Bill Connelly takes a look at what we can learn from defensive box score stats and general rates of havoc.
04 Jan 2008
by Aaron Schatz
We're going to do things a little differently this year, sorting the playoff previews by day rather than by conference, in part because I'm running a little behind on time. The Sunday previews probably will not run until Saturday morning. (I'm putting together playoff material from all over the place -- we've got the stat notebook at ESPN, the brief, more standard preview at the New York Sun, and this one which combines the two and adds additional ideas -- plus I took New Year's off for family time so I'm working off a four-day week. It's a bit hectic, and I apologize in advance if this year's previews seem a little disjointed.)
For those who may be visiting this site for the first time to read this preview, we explain our stats at the bottom of the page, or click this link. Each preview also includes a link to the open discussion thread for that game.
Remember that any stats from game charting are incomplete and fairly subjective. Also, the trendlines in the week-to-week charts are heavily influenced by the wacky Week 17 games.
Walpole! Xaverian! King Philip! It's a South Shore rumble! Only on NBC!
OK, seriously, for those of you in the other 49 states... The first game of the weekend is all about emotion. Washington is riding an emotional four-game winning streak, playing their hearts out every game in memory of their late teammate Sean Taylor. Seattle rides the emotions of their fans, who cause more opponent false starts at Qwest Field than at any other NFL stadium.
Both Seattle and Washington are known for their big-name running backs, but both teams are actually more efficient when passing the ball. Our DVOA ratings rank the Seattle offense ninth passing and 22nd rushing, while the Washington offense is 12th passing and 23rd rushing.
Clinton Portis has put up some nice rushing totals in recent weeks, but he's still averaging less than four yards per carry (with a -2.3% DVOA and 40 percent Success Rate) during the Redskins' four-game winning streak. Seahawks' star Shaun Alexander is a shell of his former self, averaging less than 3.5 yards per carry this season while his backup Maurice Morris gains 4.5 yards per carry. Alexander has a 38 percent Success Rate, 52nd among 56 ranked running backs. Behind the same offensive line, Morris ranks 10th with a 52 percent Success Rate.
Seattle ranked 30th in Adjusted Line Yards on runs up the middle, which was the strength of the Washington run defense: they were sixth preventing runs up the middle, but below-average in the four other directions we measure. Even if Alexander or Morris break through, don't expect lots of big gains. The Washington secondary knows how to tackle, and no defense allowed fewer rushing yards to come 10 or more yards past the line of scrimmage.
As for the passing game, Matt Hasselbeck will continue to be a steady above-average but sub-superstar quarterback, while (Walpole's Own) Todd Collins will try to continue his magical season for Washington. The longtime backup had not started a game in 10 years, but he's led the Redskins to all four of their recent wins following an injury to starter Jason Campbell. Collins is averaging 7.3 net yards per pass, while Campbell averaged just 5.9.
Both teams have above-average defenses, roughly equivalent with different strengths. The Seahawks depend on their front seven to stop the run and rush the passer; they rank fourth in Adjusted Line Yards, and seventh in Adjusted Sack Rate. Behind that, they play the kind of Cover-2 zone that veterans like Collins often excel at manipulating. The Redskins have much better cornerbacks who play a lot of man coverage, but are just average against the run and have one of the worst pass rushes in the league (26th in Adjusted Sack Rate).
(It's interesting to note that, in his limited time before getting injured in Week 7, Carlos Rogers was much better than Shawn Springs. Rogers had a 71 percent Success Rate and allowed just 4.1 yards per pass. Springs this season has a 56 percent Success Rate and 6.9 yards per pass.)
35 percent of passes against the Redskins defense were thrown to the middle of the field, which led the NFL. However, it is unlikely the Seahawks will try to take advantage; they threw just 19 percent of passes to the middle of the field, less than any other team in the NFC. That's strange since the Seahawks were so successful when they threw to the middle of the field (third in the NFL in DVOA).
Washington ranked just 20th in DVOA against number-one receivers, but was second against number-twos and fifth against "other receivers." This would be useful if we knew which Seattle receiver the Redskins might consider "number one," but I have no idea if that means Deion Branch or Bobby Engram.
A random, interesting bit of information: Washington's defense had the best DVOA in the NFL in the first half of games, but ranked just 25th in the second half of games.
Of the four games this weekend, this is the one most likely to give us an exciting kickoff return. The Seahawks and Redskins are both above-average on kickoffs and below-average on kickoff returns. Otherwise, special teams don't look like much of a factor.
On the surface, it sure seems like Seattle was the much better team, but our advanced stats show that's not necessarily the case. Seattle played the easiest schedule in the league according to average DVOA of all 16 opponents. Washington played the fourth-hardest schedule, and the hardest schedule of any team to make the playoffs. The difference between the two really wasn't that large, so the winner of this game comes down to execution and which you believe in more: home-field advantage or the emotional power of memorializing a fallen teammate.
Since midseason, no team has improved more than Jacksonville, and no team has declined more than Pittsburgh. The signature win in Jacksonville's recent string of success came three weeks ago, when they beat the Steelers 29-22 in Pittsburgh. Now they return to the Steel City for a rematch.
Our DVOA ratings show the path of each team's trajectory. After nine weeks, Jacksonville ranked 13th on offense and 19th on defense. Since Week 10, the Jaguars are second on offense (behind New England) and sixth on defense.
On the other hand, the Steelers ranked fourth on offense and second on defense at midseason. Since Week 10, the Steelers are 18th on offense and 14th on defense.
Pittsburgh's troubles have been a combination of sloppy defense and injuries. The defense has struggled without defensive end Aaron Smith, who tore his biceps in Week 13, and safety Troy Polamalu, who has missed four of the past six games with a knee injury but should be available for the postseason. On offense, the Steelers are now down to their third-string left tackle, Trai Essex. This is a problem because even with the regular left tackle for 12 games and the backup for four, only one offensive line was worse in Adjusted Sack Rate (San Francisco).
The general feeling among fans is that the Steelers' running game will suffer because of the loss of Willie Parker. This is true in the sense that it is good to have two backs with different skill sets. However, Najeh Davenport was far more effective than Parker this season. Davenport gained 4.7 yards per carry and his success rate of 52 percent ranked ninth among backs with at least 75 carries. Parker gained just 4.1 yards per carry and his success rate of 42 percent ranked 42nd. Davenport is a better back to use against the Jaguars in particular, because they struggle stopping runs up the middle (28th in Adjusted Line Yards).
The Jaguars are also known as a run-first offense, and their running game has been spectacular in recent weeks. Veteran Fred Taylor is averaging 6.6 yards per carry over his past seven games, and Taylor and Maurice Jones-Drew combined for 216 yards on 37 carries in the first meeting of these teams. The Jaguars are good running everywhere except, it turns out, in the red zone, where they are just average. That's a problem because Pittsburgh is the league's best defense against the run in the red zone.
If the Jaguars want to break a big run, they need to go left. The Jaguars ranked sixth in Adjusted Line Yards on runs left end, but just 31st on runs right end. The Steelers defense was the worst in the league on runs left end, but the best in the league on runs right end. Oddly, the Jaguars had all that success running against the Steelers in the first game despite having just one running back carry listed as left end, a five-yard gain by Fred Taylor on a second-and-10 in the first quarter.
The surprise this season is not the quality of the Jacksonville running game, but rather the quality of the passing game. Before this season, David Garrard had 539 career pass attempts with 18 touchdowns and 11 interceptions. This year, Garrard has 325 pass attempts with 18 touchdowns and just three interceptions.
Both Pittburgh and Jacksonville are aggressive through the air when they aren't running the ball. Pittsburgh (23.1 percent) and Jacksonville (22.5 percent) were first and fourth in the percentage of deep passes (longer than 15 yards through the air, no matter if complete or incomplete). Only Peyton Manning, Tony Romo, and Derek Anderson gained 20 or more yards on a higher percentage of their passes than Garrard did. Garrard also has a great weapon when he needs to dump the ball off or run a screen, but that weapon may not be as useful against Pittsburgh. Maurice Jones-Drew was second among all running backs in receiving value, but Pittsburgh is the league's best defense at preventing success by running backs in the passing game.
With all this talk about Garrard, people may forget that Ben Roethlisberger was pretty good this year too: sixth in DVOA, ninth in DPAR. And he has the best receiver on either team: Santonio Holmes, who had a huge breakout sophomore year, ranking sixth among all wide receivers in DVOA and 11th in DPAR.
It will be interesting to see whether the Steelers line Holmes or experienced veteran (and technical route-running master) Hines Ward opposite Jacksonville cornerback Rashean Mathis, because if the early game charting data is to be believed, the 2006 All-Pro has completely regressed this season. Last year, Mathis had a 57 percent Success Rate and allowed 6.3 yards per pass. In our data so far in 2007, Mathis has a 47 percent Success Rate and allows 7.3 yards per pass. If you want standard stats instead, Mathis has dropped from 21 passes defensed and eight interceptions to six passes defensed and one interception. (I don't have data in yet from Week 15 to know who Mathis was usually covering in that game.)
Perhaps because they are good throwing deep, both of these offenses excel on third-and-long. The Jaguars have the best DVOA in the league, while the Steelers rank third. Both defenses struggle in the same situation: Pittsburgh is 23rd in defensive DVOA on third-and-long, while Jacksonville is 24th.
Jacksonville's offense likes to go three-wide, but Pittsburgh ranks second in the league in DVOA against "other receivers" besides the top two. Brian McFadden's charting numbers fall right in line with the DVOA stats: data collected so far gives him a 75 percent Success Rate with an awesome 3.9 yards per pass. McFadden gave up a first down on just four of the 32 charted passes for which I have data.
There is one place where the Steelers are far more conservative than the Jaguars: fourth down. According to the Football Outsiders Aggressiveness Index, Jack Del Rio was one of the league's most aggressive coaches going for it on fourth down, while Mike Tomlin was the league's most conservative coach on fourth down. Jacksonville ran 31 regular plays on fourth down. No other team ran more than 24, and the Steelers ran only 12. (Of course, Tomlin picked the worst possible time to get aggressive, with that ridiculous "wide receiver sweep" on fourth-and-goal against the Patriots.)
According to the Football Outsiders game charting project, the Steelers rushed five or more defenders more often than any other team, but they were only average when it came to rushing six or more defenders. San Diego is the only defense that rushes exactly five more than Pittsburgh does. According to the early game charting data, Jacksonville gains more yards per pass against five rushers (8.0) than against four (6.6) or six-plus (3.8).
(And if it isn't working, Dick LeBeau isn't going to change things. This is anecdotal, not analytical, but last night I charted the second half of the Patriots-Steelers game, and I thought I was living through "Groundhog Day." Every single play, the Steelers brought five. Play after play, Tom Brady simply threw to wherever the hole was left by the fifth pass-rusher. The identities of the blitzers changed each play, but the number never did. Every play was five-wide, five rushers, five blockers, Patriots completion. The Jaguars offense doesn't work like clockwork in the same way, but it works pretty well. Come on, Dick, change it up a little, for crying out loud... on the other hand, to say something positive about Pittsburgh, watching the Steelers block for a left end run with both wide receivers and the tight end on the left side is like watching ballet. Everything is perfect.)
What about special teams? Well, Jeff Reed had a nice year on field goals, and their punt returns were once again awful. On the Jacksonville side, it is important to note the difference once Josh Scobee returned from his injury at midseason. The change on field goals was fairly straightforward: Scobee was worth -0.1 points on field goals, while Carney was worth -3.7 points. The change on kickoffs was a little more complicated. Based solely on kickoff distance, assuming average kickoff coverage, Carney was worth -1.0 points while Scobee was worth 2.2 points. However, looking at net kickoffs, Carney was worth 10.1 points while Scobee was worth -5.3 points. It's like the Jaguars' special-teamers forgot how to tackle when Scobee got healthy. Carney had five touchbacks in nine games, two of which were in Denver, but the Jaguars allowed just three returns over 25 yards and none over 32. Scobee had 12 touchbacks in seven games, including two in the snowy conditions in Pittsburgh, but the Jaguars allowed 11 returns of 25 or more yards, two of which were touchdowns by Andre' Davis last week. What the hell?
There's no question that Jacksonville is the wild card team most likely to upset a division champion and move on to the second round of the playoffs. But many fans are talking about the first game between these teams as if it was a Jacksonville blowout. In fact, the Jaguars only won by a touchdown, and they let the Steelers come back from a 15-point deficit before taking their final lead in the last two minutes. This game has a good chance of being just as close and exciting, and a Jacksonville win is not fait accompli.
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.) Red zone DVOA is also listed. These numbers are all regular season only. WEI DVOA is WEIGHTED 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).
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.
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."
68 comments, Last at 06 Jan 2008, 6:02pm by Boss Hog