Extra Points
News and commentary from around the Web

Who Is the World's No. 1 Quarterback?

Our friends at Sports Info Solutions have an article that takes Bill James' "Starting Pitcher Rankings" and translates them to quarterbacks. The basic idea is to use their Total Points stat to rank quarterbacks over a three-year period, although with more recent games given more weight. Patrick Mahomes is No. 1, but the more interesting findings are below the top three. In particular, Derek Carr is No. 4. That sounds crazy, but it's not out of line with our stats; we've got Carr eighth in passing DVOA for 2019 and third so far in 2020. You might also argue with how low they have Lamar Jackson, which may be overweighting his fumbles.

View Full Article

Comments

26 comments, Last at 07 Nov 2020, 10:25pm

2 According to this list, the…

According to this list, the best two backup QBs are ranked 19th & 20th. The worst 2 starting QBs are ranked 95th & 88th out of 95. The Football Team is starting its best quarterback, as are the Bears. But only 1 AFC East team is starting its best quarterback.

3 Honestly that sounds... not…

Honestly that sounds... not that inaccurate. The two best backup QB's being around 20th best in the league seems just about right. The two worst starting QB's not even deserving to be in the league, sound about right. I mean Goff his rookie year was nowhere near the 32nd best QB in the world, not even close.

5 The best backups appearing…

The best backups appearing around 20th makes sense, but it strikes me as way off at the bottom of the list - too many regular starters down in the 80s and 90s. Like, Nathan Peterman is literally unplayable, but he ends up at 74 here - higher than Tyrod Taylor, Nick Foles, Daniel Jones, Eli Manning, Andy Dalton, Carson Wentz, Joe Flacco, Cam Newton, Mitch Trubisky, Dwayne Haskins, Alex Smith, and Sam Darnold.

If you compare to DVOA (I looked at 2018 & 2019), you can see that a significant majority of non-qualifying QBs (mostly backups and guys who get benched) finish with a worse DVOA than all but the worst 3-4 starters.

It seems like they're trying to average performance across games with this list - so if a backup comes in to a game and throws 7 passes without making a mistake, is that inflating their average well above their actual combined play-by-play efficiency (like what DVOA would capture)?

7 Their system seems more…

Their system seems more comparable to DYAR. If you are above average, then you get more DYAR the more you play. However, if you are below average you accumulate negative DYAR the more you play so you are actually "better" if you play less.

14 Based on their explanation,…

Based on their explanation, I don't think this is totally a counting stat issue. It sounds like they are assigning a value to a QB's performance for each game he plays in and averaging those together (with more weight applied toward more recent games). *If* I understand it correctly, if there isn't like a minimum attempt threshold or weighting or something like that, a bunch of backup QBs could get credit for decent performances in games where they come in and throw a small number of passes. Those games would pump up their number relative to what it would actually be if it was more reflective of their performance in extended playing time.

10 They may be wrong on…

They may be wrong on Peterman, but is he really all that much worse than those other quarterbacks? I mean, I think he is worse, but Flacco has been horrible for years now, Eli was always massively overrated, Foles has looked terrible recently, Haskins is flat out awful, Smith is mediocre at best, Trubisky is garbage, etcetera. I honestly believe that if those teams had Nathan Peterman come in for them, that they would not do significantly worse.

Again, I think he is awful, but he still has a job as a backup for Gruden and the Raiders. And Gruden was a QB himself, who seems to have gotten the most out of Carr. I think we might be overstressing that 5 interception half.

15 Peterman had even worse…

Peterman had even worse efficiency numbers when he got more playing time the year after his 5 INT half! He had negative ANY/A on 88 dropbacks in 2018. His career ANY/A is 0.21 on 138 dropbacks - he should be last. (I did lay out in more detail above what looks to me to be the problem.)

4 The more I look at that…

The more I look at that article, the more I like it, even just because it confirms my own biases. If Mahomes, Wilson, and Rogers are each worth +6-6.5 extra points per 60 snaps, roughly a game, how many wins does that equal above replacement? I suspect quite a few. Also, how good is replacement anyway, since that's points over average, not the guy you could expect to find as a cheap free agent. That number might be -2.0-3.0 or so, so Mahomes, Wilson, and Rogers might be more accurately worth about 8-9+ points per 60 snaps over replacement play.

It also leads us to some somewhat rough tier lists.

MVP(6-6.5+ points per 60)

Mahomes, Wilson, Rogers

ELITE (2-6+ points per 60)

Carr, Brees, Brady, Watson, Ryan

Very Good (1-2+ points per 60)

Allen, Rivers, Goff, Cousins, Tannehill, Stafford

Above Average(0-1.0+ pp60)

Prescott, Murray, Bridgewater, Jackson

And so on. Of course, the problem with tier lists is that we've artificially made strict cutoffs. In reality, eyeballing the list I can see a roughly normal distribution of scores, which we would expect. We also have to understand that any statistical analysis cannot capture the complexity of real football, and it's quite likely that some QB's have systemic advantages over others, for various reasons.

But I would love to see expected Wins Above Replacement using these numbers. Someone should be able to do the analysis.

11 Systematic advantages be damned!

They're too hard to quantify statistically.

I see Goff as a system QB who would fare poorly in any read-option offense.  Murray could likely run the Ram's offense, but could Goff run Arizona's?   NO.

We know there are inherent inequalities in the metrics, but I don't see any way to account for all the confounding variables.

 

6 I'm sorry, but

Any list with Prescott and Jackson as league average quarterbacks is just wrong. In particular, Jackson is a much better player than his rookie year. I can appreciate what they're trying to do, but this has a few big flaws. Would any NFL coach or exec put Derek Carr as a top-five QB? I highly doubt it.

9 IIRC PFF always loved Derek…

In reply to by KaosTheory

IIRC PFF always loved Derek Carr. Not "fifth best QB," love, but they had him as the 14th best QB in the league before the season, and the 12th best QB after six weeks. 

My understanding is that the Raiders offense has a lot of easy short completions, which inflates Carr's stats. However, he is also a legitimately very good quarterback, especially with those short accurate throws, so people consistently underrate him as well. 

Any statistical analysis would need to account for those short completions maybe being rated as more important/valuable than they really are, and biasing against them, but I don't think their analysis is all that bad.

And Jackson is dragged down by his rookie season, so that one gets a mulligan in my opinion.

8 Another Metric?

QBR and DVOA are enough.  I looked at Kyler Murray and he's about the same ranking on those two and your new metric.  

Conversion on 3rd down is too confounding a variable and should not weigh heavy in to any ranking.  It fluctuates year to year and is way too dependent on the skill positions around the QB.  

If anything, running yards need to be factored in more because the NFL is moving toward a mostly mobile QB league.  There will always be a few statues who can get by on their fastball (Rodgers) but they will be the outlier, not the norm.

Regarding Lamar Jackson, he's lost value this year because his passing is down and he's not running nearly enough given the opportunities there.  He's not a good enough passer to neglect the running yards that are there.  As strange as it sounds, he needs to take a page from Kyler Murray and go on more designed runs because he's probably better than any RB on the Raven's roster.  

 

 

23 FWIW, of QB's with 20,000+…

In reply to by DIVISION

FWIW, of QB's with 20,000+ passing yards, Rodgers ranks 13th in rushing yards/game and 10th in career rushing yards.  I wouldn't call him a statue, though at age 36 it looks like he might be slowing down a bit this year.

12 Interesting project to…

Interesting project to compare this list with PFF's subjective analysis of QB performance, at least after week 6 this year. 

https://www.pff.com/news/nfl-ranking-all-32-starting-quarterbacks-after-week-6-2020-season#:~:text=Ranking%20all%2032%20NFL%20starting%20quarterbacks%20after%20Week,5%205.%20Ryan%20Tannehill%2C%20Tennessee%20Titans.%20More%20items

The list looks almost identical. Allen is 9th there, but PFF has him 8th. Jackson moves up, Carr moves down. Rivers moves way down. But for the most part the list is shockingly similar.

18 Josh Allen...

...is great entertainment if your NOT a fan of the Bills.

He could make a great play or a disastrous one from moment to moment.

That Rams/Bills game epitomizes him.  Awesome to behold, but alternately like a trainwreck.  

I hope he's not streaky his entire career because Bills fans have suffered enough.

22 Narratives

In reply to by DIVISION

Ah but even then people get a bit excited with the Josh Allen narrative. He played really well that game and has had about 5 odd moments all season so far.  That's fine. At some point you have to move along. 

(the worry would be that teams are adjusting to him and shutting him down but he's still managing the games pretty well.)

25 Roethlisberger at #27 seems bizarre

Nothing to add here, I just have watched most of his games and although I understand the numbers are the numbers, he really doesn't seem like a bottom-5 QB to me.