Writers of Pro Football Prospectus 2008

Most Recent FO Features

HarrisChr11.jpg

» Film Room: Chris Harris

Is Harris one of the league's top cover corners, or a product of the system in which he plays? Cian Fahey says the answer lies somewhere in the middle.

25 Jun 2013

2012 Play-Action, Defense

by Mike Ridley

Recently, we looked at which offenses were the best and worst against play-action. Today, we’re going to switch sides and look at it from the defensive point of view. As we pointed out last year, it’s an underappreciated statistic, but it can prove useful in diagnosing deficiencies on defense.

The team that was most often burned by play-action in 2012 was the New Orleans Saints. Although the Saints already had a fairly poor pass defense (their 14.0% DVOA against regular passes would’ve ranked 24th), their 45.0% DVOA against play-action dropped their total pass defense down to 28th. A large factor in this league-worst DVOA was the 10.0 yards per play the Saints allowed on play-action, worst in the league by 12 percent.

A closer look into the Saints failures against fake handoffs shows they allowed a league-high 40 plays of 15 or more yards on play action, averaging an astounding 28.9 yards. Chief among the offenders was Patrick Robinson, who, thanks to our game charters, was listed as in coverage on a whopping nine of those 40 plays. Jabari Greer chipped in his fair share, as well, getting burned on six additional plays (including one where Robinson was listed as "DEFENDER2"). Our good charting friends "Blown Coverage" and "Hole in Zone" made the biggest impact, though. In the eight plays those reasons were listed, the offense averaged 43 yards per play.

The defense that did the worst against play action relative to overall pass defense was the Baltimore Ravens. The Ravens had a pass defense DVOA of -5.2% against regular passes, which would’ve been good enough for 10th in the NFL. However, against play-action, the unit crumpled, compiling a DVOA of 37.0%, fourth-worst in the league. This happened only a year after the Ravens posted the third-best DVOA against play-action. On play-action, offenses would attack the Ravens deeper down field, with the average pass traveling nearly three yards further downfield. The receivers were also able to do more once the ball was in their hands, with their Yards After Catch going up by 60 percent. This culminated in the Ravens having successful plays on defense (as defined by Football Outsiders) only 52 percent of the time, compared to 58 percent on regular passes, which was the seventh-best mark in the league.

Just as was the case between 2010 and 2011, there is no correlation between performance against play-action for 2011 and 2012. After a year-to-year correlation of -.06 for 2010-2011, the DVOA for pass defense against play-action had a correlation of -.03 for 2011-2012. Does this mean that running these numbers is completely meaningless? Honestly, we're not sure. There was correlation with defense against play-action between 2009 and 2010, with a correlation coefficient of .28.

The correlation between frequency and effectiveness is somewhat more prevalent, at least for those who faced play-action the most. Of the five defenses who saw the most play-action (Oakland, Dallas, Cincinnati, Jacksonville and Philadelphia), only Cincinnati wasn’t among the eight worst teams by DVOA against play-action.

Below is the play-action data from 2012 for all teams, in descending order starting with teams that had the biggest DVOA difference between play-action and regular passes. The columns in green represent performance with play-action on actual passes only, with sacks, scrambles and intentional grounding removed.

2012 Play-Action Pass
Statistics, Defense
Defense PA% Yds/Play

with PA
DVOA

with PA
Yds/Pass with PA

(Actual passes only)
DVOA with PA

(Actual passes only)
Yds/Play

no PA
DVOA

no PA
Yds/Play

Dif.
DVOA

Dif.
BUF 20% 6.5 -11.2% 6.4 -13.9% 6.4 23.6% 0.0 -34.9%
DET 20% 6.2 -3.9% 6.1 -6.2% 6.8 23.2% -0.6 -27.0%
KC 20% 8.7 4.6% 8.9 4.0% 6.9 30.9% 1.8 -26.4%
ATL 22% 6.8 -20.1% 6.9 -21.1% 6.7 5.7% 0.1 -25.8%
MIN 21% 6.2 2.2% 6.3 1.5% 6.3 21.5% 0.0 -19.2%
WAS 21% 6.6 -6.8% 6.6 -7.3% 6.8 8.1% -0.2 -14.9%
GB 19% 7.8 -10.4% 7.8 -13.7% 5.5 -3.0% 2.2 -7.4%
PIT 23% 6.0 -0.2% 6.0 -0.5% 5.5 6.3% 0.5 -6.5%
CAR 23% 7.4 -0.2% 7.6 0.0% 5.7 1.2% 1.7 -1.4%
JAC 25% 7.5 27.2% 7.7 23.8% 7.1 23.8% 0.4 3.5%
PHI 24% 7.7 30.2% 8.0 38.0% 6.6 24.1% 1.1 6.1%
Defense PA% Yds/Play

with PA
DVOA

with PA
Yds/Pass with PA

(Actual passes only)
DVOA with PA

(Actual passes only)
Yds/Play

no PA
DVOA

no PA
Yds/Play

Dif.
DVOA

Dif.
NYJ 21% 6.6 3.8% 6.5 1.8% 5.7 -4.3% 0.9 8.1%
NE 23% 8.7 20.5% 8.9 20.5% 6.6 12.2% 2.1 8.4%
SD 19% 7.7 19.1% 7.6 17.2% 5.9 9.9% 1.8 9.3%
OAK 26% 8.0 32.6% 7.8 27.7% 6.7 22.8% 1.3 9.7%
CLE 20% 7.4 20.4% 7.4 18.5% 6.2 9.3% 1.3 11.1%
NYG 21% 8.8 19.1% 8.9 14.8% 6.8 4.9% 2.0 14.2%
HOU 16% 6.9 3.3% 7.1 5.9% 5.8 -13.1% 1.1 16.4%
CIN 26% 7.5 8.9% 7.4 5.8% 5.1 -8.5% 2.3 17.4%
CHI 22% 6.5 -14.9% 6.6 -17.4% 5.1 -32.6% 1.4 17.6%
MIA 19% 8.2 24.5% 8.2 22.2% 6.0 6.2% 2.1 18.3%
TB 23% 8.4 33.3% 8.5 35.2% 7.1 14.1% 1.2 19.2%
Defense PA% Yds/Play

with PA
DVOA

with PA
Yds/Pass with PA

(Actual passes only)
DVOA with PA

(Actual passes only)
Yds/Play

no PA
DVOA

no PA
Yds/Play

Dif.
DVOA

Dif.
TEN 22% 6.8 27.0% 6.8 24.2% 6.6 7.0% 0.3 20.0%
SEA 20% 6.9 3.3% 7.1 3.2% 5.2 -22.2% 1.6 25.4%
DEN 17% 6.4 10.5% 6.5 10.9% 5.1 -15.0% 1.4 25.6%
SF 18% 6.7 13.3% 6.8 13.2% 5.2 -13.7% 1.5 27.1%
DAL 26% 8.3 37.3% 8.3 36.6% 6.2 9.2% 2.1 28.1%
ARI 23% 8.4 2.5% 8.8 3.2% 5.3 -27.8% 3.1 30.3%
IND 20% 8.7 43.4% 8.8 43.3% 6.4 12.9% 2.3 30.5%
NO 23% 10.0 45.0% 10.0 44.5% 6.9 14.0% 3.1 31.0%
STL 23% 8.1 27.0% 8.1 25.6% 5.5 -13.7% 2.6 40.7%
BAL 23% 8.7 37.0% 8.8 37.9% 5.6 -5.2% 3.1 42.2%
NFL 21% 7.6 13.8%
7.6

12.9%
6.1 3.8%
1.5

9.9%

Posted by: Mike Ridley on 25 Jun 2013

19 comments, Last at 17 Oct 2013, 12:11pm by Freddy

Comments

1
by Illuminatus! (not verified) :: Tue, 06/25/2013 - 3:51pm

That's a very surprising stat for Buffalo. Considering how monumentally crappy they were against the run, to come out with as one of the best teams vs. PA is surprising. Maybe they were so bad against the run because they were always playing pass first, which correspondingly made them strong against PA?

4
by Karl Cuba :: Tue, 06/25/2013 - 5:54pm

Or were they so concerned with trying to assuage the deluge of rushing yards that they bit on every play fake?

5
by JoeyHarringtonsPiano :: Tue, 06/25/2013 - 6:22pm

I think what you mean is, they had no reaction to a handoff action, so they were terrible at defending the run, but the upside is, their pass defense didn't suffer when the handoff turned out to be fake.

10
by Karl Cuba :: Wed, 06/26/2013 - 11:51am

Yeah, brain fart on my part.

13
by Illuminatus! (not verified) :: Wed, 06/26/2013 - 2:32pm

But they did more than just "not suffer," they improved dramatically. They are listed as 22nd in the league vs. the pass by DVOA overall with 13.1%, but at -13.9% DVOA vs. play action with an actual pass.

That's just an enormous swing, and it shows how bad they must have been against all other passes if they had this very strong D vs. PA (which they faced at a slightly below average clip, but still) and ended up so bad overall.

15
by dcaslin :: Wed, 06/26/2013 - 6:30pm

That's my guess on the Ravens too. Old SS who's lost a step, plus worst D-line in years plus retiring and injured MLB's = team that's going to overcommit to stop the run and thus get killed on the play action pass. I'm optimistic 2013 will be better...

2
by speedegg :: Tue, 06/25/2013 - 4:15pm

Interesting stuff. I wonder what PA % say about bad safety play and/or bad linebacker play. Teams with bad or aging safeties are pretty high in DVOA (Dallas, New Orleans, Baltimore, Oakland, Jacksonville, etc), but Atlanta is REALLY good against PA and their safeties aren't that good, so it could also be scheme related. Buffalo has a good safety in Jairus Byrd so that might be part of the reason they're so good against PA.

Also, seems like teams will address their safety issue. SS Roman Harper was a liability in coverage for the Saints, so the Saints drafted Kenny Vaccaro. Tampa Bay drafted mark Barron, but they are one of the bad teams, so maybe it might take a season or two to improve.

11
by bravehoptoad :: Wed, 06/26/2013 - 12:07pm

It does seem like there are some really good run defenses down on the bottom third of that chart. Maybe that's partly why they weren't so good against play action? More likely to sell out against the run?

17
by speedegg :: Thu, 06/27/2013 - 1:31pm

That partially goes with scheme. Coach Carroll said at a Nike coaches clinic that in his 4-3 Under defense one of biggest weakness is the MLB against play action. It's a 1-gap attacking defense, so everyone needs to hit their gaps when the ball is snapped. I think he wrote that it's "reading on the run", when the O-line and HB run the linebackers step up and attack their gaps.

Also, play action is different from RUN action. For a run, the O-line drives into the D-line to push them back, for a pass the O-line takes a step back and tries to form a pocket around the QB. So defenders are taught to read "low hat" (O-line keeps helmets down and drive into D-line) or "high hat" (O-line stands up to...well, take a hit from the D-line driving into them). Some teams O-lines got lazy in their run action (cough! Eagles! cough!) so the opposing defense weren't fooled.

It also depends on the defensive play call. Teams will call quarters to defend the run and the deep pass, because the safeties have run/pass responsibilities and you have 4 deep defenders. Greg Cosell, senior producer at NFL Films, went on Eagles TV to explain the Tampa Bay play action passing attack and also explained quarters coverage: Breaking Down Tampa Bay's passing attack

3
by Peregrine :: Tue, 06/25/2013 - 4:48pm

Huh. The Falcons had the best pass defense against play action? Learn something new every day.

And I'd say that the starting safeties, William Moore and Thomas DeCoud, are a relative strong suit for the defense. In fact, the safeties are the only defensive unit that isn't a glaring question mark going into 2013.

16
by speedegg :: Wed, 06/26/2013 - 8:01pm

I would say it depends on how you define "relative strong". DC Mike Nolan does things to minimize their liability in deep coverage like blitzing them, have them occupy underneath zones, or have other defenders drop into deep coverage.

Given that Atlanta went mostly defense in the draft (CB, CB, DE, TE, Safety, etc) that kind of tells you what they think and where Nolan wants to go with scheme.

6
by Robert Ferringo (not verified) :: Tue, 06/25/2013 - 8:08pm

Is there a way to correlate the defenses against the expected value of their opponents' play-action offense?

For example, I see that Buffalo has the top play action defense. It must've helped playing the Jets and Dolphins twice apiece, given that by FO's rankings they owned two of the worst play-action offenses.

So it's kind of a chicken-and-the-egg situation. Were the Bills good against play-action because of something about their defense, or did they have good play-action numbers because the offenses they faced sucked at it?

7
by Scott C :: Tue, 06/25/2013 - 11:58pm

There are statistical tools to analyze such entanglement, but I haven't seen those more advanced tools mentioned at FO.

8
by Mike Ridley :: Wed, 06/26/2013 - 9:48am

Miami's DVOA was actually the third highest on PA. However, they ran it only 16% of the time. Looking at traditional (or outdated metrics) such as QB rating, Tannehill had the largest increase between PA QB rating and non-PA QB rating, so I don't think that playing the Dolphins twice had anything to do with the Bills exceptional performance.

14
by Illuminatus! (not verified) :: Wed, 06/26/2013 - 2:33pm

Buffalo doesn't have the best PA defense in the league, it just has the widest spread between it's PA defense and it's regular defense.

9
by Dusty (not verified) :: Wed, 06/26/2013 - 11:51am

It is not suprising that Oakland, Dallas, and Cincinnati were the 3 most targeted teams with play action. Each of those teams has a reputation for being undisciplined. Play action is the best way to expose that.

12
by Karl Cuba :: Wed, 06/26/2013 - 12:17pm

Looking at both years' data it becomes clear that the 49ers suck at defending play action, which is the opposite of what I expected. The safeties and linebackers must be flying downhill at the first whiff of a handoff.

18
by LionInAZ :: Wed, 07/03/2013 - 10:02pm

Similar reaction for me on the Lions -- suck on play-action offense, yet pretty good on defending against it. Must be a coaching issue.

19
by Freddy (not verified) :: Thu, 10/17/2013 - 12:11pm

There are statistical tools to analyze such entanglement, but I haven't seen those more advanced tools mentioned at FO.