Just in case Sunday night did not make it clear, defense rules the NFL in 2016. The best teams in overall DVOA all combine top-five defenses with below-average offenses. The Eagles now lead the league in both defense and special teams and are back to No. 1 overall.
09 Jan 2009
by Aaron Schatz
There aren't as many matchups in the NFL with as much history as the New York Giants and the Philadelphia Eagles. Chuck Bednarik jacking up Frank Gifford. The Miracle of the Meadowlands. Randall Cunningham's 91-yard punt. Plaxico's overtime touchdown off a Philadelphia big blitz. Winston Justice folding against the Giants pass rush on national television.
There aren't as many matchups in the NFL with as little history as the Carolina Panthers and the Arizona Cardinals. Well, unless your name is Steve Smith and the topic of today's history lesson is "65-yard touchdowns."
For those who may be unfamiliar with the Football Outsiders stats, they are explained at the bottom of the page. Scroll down or click this link.
This year, we're going back to the old school for our in-game discussions. You can use these preview threads to discuss things before and then during each game. Just remember to switch over from NFC to AFC when the NFC game is over.
If you have FO Premium, you can click here to see all the matchup of DVOA splits for this game.
Back in late October, the Cardinals went to Charlotte and took a 17-3 lead over the Carolina Panthers. That was midway through the third quarter; the Panthers came back for a 27-23 victory and that was the beginning of Arizona's second-half collapse. They ended up in the playoffs anyway, and they've reversed that collapse a bit the past two weeks. Is it enough to win a rematch against a rested Panthers team that holds a number of matchup advantages over Arizona?
We know that the Cardinals can pass in pretty much any situation, to both sides of the field. So what are the strengths and weaknesses of the Panthers defense?
Arizona gained the same average yardage on passes to the left and right, and threw to each side roughly the same percentage of the time. However, the Panthers gave up more yards per play on the left (7.3) than the right (6.0) even though opponents threw far more often to the right (45 percent of passes) than they did to the left (31 percent of passes). Chris Gamble is normally on the offense's right, so perhaps opponents did not get the memo that he improved significantly this season. Arizona certainly didn't get it -- in Week 8, we have 12 passes listed with Gamble in coverage and only three with Lucas in coverage. For the season, game charting stats show Gamble allowing just 6.0 yards per pass, as opposed to Lucas allowing 8.6 yards per pass. (Of course, Gamble also did better than Lucas in the charting stats in 2007, so perhaps someone should have distributed the memo in the offseason.)
The Panthers defense also excels against short passes in the middle of the field, home of Pro Bowl linebacker Jon Beason. The Panthers allowed just 5.8 yards per pass on short middle passes (fourth in the NFL) with a DVOA of -12.1% (third in the NFL). This is where the Cardinals will really miss Anquan Boldin if he can't go, and even if he isn't fully healthy. Boldin caught 79 percent of intended passes in the short middle and gained 10.2 yards per pass. The rest of the Arizona offense combined for a catch rate of just 59 percent and an average of only 4.7 yards on short middle passes.
The Panthers had a strong pass rush this year, but the Cardinals had excellent pass blocking, and both units got stronger in the second half of the season. Since Week 9, Arizona had the fifth-best Adjusted Sack Rate on offense, while Carolina had the third-best Adjusted Sack Rate on defense. Carolina blitzed at a rate pretty close to league average. They use a lot of zone blitzes -- only Baltimore and Philadelphia used the strategy more frequently -- and they use them well, allowing a yard and a half less per play with a zone blitz. Kurt Warner had some trouble with the zone blitz, two yards less per pass than he gained otherwise, and generally responded to zone blitzes with little dumpoffs. Surprisingly, Tyler Brayton seems to be the lineman dropped into coverage as often as Julius Peppers, if not more often.
We thought the run was an afterthought for the Cardinals -- as I noted last week, this is a team that passed over half the time even when it had a lead of more than a touchdown in the second half -- but Ken Whisenhunt apparently rediscovered this part of the playbook while preparing for the Falcons. Like their division rivals in Atlanta, the Panthers actually had a poor rush defense, ranking 24th in both DVOA and Adjusted Line Yards. However, you wouldn't know it if all you watched was the first game between Arizona and Carolina. Anquan Boldin took an end around for 30 yards, but otherwise, the Panthers held the Cardinals to 20 yards on 13 carries.
Unfortunately for the Cardinals, the relative strength of their poor running game is the same as the relative strength of the poor Panthers run defense. Arizona is above-average in Adjusted Line Yards in two directions: right end and right tackle. Carolina's defense happens to be above-average in the same two directions. Both Arizona's offense and Carolina's defense rank 31st in Adjusted Line Yards on runs up the middle. It isn't all bad news -- there are two somewhat positive things to say about the Arizona running game:
This is where Carolina has the advantage on Arizona. The Panthers offense was far better than the Cardinals defense this year, particularly in the second half of the season.
It starts with the running game, of course. The Panthers ran on 52 percent of plays, more than any team except Baltimore and Atlanta. They pulled a rare double, finishing first in the league running in power situations, but also first in the league in the percentage of their rushing yards that came more than 10 yards past the line of scrimmage. No other team since at least 1997 has led the league in both categories, on either offense or defense. DeAngelo Williams averaged at least 5.2 yards per carry in all five of the directions we track. Jonathan Stewart wasn't quite that good but he averaged at least 4.3 yards per carry in every direction except left end.
Arizona's run defense was about league average, and it will certainly have its hands full with Williams and Stewart. The Cardinals actually ranked first in the league in ALY against runs right tackle, but when you look at their other numbers -- 29th against runs up the middle, 19th against runs right end -- it seems pretty clear this is a statistical fluke.
The passing game in Carolina is pretty much all about Steve Smith. As noted last week, the Cardinals rank 29th in DVOA against number one receivers, but opponents actually threw to their top receivers less often than the NFL average. That won't happen with the Panthers, because no team threw more often to its number one receiver.
Roderick Hood is one of those players that Football Outsiders constantly promotes as underrated, but he has a poor record against Smith. Smith doesn't stay on one side or the other, so when these teams first met, the Cardinals specifically moved Hood around to cover him. Our game charters marked Hood as the defender on all eight passes to Smith that day. Smith dominated Hood with three first downs and two touchdowns, including one for 65 yards. If that sounds familiar, it's because Smith also burned Hood for a 65-yard touchdown when these two teams met in 2007. It will be interesting to see if the Cardinals decide to stick Hood on Smith all the time, or switch to Rodgers-Cromartie, or just have the two cornerbacks stay on their usual sides and let each one cover Smith part of the time.
The Cardinals switch between a 4-3 and 3-4 alignment, and often rush their outside linebackers. Only three teams rushed five defenders more often. That's a bit of a problem against Carolina, which was one of the few offenses that wasn't hurt by blitzes. Carolina averaged the same 6.7 yards per play against five pass rushers that it did against four pass rushers, and it was even better -- 9.2 yards per play, fourth in the NFL -- against a big blitz of six or more defenders. In Week 8, the Panthers didn't have a lot of success against the blitz, but the success they did have was huge -- that 65-yard touchdown pass to Smith came when Arizona sent six on third-and-2.
Just in case we haven't brought out enough statistics to show how tough this matchup is for the Arizona defense, here are some more:
If this game stays close, don't be surprised to see the Panthers come from behind. The Panthers' offense ranks second in DVOA when tied or losing, but 14th when winning by a touchdown or less, and 24th when winning by more than a touchdown. Arizona's defense is above-average when the Cardinals are tied or losing, but 24th when the Cardinals are winning by a touchdown or less, and the worst in the league with a lead of more than a touchdown.
Carolina has a big advantage here in two areas: kickoffs and punt returns. The Panthers were the only team in the NFL this year to employ a kickoff specialist, and Rhys Lloyd delivered with an NFL-leading 30 touchbacks. Nobody else had more than 22. Even when the other team managed to return a kickoff, the Panthers allowed an average of just 22.5 yards, eighth in the league.
Carolina's Mark Jones was a consistent return man, getting yardage on nearly every return and never muffing or fumbling on either a punt or kickoff. He was returning punts than kickoffs, but the Cardinals were lousy preventing returns on both. As noted last week, despite reasonable kicking from Neil Rackers and Ben Graham, the Cardinals gave up more value on kickoff returns than any team except Kansas City, and more value on punt returns than any teams except Minnesota and Washington. Combined, the Arizona coverage teams gave up an estimated 22 points worth of field position compared to the NFL average.
Looking through the matchups in this game, I think the Cardinals were lucky to come within four points back in Week 8 -- and that was before the Cardinals collapsed in pretty much every way over the second half of the season. Upsets do happen -- after all, I thought Atlanta was going to beat the Cardinals last week -- but Carolina is the easiest favorite to pick this weekend.
If you have FO Premium, you can click here to see all the matchup of DVOA splits for this game.
The Giants have really played two seasons. Since mid-November, they've been without star receiver Plaxico Burress, and they've mostly been without starting running back Brandon Jacobs. The Giants have declined without those players, but it may surprise you to learn which side of the ball the decline is on. The problem in New York this December has been defense, not offense.
New York led the league with 26.2% offensive DVOA when Burress went out in Week 12 (he was injured and only thrown one pass, then shot himself the following week). Over the past six games, the Giants have an offensive DVOA of 18.8% -- that ranks eighth because of a few offenses that got really hot at the end of the year, but it really isn't that big a decline.
On the other hand, the Giants defense has dropped from -11.2% DVOA through Week 11 -- seventh in the NFL, and remember, that period includes their Monday night meltdown against Cleveland -- to 6.8% DVOA over the past six games, 15th in the NFL.
I have to admit that before I began working on this preview, I had no idea that the Giants' defense had so many problems over the past month and a half. As far as I can tell, nobody is talking about this, and there weren't any particularly notable injuries. (Sammy Knight went on IR, but come on, how much did this team depend on a part-time veteran safety?) Let's try process of elimination to see what the heck is going on.
Splitting things up into passing and rushing makes that pretty obvious. Since Week 12, the Giants run defense has improved slightly, allowing .05 fewer yards per carry and going from -5.5% DVOA to -10.9% DVOA. On the other hand, the pass defense has gone from -15.7% DVOA (seventh in the NFL for Weeks 1-11) to 22.1% DVOA (23rd in the NFL for Weeks 12-17).
Next question: Is it the pass rush or the coverage? It seems to be both. New York's Adjusted Sack Rate was 8.1 percent in Weeks 1-11, but 6.1 percent since then. That's an issue, but clearly that's not solely responsible for the big decline in the pass defense. Justin Tuck has suffered from the flu, but he still has 3.5 sacks in the past six games, plus a tackle of Donovan McNabb on a "run for -4 yards." Count that as a sack, and the numbers aren't much different from his 8.5 sacks in the first 10 games. The bigger issue is the tackles -- for example, Fred Robbins has not had a sack since Week 7.
Is the pass coverage suffering on any specific down? Actually, it is two specific downs.
|New York Giants Pass Defense by Down, 2008|
The Giants are giving up more yardage on second down, but that's against a harder schedule these last few weeks, so the DVOA evens out. The same is not the case on first down or third down, where the Giants have declined significantly even after taking into account their harder schedule.
Is it directional, or related to a single cornerback? I don't think so. Since Week 12, the Giants are giving up more yards per pass in all six directions (six = three directions, split into short and deep passes) except for deep middle, and a higher completion rate in each direction except for deep middle and short right (so huzzah for Danny Clark, I guess).
The problems with pass defense make even less sense when you consider that New York's cornerback seems to actually be playing better since midseason. Many top-drafted cornerbacks don't fully mature until their third or fourth seasons in the league, and that seems to have been the case with New York's Corey Webster. Once thought a bust, Webster came on in last year's playoffs. This year, in his fourth season, the Football Outsiders game charting project says he's been one of the top corners in the league. His 63 percent Success Rate ranks sixth among cornerbacks with at least 40 charted passes, and his 5.2 yards allowed per pass ranks third. Webster also tied for the league lead with 24 passes defensed. Since Week 12, Webster is listed with a miniscule 3.9 yards allowed per pass and a 68 percent Success Rate. He's certainly not the problem.
New York's other starting corner is Aaron Ross, who struggled in man coverage in his second season (8.2 yards allowed per pass, 53 percent Success Rate). Those numbers were a bit worse after Week 12, but if Ross was the problem, he would have been a big problem in the first half of the season too. Ross missed the final game of the regular season with a concussion but will return against Philadelphia. If Giants fans looked at the charting numbers, they might want Ross to stay out. Nickelback Kevin Dockery had phenomenal numbers this year, even better than Webster -- 4.4 yards per pass and a 67 percent Success Rate. Of course, two years ago, Dockery came out as the worst corner in the game by charting stats, and he was awful last year too, so there's clearly some nickelback sample size tomfoolery going on here. Adding adjustments for opponent strength will fix this somewhat in the offseason -- look through the charting stats and you see that Dockery spent most of his time covering guys like Josh Morgan, James Thrash, and Antonio Chatman.
There's one more thing left to try. Let's look at the Giants' pass defense by "type of receiver."
|New York Giants Pass Defense DVOA by Receiver, 2008 (with Rank)|
|vs. #1 WR||vs. #2 WR||vs. Other WR||vs. TE||vs. RB|
Ah. I think we may have found the problem. The issue is less about the cornerbacks and more about linebackers, holes in zones, and running backs left uncovered in the flat.
Number one receivers are gaining a half yard more per pass against the Giants since Week 12, but the struggles against number one receivers is more about turnovers. The Giants had five interceptions on passes to number ones in the first 10 games, but none since. Actually, this is an overall problem. The Giants had 14 interceptions and one DPI in the first ten games. In the last six games, they have two interceptions and four DPIs.
Once you get past the starting wideouts, everyone has been catching more passes and gaining more yards. "Other wide receivers" have gone from 6.2 to 8.5 yards per pass. Tight ends have gone from 5.8 to 7.0 yards per pass. Running backs have gone from 5.1 to 7.4 yards per pass. Quarterbacks have gone from completing 55 percent of passes to players at these three positions in Weeks 1-11 to 72 percent since Week 12.
These are not good weaknesses to have against the current Eagles offense. You know about Brian Westbrook, of course -- he combined for nine catches, 105 receiving yards, and a touchdown in two games against the Giants this year. The Eagles love to use multiple-receiver sets, and only three teams threw more often to "other wide receivers." At tight end, drop machine L.J. Smith will return from injury for Philadelphia but has lost his job to a much better receiver, Brent Celek. Smith this season had a -15.4% DVOA, 35th among tight ends, with a catch rate of 58 percent. Celek had 21.9% DVOA, eighth among tight ends, with a catch rate of 71 percent -- but barely played in the two regular-season games against New York, with only one pass thrown his way.
This "When the Eagles Have the Ball" section has primarily been about the Giants, but maybe we should say something about Eagles quarterback Donovan McNabb. How about this: He's been one of the top five quarterbacks in the NFL for 15 out of 17 weeks this season. Last week's preview had a table showing how Philadelphia's decline during the Week 11-12 slump was almost all passing. That also means that McNabb, who finished 13th in DVOA among quarterbacks, would rank fifth if we did not include those two games (actually, in his case, one and a half games).
What about the run? The Eagles have been doing it a little more often, particularly in short-yardage situations, which has helped solve their problems on third-and-short. Given Brian Westbrook's injuries, the Eagles may want to use a little bit more Correll Buckhalter in this game. In the first two games between these teams, Buckhalter got only two carries -- but over the course of the season he had a slightly higher DVOA and a slightly higher Success Rate than Westbrook did. Apparently he will play some fullback this week, so we'll have to see how the Eagles use him and whether that includes some carries.
One run the Giants should have no problem stopping is the draw. You might think of Philadelphia as a good draw team, but they actually gained a below-average 4.8 yards on draws. The Giants gave up a lot of yards to draws in 2007, but this season they actually had the best defense against draws according to the game charters, allowing just 2.8 yards per draw play. Eight of the 26 draws charted against the Giants were stopped for a loss or no gain.
OK, so, I pointed out that the real problem during New York's season-ending 3-3 stretch was the defense, not the offense. However, the offense has had some problems too. There are two of them in particular.
First, it is really obvious that Brandon Jacobs has not been 100 percent over the last few weeks. Through the first ten games, Jacobs had 5.4 yards per carry and a 54 percent Success Rate. He played hurt in three of the final six games, and in those games he averaged just 3.8 yards per carry with a 44 percent Success Rate. The problem is Jacobs, not the offensive line, because Derrick Ward's Success Rate stayed the same after Week 12 while his yards per carry went up. If Jacobs is truly returning to the lineup healthy this week, it gives the Giants a huge boost. In the first Eagles game, Jacobs had 126 yards on 22 carries. In the second Eagles game, he had 52 yards on 10 carries, but 23 yards came on one run and his Success Rate that day was just 40 percent.
Doug Farrar has done a good job of explaining how the Eagles' run defense improved significantly this season, but they were not quite as strong on the inside compared to the outside. Philadelphia ranked sixth in Adjusted Line Yards overall, but the Eagles were 17th against runs up the middle. The Giants led the league in Adjusted Line Yards on runs up the middle.
Second, the Giants have had trouble throwing deep passes (16+ yards through the air) without Plaxico Burress. Through Week 11, the Giants completed 45 percent of deep passes and gained 11.5 yards per pass attempt (including pass interference gains). Since Week 12, they completed just 34 percent of deep passes with 8.7 yards per pass attempt. Domenik Hixon has caught just four of 16 intended deep passes since Week 12; Amani Toomer has caught just two of ten (and drawn a DPI flag).
(For those curious, the Giants attempted roughly six and a half deep passes per game both before and after Burress' departure.)
One thing that hasn't changed since Burress left the lineup is Steve Smith's place as the third-down machine in the slot. Remember above when I said only three teams threw more passes to their "other" wide receivers than the Eagles did? The Giants were one of those teams -- in fact, they were number one. Ironically, Smith has a DVOA rating of just 5.6% on third downs, and has a higher catch rate on both first and second down than he does on third down.
Philadelphia's best weapon against the Giants is the fact that Eli Manning struggles against the blitz. Only four teams send five or more pass rushers more often than the Eagles, and no team sends six or more pass rushers more often. The Eagles do a great job of stopping the offense when they blitz, and Manning's numbers get worse with more pass rushers (although he did have a couple of big gains on the few plays that opponents sent seven defenders).
|Giants Offense and Eagles Defense by Number of Pass Rushers, 2008|
|Yards Per Pass (or Scramble)||Frequency of Pass Plays|
|Rush 4||Rush 5||Rush 6+||Rush 4||Rush 5||Rush 6+|
In the first two games between these teams, our charters marked nine plays where the Eagles rushed six or more against Manning. One of those plays came on the goal line, and Manning hit Kevin Boss for a touchdown. The other eight passes were incomplete.
Another example of how well the Philadelphia blitz packages work: In the Football Outsiders game charting project, we mark sacks in four categories: overall pressure, rusher untouched, blown block, or quarterback fault (for example, the quarterback slipping on the grass on his own). Philadelphia leads the NFL in sacks marked as "overall pressure," and only Buffalo had more sacks marked as "rusher untouched."
However, Manning isn't necessarily weak against all of Philadelphia's strategies. Philadelphia uses zone blitzes more often than any other defense, and Eli Manning was roughly league average against the zone blitz (6.3 yards per play without zone blitz, 5.6 yards per play against it).
The Philadelphia pass rush probably deserves some of the credit for the great charting numbers put up by the Philadelphia cornerbacks. After adding another couple of weeks of data to game charting, Sheldon Brown still has the best Success Rate of any starting cornerback, 73 percent. Asante Samuel allowed 6.2 yards per pass with a 56 percent Success Rate, roughly the same as his numbers in New England a year ago. The Eagles started out the year with Lito Sheppard at nickel, and he struggled again -- 10.8 yards per pass and 50 percent Success Rate -- so he was replaced by Joselio Hanson. It was a good move, as Hanson allowed 5.4 yards per pass with a 66 percent Success Rate, but it doesn't say anything about this year's two Eagles-Giants games, because the change was made before the first meeting in Week 10.
Overall, both New York and Philadelphia were a little bit above average on special teams this year, although the Giants have one specific weakness: kickoffs. This is why John Carney's selection for the Pro Bowl was so ludicrous: While Carney was steady kicking field goals, he's too old to do the other half of his job, and it created field position problems for the Giants. Carney's average of 60.4 yards per kickoff was the second-lowest in the NFL, and he had a grand total of three touchbacks; by comparison, Philly's David Akers had 17. New York had average kickoff coverage, which was the only reason they didn't finish last in the Football Outsiders kickoff numbers. It just so happens that kickoff returns were the strongest part of Philadelphia's special teams, so the most likely player to turn this game on special teams will be rookie Philadelphia return man Quintin Demps. During the regular season, his average return against the Giants ended at the Philadelphia 30-yard line. Unfortunately, one of those returns also ended up in the hands of the Giants after a fumble. He'll try not to do that this time.
I'm not sure what has happened to the Giants pass defense, and their weaknesses over the past six weeks are in the right places for Philadelphia. However, the Giants also have the advantages of home field, a week of rest, and a healthy Jacobs. Mike Tanier, with his usual dose of Eagles fan pessimism, picked New York in Walkthrough. Our premium picks equation, with the usual DVOA Eagles optimism, picked Philadelphia. I fall between the two and say this game is about as close to a pick 'em as we can get.
DVOA (Defense-adjusted Value Over Average) breaks down each play of the season and compares it to the NFL average based on situation and opponent. You'll find it explained further here. Since DVOA measures ability to score, a negative DVOA indicates a better defense and worse offense, and a positive DVOA indicates a better offense and worse defense.
Each team is listed with DVOA for offense and defense, total along with rush and pass, and rank among the 32 teams in parentheses. (If the DVOA values are difficult to understand, it is easy to just look at the ranks.) We also list red zone DVOA and WEIGHTED DVOA (WEI DVOA), which is based on a formula which drops the value of games early in the season to get a better idea of how teams are playing now (explained here). These numbers are all regular season only, except for WEIGHTED DVOA which includes the first round of the playoffs.
SPECIAL TEAMS numbers are different; they represent value in points of extra field position gained compared to NFL average. Field goal rating represents points scored compared to average kicker at same distances. All special teams numbers are adjusted by weather and altitude; the total is then translated into DVOA so it can be compared to offense and defense. Those numbers are explained here.
Each team also gets a chart showing their performance this year, game-by-game, according to total DVOA. In addition to a line showing each game, another line shows the team's trend for the season, using a third-power polynomial trendline. That's fancy talk for "the curve shifts direction once or twice." Note that even though the chart appears in the section for when each team has the ball, it represents total performance, not just offense.
100 comments, Last at 12 Jan 2009, 12:46am by DVOA Fan