not, I've gone ahead and done the heavy lifting. You're welcome. And as an infamous "Brain" once said, "You won't offend me with cash."
(*Ed. Note: Actually, if you put the lists together yourself, they would look slightly different because of an error that has misvalued roughly 6-8 interceptions per season. This article contains fixed numbers, which haven't yet been added to our stats pages or the player pages.)
Without further ado, below is the shocking answer to our first question.
|Best Pass DVOA, Season, 1991-2012
(min. 100 passes)
As I was watching the Hall of Fame game the other night, and the Cowboys quarterbacks coach got a fleeting moment of camera time, little did I know I was face to face with the single-season record holder for pass DVOA. Now, we can crack wise here in any number of ways, but considering that few of us have contemplated Wade Wilson's career over the past decade, this actually seems like a great time to stroll down memory lane. Outside of Vikings fans and NFL history buffs, who remembers that he led the league in completion percentage and made the Pro Bowl with Minnesota in 1988? Me neither. DVOA doesn't go back that far yet, but just from knowing a little bit about how other advanced stats relate to DVOA, I'm pretty confident in guessing that Wilson's DVOA in 1988 won't be as good as the one in the table; and that makes me start to wonder if I should have appreciated his talent more when I was a kid. (So much for not cracking wise.)
In all seriousness though, it seemed obvious at first that Wilson's spot at the top was simply a byproduct of FO having a lower-than-standard qualifying threshold for attempts: His 169 pass plays in 1992 were more than 100, but less than 224. Yes, if we bump up the threshold, both he and Boomer Esiason drop out of the top 10, replaced by Mark Rypien's 41.9% DVOA for the best team ever (so far) and Philip Rivers' 41.7% DVOA in 2009. However, consider the reverse. If we drop the NFL's threshold to 100, Wilson would have finished 1992 ranked first in adjusted net yards per attempt (ANY/A), eking out the win over another member of our list, Steve Young. (Do you remember that Wilson was the Vikings starting quarterback in 1987 when Minnesota scored one of the biggest upsets in playoff history? I had forgotten, and would have preferred it to remain that way.)
If we look at all quarterback seasons from 1991-2012 using the 100-pass threshold, Wilson's 8.12 ANY/A in 1992 ranks 16th. In the end, I may not agree that Wade Wilson's nine-game exploits in 1992 represented the most efficient passing season of the past 22 years, but I do think it provides a nice diversion on a Wednesday afternoon.
Onto the answers for Question 2 and Question 3, which you'll find in the tables below:
No Wilson-esque surprises here. In this sea of predictable results, one thing may stand out to some readers: Football Outsiders is biased in favor of the New England Patriots. Of course, an alternative explanation of Tom Brady's dominance in pass DYAR over the past six seasons could be simply that it's one of the greatest stretches of valuable quarterback play in NFL history. With a ragtag receiving corps at his disposal this upcoming season, the proof of that clam chowder will be in the eating. Regardless, it's still quite a feat that Brady set the pass DYAR record in 2007, lost 2008 to a shredded knee, and then proceeded to rattle off four more of the top 10 seasons ever upon returning. (Nod to Colts and Broncos fans: Yes, Peyton Manning's comeback season in 2012 was also quite a feat as well.)
Meanwhile, Michael Vick has been the Tom Brady of quarterback rushing value over the past decade or so, but there's a far more interesting conversation to be had once we look at the right side of the next table, which gives you the answers to Question 4 and Question 5.
(UPDATE: For your convenience, I've added links to the PFR box score for each game. If you want to reminisce or see what the quarterback's standard stats looked like, just click on the week associated with that performance in the table.)
As Vince discussed after Week 6 of last year, there have been only eight games of 50 rush DYAR or more by a quarterback. Vick produced three of them, along with four of the top 10 games since 1991. You know who has produced five of the top 10? The Minnesota Vikings, and their repeated appearances on these lists is really starting to freak me out. There's Daunte Culpepper running twice through the okra patch. There's Joe Webb, another forgettable Vikings backup destined to be rediscovered by some intrepid FO writer 10 years from now. Vick's crowning achievement came against the Vikings in 2002. As did the performance by [Redacted] rookie Robert Griffin in Week 6. (Not yet in the table is Steve Young's performance against the 1988 Vikings, which produced this iconic run. Seriously, what's up with Minnesota's run defense against quarterbacks?)
It's actually a nice bit of symmetry that both Vick and Griffin have had their best rush DYAR against the same opponent because they might be the only two quarterbacks who populate the 2023 version of the table. Griffin may only have one game among the best 10 ever (so far), but as a rookie he also amassed at least 35 rush DYAR in Week 2 (45), Week 3 (40), and Week 17 (37). Of course, the real problem for Vick keeping this record for very much longer is that, unlike when he seemed to be the only dog in the fight 10 years ago, today's NFL includes more than just Griffin. Russell Wilson also made the best-ever (so far) table in his first season, while Gang of Four compatriot Colin Kaepernick almost made it in his first half as San Francisco's No. 1 quarterback (37 rush DYAR in Week 10). Elder statesman Cam Newton also just missed the table with 38 rush DYAR in Week 14 last year.
Other than noting that Brady has three of the 10 games in addition to five of the top 10 seasons (so far), I'll leave discussion about single-game pass DYARs to the comments section. Between another Vikings quarterback ranking No. 1 and two of the other games coming against Minnesota's defense (Scott Mitchell's in 1995 and Drew Brees' in 2011), I would really like to finish writing this article without needing a despojo shortly thereafter.
So, let's move on to the best careers since 1991 according to pass DYAR and rush DYAR. It's gotta be Brady and Vick at the top, right? Below are the answers to Question 6 and Question 7. (An asterisk means the quarterback is still active.)
Personally, I would have given half-credit for answering that trick question with "Tom Brady," but Mr. Selig overruled me. His exact words were, "Kid, close only counts in horseshoes and Doug Eddings' strike zone."
By virtue of having a two-year head start, and Tom Brady not becoming ROBO-QUARTERBACK until his fifth season in the league, Peyton Manning currently holds the career pass DYAR record. To wit, since 2004, they've played the same number of seasons and are virtually tied: 14,637 pass DYAR for Manning; 14,339 for Brady. Add in the man who retires more often than a pit crew, and you basically have the three individuals that have dominated NFL conversation over the past two decades. I don't think that's a coincidence.
The rest of the pass DYAR rankings are an equal mix of quarterbacks who began applying their skills to NFL football around the same time FO did, and the ghosts of 90s playoff teams. (Also, four former members of the Minnesota Vikings; but I digress.) The most striking thing about this group, however, is that none of them seem to have a legitimate shot at overtaking Manning (or maybe Brady) one day. Tony Romo is over 15,000 pass DYAR behind and on the wrong side of the quarterback aging curve. Ditto Aaron Rodgers, although he's a "young" 30 years old in terms of NFL wear and tear. Even Matt Ryan, who's currently sitting at the peak of the curve, probably needs to average something absurd like 2,000 pass DYAR over the next 10 years. Two problems with that: (1) His career high is 750 pass DYAR less than that; and (2) It assumes Manning and Brady end their careers around 25,000.
When it comes to career rush DYAR, the amazing thing about Vick is obviously not that he's the best ever (so far). No, it's that the No. 2 active player, Aaron Rodgers, is nearly 900 rush DYAR behind him. Granted, Newton (or Griffin or Wilson or Kaepernick) could probably sleepwalk his way to the record by the time we write "FO 20th Anniversary Stats" in 2023, but it's still pretty impressive that Vick has such a lead with one foot in the Daunte Culpepper Commemorative Scoring Tent sponsored by Al and Alma's.
The other deceptively quick thing I'll say about our career rush DYAR rankings is that they certainly have an Ernest Thomason-esque quality to them. When a great like Tom Brady is only the 19th-most valuable running quarterback of the past 22 years, it's definitely time for the CQBAA to start a petition -- which happens to provide a nice segue to this column's final act.
You may have noticed that, in addition to FO's Fleet-Footed Foxboro Flame, nine other quarterbacks in the top 22 for career pass DYAR also appear in the top 22 for career rush DYAR. It seems that, even before the recent trend represented by professional football's Gang of Four, the most valuable quarterbacks weren't just elite passers. Aaron Rodgers ranks among the top 10 in both categories, and he's been Green Bay's starter since 2008. Steve McNair's rankings are even more impressive, to the point of considering whether the public vastly underrated his career. (Not to Wade Wilson levels, mind you, but still.)
This raises one final extra-credit essay question for today's pop quiz: How will the rankings change when we combine pass DYAR and rush DYAR into an aggregate value we'll call "total DYAR?" Intuitively, you probably wouldn't, given that the former is greater than the latter by a factor of 10. In that case, let me make the question a little tougher.
As readers of FO, you're almost certainly familiar with our friends at Pro Football Reference. Five years ago, Doug Drinen saw a glitch in the matrix when he produced all-time rankings based on the total approximate value (AV) players had amassed over the course of their careers: It doubled as a list of who played the longest. To strike a balance between performance and longevity, rather than simply adding up a player's yearly AVs, he instead gave the player 100 percent credit for his best year, 95 percent credit for his second-best year, 90 percent credit for his third-best year, and so on. In Drinen's system, a shorter-career, higher-peak player will have a higher weighted career AV than a longer-career, lower-peak player even if their unweighted career AVs are similar. For instance, Mark Brunell (120 AV) had a slightly more valuable career than Kurt Warner (113 AV) if you just add up their yearly AVs . With weighting, however, the gap completely disappears because Warner climbed two stratospheric peaks in 12 years while Brunell hung on like grim death long after his most valuable years were behind him.
A short time later, Drinen's PFR colleague, Neil Paine, took a different approach in the specific context of ranking quarterback careers. Namely, instead of applying Drinen's weights, he just took the average value produced by a quarterback in his best six seasons. (Those with five or fewer seasons got zeroes plugged in for missing seasons.) The best part: Paine used DYAR's defenseless cousin YAR where possible, and used PFR's other advanced stats to estimate seasonal YAR prior to 1994.
So now the extra-credit question is, "How will the rankings change when we look at career total DYAR and how will they change when we also look at two additional measures of career value: (1) weighted career total DYAR and (2) average DYAR in the player's best six seasons? Quickly write down your answers -- Mr. Selig is running late for his yoga class -- and consult the table below to see how you did.
As you can see, when we aggregate up to total DYAR (i.e., the table on the left), McNair climbs four spots from where he was in the pass DYAR rankings shown earlier. Three other quarterbacks who ranked highly in rush DYAR earlier move up one spot here (Jeff Garcia, McNabb, and Culpepper), while four statues move down one spot (Warner, Drew Bledsoe, Eli Manning, and Matt Schaub). Otherwise, the total DYAR table's basically the same as the pass DYAR table. In fact, the same 22 quarterbacks appear in both.
Similarly, when we apply the weighting system (i.e., the table in the middle), total DYARs drop but the rankings remain almost identical to the unweighted table on the left. To my eyes, there are only four significant changes, all of which are related to "peak vs. longevity." First, the gap between Peyton Manning and Tom Brady shrinks significantly; Brady's last six years have been that sick. Second, Brett Favre rightly gets penalized for dragging us through multiple retirement sagas. Third, Rodgers essentially switches places with Ben Roethlisberger because his higher peak more than makes up for a slightly shorter career. Finally, Culpepper's big bang of a career allows him to cruise up five rows.
The big changes show up when we average a quarterback's six best seasons to judge the total value of his career (i.e., the table on the right). Brady just barely leapfrogs Manning into first place. More importantly, though, he clearly separates himself from this column's third wheel, Drew Brees. Meanwhile, Rodgers and Roethlisberger continue to go in opposite directions. Considering that none of these systems know how many Super Bowls they've won, this makes perfect sense. And regardless of whether we're weighing better seasons more than worse seasons or focusing only on the best six seasons, I think we end up with a more accurate ranking in this case than when we simply add up yearly DYARs.
Finally, here are two Titans-related observations before class ends. Moving from left to right, Matt Hasselbeck goes from "just off the list" at 23rd to "on the list" at 22nd to "near the middle of the list" at 16th. I guess that trajectory is due to his deceptive speed and having five good seasons in a 14-year career. On the other end of the spectrum, Paine's six-year average really doesn't like Steve McNair, to the point that I think it makes Drinen's weighting system the best method here. Over the course of a 13-year career, McNair finished in the top half of the total DYAR rankings nine times. To me, that's above-average performance and above-average longevity.
And with that, Mr. Selig has informed me that, as much as he'd like to watch us keep batting this around forever, time has expired. Please turn in your quizzes. We'll see you later this week, when the topic will shift to "worst quarterback performances since 1991." There may or may not be a quiz.