Possibly the closest Super Bowl matchup in history also poses the question: how much does it mean when certain aspects of an NFL team improve dramatically in the second half of the season?
08 Jan 2010
by Aaron Schatz
This year's NFC playoff field is extremely balanced, and any of the six teams makes sense as NFC Champion. The teams that were hottest in the first half of the year, New Orleans and Minnesota, seem to have cooled down. The wild card teams have been playing extremely well in recent weeks, or at least they were until Philadelphia laid an egg in Dallas last week. The Cowboys have finally gotten over their December issues of recent years, and while the Cardinals are statistically the worst of the six teams, we know from last year that they can turn it on and be as good as any other team in the league over a month-long period.
So what factor might end up being the deciding factor in the NFC this year? It could be consistency. Arizona made it to the Super Bowl last year partly because Carolina performed a complete face-plant in the second round. Most of the teams in the NFC bracket have had wild swings over the course of the past season, but one team -- and it's not who you might think -- has been markedly variance-free.
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.
All readers can click here for in-game discussion on our message boards. If you have FO Premium, you can click here to see all the matchup of DVOA splits for this game.
For the Philadelphia-Dallas game, we have a special present for readers. Since the first half of the first Philadelphia-Dallas game was absent from our game charting, I prepared for this preview by charting the first halves of both games, one after the other. To share with everyone a sample of the kind of data we get from charting, we're going to make these spreadsheets available for download absolutely free. Click here to get both first halves of charting, and feel free to make your own observations in the comments.
One fact I'm not sure where to fit in: It is interesting to note that both the Eagles and Cowboys were worse in the red zone than they were overall, on both sides of the ball.
First of all, I just have to say... Wheeeeeee! Look at that graph. Man, is that a roller coaster ride, or what? Now you see what it is like to be Mike Tanier.
As for when the Eagles have the ball, I don't know which would surprise fans more -- that DVOA ranks the Eagles offense only 10th in the league, or that DVOA ranks the Cowboys defense only 10th in the league. However, the difference between them is that the Cowboys defense has been improving in recent weeks. Through Week 9, Dallas ranked 23rd in pass defense DVOA and 15th in run defense DVOA. Since Week 10, the Cowboys rank ninth in both categories.
Dallas has the league's biggest gap between defense against shotgun formations (6.4 yards per play, 18th in DVOA) and defense against quarterbacks under center (4.3 yards per play, fourth in DVOA). This could be a problem since Philadelphia uses shotgun on an NFC-leading 47 percent of plays. However, the Cowboys defense went against this trend in the two regular-season meetings with Philadelphia, allowing an average of 5.1 yards per play with Donovan McNabb under center but just 4.7 yards per play with him (or Michael Vick) in shotgun.
Still, it might make most sense to spread out the Cowboys defense by throwing to as many different receivers as possible. Based on our DVOA numbers, the Cowboys are much better covering starting wide receivers than they are covering slot receivers, tight ends, or running backs. They had a particularly hard time last week covering tight end Brent Celek. In two games against the Cowboys, DeSean Jackson and Jeremy Maclin have combined to catch just 38 percent of passes, each averaging about 40 yards per game. In particular, the Eagles haven't been hitting on those deep strikes to Jackson. In the two Dallas games, McNabb has completed only three of 16 passes to Jackson and Maclin that went 10 or more yards in the air, and the only one longer than 20 yards was a 31-yard completion to Maclin last week in the second quarter. Jackson had a 32-yard catch later on the same drive, but that catch was 8 yards in the air and 24 yards after the catch. In total, Maclin and Jackson have combined for three pass plays longer than 15 yards in two games against the Cowboys. Celek had four pass plays longer than 15 yards just in last week alone, and all the other Eagles have combined for eight pass plays longer than 15 yards in the the two Cowboys games.
If the Eagles are going to send everyone out into patterns, that means they'll have to depend on their five linemen to protect against the Dallas pass rush. Left tackle Jason Peters will be very important, as he'll generally need to take on DeMarcus Ware one-on-one. Peters was injured and missed a few plays in the Week 9 game, and when Todd Herremans moved over to left tackle, he had major problems with Ware. The Eagles helped both Peters and Herremans somewhat by lining up two tight ends on the right side and leaving one in to block, with everyone shifting left and the left guard helping the left tackle. However, that second tight end, Alex Smith, is a receiving-first tight end just like Celek. The Eagles really could have used a blocking-first backup tight end this season.
While the Eagles line has done a good job protecting McNabb, he tends to eat the ball when pressured. Philadelphia ranked second in fewest quarterback hits but 20th in Adjusted Sack Rate. Including plays cancelled by penalty, the average NFL quarterback this season had 10 more quarterback hits than sacks. McNabb took 37 sacks but only 19 quarterback hits after passes. The Cowboys, meanwhile, didn't have quite the pass rush that they had in 2008, dropping from first in Adjusted Sack Rate to 12th.
The Eagles have a very efficient running game, when they feel like using it. Of course, since they're the Eagles, that's not very often. Philadelphia was dead last in the NFL with just 301 running back carries. If the Eagles do want to run, the weakness of the Cowboys run defense is on the offensive right, at least according to our Adjusted Line Yards stats.
There may be a lot of talk in the pregame about how Brian Westbrook has long been the focal point of the Philadelphia offense, but it's hard to tell if the Eagles have really gotten anything out of getting him back from knee injuries and post-concussion syndrome for the last two weeks of the season. Before he missed eight straight games, Westbrook averaged 4.8 yards per carry and 6.2 yards per pass. In the two games since his return, he's averaging 3.5 yards per carry and 2.8 yards per pass. He isn't bringing much to the table that LeSean McCoy doesn't bring as well, and unlike Leonard Weaver he's not a complimentary piece who is unique on the Eagles roster.
The most remarkable part of last week's Dallas victory was the way the Eagles completely abandoned their usual blitz-heavy defensive style. In plays we've charted through Week 13, the Eagles sent a big blitz of six or more pass rushers more often than any other team (23 percent of the time) and had a play marked as a zone blitz more often than any team except the Jets (13 percent of the time). In the first game with Dallas, the Eagles sent a big blitz nearly half the time and five or more 68 percent of the time, with six plays marked as zone blitzes. But when I charted the first half of last weekend's game, I marked only one play as a zone blitz and only two plays with more than the conventional four pass rushers, one with five and one with seven.
Why did the Eagles change their strategy? It turns out that in the first game with Dallas, their best defensive plays were the ones where they sent fewer pass rushers. In Week 9, the Eagles allowed only 1.7 yards per play on the 12 plays where they sent just three or four pass rushers. They allowed 6.0 yards per play on the 18 plays where they sent six or more pass rushers. But worst of all, they allowed 19.3 yards per play on eight plays with five pass rushers, including two conversions on third-and-14 and a huge 64-yard pass to Patrick Crayton that found a hole in zone coverage.
My guess is that the Eagles saw this on film, and decided that the best way to beat the Cowboys was to send a conventional pass rush and hang back in coverage against Tony Romo. It completely backfired. Romo took advantage of holes in zone coverage, Asante Samuel misjudging his attempts to jump routes, and cornerbacks playing back and leaving room for wide receiver screens.
The best strategy for Philadelphia is probably to rotate standard rushes with big blitzes. Send four every time, and a quarterback knows what to expect. In addition, Romo was at his best when facing five pass rushers -- both in the first game between these teams and for the season as a whole.
Miles Austin has enjoyed a breakout season this year as the Cowboys' number one receiver, and it is interesting to note that while the Cowboys bounced him back and forth between the left and right sides in the first game, they utilized him almost strictly on the right side in the second game. That means Asante Samuel, not Sheldon Brown, will generally be covering Austin. Our charting numbers so far have Brown as one of the top cornerbacks in the league, with a 70 percent Success Rate, matched only by Darrelle Revis and -- get ready for a shock -- Jabari Greer of New Orleans. Samuel still has an above-average 63 percent Success Rate in plays charted so far, and he has allowed fewer yards per pass than Brown (5.6 compared to 5.9).
As Mike Tanier pointed out earlier this week, the Cowboys have started setting up their pass plays by running draws, and they did this a lot against the Eagles in Week 17. They also put themselves into advantageous situations by getting consistent yardage on first down. Dallas ranks fourth overall in offensive DVOA on first down -- including first running the ball -- while the Philadelphia defense is much worse on first down (17th in DVOA) than it is on second (first) or third (fourth). The Cowboys also had the lowest three-and-out rate in the league, going three-and-out on just 16.2 percent of drives.
Fullback Deon Anderson and tight end John Phillips could play a big role for Dallas in this game. Most teams gain more yardage on runs with just one back instead of two (partly because teams tend to go single-back in longer-yardage situations). However, the Cowboys average a yard more on runs with two backs, and the Eagles are one of just eight defenses to give up more yardage on runs with two backs (3.9) than on runs from single-back sets (3.6). Calling Phillips a tight end seems a bit silly, as I've never seen him actually line up in a tight end position anywhere other than the goal line. The Cowboys seem to have a very strict fullback rotation, where Anderson plays fullback when there are two wide receivers on the field and Phillips plays fullback in a three-tight end personnel set that also features Jason Witten and Martellus Bennett.
Philadelphia had a good year on special teams overall; they were second in the FO ratings, although Cleveland was so far ahead that the Eagles finished closer to 18th place than they did to first. The Eagles are a good example of why it is a bit silly for the NFL to name one combined punt and kickoff returner for the Pro Bowl -- these are two different jobs with different sets of skills, usually done by different players. The Eagles were first in the league on punt returns, thanks to Pro Bowl return man DeSean Jackson, but 27th in the league on kickoff returns, where the job rotated between Jeremy Maclin, Macho Harris, and the now-injured Ellis Hobbs (who was the best of the three).
The Cowboys had very specific strengths and weaknesses on special teams. They ranked third in net kickoff value, thanks in large part to rookie kickoff specialist David Buehler, who led the league with 29 touchbacks. The Cowboys also have a strong punt returner in Patrick Crayton. The biggest weakness of the Cowboys' special teams was placekicker Nick Folk, but that area hasn't been a problem since the Cowboys ditched Folk and signed Shaun Suisham. The Redskins really overreacted to Suisham going through a small midseason slump, and our numbers combining his time in both Washington and Dallas rate him as league-average this season. It's not like the Cowboys want to be stuck running him out there for a 52-yarder, down by three in the final minutes, but he's not really a problem either.
Fans look at the Cowboys over the past few Decembers, they look at their recent playoff record, and they look at the poorly-regarded Wade Phillips on the sideline, and they think the issue in this game is whether the Cowboys can play consistently good football. But based on what we've seen in 2009, the question isn't which Cowboys team will show up. It's which Eagles team will show up. Take a look at the Dallas week-to-week graph; the Cowboys have been extremely consistent this year. They're third in variance, and rank in the top ten for both offense and defense separately. The Eagles, on the other hand, rank 25th in variance. You saw the reasons last week when these teams played for the NFC East title. Nobody in the NFL is better at studying quarterback tendencies and jumping routes than Asante Samuel, but if he misses, you've got problems. Twice last week, Samuel tried to jump routes and gave up big plays, a touchdown pass to Jason Witten and a 40-yard pass to Miles Austin. Philadelphia's zone blitzes are often hit-or-miss -- if the pass rush doesn't get to Romo in time, you end up with something ridiculous like Jason Babin trying to cover Miles Austin in the middle of the field (see: 11-yard gain with 3:28 left in the first quarter). We know McNabb's accuracy seems to come and go, and those deep bombs to a streaking DeSean Jackson are hit-and-miss. If McNabb hits one of those last week, if he doesn't drop a snap in the red zone, if one of those Samuel route-jumps gets a pick, it's a different game. This week, it probably will be. That different game will be a closer game -- but the Cowboys are still the favorite, because even when we ignore the issue of consistency, they've been the slightly better team over the past few weeks.
All readers can click here for in-game discussion on our message boards. If you have FO Premium, you can click here to see all the matchup of DVOA splits for this game.
Unlike the Eagles-Cowboys matchup, looking at game charting from Week 17 isn't going to tell us much about Cardinals-Packers II. The Cardinals played their second- and third-string quarterbacks for most of the game, and used a vanilla defensive strategy because they knew they were playing Green Bay again this week and didn't want to show their hand. If you toss out Week 17, the DVOA gap between these two teams gets a little smaller. But it doesn't get that much smaller. Green Bay is still the hottest team in the NFL going into this year's playoffs, and Arizona is still the weakest team in the NFC. (They look a lot stronger than they looked a year ago, but then again, this year you can't really argue that the Cardinals were sleeping through the final month of the season -- just the final game.)
What's most surprising about Arizona, at least if you believe our DVOA stats, is that this is probably the most balanced team in the league. Unfortunately, it's extremely balanced at the level of "kinda good." The Cardinals rank between 11th and 16th in DVOA in all five phases of the game -- run offense, run defense, pass offense, pass defense, and special teams -- and taking out their Week 17 stats wouldn't change that.
Note that neither one of these teams is quite as good as it has looked in 2009. Based on average DVOA of opponent, Arizona had the easiest schedule in the NFL this year, and Green Bay was next to last.
The Packers have improved since midseason on both sides of the ball, and on offense the real stunner has been the improvement of the offensive line. Veteran right tackle Mark Tauscher missed most of the first half of the year with a knee injury, came back (probably too early) against Tampa Bay in Week 9, missed Week 10, and then came back for good in Week 11. This wasn't the only injury the Packers had early on. Left tackle Chad Clifton missed four out of five games from Week 3 to Week 8, which forced a number of other linemen to play out of position.
In Weeks 1-10, the Packers offense was 14th passing and ninth rushing. Since Week 11, the Packers rank fifth passing and first rushing. Green Bay's Adjusted Sack Rate has dropped by more than half -- Aaron Rodgers took 41 sacks in the first nine games of the year (ASR: 12.1%) but just 10 sacks in the last seven games (ASR: 4.2%). For the season, the Packers rank 30th in ASR but if we only counted the second half of the year, they would be in the top ten.
Note that there hasn't been a similar improvement in run blocking, at least according to Adjusted Line Yards; instead, Ryan Grant has had more long gains over the second half of the year. That hits a particular Cardinals weakness, as Arizona ranks 10th in ALY but 26th in Second-Level yards per carry and dead last in Open Field yards per carry. Even worse, this weakness has gotten worse since mid-year. Arizona's run defense has dropped from fourth in DVOA (Weeks 1-9) to 25th (Weeks 10-17). The Cardinals allowed just 4.0 yards per carry in their first eight games, with 1.07 Open Field Yards per carry. Since Week 10, they have allowed 5.2 yards per carry and 1.52 Open Field Yards per carry.
The Packers offense has also been stellar on third downs this season. They rank 12th in offensive DVOA on first down and 10th on second down, but first overall on third down. That's not just Rodgers, as the Packers rank second in rushing DVOA on third downs and third when running in short-yardage power situations. The Cardinals defense has also been better on third downs than overall, but not by quite the same extent (12th overall, ninth on third downs).
On defense, the difference between Arizona this season and last season is pass defense, where the Cardinals improved from 25th in DVOA to 12th. The Cardinals also got rid of one of a weakness that had plagued them for years: shotgun formations. As we noted in Football Outsiders Almanac 2009, Arizona had been one of the league's worst defenses against the shotgun for three seasons. This year, the Cardinals ranked 14th in defensive DVOA against the shotgun, the exact same rank as they had against standard formations.
Although I don't feel Dominique Rodgers-Cromartie was worthy of a Pro Bowl selection, he is clearly the better of Arizona's two starting cornerbacks. Game charting so far shows DRC with a 67 percent Success Rate and 6.1 yards allowed per pass, while Bryant McFadden has a 55 percent Success Rate and 7.0 yards allowed per pass. The cornerbacks generally play left-right instead of either one of them concentrating on the opponent's best receiver, which explains why DRC can have better charting numbers even though the Cardinals rank 26th in DVOA against number-one receivers and second in DVOA against number twos. That should mean a big game for Donald Driver (covered by McFadden on the left) instead of Greg Jennings (covered by DRC on the right). If DRC can't go because of the bruised knee suffered in last week's game, the starter opposite McFadden will probably be Ralph Brown, whose charting numbers so far feature a 57 percent Success Rate and 7.3 yards allowed per pass.
Arizona needs to watch out for the big blitzes, as Aaron Rodgers averages 8.4 yards per pass when opponents send six or more pass rushers, but just 6.3 yards per pass against four or five pass rushers. Even more impressive is Rodgers' success against zone blitzes; Rodgers has averaged 13.0 yards per pass this season on plays we have charted as zone blitzes. (Collected charting currently only goes through Week 13, so that doesn't even include any plays from his last four weeks of impressive play.)
The Packers only used play-action on 12 percent of plays, 29th in the NFL, but they may want to consider using it more against the Cardinals, who gave up 1.5 yards more pass when opponents used play action.
While the passing game improved for Arizona this year on the defensive side of the ball, the improvement on the offensive side of the ball came in the running game. The Cardinals went from a dismal 28th in rushing DVOA in 2008 to 22nd in the first half of 2009, and then 12th from Week 10 onwards. A big part of the change has been increased reliance on -- and improvement by -- rookie Beanie Wells, although Hightower has had a couple sweet highlight-length runs over the past few weeks and thus has a higher yards per carry average:
|Arizona Running Game, 2009|
|Beanie Wells Weeks 1-9||70||310||4.43||49%||-7.5%|
|Beanie Wells Weeks 10-17||108||474||4.39||56%||7.1%|
|Tim Hightower Weeks 1-9||78||277||3.55||46%||-11.4%|
|Tim Hightower Weeks 10-17||65||315||4.85||42%||-2.9%|
The Cardinals' running game is especially (and surprisingly) strong in the red zone, where it ranks first in the league. (The Cardinals are fifth passing in the red zone, and combined they are first in red-zone DVOA overall).
Running is going to be tough against the Packers, however, because the improved Packers defense has been at its best against the run. The Packers ranked third in DVOA against the run and had the league's best run defense on first downs, allowing just 3.5 yards per carry with a Running Back Success Rate (runs of at least 4 yards on first-and-10) of just 37 percent. Their run defense should also be able to counter the Cardinals in the red zone, where it ranks third in the NFL, although the Packers do have trouble stopping the pass in the red zone (22nd in DVOA).
Green Bay's run defense has improved in recent weeks, but not as much as its pass defense. Conventional (and FO) wisdom said the Packers were in trouble after they lost Aaron Kampman and Al Harris for the season after Week 11, but instead they just got better. Leaving out last week's game against Arizona, the Green Bay pass defense improved from -7.0% DVOA through Week 11 to -16.1% DVOA in Weeks 12-16.
Nobody seems to have noticed, but the Cardinals passing game slowed down significantly this year -- especially compared to last year's playoffs, but also compared to last year's regular season. The Cardinals dropped from 7.1 to 6.5 net yards per pass and from eighth to 14th in passing DVOA. Arizona's top four wide receivers all saw their stats drop this year, especially in terms of Yards After Catch. Larry Fitzgerald, for example, had basically the same catch rate and the same number of pass targets, but more than 300 fewer yards because of a drop of 1.5 YAC. The receiver whose stats dropped the least was Steve Breaston. That's good news for the Cardinals, since he will likely be forced into a larger role in the offense due to injuries suffered last weekend by Anquan Boldin.
|Arizona Regular Season Receiving Stats, 2008-2009|
Numbers from our game charting project suggest that the Packers, like the Cardinals, should avoid using too many blitzes in this game -- but for a different reason. It isn't that Kurt Warner has been particularly effective against the blitz this year (his pass yardage average is generally the same no matter how many pass rushers) but rather that the Packers defense has given up some big plays when blitzing. On charted plays, the Packers have given up an average of 8.0 yards when sending six or more, and they've given up an average of 7.2 yards when zone-blitzing. I went and checked, thinking this might be a specific issue with Tramon Williams, but it isn't -- pretty much every defensive back and linebacker in Green Bay is the listed defender on at least one big pass play against a Packers blitz.
In fact, based on FO charting stats so far, Williams has actually done a better job on coverage than Charles Woodson has. Now, understand that this doesn't adjust for which receiver each cornerback is covering, and I have my doubts about how accurate our numbers are when they are this different from conventional wisdom... but Williams comes out with 6.2 yards allowed per pass and a 62 percent Success Rate, while Woodson has 8.4 yards allowed per pass and a 52 percent Success Rate. After the Baltimore game, many observers (including me) expected opponents to start throwing deep on Williams constantly, hoping for the return of Admiral Armbar, but it actually hasn't been a problem. Williams has only been called for Defensive Pass Interference four times this season, three of those in the Baltimore game, and he hasn't drawn a single penalty since.
Our stats also suggest that the Cardinals should do more to put Kurt Warner in the shotgun. In shotgun formations, the Cardinals ranked 11th in offensive DVOA while the Packers ranked eighth on defense. In standard formations, the Cardinals ranked 22nd in offensive DVOA while the Packers ranked third on defense.
The Packers' Achilles heel comes on special teams, where they ranked dead last according to Football Outsiders metrics. Green Bay's biggest problem is punter Jeremy Kapinos, who averaged a league-low 34.1 net yards per punt. Kapinos was the only punter this year who had more touchbacks (10) than fair catches (7) and he had only four punts this year that ended inside the opponent's 10-yard line. (One of these actually came last week against Arizona, a punt Steve Breaston caught on the 1 and returned to the 2.) However, Kapinos isn't the only problem; the Packers were below-average in all five areas of special teams. Arizona had a good year on special teams, and their best area was the same as Green Bay's worst (net punting). It's not likely that this game turns into a defense-heavy field position battle, but if it does, the punting units definitely give the Cardinals an edge.
Based on statistics, this is the biggest mismatch of the first round, but we all know the trouble that the Cardinals caused me when I wrote that about them last year. So when I say that Green Bay is the most likely of the four road teams to win this weekend, I do it with a little trepidation. The Packers should be better on both sides of the ball, especially if DRC and Boldin can't play. But if the game is close, all it takes is one or two special teams miscues to blow a win and spoil FOX's dream of Favre vs. the Pack III.
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.
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 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).
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.
38 comments, Last at 11 Jan 2010, 1:59am by t.d.