Stat Analysis
Advanced analytics on player and team performance

Failed Completions 2020

Chicago Bears QB Mitchell Trubisky
Photo: USA Today Sports Images

As we begin to roll out our detailed stat reviews of the 2020 season, it's important to remember just how strange last year actually was. No offseason to speak of, socially distanced practices, games moved all willy-nilly around the calendar as everyone struggled to deal with the global pandemic. Under these conditions, it's entirely possible that strange results or reversal of trends are at least partially caused by that more than anything else. It's possible that, when we look back historically, 2020 will stick out like 2011 or 1987, other years impacted by effects off the field.

Take the passing game, for instance. 2020 was, by many metrics, the most efficient passing year in league history. The NFL set new records for completions, touchdowns, passing first downs, and ANY/A, to name just a few. Part of that is the continued evolution of the passing game, of course, but 2020 had some factors to it that could artificially boost passing numbers. The lack of offseason practice seems to disproportionally impact defenses, as we saw after the 2011 lockout. The lack of crowd noise made it easier for offenses to function, particularly on the road. Offensive holding and pass interference calls dipped dramatically, allowing quarterbacks more time to work in the pocket. The cards aligned themselves for a positive passing season.

With that environment in place, perhaps we shouldn't be surprised to see the NFL's failed completion numbers drop to levels we haven't seen in a decade.

Every year, we here at Football Outsiders study failed completions around the league. A failed completion is any completed pass that fails to gain 45% of needed yards on first down, 60% on second down, or 100% on third or fourth down. You can see last year's study on the subject here.

In 2020, quarterbacks completed 11,756 passes, with 2,804 failed completions. That ends up being a rate of 23.9%. It's the lowest mark since we began writing these yearly articles in 2013, and over a percentage point drop from where it was in 2019. The league's failed completion rate has risen over time as dumpoffs and screens have replaced slamming a fullback into a loaded box as the give-up play du jour, but 2020's numbers have much more in common with the late 2000s than anything we've seen in recent years.

Image 1

I suspect that 2020 will prove to be an outlier for all the reasons mentioned above, but if quarterbacks in 2021 want to keep pushing the ball a little further downfield on each play, I certainly would not complain.

This is normally where we put the disclaimer that not every failed completion is created equally, that for this article we make things binary, simply summing up successes and failures. A 7-yard completion on third-and-10 is better than a 2-yard completion, especially in field goal range, and that is reflected in DVOA—but for the purposes of this article, a failure is a failure. That remains true for all of the statistics and tables coming up, but there's no way to talk about the top of 2020's table without talking about degrees of success and failure. Let's get to the charts.


Quarterbacks

In the following table, the 36 qualified quarterbacks of 2020 are ranked by ascending failed completion rate (FC%). We also included failed completions as a percentage of attempts (very little change in the rankings) as well as the average ALEX (all downs) for the season. "Passes" here refers to all passing plays, including sacks.

Quarterbacks, Failed Completions, 2020
Rk Player Team Passes Comp Failed FC% FC%ATT Rk ALEX Rk
1 Mitchell Trubisky CHI 318 199 31 15.6% 9.7% 1 -0.1 10
2 Patrick Mahomes KC 613 390 62 15.9% 10.1% 2 -0.5 19
3 Josh Allen BUF 606 396 65 16.4% 10.7% 3 -0.4 13
4 Dak Prescott DAL 233 151 26 17.2% 11.2% 5 -0.8 23
5 Deshaun Watson HOU 597 382 69 18.1% 11.6% 7 0.3 3
6 Ryan Tannehill TEN 515 315 58 18.4% 11.3% 6 -0.1 9
7 Baker Mayfield CLE 513 305 57 18.7% 11.1% 4 0.1 8
8 Kirk Cousins MIN 560 349 67 19.2% 12.0% 8 -0.7 22
9 Ryan Fitzpatrick MIA 290 183 37 20.2% 12.8% 10 -0.5 20
10 Russell Wilson SEA 612 384 82 21.4% 13.4% 12 0.2 6
11 Aaron Rodgers GB 552 372 80 21.5% 14.5% 20 -0.4 14
12 Matthew Stafford DET 568 339 73 21.5% 12.9% 11 0.1 7
13 Jared Goff LAR 578 370 82 22.2% 14.2% 15 -1.6 30
14 Andy Dalton DAL 363 216 50 23.1% 13.8% 13 -1.6 29
15 Tom Brady TB 650 401 93 23.2% 14.3% 17 0.9 1
16 Joe Burrow CIN 437 264 62 23.5% 14.2% 16 -0.4 16
17 Derek Carr LV 549 348 82 23.6% 14.9% 23 0.5 2
18 Tua Tagovailoa MIA 316 186 44 23.7% 13.9% 14 -0.4 15
19 Gardner Minshew JAX 360 216 52 24.1% 14.4% 18 -0.4 17
20 Matt Ryan ATL 675 407 100 24.6% 14.8% 21 -0.2 11
21 Philip Rivers IND 572 369 91 24.7% 15.9% 27 -1.6 31
22 Carson Wentz PHI 497 251 62 24.7% 12.5% 9 0.2 5
23 Drew Brees NO 406 275 68 24.7% 16.7% 30 -2.4 34
24 Lamar Jackson BAL 405 242 60 24.8% 14.8% 22 -0.3 12
25 Nick Mullens SF 350 211 53 25.1% 15.1% 24 -2.5 35
26 Kyler Murray ARI 587 375 95 25.3% 16.2% 28 -0.7 21
27 Cam Newton NE 398 242 63 26.0% 15.8% 26 -2.3 33
28 Justin Herbert LAC 630 396 106 26.8% 16.8% 31 -1.2 26
29 Drew Lock DEN 470 254 68 26.8% 14.5% 19 0.3 4
30 Ben Roethlisberger PIT 640 399 109 27.3% 17.0% 32 -1.1 25
31 Daniel Jones NYG 498 280 78 27.9% 15.7% 25 -1.4 27
32 Sam Darnold NYJ 404 217 67 30.9% 16.6% 29 -0.8 24
33 Teddy Bridgewater CAR 529 340 106 31.2% 20.0% 36 -1.5 28
34 Nick Foles CHI 334 202 65 32.2% 19.5% 33 -0.5 18
35 Alex Smith WAS 276 168 55 32.7% 19.9% 35 -3.8 36
36 Dwayne Haskins WAS 263 148 52 35.1% 19.8% 34 -2.1 32

Pittsburgh's Ben Roethlisberger led the league with 109 failed completions, just beating Justin Herbert and Teddy Bridgewater (106 each) to the line. This is partially a factor of volume, as only Tom Brady and Matt Ryan had more dropbacks than Roethlisberger in 2020. Still, a failed completion rate of 27.3% is not at all good. Roethlisberger was near the league lead with 110 failed completions in 2018, but with a healthier 24.3% failed completion rate, leaving him much closer to the middle of the table. His failed completion rate actually improved as the season went along, but that's more of a factor of Roethlisberger simply completing fewer passes down the stretch. These are very concerning numbers for a 39-year-old passer without a quality backup behind him at the moment.

And we say "quality" for a reason. This offseason, Pittsburgh went out and acquired Dwayne Haskins for their quarterback room. Haskins ended up with the worst failed completion rate in the league at 35.1%. At the moment, Haskins would be the third-string passer behind Mason Rudolph … who was worst in the league with a 33.5% failed completion rate in 2019. Pittsburgh right now is the absolute center of the universe for dink-and-dunk passers; it may not be the worst quarterback room in the league, but it is at least in the discussion.

It's nowhere near as bad as Washington was in 2020, mind you. Not only did Haskins have the highest failed completion rate in the league, but he only just finished ahead of teammate Alex Smith; the two flop places if you sort by failed completions as a percentage of attempts as opposed to a percentage of completions. That has never happened before; no team has managed to get both of the bottom two passers at the same time. It has been close—Case Keenum and Jared Goff had two of the bottom three slots in 2016 for Jeff Fisher's Rams—but ending up with the worst two passers at the same time is impressive in a perverse sort of way. It's fitting that Smith would be last in the league in ALEX, considering the stat was named for him; his -3.8 mark was the worst in over half a decade, and you can see the direct line from lack of arm strength to low ALEX to failed completions right there. Haskins' ALEX wasn't anything to write home about either, though it was his poor accuracy more than anything else that deflated both his completion total and his receivers' ability to turn some of those failures into successes with yards after the catch. Either way, both players are gone now, and the Football Team had a losing record, meaning they're in decent position to pick up a new franchise quarterback in the draft … oh. Well, there's always next year.

The rest of the bottom of the list is more or less what you would expect. Teddy Bridgewater learned well from his time in New Orleans, never finding a dumpoff or checkdown he wouldn't throw to. Sam Darnold regressed some even from his mononucleosis-affected 2019 season; you can blame Adam Gase for some of this but not all. And then there's Nick Foles, the only man ever to cross the 40% failed completions rate barrier, bringing his patented brand of football to Chicago. And speaking of the Windy City…

There is, as you might suspect, a solid correlation between a low failed completion rate and a high DVOA. While a failed completion is far from the worst outcome of a play, too many of them will kill drives and squash your efficiency scores. You have MVPs, All-Pros, and recent Pro Bowlers flooding the top 10 in this year's rankings, and deservedly so.

The best failed completion rates don't always go to the very best players, but it's not a spot you expect to see many scrubs in. Since 1983, the average DVOA for the player with the best failed completion rate is 18.9%. Since 2001, that has gone up to 22.8%, as the failed completion has become more of a thing. Tom Brady has had the league's lowest failed completion rate three times, Peyton Manning twice. Hall of Famers such as Kurt Warner, Dan Marino, Jim Kelly, and Dan Fouts have all been on top of the world at one point or another. Even the lesser names to top this table generally do so in career seasons—Matt Cassel led the league in 2010, his one Pro Bowl season; Jay Schroeder led the league in 1990, when he also topped all passers in DVOA. At the very least, the leader in this category generally is a plus for their team.

So how on Earth did Mitchell Trubisky end up perched atop this year's table? In his first three seasons, Trubisky has put up failed completion rates of 36.2%, 24.6%, and 28.8%, only to nearly completely cut them out of his game in 2020. It's not a factor of him replacing failed completions with incomplete passes, either; he ranks third in successful completion percentage. It's not a factor of him pushing the ball deeper; his average depth of target on completions dropped to 5.3, the lowest mark of his career. It wasn't a factor of a change in the offense; we saw Nick Foles run out with basically the same team and the same playcallers and double Trubisky's failed completion rate.

And it wasn't a factor of him suddenly becoming a solid quarterback, either. Trubisky's -7.5% DVOA makes him the sixth quarterback in our 38-year database to have both the best failed completion rate and a negative DVOA. It's a difficult duo to pull off!

Worst DVOA for Failed Completion Rate Leaders, 1983-2020
Year Player Team FC% DVOA
1987 Vinny Testaverde TB 8.5% -24.4%
2009 Jake Delhomme CAR 18.5% -23.1%
2016 Matt Barkley CHI 10.9% -10.9%
2020 Mitchell Trubisky CHI 15.6% -7.5%
1994 Jeff Blake CIN 16.0% -1.9%
2002 Donovan McNabb PHI 18.5% -0.8%
1984 Jim Plunkett LARD 12.0% 1.9%
1985 Phil Simms NYG 10.2% 3.1%
2000 Gus Frerotte DEN 14.4% 5.1%
1989 Jim Kelly BUF 11.0% 6.4%

Trubisky did see his ALEX rise from -1.2 last year to -0.1 this year, and that does help a lot. There's a significant correlation between ALEX and successful completions; when you throw the ball closer to the sticks, it's easier for those plays to gain significant yardage. And Trubisky did show a bit of improvement from 2019, so it's not a surprise that his numbers saw some increase. But no, what we're seeing here has less to do with Trubisky than it does with the binary nature of failed completions.

A bomb from the shadow of your own goalposts on third-and-forever that results in a fourth-quarter go-ahead score is an exceptionally valuable play, and it counts as one success in these rankings. A 2-yard reception on second-and-3 is not a particularly valuable play, but it counts as one success in these rankings. For the purposes of DVOA and DYAR and everything else, there's a world of difference between those two outcomes, but failed completions are determined by a binary yes-and-no check. And Trubisky had far more of the latter than the former.

On successful completions, Trubisky had a 125.6% DVOA. That's the second-worst total in the league among qualified passers, just barely squeezing above Alex Smith's 120.4%. Trubisky's average successful pass came with 7.8 yards to go, tied for shortest in the league, and picked up an average of 11.6 yards, second-least in the league behind Tua Tagovailoa. In short, Trubisky was the same dink-and-dunker he has been during his career; he just ended up throwing the ball in better situations and just barely getting past the success mark more often than not. Seventeen of his 168 successful completions gained exactly as much yardage as they needed to count; 16 more were 1 yard over the mark. Essentially, Trubisky worked his butt off for a C- average. That's a passing grade, but a disappointing one.

For the record, the players with the highest DVOA on successful completions were all in San Francisco. Nick Mullens led all qualified passers with a 169.6% DVOA; Jimmy Garoppolo would have topped that with 172.2%, and C.J. Beathard's 157.5% would have finished just outside of the top 10, but they both failed to qualify. In a world where every pass was a success, the 49ers would have done quite well. It's just a shame about all those incompletions, sacks, and interceptions.

While Trubisky's jump from 28.8% to 15.6% was the best in the league, five other passers improved their failed completion rate by at least 5% between 2019 and 2020. Patrick Mahomes' numbers took a slight hit in 2019 due to injury but rebounded in 2020 to where he had been before that. A strong running game and play-action focus helped Baker Mayfield to the best season of his career. Kirk Cousins set career highs for touchdowns and touchdown rate. Gardner Minshew put up league-average numbers and should find a role as a backup somewhere next season. And then there's Josh Allen.

Here's a look at successful completion percentage—the percentage of all pass attempts that end as successful completions. This is where we count all failed completions as incomplete passes, removing some of the empty passing calories from completion percentage around the league:

Successful Completion Percentage, 2020
Player Team Att Comp. Cmp% Rk Failed Success% Rk
Josh Allen BUF 572 396 69.2% 6 65 57.9% 1
Deshaun Watson HOU 542 382 70.5% 3 69 57.7% 2
Mitchell Trubisky CHI 297 199 67.0% 15 31 56.6% 3
Dak Prescott DAL 221 151 68.3% 10 26 56.6% 4
Aaron Rodgers GB 522 372 71.3% 1 80 55.9% 5
Patrick Mahomes KC 587 390 66.4% 17 62 55.9% 6
Kirk Cousins MIN 513 349 68.0% 11 67 55.0% 7
Ryan Fitzpatrick MIA 266 183 68.8% 7 37 54.9% 8
Russell Wilson SEA 554 384 69.3% 5 82 54.5% 9
Ryan Tannehill TEN 480 315 65.6% 23 58 53.5% 10
Drew Brees NO 387 275 71.1% 2 68 53.5% 11
Jared Goff LAR 547 370 67.6% 13 82 52.7% 12
Baker Mayfield CLE 475 305 64.2% 31 57 52.2% 13
Derek Carr LV 512 348 68.0% 12 82 52.0% 14
Philip Rivers IND 539 369 68.5% 9 91 51.6% 15
Kyler Murray ARI 546 375 68.7% 8 95 51.3% 16
Matthew Stafford DET 520 339 65.2% 25 73 51.2% 17
Tom Brady TB 604 401 66.4% 18 93 51.0% 18
Joe Burrow CIN 401 264 65.8% 21 62 50.4% 19
Gardner Minshew JAX 326 216 66.3% 19 52 50.3% 20
Andy Dalton DAL 332 216 65.1% 27 50 50.0% 21
Tua Tagovailoa MIA 286 186 65.0% 28 44 49.7% 22
Matt Ryan ATL 623 407 65.3% 24 100 49.3% 23
Justin Herbert LAC 591 396 67.0% 14 106 49.1% 24
Cam Newton NE 366 242 66.1% 20 63 48.9% 25
Lamar Jackson BAL 373 242 64.9% 29 60 48.8% 26
Nick Mullens SF 326 211 64.7% 30 53 48.5% 27
Teddy Bridgewater CAR 488 340 69.7% 4 106 48.0% 28
Ben Roethlisberger PIT 607 399 65.7% 22 109 47.8% 29
Daniel Jones NYG 446 280 62.8% 32 78 45.3% 30
Alex Smith WAS 252 168 66.7% 16 55 44.8% 31
Nick Foles CHI 310 202 65.2% 26 65 44.2% 32
Carson Wentz PHI 435 251 57.7% 36 62 43.4% 33
Drew Lock DEN 438 254 58.0% 35 68 42.5% 34
Sam Darnold NYJ 361 217 60.1% 34 67 41.6% 35
Dwayne Haskins WAS 241 148 61.4% 33 52 39.8% 36

We have only run this table twice before, and Drew Brees was atop it each time. Brees' arm turning into a pumpkin ends that reign, which is not overly surprising. What would have been surprising before the year began was the man who replaced him.

Josh Allen has had decent failed completion numbers throughout his career; we joked in 2018 that he was so inaccurate that he simply didn't rack up many completions at all, short or otherwise. His perennially high average depth of target and ALEX also contributed; Allen has never been the kind to throw the sort of checkdown that leads to a failed completion. But no more mockery, at least for now—Allen more than shut up all his critics last year with one of the most dramatic improvements we have ever seen from a quarterback. Allen's successful completion percentage jumped from 44.3% in 2019 to 57.9% last season, by a wide margin the biggest improvement of the year. Considering their respective playstyles, Allen sitting in Brees' throne feels wrong on several fundamental levels, but he didn't just replace Brees, he surpassed him—his successful completion percentage tops what Brees was able to do in 2018 or 2019. Josh Allen, accurate quarterback. Who woulda thought?

Some final quick hits:

  • Trubisky, Mahomes, Mayfield, and Ryan Tannehill each jump double-digit spots in completion percentage once you remove the unsuccessful completions from each quarterback. Conversely, Bridgewater, Smith, and Herbert go plummeting down the rankings as their standard stats get a bit exposed.
     
  • Speaking of Herbert, he comes out the worst of the three primary rookies from 2020. Joe Burrow squeaks just ahead of Tua Tagovailoa in both failed completion rate and successful completion rate, with Herbert bringing up the rear each time. Things may not be looking up for Herbert next season, as new offensive coordinator Joe Lombardi comes out of the Saints' system. Expect plenty of dumpoffs to Austin Ekeler in the future.
     
  • Aaron Rodgers would like to remind you that he's back. The MVP saw his successful completion rate jump 10 points last season; his league-leading completion percentage was a little inflated by the Packers' offensive system, but he still finishes in the top five.
     
  • Not strictly related to failed completions, but last year's article spent a long time comparing Tom Brady to Jameis Winston as the Buccaneers changed quarterbacks. We mentioned that Brady had more failed completions than Winston thanks in part to throwing the ball deep less often. Well, Brady's ALEX increased from -1.1 to a league-leading 0.9 this year, and he saw his failed completion rate fall from 27.9% to 23.2%. It turns out receivers help a quarterback, and that Brady will never, ever age.

Receivers

What about the receivers on the other end of those failed completions? It's worth taking a look at that, even though appearances here generally have more to do with usage and scheme than a receiver's individual talents. We exclude running backs from these tables; they would otherwise dominate due to their roles on checkdowns and emergency outlets. For the record, however, Alvin Kamara led all players with 35 failed receptions (Drew Brees lost the ability to throw beyond the line of scrimmage in 2020). Mike Davis, J.D. McKissic, and Nyheim Hines also topped 25.

Most Failed
Receptions (WR/TE)
  Lowest Failed
Reception Rate (WR/TE)
  Highest Failed
Reception Rate (WR/TE)
Receiver Team Failed Receiver Team Rec Failed Pct Receiver Team Rec Failed Pct
Robby Anderson CAR 29 Sammy Watkins KC 37 1 2.7% Evan Engram NYG 63 24 38.1%
Curtis Samuel CAR 25 Rashard Higgins CLE 37 1 2.7% Anthony Miller CHI 49 17 34.7%
Logan Thomas WAS 24 Gabriel Davis BUF 35 2 5.7% Golden Tate NYG 35 12 34.3%
Evan Engram NYG 24 Darius Slayton NYG 50 3 6.0% Logan Thomas WAS 72 24 33.3%
DeAndre Hopkins ARI 24 D.J. Moore CAR 66 4 6.1% KJ Hamler DEN 30 10 33.3%
Robert Woods LAR 23 Allen Lazard GB 33 2 6.1% Curtis Samuel CAR 77 25 32.5%
Keenan Allen LAC 23 Will Fuller HOU 53 4 7.5% Larry Fitzgerald ARI 54 17 31.5%
Diontae Johnson PIT 23 Travis Fulgham PHI 38 3 7.9% Jamison Crowder NYJ 59 18 30.5%
J.Smith-Schuster PIT 22 Travis Kelce KC 105 9 8.6% Robby Anderson CAR 96 29 30.2%
Terry McLaurin WAS 20 A.J. Brown TEN 70 6 8.6% Isaiah McKenzie BUF 30 9 30.0%
Stefon Diggs BUF 20 Keelan Cole JAX 55 5 9.1% Laviska Shenault JAX 58 17 29.3%
Davante Adams GB 19 John Brown BUF 33 3 9.1% David Moore SEA 35 10 28.6%
Cooper Kupp LAR 19 M.Valdes-Scantling GB 33 3 9.1% Zach Ertz PHI 36 10 27.8%
Min. 30 receptions

Between Davis, Robby Anderson, and Curtis Samuel, it's amazing the Panthers had any successful completions in 2020; I don't believe we've ever had a team with three players hit 25 failed receptions before. Sometimes players end up on this leaderboard out of sheer volume; players such as DeAndre Hopkins, Keenan Allen, Stefon Diggs, and Davante Adams end up with failed completions because they all catch a zillion passes and are the go-to guys when a play breaks down. Not Anderson and Samuel, however, who each appear on the highest failed reception rate table as well. This has more to do with Teddy Bridgewater's arm talent than Anderson or Samuel's skills; seeing Joe Brady's offense run by a quarterback with a live arm would be interesting indeed.

Hopkins returns to the failed receptions leaderboard for the second year in a row despite switching teams. He's joined by Cooper Kupp (last year's leader) and Robert Woods as returnees as the NFC West knew how to throw a checkdown. Whether Kupp or Woods will return next year now that the Rams have swapped Jared Goff for Matthew Stafford will be interesting to watch; Stafford has consistently had a lower failed completion rate than Goff, though it was very close this year.

Travis Kelce deserves some sort of award for having single-digit failed receptions on 105 catches; he's the first receiver since DeAndre Hopkins in 2015 to hit the lowest failed rate table with over 100 catches. Working with Patrick Mahomes does make it easier, as Sammy Watkins will attest, but to be that effective on that volume is rarified air. Putting up great numbers with Bridgewater is a taller task, so congratulations to DJ Moore being the only player to repeat on this leaderboard from last season.

We shall also use this opportunity to once again complain about Evan Engram making the Pro Bowl over Robert Tonyan. The top non-running back in highest failed reception rate is almost always going to be a tight end, but this is the third year in a row Engram has been among the league leaders. Engram's raw numbers are some of the most inflated in the league, padded by meaningless catches in failed situations.


Defenses

Finally, let's look at the defenses' ability to create failed completions, with a comparison to how these units fared in 2019.

Defenses: 2020 Failed Completions Compared to 2019
Rk Team Comp Failed FC% 2019 Rk 2019 FC% Diff Rk
1 WAS 330 96 29.1% 11 26.4% 2.7% 4
2 LAR 347 100 28.8% 16 24.4% 4.4% 2
3 TB 426 119 27.9% 7 28.4% -0.5% 12
4 LAC 340 92 27.1% 2 30.2% -3.1% 22
5 DEN 374 101 27.0% 6 28.4% -1.4% 19
6 PHI 364 98 26.9% 10 26.7% 0.2% 10
7 SF 341 90 26.4% 1 31.8% -5.4% 28
8 CHI 350 92 26.3% 15 25.1% 1.2% 8
9 CAR 398 104 26.1% 22 23.6% 2.5% 6
10 PIT 298 77 25.8% 5 29.3% -3.5% 25
11 DAL 328 84 25.6% 14 25.4% 0.2% 11
12 ATL 425 108 25.4% 30 20.8% 4.6% 1
13 SEA 450 112 24.9% 25 22.7% 2.2% 7
14 NO 333 81 24.3% 13 25.9% -1.6% 21
15 BAL 380 91 23.9% 9 27.4% -3.5% 24
16 JAX 370 88 23.8% 29 21.1% 2.7% 5
17 GB 346 81 23.4% 26 22.4% 1.0% 9
18 BUF 369 86 23.3% 3 30.2% -6.9% 32
19 ARI 365 85 23.3% 19 24.2% -0.9% 15
20 NYG 333 76 22.8% 21 23.8% -1.0% 17
21 IND 369 84 22.8% 4 29.4% -6.6% 31
22 MIA 370 84 22.7% 32 19.2% 3.5% 3
23 CLE 371 83 22.4% 23 23.3% -0.9% 16
24 OAK 418 90 21.5% 27 22.3% -0.8% 13
25 NE 358 77 21.5% 8 27.7% -6.2% 29
26 CIN 340 73 21.5% 24 22.7% -1.2% 18
27 MIN 343 72 21.0% 28 21.8% -0.8% 14
28 KC 349 72 20.6% 20 23.9% -3.3% 23
29 HOU 377 77 20.4% 18 24.3% -3.9% 26
30 NYJ 388 77 19.8% 12 26.2% -6.4% 30
31 TEN 423 82 19.4% 17 24.4% -5.0% 27
32 DET 383 72 18.8% 31 20.2% -1.4% 20

As a reminder, defensive numbers are far less sticky from year-to-year than the quarterback stats. This year, there was a surprising 0.41 correlation, but that number's typically down closer to the 0.25 range. A lot of this depends on the types of quarterback on the schedule before you even get to year-over-year turnover.

That being said, we can still learn something from this table; it's still correlated with good pass defense in general. What you're seeing here is a list of teams that force the most dumpoffs, checkdowns, and low-ALEX plays.

You're also seeing the end of the Matt Patricia era in Detroit. The "defensive guru" had his defenses finish 32nd, 31st, and 29th in his three years with the Lions, and 30th the year before in New England. For a stat that tends to fluctuate wildly, that's a damning indictment. Plenty of teams have trotted out bad defenses. Few have trotted out some as consistently bad as Patricia has over the past four years.

Comments

28 comments, Last at 22 Mar 2021, 10:27am

1 Hall of Famers such as Kurt…

Hall of Famers such as Kurt Warner, Dan Marino, Jim Kelly, Phil Simms, and Dan Fouts ...

One of these players is not like the others.

3 ”It wasn't a factor of a…

”It wasn't a factor of a change in the offense; we saw Nick Foles run out with basically the same team and the same playcallers and double Trubisky's failed completion rate”

Interesting, as a Bears fan I thought they ran a totally different offense in Trubiskys last 7 games. A lot more PA rollouts as opposed to the RPO game they were running his first two/three games.

im also curious on a separate note, would a QBs DVOA be affected by poor OL play? Poor receiver play, play calling etc? I always thought Trubisky was a decent QB until he racked up multiple shoulder injuries in 2018-2019. Pace didn’t help him either with the OL regression and lack of moves made at the position. Not making excuses for the kid I just believe he is one of the most polarizing QBs in the league. So much talent but little to show for it IMO. Maybe I’m just a homer. Thanks 

5 I actually agree that what I…

I actually agree that what I saw of the Bears in the second half of the season looked very different from when they started. Not that I would suggest retaining him, but it definitely showed that maybe they weren't using their QB in ways that complimented his skill set?

6 Everyone is going to have a…

Everyone is going to have a different take, I truly believe his shoulder injuries hurt his confidence in the pocket. I would be interested to see where Trubisky ranked in DYAR and DVoA the first 9-10 games of 2018. I think after hurting his shoulder, and playing through injury, he became much more paranoid in the pocket which has led to his mechanics to suffer. Poor footwork, forcing balls, a little bit of the Cutler Syndrome without the rifle arm

8 Trubisky's first three-games…

Trubisky's first three-games /  last six games splits are interesting, though I didn't want to put too much weight on a three-game sample size.

Trubisky's failed completion rate actually increased later in the season; it was 13.7% in Weeks 1-3 and 16.2% in Weeks 12-17.  That is, at least in part, a factor of the classic "you must have a completion to have a failed completion" problem; Trubisky's successful completion percentage rose from 51.2% to 58.9%.

And you're right; the Bears used play-action about twice as much with Trubisky than with Foles (though I don't have Week 1-3 splits right here), and significantly more than they did in 2019.  That's a solid part of the explanation why Trubisky's DVOA increased from last season.  I'm not sure I fully buy "more PA = fewer failed completions", but giving Trubisky plays he feels more comfortable with, and having that lead to better results, does make quite a bit of sense. 

It could, however, also be due to the fact that those last six games included matchups against Detroit, Jacksonville and Houston, three of the bottom four pass defenses last year.  "Worse coverage = fewer failed completions" makes a lot of sense, too, and Trubisky got to skip the regular season games against the Saints, Rams, Buccaneers, etcetera.  Strength of schedule likely explains why Trubisky was #1 as opposed to just a surprising name in the top ten; it doesn't explain everything away but it's worth remembering.

 

As for the second part of the question, yes, everyone's DVOA can be affected by everyone else.  DVOA and DYAR just rely on the results of the plays themselves.  It doesn't know if an incomplete pass is incomplete because the receiver dropped a perfectly thrown ball or if the pass was 20 feet over his head; it doesn't know if a quarterback took a sack because the left guard forgot how to play football or if he held the ball for 30 seconds before getting clobbered.  That's one of the reasons we do all these off-season stat articles; to help provide context for the numbers, and one of the reasons we think Allen Robinson is very good despite regularly finishing outside the top 40 in DVOA; we've seen the quality of passers throwing to him!

That being said, quarterback DVOA does tend to be stickier from year-to-year, and vary less based on surrounding talent than, say, receivers do.  They just have much more control over what happens on any given play and have a large sample size of plays to work from.  I'd put his injured shoulder higher on the list of excuses for Trubisky than the quality of players around him.

 

I think I'd rather have him under center in 2021 than Andy Dalton, at the very least.

12 Thanks Bryan for the insight…

Thanks Bryan for the insight,

I think the signing to the Bills is very eye opening, even the Bills GM called the signing a "reset" for Trubisky and doesn't expect him there next season. I truly believe he played through a bad shoulder injury TWICE, and it affected him along with the criticism from Chicago media. If we went back into time before the 2017 NFL draft, I still would have told you he could have been at the same level with Mahomes and Watson. But its amazing how something like a persons personality in the wrong market/situation can make the world of difference. Hopefully he can be a starter at some point with another team in the league. 

24 Well, the Bills now have 2…

Well, the Bills now have 2 of the 3 highest success rate QBs of 2020, the other guy is in deep doo-doo and the next guy is recovering from having his leg nearly snapped off.

The world has become very weird.

7 Success on 3rd down

I wonder if it's worth adjusting the definition of success on 3rd downs when a team goes for it 4th down. It seems like as going for it on 4th becomes more of an accepted practice, coaches are starting to call plays on 3rd-and-long/medium with the primary goal of closing the gap so they have a higher chance of converting 4th down. If the offense closes the gap and goes for it, I think the coaching staff would consider it a success.

My proposal would be a third down play is a success if: a) it converts to a 1st down OR b) it gains 60% of needed yards AND the offense goes for it on 4th down

60% is arguable - I just pulled it from the definition for success on 2nd down. Perhaps it wouldn't change the overall conclusions all that much since this isn't happening on every series or every drive, but I think it would more accurately reflect the way games are being played now.

9 I did a ~very~ quick search…

I did a ~very~ quick search in the play-by-play data here, so I may be off by a few in edge cases that I didn't consider, but here we go.

If we used your proposal, that would flip 189 of the 2,804 failed completions to successes, or about 6.5%.  Matthew Stafford would be the biggest beneficiary, gaining 11 successes, followed by Philip Rivers and Drew Brees with nine apiece.

Of those 189, 106 led to ensuing first downs, with the Saints (8), Rams (7), Packers and Colts (6 each) being the teams who most frequently set up fourth-down conversions with short passes.

Whether those numbers are large enough to justify changing the definition of a successful completion is somewhat a matter of opinion.  I would argue that not all of these would be considered successes by the coaching staff -- sometimes, you come up short and are forced to go for it on fourth down thanks to game situation.

13 Going for it

To me, the big-picture questions are:

1. Just looking at game situation, would the team have gone for it with an incompletion? If yes, then it's hard to say that a nearly successful pass on 3rd down should count as a success--time is probably an important factor too. If no, then I would say it is a success--it gave the team a much likelier chance to keep the ball and be successful. (That's how we judge success rate on 1st and 2nd down.)

2. Did the completion make a potential FG attempt less than, say, 55 yards? (To me, changing a 43 yard attempt to a 37 yard attempt has value, but very little. Changing a potential 58 yard attempt to a 51 yard attempt has a lot more value, esp. considering that an incompletion may have resulted in a punt.)

3. Did the completion gain at least 4 yards? (In other words, a 2 yard completion on 3rd and 3 should still be a failure; 9 yards on 3rd and 10 should be a success if it leads to a successful 4th down conversion.)

Maybe the bar should be: 3rd down positive gain that leads to a successful 4th down conversion AND yardage gained on the 4th down play would not have been enough to convert the 3rd down play.

-------------------------------------------------------------------------------------------

Based on Bryan's stats, I would say that 8 of Brees' "failed completions" were probably instrumental in 4th down success. I have to believe that the way the Saints frequently use Taysom Hill is a factor in 3rd down play-calling. I'm sure the Panthers acted the same with Cam Newton; I suspect that the Bills, Ravens, and Cardinals feel similarly.

17 Instead of developing a more…

In reply to by Joseph

Instead of developing a more and more complicated definition, you could just switch to EPA (where a play is successful if it has positive expected points added) or WPA. The situation-specific subtleties have already gone into defining expected points & win probability.

21 Thanks much for taking a…

Thanks much for taking a look!

Most of the football I watched this year involved the Packers and it definitely seemed to me like they were doing this with some regularity, so I'm mollified to see that they popped up after a quick dig through the data. As mentioned in a couple of other replies, there are definitely some other ways the definition could be refined. Game situation might be another one.

6.5% of the failed completions seems like it's right on the edge of being big enough to want to account for in my mind. You could definitely argue either way.

10 3rd down success %

I would argue closer to 75%-80% of the necessary yardage plus going for it. Going from 3rd and 10 to 4th and 4 (60%) doesn't seem like a success--it simply puts a coach into no-man's-land. Going from 3rd and 10 to 4th and 2 (80%) seems like a coach might be more likely to go for it if the team is <60 yds from the end zone. 4th and 4 is much more likely to be a punt/FG, or a reluctant 4th down attempt b/c the team is just about 40 yards from the end zone. While we can argue that they should go for it at that point, it feels like the best of less optimal choices. 4th and 2 from the same scrimmage line seems like a clear go-for-it scenario, where the success rate of going for it is much higher than a successful punt or FG. Anyway, just my 2 cents.

11 I'll stick with DVOA, its too hard to define a good play

To define good plays vs not good plays is creating a binary system as the article explains.  This article is interesting and states that (of course) there is a large correlation between having a good DVOA and minimizing bad completed pass play percentage.  Kudos to Mitch Trubisky for minimizing bad completed pass plays and maximizing good completed pass plays and still ending up with a -7.5% DVOA.  The reason why was well explained in the article.  I doubt that Buffalo will mention this article during the Trubisky signing press conference, but I am amused by the timing of this article and the signing.

I asked Aaron in 2019 how Lamar Jackson running for 15 yards against Seattle on 3rd and 17 could be considered a bad play.  The Ravens went for 4th and 2 and scored a TD.  Aaron's response was that you get partial credit for the 15 yard run.  Once I learned this, DVOA made so much more sense, it is not a binary system on any given play.  DVOA recognizes that the 15 yards was valuable. 

Trying to define "good plays" is very difficult.  In the posts above, you are trying to make a good play out of being short on 3rd down simply by the coach going for 4th down.  So the coach goes for 4th down, and the team commits a false start or delay of game, now the coach needs to punt.  That 3rd down play that is short no longer looks very good, the player really did need to pick up that 1st down after all.  Coaches decisions should not retroactively determine whether the prior play was a good play or not.  

We saw in the playoffs when Tampa gained 8 yards to make it 2nd and 2 at the two minute warning.  GB intentionally jumped offside on the next play to give Tampa the 1st down (Tampa should have declined the penalty).   An 8 yard play was better than a 10 yard play, this is difficult to account for, as is the guy that goes down at the 1 instead of scoring a TD to run out the clock (bad play Todd Gurley by scoring).

Yes, someday when analytics are more advanced we will be giving runners credit for not reaching the ball forward when they catch a 9 - 10 yard completion on 1st down, as 2nd and less than a yard is better than 1st and 10.  We will penalize runners that do not go down but instead go out of bounds when you need to run out the clock.  We will penalize Gurley for scoring, and credit those in that situation that go down.

 

 

14 DVOA versus success rate

I don't disagree with you. My point is that, with the growing tendency to go for it on 4th down, if FO should adjust its baselines where a certain % (<100%) might be considered a success--just like the baselines for 1st and 2nd downs. On the flip side, I don't think a player/team should get credited for a successful play (in a binary system like success rate) if the outcome is still a punt. As I mention in comment #13--probable punt that turns into FG attempt may need to be counted as a success (regardless of whether the FG is converted). But maybe it shouldn't be. If teams are beginning to consider 3rd down as a stepping stone to going for it on 4th down, at some point we might consider if we are measuring success correctly.

18 A team gains two yards on…

A team gains two yards on third and three, then goes for  it on 4th down. Was that third-down play really successful? It doesn't feel like it should be.

Maybe instead of a percentage of needed yards, use yardage totals? Any, say, 7+ yard gain that results in a fourth and 1, whether or not the team goes for the fourth down (because they probably should have). That feels "successful" to me

19 Yardage totals

I agree. That's why my thought is 75-80%. Two yards on 3rd and 3 wouldn't be enough. It would mean at least 3 yards on 3rd & 4, or 4 yards on 3rd and 5. Expressed like this, I would say then at least 80%--IF FO were to change their definition. If they didn't, no big deal. I don't think it should have to result in 4th and 1, though. 

20 I don't think either of you…

In reply to by Joseph

I don't think either of you are wrong here. Given the way games are called now, you'd think the offense would need 4th-and-2 or shorter for 3rd down to be successful, but getting to 4th-and-2 from 3rd-and-4 isn't a win for the offense. There's likely not a way to define success or failure in a perfect, binary way that doesn't eventually just boil down to something as complicated as EPA or DVOA. As Bryan's reply above states, we're really chasing after edge cases here. Sometimes simple is good.

28 Edge cases

Overall, we probably are chasing edge cases, considering the Saints led the league with 8 "failed" 3rd down completions that led to 4th down conversions--that's one every other game.

On the other hand, I wonder if it is something we should start tracking on some level, to see if this tactic increases. I look at it like the Wildcat trend: you can't use it a lot; but if you have the right personnel, it can be something that you practice and use in certain game situations. Some teams will use it more than others, but it's not a "trick" play either.

15 Is there a big difference…

Is there a big difference for Roethlisberger early in the season and later in the season with and after the Baltimore game who figured the offense out and which game plan everyone after copied. 

The Steelers never adjusted, so I wonder if that game was really a turning point ofbthe season.

16 Not particularly; at least,…

Not particularly; at least, not in this stat.  Roethlisberger had a 27.8% failed completion rate before Week 12 and a 26.4% failed completion rate afterwards.  His successful completion rate dropped from 48.5% to 46.5%.

His numbers certainly dropped off in those final games, but it was more a factor of throwing incomplete passes rather than completing poor ones.

22 Failed drops

would it be possible to get numbers (qb and wr) for drops on passes that would have never been successful if caught?   just curious....

23 I wonder what caused the …

I wonder what caused the (relatively) big spike in the late 80s/early 90s  (and, the similar drop in WR YPC over the same period). Was it simply teams quickly jumping on the "WCO Bandwagon", as it were? Normally such dramatic rate changes are indicative of rules changes or other more concrete things you can point towards.

25 Quite a bit of it is Walsh…

Quite a bit of it is Walsh disciples percolating throughout the league.  In that 1989-1994 spike, you saw:

*Mike Holmgren take over the Packers
*Mike Shanahan take over the Raiders
*Jeff Fisher take over the Oilers
*Dennis Green take over the Vikings
*Bruce Coslet take over the Jets

All of them had worked with Walsh; all of them brought WCOs to their new jobs.

In 1987, only two teams (the 49ers and Bengals) ran what we'd call true west coast systems.  By 1993, that was up to over a quarter of the league, with plenty of more using west coast elements in their offense.  Imitation is the sincerest form of flattery and all.

26 I don't remember ever seeing…

I don't remember ever seeing the difference in subsequent 1st down % of 2-and-5 vs. 2-and-6 that would make a 4 yard completion on first down a failure and a 5 yard completion a success.  Is that % available?

27 You are comparing a binary system vs DVOA

DVOA gives partial credit, so a 4 yard pass on 1st down is a minor failure, a 5 yard pass is a minor success.  Unlike a student who must pass a test with a 70%.  Those that get a 70% pass, those that get a 69% fail, those that get 20% fail.

Reading this article Trubisky was like the student that did the minimum to get by.  He got a 70-73% on most of the tests.  Overall he was like the below average student, but was fortunate enough to "succeed" on a tremendous amount of completed passes and "fail" on very few.

Just as you would take the student that scored over 90% on 8 tests and 65% on 2 tests over the student that scored 70% on all 10 tests, you would take a great many QB's over Trubisky.

We know when we watch the game we are fairly indifferent when our team gets 4 yards vs 5 yards on a first down play.   The difference in that yard on one particular play will virtually never be the difference between winning and losing.  Of course the difference between getting one yard less on EVERY first down play in the game would add up.