Writers of Pro Football Prospectus 2008

18 Feb 2016

Failed Completions 2015

by Scott Kacsmar

The Super Bowl is over, which means it is time to start cranking out stat studies from the previous season. One of the quickest categories we can analyze from the play-by-play data is failed completions. These are any complete passes that fail to gain 45 percent of needed yards on first down; 60 percent on second down; or 100 percent on third or fourth down. You can see last year's study here.

Not every failed completion is created equally, but for this article we make things binary, summing up successes and failures. For the purposes of DVOA, there are fractional points involved where a 10-yard completion on third-and-12 would generate some partial success.

Going back to 1989, we have seen the NFL's failed completion rate (as a percentage of all completions) slowly rise with the growing love of the short passing game, but 2015 set a record with the highest rate yet at 25.1 percent. That is a grand total of 2,893 failed completions. The most failed completions in a game this season was 13, done four times (Jets vs. Eagles, Ravens vs. Jaguars, Rams at Bengals, and 49ers vs. Bengals).

Failed completions are obviously most significant on third and fourth down, which was one of the reasons we created Air Less EXpected (ALEX) as a stat this season. We will look at ALEX as it relates to failed completions for the first time here. Data is only for the regular season and the total number of completions is slightly higher than the official NFL total since Football Outsiders includes backward (lateral) passes.

Not only did we have a record-high season in failed completion rate (FC%), but we have the first quarterback with at least 100 completions since 1989 with an FC% of more than 40 percent. If you recall which quarterback had a horrible season converting on third downs (he's the one with a lot of dark red in the table in this article), you can probably guess who it was. In the following table, the 37 qualified quarterbacks are ranked by ascending FC%, but we also included failed completions as a percentage of attempts (which causes very little change in the rankings) as well as the average ALEX (all downs) for the season.

Rk Quarterback Passes Comp. Failed FC% FC %ATT Rk ALEX Rk
1 Cam Newton 494 296 44 14.9% 8.9% 1 1.4 3
2 Carson Palmer 532 342 52 15.2% 9.8% 2 2.1 1
3 Jameis Winston 533 312 60 19.2% 11.3% 3 0.5 6
4 Blake Bortles 603 355 69 19.4% 11.4% 4 0.3 8
5 Brian Hoyer 369 224 45 20.1% 12.2% 5 0.5 7
6 Andy Dalton 386 255 52 20.4% 13.5% 9 -0.2 13
7 Brock Osweiler 275 170 35 20.6% 12.7% 7 0.0 12
8 Marcus Mariota 368 230 49 21.3% 13.3% 8 0.1 11
9 Russell Wilson 481 329 71 21.6% 14.8% 12 -0.5 17
10 Ben Roethlisberger 469 320 70 21.9% 14.9% 14 1.5 2
11 Matt Ryan 614 407 90 22.1% 14.7% 11 -1.2 24
12 Andrew Luck 291 162 36 22.2% 12.4% 6 0.6 5
13 Matthew Stafford 592 398 90 22.6% 15.2% 15 -2.4 35
14 Tom Brady 621 402 96 23.9% 15.5% 18 -1.2 27
15 Philip Rivers 661 438 105 24.0% 15.9% 21 -2.0 33
16 Jay Cutler 482 311 75 24.1% 15.6% 19 -0.3 15
17 Ryan Fitzpatrick 562 335 81 24.2% 14.4% 10 0.3 9
18 Matt Hasselbeck 257 157 38 24.2% 14.8% 13 -1.2 26
19 Eli Manning 617 387 98 25.3% 15.9% 20 -0.8 18
Rk Quarterback Passes Comp. Failed FC% FC %ATT Rk ALEX Rk
20 Derek Carr 572 350 91 26.0% 15.9% 22 -1.1 22
21 Drew Brees 626 429 113 26.3% 18.1% 27 -1.3 28
22 Johnny Manziel 222 129 34 26.4% 15.3% 17 -1.1 23
23 Kirk Cousins 543 379 100 26.4% 18.4% 29 -0.8 20
24 Ryan Mallett 242 136 37 27.2% 15.3% 16 -0.2 14
25 Aaron Rodgers 571 347 96 27.7% 16.8% 23 -0.8 19
26 Alex Smith 471 308 86 27.9% 18.3% 28 -2.4 36
27 Joe Flacco 409 266 76 28.6% 18.6% 30 -1.4 30
28 Ryan Tannehill 585 364 105 28.8% 17.9% 26 -0.5 16
29 Sam Bradford 531 346 100 28.9% 18.8% 31 -2.0 34
30 Josh McCown 292 186 55 29.6% 18.8% 32 -1.4 29
31 Tyrod Taylor 380 242 72 29.8% 18.9% 33 1.0 4
32 Peyton Manning 330 198 59 29.8% 17.9% 25 0.3 10
33 Colin Kaepernick 244 144 43 29.9% 17.6% 24 -0.8 21
34 Teddy Bridgewater 447 292 95 32.5% 21.3% 35 -1.6 32
35 Matt Cassel 202 119 39 32.8% 19.3% 34 -1.2 25
36 Blaine Gabbert 282 178 63 35.4% 22.3% 36 -2.6 37
37 Nick Foles 335 190 78 41.1% 23.3% 37 -1.5 31
- 2015 NFL Totals 18258 11534 2893 25.1% 15.8% - -0.6 -

Maybe you forgot him since he was benched for Case Keenum, but Nick Foles (41.1 percent) now has the highest FC% since 1989. He beat out Anthony Wright, who was at 39.6 percent for the 2005 Ravens. For reference, Keenum was about average at 24.7 percent this season, so this says more about Foles than it does Tavon Austin and the Rams. Foles also ranked 32nd with an FC% of 28.0 percent in 2014.

Cam Newton and Carson Palmer had career seasons and were the top two MVP candidates. Both look good here, which is nothing new for Newton -- his worst year in this category was 2014, and even then his FC% was just 17.9 percent. His vertical nature combined with some erratic accuracy makes him a likely candidate for a low FC%, which we also may be seeing with Jameis Winston in Tampa Bay if his rookie season is any indication of future performance. Blake Bortles had one of the year's biggest improvements in his second season, going from 35th in FC% (30.4 percent) as a rookie to fourth this season. (Some of that change can also be attributed to the switch at offensive coordinator from Jedd Fisch to Greg Olson.)

Matt Ryan managed the best FC% for anyone with a below-average ALEX this season. Tyrod Taylor and Peyton Manning are the only quarterbacks to finish in the top 10 in ALEX, but below average in FC%. Neither played in an offense with much YAC, but that is especially surprising in Taylor's case since his game was so aggressive. We will see later which receivers impacted this. When we get to looking at YAC+ for 2015, we should be able to come up with an expected FC% for each quarterback and receiver.

The next quarterback table compares FC% to ALEX (all passes) by down where third and fourth down are grouped together. Rankings are from best (green) to worst (red), and the darker the color, the more standard deviations that quarterback is from the average.

We have an extreme outlier here in Nick Foles, whose second-down FC% of 47.9 percent is nearly double the average of 25.6 percent. Not surprisingly, he has the lowest ALEX on second down. Some of the season's best passers (Newton, Palmer, and Ben Roethlisberger) avoid any red marks here. Newton is the only player to rank in the top five in all six splits. The other three passers with all green are a surprising trio: Brock Osweiler, Blake Bortles, and Brian Hoyer.

On Twitter, I had a lot of people expecting Teddy Bridgewater to bring up the rear like Foles has, and he is near the bottom. Through two seasons, he definitely seems to be in the "checkdown Charlie" category of passers, and having the highest FC% on first down is not good for an offense. Naturally, the Vikings finished 26th in first-down DVOA.

Someone has to be on the receiving end of these plays. I looked at the failed completions for everyone with at least 30 receptions. The following table excludes running backs since they catch so many short passes that they would dominate these lists. Duke Johnson led all running backs with 32 failed receptions, and Justin Forsett had the highest FC% at the position at 64.5 percent.

Most Failed Completions (WR/TE) Lowest Failed Completion Rate (WR/TE) Highest Failed Completion Rate (WR/TE)
Receiver Team Failed Receiver Team Rec. Failed Pct. Receiver Team Rec. Failed Pct.
Jarvis Landry MIA 35 Dorial Green-Beckham TEN 32 0 0.0% Eddie Royal CHI 37 19 51.4%
Antonio Brown PIT 30 Devin Funchess CAR 31 1 3.2% Tavon Austin STL 53 22 41.5%
Delanie Walker TEN 25 Rishard Matthews MIA 43 2 4.7% Mychal Rivera OAK 32 13 40.6%
Brandin Cooks NO 24 Mike Evans TB 74 4 5.4% Vance McDonald SF 30 11 36.7%
Randall Cobb GB 24 Vincent Jackson TB 33 2 6.1% Richard Rodgers GB 58 20 34.5%
Golden Tate DET 23 Ted Ginn CAR 44 3 6.8% Chris Hogan BUF 36 12 33.3%
Julio Jones ATL 22 Allen Robinson JAC 80 6 7.5% Jarvis Landry MIA 111 35 31.5%
Tavon Austin STL 22 DeAndre Hopkins HOU 111 9 8.1% Charles Clay BUF 51 16 31.4%
Jason Witten DAL 21 A.J. Green CIN 86 7 8.1% Jared Cook STL 39 12 30.8%
Jordan Matthews PHI 20 Markus Wheaton PIT 44 4 9.1% Randall Cobb GB 79 24 30.4%
Richard Rodgers GB 20 Terrance Williams DAL 52 5 9.6% Coby Fleener IND 54 16 29.6%

Jarvis Landry led the NFL with 35 failed completions after finishing second in 2014 with 26. His rate actually went up a little too, from 31.0 percent to 31.5 percent. He is a bit of a stud in fantasy football with PPR scoring, but his short gains have not really translated into real-life success for Miami's passing game. Landry's 174 DYAR over his first two seasons is less than the 235 DYAR that teammate Rishard Matthews had in 2015 alone. You can see Matthews had the third-lowest FC% of 2015 as he was very efficient with his targets. Only two deep-threat rookies (Dorial Green-Beckham and Devin Funchess) had a lower rate. Vincent Jackson and Terrance Williams are the only players that ranked in the top 10 for lowest FC% in 2014 as well. Williams' numbers (five failures on 52 receptions) were actually also achieved by Tyler Eifert and Michael Floyd, but we only showed Williams for formatting reasons.

Even with new quarterbacks, Chris Hogan and Tavon Austin each made his second straight appearance among the 10 highest failed completion rates. Hogan and Charles Clay both appear for Buffalo, though what surprised me more was Sammy Watkins going from 7.7 percent FC% last year (best among 2014 rookies) to 13.3 percent this year. If you followed Watkins' year, you know he had a ridiculous average depth of target (16.8 yards), but he still had eight failed catches among his total of 60. Just recently there was a story about DeAndre Hopkins wanting to improve on his YAC, which we highlighted as being almost absent in his game this past season. Allegedly Watkins was the one who told Hopkins about the two former Clemson receivers' ranking last this past year. But Hopkins still finished with the eighth-lowest rate here.

The correlation between a receiver's FC% and his ALEX (all targets) was minus-0.84. Compare that to the minus-0.60 correlation between a quarterback's FC% and his ALEX (all passes). I think this speaks to how a quarterback throws to a variety of targets, and some of them are better at turning short passes into longer gains. For an individual receiver, his role determines a lot of his receiving stats, and you are either capable of making plays after the catch or you are not. This is why we view YAC as more of a receiver stat than a quarterback stat.

Finally, let's look at the defenses' ability to create failed completions, with a comparison to how these units fared in 2014.

Defenses: 2015 Failed Completions Compared to 2014
Rk Team Comp. Failed FC% 2014 Rk 2014 FC% Diff Rk
1 DEN 344 111 32.3% 4 28.1% 4.2% 4
2 CAR 390 122 31.3% 2 28.5% 2.8% 7
3 CIN 416 117 28.1% 8 26.6% 1.5% 11
4 BAL 350 98 28.0% 10 26.2% 1.8% 10
5 DET 361 101 28.0% 5 27.9% 0.0% 22
6 STL 399 111 27.8% 6 27.2% 0.6% 19
7 SD 328 91 27.7% 9 26.3% 1.5% 12
8 ARI 343 95 27.7% 7 26.8% 0.9% 16
9 DAL 330 91 27.6% 14 25.6% 2.0% 9
10 SEA 333 91 27.3% 11 26.1% 1.2% 15
11 ATL 371 100 27.0% 32 17.0% 10.0% 1
12 IND 354 94 26.6% 22 23.1% 3.5% 6
13 MIN 359 94 26.2% 3 28.3% -2.1% 27
14 NE 361 93 25.8% 16 24.9% 0.9% 17
15 NYJ 343 87 25.4% 17 24.1% 1.3% 14
16 KC 349 88 25.2% 23 23.0% 2.3% 8
Rk Team Comp. Failed FC% 2014 Rk 2014 FC% Diff Rk
17 HOU 326 82 25.2% 18 23.8% 1.3% 13
18 SF 376 94 25.0% 29 20.3% 4.7% 2
19 JAC 391 96 24.6% 19 23.8% 0.7% 18
20 CHI 322 79 24.5% 28 20.5% 4.0% 5
21 GB 322 76 23.6% 31 19.2% 4.4% 3
22 WAS 354 82 23.2% 15 25.2% -2.1% 26
23 OAK 404 92 22.8% 1 29.7% -6.9% 32
24 NYG 423 95 22.5% 26 22.5% 0.0% 23
25 CLE 322 72 22.4% 12 26.0% -3.6% 30
26 BUF 348 77 22.1% 13 25.8% -3.7% 31
27 MIA 350 77 22.0% 27 21.9% 0.1% 21
28 PHI 394 86 21.8% 21 23.2% -1.4% 25
29 PIT 402 87 21.6% 24 22.8% -1.2% 24
30 NO 372 76 20.4% 30 20.3% 0.1% 20
31 TB 378 76 20.1% 20 23.5% -3.3% 29
32 TEN 319 62 19.4% 25 22.8% -3.3% 28

It is always interesting when an obscure stat such as this one spits out results like this, with the top two teams earning No. 1 seeds and meeting in the Super Bowl, and leading the league in defense. Denver has been solid in this stat for a few years now, but no one has dominated failed completions like Ron Rivera's Carolina defense. The Panthers were first with an excellent 34.0 percent in 2013, their breakout year under Rivera, when they were the only team in the league above 30 percent. Even in 2014 when the Panthers had a rough year overall, the defense managed to finish second in FC%. Denver and Carolina are the only teams above 30 percent this season.

Rivera comes from Lovie Smith's coaching tree in Chicago with the Cover-2 defense, and he also spent time with the 3-4 in San Diego. Maybe that style of keeping the play in front of you works out well here. Then again, you see Tampa Bay (Smith) and Tennessee (Dick LeBeau's "tackle the catch" principles going stale) bringing up the rear, with fired head coaches in both cities.

The Falcons made the biggest leap behind new coach Dan Quinn this season, but that was not too hard considering they were 32nd in 2014. Still, their improvement of 10 percentage points was more than double that of the next closest team (San Francisco). The biggest decline happened in Oakland after the Raiders surprisingly led the league in FC% in 2014 despite only ranking 28th in pass defense DVOA.

We'll end on a quasi-positive note: after finishing 30th in FC% in 2014, the Saints were again 30th in 2015. Hey, at least it was not 32nd like seemingly every other category for that Rob Ryan defense.

Posted by: Scott Kacsmar on 18 Feb 2016

4 comments, Last at 23 Feb 2016, 2:56am by Scott Kacsmar

Comments

1
by MilkmanDanimal :: Thu, 02/18/2016 - 4:34pm

This is a really lovely encapsulation of Tampa's season; Winston third-best for failed completions, Mike Evans and Vincent Jackson both in the top ten for lowest rate of failed completions as receivers, defense #31, giving up first downs at a lousy rate. Pretty much matches the proverbial eyeball test from last year.

2
by Aaron Brooks Go... :: Fri, 02/19/2016 - 9:59am

The comment regarding who's responsible for failures and why YAC is a receiver phenomenon was a useful observation.

3
by tomb :: Mon, 02/22/2016 - 9:51pm

is this analysis available during the season? it can be very useful.

4
by Scott Kacsmar :: Tue, 02/23/2016 - 2:56am

Not in a formal post, but I would say failed completions get brought up frequently when notable for single games in our weekly columns (Quick Reads, Clutch Encounters) during the season.