Writers of Pro Football Prospectus 2008

25 May 2017

2016 Passing Plus-Minus

by Scott Kacsmar

When a record is set in a stat like completion percentage, we need some extra context to judge just how impressive that feat is in today's passing climate, where failed completions are on the rise. After all, the league-wide completion percentage was at an all-time high for the fourth year in a row in 2016, finishing at 63.0 percent. NFL passing totals are stunningly similar over the last two seasons, and we had to go out to three decimal places to show the tiny difference.

  • 2015: 11,527 completions in 18,298 attempts (62.996 percent).
  • 2016: 11,526 completions in 18,295 attempts (63.001 percent).

Things especially look suspicious when Sam Bradford is the one setting the record as he did last season with a 71.6 completion percentage on a middling Minnesota offense while finishing just 17th in DVOA. Drew Brees held the previous record at 71.2 percent in his stellar 2011 season for New Orleans.

Brees also had four of the top five seasons in the first decade's worth of passing plus-minus data, which we looked at a year ago. In case you forgot, passing plus-minus estimates how many passes a quarterback completed compared to what an average quarterback would have done, given the location of those passes. It does not consider passes listed as "Thrown Away," "Tipped at Line," or "Quarterback Hit in Motion." Player performance is compared to a historical baseline of how often a pass is completed based on the pass distance, the distance required for a first down, and whether it is on the left, middle, or right side of the field. Note that plus-minus is not scaled to a player’s total attempts.

Was Bradford's season really that underappreciated, or did the quarterback who ranked 28th in failed completion rate and 33rd in ALEX just dink-and-dunk his way to it?

2016 Passing Plus-Minus

These results are for the 2016 passers with at least 200 pass attempts. Again, the total number of passes will differ from the NFL total due to the removal of certain passes. To help express plus-minus as a rate stat, C%+ is also included. Numbers will be expressed with a + or - sign.

2016 Passing Plus-Minus
Rk Player Team Passes C%+ +/- Rk Player Team Passes C%+ +/-
1 Drew Brees NO 630 +5.8% +36.5 17 Alex Smith KC 464 +0.2% +0.8
2 Matt Ryan ATL 504 +6.6% +33.1 18 Trevor Siemian DEN 430 +0.1% +0.4
3 Sam Bradford MIN 517 +5.5% +28.4 19 Philip Rivers SD 519 -0.1% -0.4
4 Kirk Cousins WAS 559 +4.9% +27.1 20 Joe Flacco BAL 621 -0.4% -2.2
5 Aaron Rodgers GB 565 +3.3% +18.5 21 Matt Barkley CHI 207 -2.6% -5.5
6 Russell Wilson SEA 497 +3.6% +18.1 22 Ben Roethlisberger PIT 491 -1.3% -6.3
7 Dak Prescott DAL 434 +3.5% +15.4 23 Derek Carr OAK 529 -1.4% -7.5
8 Andy Dalton CIN 506 +3.0% +15.0 24 Carson Wentz PHI 560 -1.8% -10.0
9 Tom Brady NE 400 +3.4% +13.7 25 Case Keenum LARM 301 -3.9% -11.6
10 Ryan Tannehill MIA 362 +3.4% +12.3 26 Brock Osweiler HOU 464 -2.7% -12.5
11 Andrew Luck IND 503 +1.3% +6.3 27 Eli Manning NYG 569 -2.7% -15.4
12 Marcus Mariota TEN 415 +1.2% +4.9 28 Colin Kaepernick SF 310 -5.2% -16.2
13 Matthew Stafford DET 551 +0.7% +4.0 29 Ryan Fitzpatrick NYJ 373 -5.6% -21.0
14 Jameis Winston TB 528 +0.6% +3.2 30 Blake Bortles JAC 581 -4.2% -24.2
15 Tyrod Taylor BUF 398 +0.7% +2.9 31 Cam Newton CAR 463 -6.1% -28.3
16 Carson Palmer ARI 553 +0.2% +1.3 Minimum 200 passes to qualify

Brees has led the league in passing plus-minus an impressive five times since 2010. His 2016 mark of plus-36.5 ranks No. 8 on the top seasons since 2006. Matt Ryan's MVP season unsurprisingly finished second, but that was mostly because he had 126 fewer aimed passes (i.e. passes that qualify for plus-minus) than Brees. It is important to note that Ryan had the highest C%+ of 2016.

Then we get to Bradford at No. 3 with a season that ranks 23rd since 2006. I have to admit, his numbers came out much better than expected. This is also true when you break down Bradford's plus-minus by the different ranges on the field.

Sam Bradford: 2016 Passing Plus-Minus Splits
Pass Range Bradford +/- Rk Note
Short (5 yards or less) +22.6 1 Min. 100 passes to qualify (31 ranked)
Mid (6 to 15 yards) -0.4 18 Min. 80 passes to qualify (31 ranked)
Deep (16 to 25 yards) +3.1 11 Min. 25 passes to qualify (31 ranked)
Bomb (26 yards or more) +3.0 9 Min. 15 passes to qualify (32 ranked)

Not surprisingly, Bradford did most of his damage on the shortest throws from 5 yards or less, and it was the medium range that gave him the most trouble with a ranking of 18th in the league. However, his deep passing was still adequate, and it's not like he was afforded much help from his offensive line, running game, or receivers. We can view Bradford's season in a better light than, say, David Carr in 2006, when he mostly dinked and dunked his way to a plus-27.3 season, but Bradford's season still doesn't hold a candle to Brees in 2016 (or 2011 for that matter). This past season, Brees was only plus-14.9 on short passes and plus-21.6 on everything else, so you can see a big difference in how he operates compared to Bradford.

Last year we asked if Kirk Cousins could improve on his strong season (plus-24.6), and he actually did with a plus-27.1 to finish fourth. He remains an interesting quarterback with a lot of varying opinions as he heads into another season on the franchise tag in Washington.

Aaron Rodgers rebounded nicely to a top-five season, but it was not always that sharp. Rodgers was at minus-1.0 thru Week 11, but posted a red-hot plus-19.4 in the final six games of the season. Andy Dalton did not seem to be anywhere near as good last year as he was in 2015, but his plus-15.0 is a career high. He also was missing A.J. Green and Tyler Eifert for large stretches of the season. Dalton has actually outclassed Andrew Luck in plus-minus in each of the last five seasons, but Luck had his best numbers yet in 2016 in what is really an underappreciated effort by the fifth-year quarterback. Dak Prescott's plus-15.4 is very good for a rookie, but Russell Wilson (plus-23.3 in 2012) still holds the best plus-minus for a first-year player since 2006.

Some notable passers finished with a negative plus-minus this season, including the entire 2004 draft class of Philip Rivers (minus-0.4), Ben Roethlisberger (minus-6.3), and Eli Manning (minus-15.4). Manning has actually been minus-10.0 or worse in the last two seasons, but Rivers had never been lower than plus-3.5 (2007) before this season. Roethlisberger might be the most surprising given he had a career-high plus-26.2 in 2015 before posting his career low. His return from injury against the Ravens in Week 9 was a season-low minus-6.5 alone, but even before Roethlisberger tore his meniscus in Miami, he was at minus-2.8.

While those quarterbacks have had better years in the past, Derek Carr is still a quarterback whose best days lie in the future. He was next to last in plus-minus in 2015 at minus-16.5, and while he improved this past year, he still finished just 23rd at minus-7.5. California-based quarterbacks did not exactly shine in this table. Colin Kaepernick did not see that usual Chip Kelly boost in plus-minus, finishing at minus-16.2 on just 310 passes.

He did not have enough attempts to qualify, but Jared Goff's rookie season is worth another (dis)honorable mention. Goff managed a minus-18.8 on just 186 aimed passes, which is good for a minus-10.1 C%+. Since 2006, only 2011 Blaine Gabbert (minus-10.5 C%+), 2008 Derek Anderson (minus-11.1 C%+), and 2009 JaMarcus Russell (minus-11.6 C%+) fared worse in C%+.

Carson Wentz only ranked 24th in his rookie season, but a lot of the blame has shifted towards his bad wide receivers. There is some merit to that as far as these numbers go. Wentz was minus-13.5 when targeting Nelson Agholor (minus-4.9) and Dorial Green-Beckham (minus-8.6), but had much better numbers to tight end Zach Ertz (plus-11.4) and slot receiver Jordan Matthews (plus-1.1). We'll see how the additions of Alshon Jeffery and Torrey Smith go, but those receivers have historically been low-percentage vertical players.

Four of the bottom seven quarterbacks lost their starting jobs with their team. Blake Bortles could be joining them soon if things don't change quickly in Jacksonville. Bortles' 2014 rookie season (minus-29.9) was the third-worst plus-minus since 2006, but he seemed to be on the right track in 2015 (minus-8.9). Things regressed in a big way last year, and he had his worst numbers to what should have been his three best pass catchers: Allen Robinson (minus-10.4), Allen Hurns (minus-7.8), and Julius Thomas (minus-3.5).

Finally, Cam Newton was coming off an MVP season, but he posted his fourth negative season in plus-minus in six years. Newton only completed 52.9 percent of his passes in 2016, and not even an adjustment for his vertical approach does him any favors here. Newton's minus-28.3 is the fifth-lowest season since 2006, sandwiched between two seasons from teammate Derek Anderson, who actually had a plus-1.7 in relief of Newton in 2016. We'll see how willing Carolina is to incorporate short passes to rookie running back Christian McCaffrey this year, but Newton still has a long way to go in his development as a consistent passer.

Did Roethlisberger's struggles bring down Antonio Brown's numbers as well? Did Dak Prescott have trouble getting in sync with Dez Bryant? How deadly was Rob Gronkowski downfield in his injury-shortened season? Those questions and more will be answered next week when we look at receiving plus-minus.

Posted by: Scott Kacsmar on 25 May 2017

21 comments, Last at 31 May 2017, 11:17am by Noah Arkadia

Comments

1
by Tomlin_Is_Infallible :: Thu, 05/25/2017 - 9:14pm

"Did Roethlisberger's struggles bring down Antonio Brown's numbers as well?"

if by struggles you mean not having a reliable downfield threat for more than about 4 games all season long............

--------------------------------------
The standard is the standard!

2
by Chuckc :: Fri, 05/26/2017 - 7:46am

Since when is Antonio Brown not a reliable downfield threat?

4
by Tomlin_Is_Infallible :: Fri, 05/26/2017 - 10:11am

since the #s say so? his AYPT are very low for a "#1" receiver. he excels when there's a Martavis Bryant (or similar) who has the defense crapping their pants. Hell, even just a non-busted-open-hands Sammie Coates made a huge difference for 4 games.

http://www.rotoworld.com/articles/nfl/69265/446/why-receiver-air-yards-m...

--------------------------------------
The standard is the standard!

5
by jtr :: Fri, 05/26/2017 - 10:33am

My first thought with Ben's low numbers is that he got killed by a number of downfield throws that clanged off of Coates' hands. He certainly would have looked a lot better if he got credit for all of those accurate bombs that got dropped.

8
by Tomlin_Is_Infallible :: Fri, 05/26/2017 - 11:27am

And then my thought is: what kind of coaching staff forces a guy with cut up hand and broken fingers to play ? I guess if you want to use him as a deep decoy you do have to actually throw the ball his way.

smh

--------------------------------------
The standard is the standard!

20
by JohnxMorgan :: Tue, 05/30/2017 - 6:21pm

does not mean he is not an effective downfield threat.

The linked post, unless I am misunderstanding it, only states that the depth of all pass targets is more predicative than the total yardage gained, that RAC and deep catch conversion rate heavily regress toward the mean, and that if a player is being targeted for a high percentage of a team's overall "air yards" but has not converted those targets into yards gained, he is likely to improve either through the season or next season.

aYTP is an average. A player who is an excellent deep threat, like Brown, but who also is the target of many short passes, like Brown, may have a small aYTP. It doesn't mean Brown is not a "reliable downfield threat." That's a misinterpretation.

Compare Brown to Mike Evans, for instance. Evans' dwarfs Brown's aYPT, 14.6 to 10.3. But, conversely, here's a simple break down.

Receptions of targets that traveled over 21 yards through the air (which is Elias's peculiar cutoff).

Brown: 13
Evans: 11

Receptions of targets behind the line of scrimmage

Brown: 27
Evans: 4

In fact, Brown had more receptions of 21+ yards last season than Coates and Martavis Bryant have had over the last two seasons. He's special, and an excellent deep threat.

21
by Noah Arkadia :: Wed, 05/31/2017 - 11:17am

That's a good and interesting point.

3
by Aaron Brooks Go... :: Fri, 05/26/2017 - 8:56am

So you don't discount either tipped passes (QB's fault) or WR drops (not QB's fault). Interesting.

6
by jtr :: Fri, 05/26/2017 - 10:39am

Those numbers on Goff convince me even more than the article from earlier this week that he's a bust. I think it's generally pretty well understood that passer accuracy is one of those things that generally can't be coached up very much. Inaccurate college passers almost always make inaccurate pro passers, and a rookie year with accuracy as bad as Goff's doesn't speak well to his future.

7
by mehllageman56 :: Fri, 05/26/2017 - 11:20am

I have generally agreed with you that accuracy can't be taught, but then I researched Joe Montana's college stats: 52 completion percentage, with a high of 54 his last year. Drew Bledsoe's rookie year was also terrible: 49.7 completion percentage. Some of whether a quarterback prospect succeeds depends upon the situation itself; Dak Prescott goes to Dallas and has a great rookie season because his flaw, throwing into tight windows, doesn't come up much due to a strong running game and really good receivers. Goff goes to a team that has an offense so simple the Niners know the plays better than the Rams on opening night. And it wasn't simplified for Goff; he wasn't going to be playing at that point. It wouldn't surprise me if Goff bombs, or makes a comeback; it wouldn't surprise me right now as a Jets fan if Geno Smith replaces Eli and leads the Giants to the Super Bowl either. I think Goff's prospects are closer to 50/50, which is still not what you want from the first pick in the draft.

9
by IrishBarrister :: Fri, 05/26/2017 - 12:32pm

Why isn't Cody Kessler included?

Just curious.

10
by MilkmanDanimal :: Fri, 05/26/2017 - 1:10pm

The table has a minimum of 200 pass attempts; since Cleveland ran through about 45 different QBs last year, I would guess nobody hit the bare minimum.

11
by Vincent Verhei :: Fri, 05/26/2017 - 3:58pm

Just for fun, here are the +/- numbers for all Cleveland passers last year:

PLAYER, Passes, +/-
6-C.Kessler, 174, +8.1
15-C.Whitehurst, 21, -0.1
29-D.Johnson, 1, -0.2
8-K.Hogan, 20, -0.5
11-T.Pryor, 9 ,-1.5
10-R.Griffin, 136 ,-7.2
13-J.McCown, 148, -8.9

12
by mehllageman56 :: Sat, 05/27/2017 - 5:40pm

So, Kessler missed the cutoff by 26 passes and would have led the +/- stat line? No wonder Cleveland didn't draft a qb in the first round, they might not need one.

Also, McCown at -8.9? Sign the Jets up for Sam Darnold right now.

13
by mehllageman56 :: Sat, 05/27/2017 - 5:43pm

Just realized the Cleveland stats are just +/-, not C+/-. Still, Kessler would have ranked 11th, ahead of Luck and Mariota. Maybe people shouldn't dismiss his chances yet.

14
by dbostedo :: Sat, 05/27/2017 - 8:40pm

"No wonder Cleveland didn't draft a qb in the first round, they might not need one."

"Maybe people shouldn't dismiss his chances yet."

I don't know if Kessler can be good or not, but I don't think I'd be using this stat to argue that. I don't think it tells you all that much either way.

15
by herewegobrownie... :: Sun, 05/28/2017 - 12:37pm

The fact that he wasn't able to make it to this qualifying passer list due to 2 close-together concussions and other health issues, are why the front office was at least willing to draft a 2nd round QB if not a 1st rounder (and Osweiler may also prove to be better.) He has the 1st team snaps by default now due to experience, but that could change soon.

More than the other management, Hue Jackson seems to mostly trust him becoming another serviceable Andy Dalton (although he did pull him in that Baltimore game - during a lead no less - when he wasn't making the long throws) but it made sense for the front office to draft a QB who they didn't see as having too many question marks for his availability.

Although every QB in town should have more ability to stay upright due to the huge OL upgrades (Cam Erving might not even make the final 53) Cody hasn't shown that much more than Colt McCoy, which is usually what you can expect from a 3rd rounder (generally doesn't hurt to try to get Russell Wilson or Tom Brady at that stage, provided that there are no better overall fits in your estimation.)

18
by mehllageman56 :: Mon, 05/29/2017 - 12:49am

Haven't watched the Browns that much, just noticed that Kessler had a passer rating in the 80s, which is more than I've seen from a young Jets qb in a while. At first I got the stat lines mixed up, so perhaps that explains my initial excitement.

Agree that keeping Osweiler around for now is a good idea for the Browns; might as well see if he can be competent.

16
by The Ninjalectual :: Sun, 05/28/2017 - 8:56pm

What exactly does this stat measure? I don't understand the point

17
by Vincent Verhei :: Sun, 05/28/2017 - 11:12pm

Completions above or below what an average quarterback would have done, adjusted for pass distance and yards to go for a first down.

19
by Scott Kacsmar :: Mon, 05/29/2017 - 2:22am

You can think of +/- as DYAR and C%+ as DVOA in the context of an adjusted completion percentage where multiple variables for each aimed pass are accounted for.