Writers of Pro Football Prospectus 2008

03 Jun 2016

Defense: Expected Failed Completions and YAC+

by Scott Kacsmar

Earlier this week, we looked at a new stat called Expected Failed Completions (EFC), but only focused on the offensive side of the ball. Now we will take a look at the defenses, where forcing failed completions is a good thing. While a defense would prefer to not give up a completion at all, the game has changed too much to expect a lot of incompletions. Only six defenses in 2015 allowed opposing passers to complete fewer than 60 percent of their passes for the season. In 2005, 20 defenses were able to do that. In 1985, when the Chicago Bears ruled the league, 26 of the 28 defenses held opponents below 60 percent. Today's flurry of short passes makes defense more of a quick pressure and tackling game, limiting the big gains.

So what is an EFC? Here is a refresher if you skipped the offensive article.

First, a failed completion (FC) is any completed pass that failed to gain enough yards to count as a successful play based on these guidelines: 45 percent of needed yards on first down, 60 percent on second down, or 100 percent on third/fourth down. You can see this year's study here. As I mentioned in that article, we can look at the data associated with calculating YAC+ to create EFC. Based on down, distance, field position, and pass length, each throw has an expected amount of YAC. By adding together actual air yards and expected YAC, we get an Expected Gain for each pass, which is then measured against the 45/60/100 baselines to determine whether the completion was a success or failure. 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.

It should also be noted that certain types of incompletions are excluded from this data, including passes intentionally thrown away or batted down at the line. While those are good defensive plays that should certainly factor into a defense's success, since our data is just breaking down completions, it is fine to keep those plays out. Here we are interested in analyzing how a defense allows and reacts to a completed pass moreso than preventing them from happening.

2015 Defenses: Expected Failed Completions

We have data going back to 2006 for this, but let's get right into the defensive EFC and FC results for 2015. The following table is large, so we will discuss its various components in smaller sections.

2015 Defenses: Actual vs. Expected Failed Completions and YAC+
Team Passes Comp. FC EFC BOC FC% Rk EFC% Rk DIFF Rk Short% Rk YAC+ Rk
DEN 519 344 112 36 3 32.6% 1 10.5% 21 22.1% 1 57.3% 4 -0.51 6
DET 506 361 100 28 9 27.7% 9 7.8% 32 19.9% 2 56.5% 5 0.00 21
CAR 599 390 123 54 6 31.5% 2 13.8% 5 17.7% 3 55.1% 8 -0.77 2
BAL 498 350 98 39 10 28.0% 6 11.1% 18 16.9% 4 54.6% 11 -0.17 13
SEA 499 333 92 38 6 27.6% 10 11.4% 15 16.2% 5 51.7% 15 -0.97 1
NE 568 361 94 36 9 26.0% 14 10.0% 26 16.1% 6 51.2% 16 -0.60 4
STL 563 399 113 50 12 28.3% 4 12.5% 9 15.8% 7 61.2% 1 0.60 27
ATL 541 371 101 43 9 27.2% 11 11.6% 13 15.6% 8 55.0% 9 -0.29 9
DAL 480 330 92 41 12 27.9% 8 12.4% 10 15.5% 9 57.6% 3 0.59 26
SF 517 376 95 37 9 25.3% 17 9.8% 27 15.4% 10 54.0% 12 -0.31 8
MIN 525 359 95 40 7 26.5% 13 11.1% 19 15.3% 11 49.0% 23 -0.17 12
KC 572 349 88 35 6 25.2% 18 10.0% 25 15.2% 12 49.6% 22 -0.56 5
ARI 530 343 96 44 9 28.0% 7 12.8% 7 15.2% 13 54.8% 10 0.47 25
NYJ 560 343 87 35 7 25.4% 16 10.2% 24 15.2% 14 48.4% 25 0.12 23
CHI 473 322 81 33 10 25.2% 19 10.2% 23 14.9% 15 57.8% 2 0.68 29
IND 546 354 94 43 10 26.6% 12 12.1% 12 14.4% 16 44.9% 30 -0.31 7
Team Passes Comp. FC EFC BOC FC% Rk EFC% Rk DIFF Rk Short% Rk YAC+ Rk
PHI 590 394 89 34 6 22.6% 26 8.6% 29 14.0% 17 49.0% 24 -0.06 17
NYG 596 423 98 39 11 23.2% 25 9.2% 28 13.9% 18 49.6% 21 0.17 24
BUF 560 348 78 30 8 22.4% 28 8.6% 30 13.8% 19 48.0% 26 -0.12 15
JAC 551 391 98 45 13 25.1% 21 11.5% 14 13.6% 20 50.9% 17 0.08 22
TB 515 378 80 30 8 21.2% 30 7.9% 31 13.2% 21 46.3% 27 -0.23 10
OAK 596 404 96 43 14 23.8% 23 10.6% 20 13.1% 22 50.5% 19 -0.05 19
HOU 503 326 83 41 12 25.5% 15 12.6% 8 12.9% 23 53.1% 13 -0.06 18
SD 472 328 92 51 11 28.0% 5 15.5% 2 12.5% 24 55.8% 7 0.85 30
CIN 605 416 118 67 12 28.4% 3 16.1% 1 12.3% 25 56.0% 6 -0.74 3
PIT 588 403 91 45 9 22.6% 27 11.2% 17 11.4% 26 50.1% 20 -0.01 20
CLE 476 322 76 40 7 23.6% 24 12.4% 11 11.2% 27 43.5% 31 0.93 31
WAS 536 354 89 50 10 25.1% 20 14.1% 4 11.0% 28 45.5% 29 -0.07 16
NO 521 372 77 42 13 20.7% 31 11.3% 16 9.4% 29 50.5% 18 1.39 32
MIA 506 350 78 46 14 22.3% 29 13.1% 6 9.1% 30 46.3% 28 -0.22 11
TEN 467 319 62 33 9 19.4% 32 10.3% 22 9.1% 31 42.6% 32 -0.13 14
GB 501 322 77 48 10 23.9% 22 14.9% 3 9.0% 32 52.8% 14 0.62 28

Failed Completions (FC and FC%)

Carolina was the best defense in the NFC last season and led the NFL with 123 failed completions. That total is actually the second-highest in our database, trailing only the 2013 Panthers, who forced 131. It sure helps to have a disruptive pass-rusher in the middle like Kawann Short, and a great linebacker corps with Luke Kuechly and Thomas Davis.

Tennessee forced the fewest FCs with 62, but 25 defenses since 2006 did less than that. The futility record belongs to the 2007 Dolphins (1-15), who only had 40 and also the worst FC rate of any defense (16.5 percent).

In terms of rate, both Super Bowl defenses were the only ones above 30 percent this season, as they were clearly the two best units. In fact, the Broncos and Panthers both rank in the top four in FC rate for the last decade. The 2013 Panthers (34.7 percent) are on top, and that great 2009 Jets defense ranks third (32.4 percent) from Rex Ryan's first season there.

Back to 2015, Tennessee finished last in FC rate, which could be viewed as an indictment of Dick LeBeau's aging scheme. Granted, Ray Horton had the title of defensive coordinator, but LeBeau will have more control in 2016 after coming over last year from Pittsburgh where he was known for his "tackle the catch" style of blitz-happy defense. When you look at the Pittsburgh data from 2006 to 2014, the Steelers averaged a ranking of 17th in FC rate and only had two seasons in the top 12. More on this later.

Expected Failed Completions (EFC and EFC Rate)

Neither the 2015 Panthers (54) nor the 2013 Panthers (61) had the most EFCs in our database. That mark now belongs to the 2015 Bengals, who had 67. Cincinnati has had some interesting EFC splits under Marvin Lewis. From 2006 to 2009, the Bengals ranked 20th or worse in EFC rate each year, and 28th or worse three times). Since 2010, the Bengals have finished sixth, second, fourth, fourth, 26th and first in EFC rate. Mike Zimmer was a great defensive coordinator with the team, but he was there for both splits from 2008 to 2013. Not to try tying this to one player, but Geno Atkins was drafted in 2010, even though he did not become a starter until 2011. He also missed half of 2013, and his slow return to form may help explain the ranking of 26th in 2014, but the Bengals have been as consistently strong as any defense in EFC rate over the last six years.

Detroit's league-low 28 EFCs led to the Lions becoming the only team to finish last in EFC rate multiple times, doing so in 2007 and 2015. However, the 2015 Lions are the only team to finish 32nd in EFC rate and higher than 16th in FC rate, finishing ninth. The Lions were much better than those other defenses at forcing shorter throws that were more likely to lead to a FC.

Bail-out Completions (BOC)

As mentioned in the offensive feature, these bail-out completions (BOC) are plays where the completion was successful, but it was expected to be a failure. That means the YAC really carried the play to beat expectations. There were only 301 BOCs in all of the 2015 season. Miami and Oakland allowed the most with 14 each, while the talented Broncos only allowed three, half that of the next-best defense. As a percentage of completions, Miami (4.0 percent) was the highest and Denver (0.9 percent) was the lowest.

Difference Between FC and EFC Rates (DIFF)

The table is sorted by the largest difference between the rates for EFC and FC. We are certainly used to seeing more consistency among offensive stats than in defensive stats. One big reason is the quarterback's involvement has a significant impact on any offense's numbers, where on any given defensive play, any one of the 11 players on the field can be the hero or the goat. For the 337 quarterback seasons with at least 200 attempts, we observed a correlation of 0.73 between EFC rate and FC rate. While not providing the full picture, YAC obviously does make up for a lot of the difference between these rates. However, when it comes to looking at the defensive numbers since 2006, the correlation drops to 0.55 (and 0.35 for the 2015 season).

My best hypothesis for this difference is that the offense's numbers are built from the same passer and mostly the same receivers, while the defensive numbers are accumulated from a schedule of varying skill. Whatever gap exists between the skill of the quarterback and his receivers is going to remain pretty constant for the offense, but the defense will see a lot more variation over 16 games.

While Denver had the highest actual FC rate in 2015, its EFC rate ranked just 21st. That difference of 22.1 percentage points is the largest among the 320 defenses since 2006. Denver also had the largest difference in 2014 (18.3 percent), and the second-largest difference in the 2012 season (17.4 percent), two other seasons where the defense was strong and talented, but not as consistently great as it was last year under Wade Phillips. Some of Denver's players have moved on this offseason, and while the 2016 defense should still be good, regression is assumed. Still, you really have to admire the effort that went into building that 2015 unit.

While Green Bay's 2015 offense had the largest difference between FC rates, its defense had the smallest (9.0 percent). The Packers are quite used to that under Dom Capers, who was hired in 2009, having ranked 20th or lower in six straight seasons. Capers' school of thought is very similar to LeBeau's, and you can see the Titans ranked 31st in difference. Things have really soured for Capers in this department. The Packers led the NFL in FC rate in 2008 and 2010, the latter being a great Super Bowl-winning defense. But since 2011, Green Bay has ranked 22nd or worse in FC rate each season. The main culprit for that has been allowing too much YAC, but at least the 2015 defense forced more short passes. Our last two stats will focus on those two areas.

Short Percentage (Short%)

The stat Short% is the percentage of a quarterback's completions that were thrown shorter than what was needed for a successful play, meaning he was going to rely on YAC to make the play gain enough yards. If a quarterback always threw the ball at least 5 yards on first-and-10, then he would never have to worry about a failed completion unless the receiver fumbled or had negative YAC.

The correlation between Short% and FC% was 0.70 in 2015 and 0.63 for the last decade. No matter how good the offense is, trying to convert a third-and-10 with a 2-yard pass is difficult.

St. Louis became the first defense in the last 10 years to force over 60 percent of passes to be thrown short of a successful play. The impressive part is the 2014 Rams also have the third-best rate in our database at 58.3 percent. The defense coming between those two just so happens to be the 2010 Saints (59.9 percent), one of only two top-10 defenses for Sean Payton in New Orleans. The common link between the three is defensive coordinator Gregg Williams, famous for Bountygate, bad facial hair, and being Jeff Fisher's boy. But Williams is also known for loving to blitz, so quarterbacks have to release the ball quickly when playing against him. Even Williams' 2007 Redskins led the league in Short%, and the 2011 Saints were second. With the talent the Rams have in the front seven (and the lack of it in the secondary), it makes a lot of sense why they would do so well in this particular stat.

Tennessee finished last in Short%, which is the second time that has happened for a LeBeau defense (2012 Steelers). This is not necessarily a bad thing if you are confusing quarterbacks into tougher throws down the field and coming away with takeaways. When looking at the 10 defenses to finish last in Short%, you find two Bill Belichick units (2009 and 2011 Patriots), a Rex Ryan team (2010 Jets), and two defenses led by Romeo Crennel (2006 and 2008 Browns), a Belichick disciple. These coaches are known for mixing things up to create confusion. However, finishing last in Short% is still a lousy way of accumulating FCs. All 10 of these defenses finished 26th or worse in FC rate.

YAC+

The last stat, and really the most important, is YAC+, which is the average YAC compared to what an average receiver would have gained in similar field position given that down-and-distance situation. Since these are defensive numbers, positive YAC+ means below-average performance.

You knew I could not finish a look at 2015 defenses without criticizing Rob Ryan and the Saints. Naturally, the Saints finished dead last in YAC+ (1.39), and it was the fourth-worst mark in our database, ahead of only the 2009 Dolphins (1.47), 2009 Browns (1.62), and 2011 Buccaneers (1.70).

Fortunes can change in a hurry though. In 2010, the two worst defenses in YAC+ were Seattle and Denver. Carolina finished 31st in Ron Rivera's rookie year (2011). Now look at where those defenses have been.

The one defense to truly marvel at here is Seattle, which has led the league in YAC+ three years in a row. Seattle's 2012 defense finished third, which can help explain why this team has led the league in fewest points allowed in four straight seasons. The scheme certainly helps, but the scheme would be nothing without some great talent, including some of the best safeties and linebackers in the league.

In fact, safety may be the key position here given the Steelers (Troy Polamalu) led the league three times in YAC+. LeBeau's "tackle the catch" is an easy target for criticism, but when the defense is very talented as the Steelers used to be, those units often did their job. Sure, the cushions on the outside made for a lot of easy completions and some available YAC, but good luck breaking those tackles to generate more than the expected YAC. The Colts (with Bob Sanders and Antoine Bethea) also led the league in YAC+ twice, both Super Bowl years for the Peyton Manning era (2006 and 2009), and they finished second in 2007. Tony Dungy's Tampa-2 did a good job of keeping the play in front of the defense, which was built for speed.

Seven of the top 20 defenses in YAC+ reached the Super Bowl, posting a 3-4 record in the big game. We conclude with a look at the top-ranked defenses in the four main stats in each season, and how that team fared that year.

No. 1 Defense in Each Category, 2006-2015
Team Year FC% Result Team Year EFC% Result
BUF 2006 30.6% No Playoffs GB 2007 13.9% Lost NFC-CG
WAS 2007 29.5% Lost NFC-WC OAK 2014 15.4% No Playoffs
GB 2008 30.0% No Playoffs JAC 2011 15.6% No Playoffs
NYJ 2009 32.4% Lost AFC-CG PHI 2009 15.8% Lost NFC-WC
GB 2010 30.1% Won SB NYG 2008 15.9% Lost NFC-DIV
PHI 2011 29.5% No Playoffs CIN 2015 16.1% Lost AFC-WC
DET 2012 29.7% No Playoffs CHI 2006 16.2% Lost SB
CAR 2013 34.7% Lost NFC-DIV CAR 2013 16.2% Lost NFC-DIV
OAK 2014 29.7% No Playoffs GB 2010 16.7% Won SB
DEN 2015 32.6% Won SB DET 2012 17.0% No Playoffs
Team Year YAC+ Result Team Year Short% Result
IND 2006 -0.53 Won SB BUF 2006 54.6% No Playoffs
TEN 2007 -1.15 Lost AFC-WC WAS 2007 54.5% Lost NFC-WC
PIT 2008 -0.90 Won SB GB 2008 57.5% No Playoffs
IND 2009 -1.18 Lost SB PHI 2009 57.1% Lost NFC-WC
PIT 2010 -1.14 Lost SB NO 2010 59.9% Lost NFC-WC
PIT 2011 -0.99 Lost AFC-WC PHI 2011 53.6% No Playoffs
CAR 2012 -0.96 No Playoffs DET 2012 55.0% No Playoffs
SEA 2013 -1.15 Won SB OAK 2013 56.0% No Playoffs
SEA 2014 -0.79 Lost SB STL 2014 58.3% No Playoffs
SEA 2015 -0.97 Lost NFC-DIV STL 2015 61.2% No Playoffs

Posted by: Scott Kacsmar on 03 Jun 2016

1 comment, Last at 05 Oct 2017, 7:13am by sachinkhanna78632

Comments

1
by sachinkhanna78632 :: Thu, 10/05/2017 - 7:13am