This week's DVOA commentary is all about worsts. Come find out where Washington stands among the worst special teams in DVOA history, whether San Diego has the biggest gap between offense and defense, and whether Baltimore or Jacksonville has the worst running game we've ever tracked.
02 Feb 2006
by Bill Moore
One of our greatest challenges is evaluating a defensive player, and understanding why a defense works and why it doesn't. Other than tackles, interceptions, and sacks, there are very few public measures of the quality of individual defensive players. So, if the mountain won't come to Mohammad*, Mohammad will come to the mountain. Here at Football Outsiders, we began a project this season that charts each game to develop our own statistics. The data is unofficial, and unfortunately not every game is yet completely charted, but we have enough information to draw certain conclusions.
Ben Roethlisberger rolled through 2004 as the third most efficient quarterback in the NFL. However, given his limited yardage per game and the dominance of Pittsburgh's running game, most viewed him as a game manager and not as part of the game's elite. Frankly, who can be blamed? The Steelers were ranked 32nd in attempts, and they played poorly in the playoffs when opposing teams made them win with the rookie's arm. However, hidden away was the little gem that their rookie quarterback was among the league leaders with 8.9 yards per attempt.
In 2005, once again, Pittsburgh ranked 32nd in pass attempts. Here we went again. "Big Ben is good, but once you force him to pass? Game over." But when the Steelers stepped onto the field in Indianapolis and Denver, Coach Bill Cowher and Offensive Coordinator Ken Whisenhunt not only were prepared to pass in they had to, they game-planned for it. Pittsburgh came out firing and shocked the Colts. Then the Steelers ripped it up against Denver in the AFC Championship game. Roethlisberger's stat line for that game: 21-of-27, 275 yards, 9.5 yards per attempt, and two touchdowns; he hit seven different receivers.
Stopping Roethlisberger is the task at hand for the unheralded Seattle Seahawks secondary, which is unheralded for a reason. The secondary is the weakest link of this Super Bowl-bound team. Seattle ranked 28th this year in receiving yards allowed. However, that doesn't tell the whole story: the Seahawks were fourth in number of passes thrown against them. In yards per pass attempt, they ranked middle of the pack. They were also 13th in opponent passer rating, although the list of quarterbacks who played against Seattle is more like a future episode of The Surreal Life than a current Pro Bowl roster.
Football Outsiders' main statistical metric, DVOA, ranks Seattle's defense 18th, in part because of the adjustments for strength of schedule. (DVOA is explained here.) However, stopping the run was the Seahawks' real strength. Seattle was ranked just 25th against the pass. When broken down on a game-by-game basis, the pass defense was highly erratic, with a mix of superb performances and absolute bombs. The unit excelled in two games against Kurt Warner and Arizona, a hobbled Philadelphia team, and Drew Bledsoe's Cowboys. Yet the Seahawks were torched by the likes of David Carr and Ken Dorsey. In the playoffs, the secondary had an average game against Mark Brunell, then ripped up Carolina by having the whole team cover Steve Smith -- simultaneously.
In the off-season, Seattle lost its starting right cornerback, Ken Lucas, to free agency, and the Seahawks hit the free agent market themselves, picking up left cornerback Andre Dyson at Tennessee's salary cap yard sale. They also moved third-year cornerback Marcus Trufant from the left side to the right.
Although Dyson was brought in to be a steady presence in the secondary, injuries made the left cornerback position a revolving door. Dyson started six regular season games, Kelly Herndon (who started opposite Champ Bailey in Denver last year) started six until he too went out with a strained MCL, and the inexperienced Jordan Babineaux started the final four games.
Herndon was picked on most by opposing quarterbacks. Herndon made a play (i.e. tackle, pass defensed, or interception) on 11 percent of passing plays, which ranked him third among the Seahawks. We dug further, using game tape in an attempt to identify who was responsible for coverage on each pass. Our game charters list Herndon as the responsible defender covering 17 percent of the passes against Seattle. That makes him the most common target on the Seattle defense, despite the fact that he played in only 12 games and was a starter in just six. In weeks when Herndon started, a quarter of all passes were directed his way. Receivers covered by Herndon caught 66 percent of these passes. Counting dropped passes as catches for the purposes of evaluating the defense, the completion rate rises to 71 percent. 38 percent of passes thrown against Herndon went for first downs or touchdowns, well above the league average of 30 percent. On third down he gave up first downs a whopping 55 percent of the time.
Herndon went down in Week 12 against the Giants, and Babineaux was no better. In fact, he was worse. Again, counting drops as complete passes, receivers covered by Babineaux had a ridiculously high 82 percent catch rate. Players that were listed as being uncovered only caught 75 percent of the passes thrown to them! A majority of throws against him were on shorter routes, potentially exploiting soft coverage: 61 percent were thrown under 10 yards, and 33 percent were five yards or less. Babineaux did make three interceptions on the year, but all were while playing in a backup role.
Not unexpectedly, Andre Dyson is clearly the left cornerback Seattle wants on the field. However, his biggest problem is staying on the field. He has been hurt not less than three times this season, missing six complete games and playing a reserve role for four. Nevertheless, receivers covered by Dyson have only a 50 percent completion percentage, and he has the highest percentage of passes defended (passes deemed incomplete due primarily to the actions of the defender). He knocked down Michael Vick's attempted pass to Brian Finneran in Week 2, ending Atlanta's comeback bid. He is not without flaws: like Herndon he allowed 38 percent of the passes against him to result in first downs or touchdowns.
If there was a revolving door on the left, there was a rock on the right. Following a second off-season with shoulder surgery, and after making the move from left to right, Seattle's former first round pick, Marcus Trufant started every game at right cornerback.
Deemed by coach Mike Holmgren early in 2004 to be part of a "shutdown corner" tandem, Trufant failed to live up to that billing as the year progressed. Trufant was providing receivers too much cushion, and receivers exploited holes and gaps.
The new season for Trufant didn't start much better. In fact, after the first game of the 2005 season, Holmgren publicly called out Trufant for not playing up to the talent level that caused Seattle to use the 11th overall selection on him three years ago. Six Jacksonville passes aimed at Trufant went for three first downs and a touchdown.
As the season went along, however, Trufant played respectably. The quarterbacks that the Seahawks faced may not have been Prime Time, but some of the receivers were. Trufant, playing a particular side of the field rather than a particular player, faced up against Larry Fitzgerald, Plaxico Burress, Anquan Boldin, Keyshawn Johnson, Torry Holt, and Santana Moss. Consequently, Trufant was tested deep. A third of the passes against Trufant went 16 yards or longer, yet receivers caught more than half of the passes thrown to them. 40 percent of the passes were converted into first downs or touchdowns. Although he is a good tackler and has good speed, when he makes mistakes it's usually in giving receivers too much space. In particular, Trufant has admitted having problems with 10- and 20-yard dig routes (routes where receiver quickly break to the inside) because he is protecting against the fly routes. That would explain the high completion percentage given the distances of the passes. Trufant may be playing too conservatively at times to avoid giving up a touchdown.
At the beginning of the season, Seattle's starting safeties were Michael Boulware and Ken Hamlin. That lasted through Week 6, when Hamlin suffered a fractured skull in an off-field incident after Seattle's victory over Houston.
Hamlin kept receivers in check with a low 42 percent completion rate. Hand in hand with a low completion rate, the yards per attempt were high, 18.5 yards. However, not all of Hamlin's incompletes were just a result of the distance of the throw. Hamlin was second on the team in percent of passes defended.
Losing Hamlin was a tough break for Seattle, not necessarily because of the quality of Hamlin, but rather the weaknesses of his replacement, Marquand Manuel. Manuel, who was a sixth round pick by the Bengals in 2002, doesn't have the same kind of cover skills. When Manuel covered them, receivers caught 73 percent of passes thrown to them, with an average per attempt of 14.6 yards. 67 percent of passes thrown his way resulted in first downs or touchdowns, including four of four on third down -- much higher than Hamlin's 16 percent.
College linebacker-turned-strong safety Michael Boulware started every game for the Seahawks this season. Considering it was his first full year at the position, it must be regarded as a success. Boulware can blitz (two sacks), cover (four interceptions) and play special teams (one field goal block). The tight ends, wide receivers, and occasional running back he covered had a modest 60 percent completion rate -- slightly below the all passing average. Boulware's third down conversation rate was the best in the secondary, but overall he allowed a first down or touchdown a team-high 43 percent of the time -- well above league average.
Here's how DVOA grades Seattle against different types of receivers:
|vs. #1 WR||vs. #2 WR||vs. Other WR||vs. TE||vs. RB|
The one bright spot is coverage against running backs coming out of the backfield. However, Roethlisberger doesn't throw very often to his backs. Willie Parker is the only real receiving threat out of the backfield, and less than six percent of Big Ben's passes were thrown his way (Jerome Bettis was the intended receiver on three tosses).
He's not quite Steve Smith, but Hines Ward is clearly the most important cog in the Pittsburgh offense. The largest proportion of Roethlisberger's passes go to the right, and many of those go to Pittsburgh's number one receiver. Consequently, we should see a lot of matchups of Ward against Andre Dyson (note that throws right go towards the left cornerback). It will be a tough assignment for Dyson; it will be even tougher if Herndon has to come in for him.
The other side gives us a likely matchup between Antwaan Randle El and Marcus Trufant. Randle El is a less disciplined route runner, and the weakest of Pittsburgh's three primary receivers. Trufant will also probably have his crack at Cedrick Wilson, who will see his fair share of field time as Randle El moves into the slot.
A final matchup to watch will be Steelers tight end Heath Miller and Michael Boulware. Miller has been an excellent new outlet for Roethlisberger, ranking eighth in DPAR and fifth in DVOA. Roethlisberger has thrown almost 10 percent of his passes Miller's way, and Boulware will have his hands full covering him.
A month ago, I would have said Seattle has the advantage when Roethlisberger passes. Seattle's secondary, if not all-star caliber, is finally healthy. They have speed and quickness to stay with the Pittsburgh receivers. However, Roethlisberger is better and the offensive live line is giving him time. If I were in the Seattle secondary, I'd be worried.
"We'll take the ball, and we're gonna score." In the role of prodigal son returning to the tundra of Green Bay, the former backup quarterback to Brett Favre issued those words exactly one play before throwing a game ending interception. Unfortunately for Matt Hasselbeck, those few words are the only thing for which he is remembered by some football fans.
News flash for those subjected to the infamous East Coast bias: Matt Hasselbeck is pretty good. Four quarterbacks have made Football Outsiders' top 15 DPAR (Defense-adjusted Points Above Replacement) in each of the last four years: Peyton Manning, Tom Brady, Trent Green and Matt Hasselbeck. Shaun Alexander may have captured the MVP award, but Hasselbeck is most likely the MVP of this team. However, he won't be going up against the likes of San Francisco or Houston in the Super Bowl. Pittsburgh is the fifth-ranked passing defense in the league, and they get the job of stopping Hasselbeck and turning Seattle into a one-dimensional team.
The Pittsburgh Steelers' passing defense languished in mediocrity for much of the early part of the decade, culminating in their 18th-ranked DVOA in 2003. However, underneath that uncommon performance was the bubbling of a youth movement. Earlier in 2003, the Steelers had made an atypical move for them, trading up to draft USC safety Troy Polamalu. They also curiously used their fourth round pick in that draft on Louisiana-Lafayette walk-on cornerback Ike Taylor. The draft didn't pay immediate dividends, as it took a while for the rookies to make a mark, but the development had begun. The Steelers used their next two drafts to continue building youth in the secondary, drafting Ricardo Colclough in 2004 and Bryant McFadden in 2005. As a result, Pittsburgh may have compiled one of the best young secondaries in the league.
The Steelers defensive backs are a hard hitting, energetic bunch, and defensive coordinator Dick LeBeau takes full advantage of that skill. Pittsburgh stacks the line with linebackers and safeties to disguise who is blitzing. However, such aggressiveness creates some apparent weaknesses, both real and imagined. For one, the Steelers don't make a lot of interceptions (just ask Pete Morelli). Despite being ranked 8th in the NFL in passing defense, the Steelers ranked just 18th in interceptions.
One would also expect a blitzing defense to be susceptible to the long ball. In fact, offensive coordinators did test the Steelers deep. Opponents threw almost eight percent of their passes against the Steelers longer than 30 yards (versus four percent against Seattle and six percent across the league). However, their completion rate on these long passes was less than 25 percent -- modestly below average. Additionally, the yards after catch on those receptions suggest that the Steelers were rarely burned. Only Marvin Harrison's touchdown catch on the first offensive play of the Steelers' regular season game against Indianapolis resulted in any significant yardage after the catch. Harrison's catch notwithstanding, the Steelers defense, as a whole, kept teams out of the end zone. They are tied for second in touchdowns given up with 18, three behind the Bears.
The aggressive defense that Pittsburgh employs requires a certain cohesiveness among the secondary. With the exception of an early battle for the starting left cornerback job, the secondary personnel has remained consistent. The following evaluates each position.
Ike Taylor is an interesting study that probably warrants its own article. Taylor attended Louisiana-Lafayette, where he was a junior walk-on as a running back and only moved to the defensive backfield for his senior year. At 6'1'' and 192 pounds, his athleticism alone caught the eye of the Steelers, who drafted him in the fourth round of the 2003 draft. Hardly surprising, his first two seasons in the league (only his second and third years ever playing the position) were a period of learning and adjustment. He was often criticized for lack of toughness and precision.
The criticism would mostly pass in 2005 as he became the full-time left corner. Taylor often matched up against the opponent's best receiver. The young defensive back drew praise from some opposing wide receivers, including Chad Johnson -- who is not known for praising opposing cornerbacks.
Whether quarterbacks were testing Taylor or just trying to get the ball to their favorite receivers is unknown, but receivers he covered accounted for 22 percent of all the passes we charted -- double that of right cornerback Deshea Townsend. Although he is praised as one on the best young cornerbacks in the game, Taylor is not exactly a shutdown corner. Receivers caught 63 percent of the balls thrown against him, and 42 percent went for a first down or a touchdown. Taylor can be beaten.
Taylor has played his best when it has mattered the most. Taylor played very respectively against Chad Johnson and Marvin Harrison, although each burned him for a single big play. He played Mushin Muhammad quite well in the first half of their meeting, only to let up when the Steelers were ahead by eighteen points. However, in that same game Taylor was beaten by Bernard Berrian for 43 yards. He virtually shut down the Cleveland and Jacksonville receivers that he faced, but was awful against New England. Patriots receivers caught 77 percent of the throws against him, culminating in 120 yards on 13 receptions, six of which were for first downs.
Taylor is unlikely to pull down an interception on Sunday. He is not exactly known for his hands. Although he only ended the season with one interception, he probably had a chance at more than ten if he could just catch the ball. He had numerous inexplicable drops -- "Koren Robinson playing defense" kind of drops. It may look familiar to the Seahawks.
Working the left side of the offense in his second year as a starter, Deshea Townsend is the greybeard of the group at the ripe old age of 30 years -- drafted one round after Hines Ward in 1998. Although he historically has struggled in supporting the run, Townsend is a solid cover man and may actually play the slot better than the corner. Once called the "slowest defensive back in all of football" by Football Outsiders' Michael David Smith, Townsend gave up very few long balls. In our new project, Football Outsiders has charted 75 percent of the passing plays against Pittsburgh. Not a single play where Townsend was listed as the defender went for a touchdown, despite the fact that close to 16 percent of passes against him were in the red zone.
Although he had his fair share of star-receiver matchups, Townsend was more often than not relegated to covering a team's number two and three receivers. Overall, slightly more than 11 percent of all passes against the Steelers were thrown his way. His coverage led to a 51 percent reception rate by those he was defending, and only 21 percent of passes thrown against him were for a first down. His conversion rate solely on third down was not as good, but respectable.
Also likely to see a bit of action at the right corner spot is rookie Bryant McFadden, who will play as the team's main nickel back as well. The Steelers went to the defensive back trough for a third year in a row when they picked up McFadden in 2005 draft. In three wide receiver sets, a formation we expect to see quite often from Seattle on Sunday, McFadden plays the outside receiver and Townsend moves into the slot.
During the regular season, McFadden's role was often limited to nickel and dime packages. That didn't stop him from getting playing time, and when he was on the field he may have been the best of the cornerbacks out there. Although his completion rate was even with the other backs, his percentage of balls defended (incompletes as a direct result of the action of the defensive back) was the highest on the team at 15 percent. Based on recent performance, including his play in the recent playoff game, many expect he will be the right cornerback of the future, swapping roles with Townsend.
Who doesn't know strong safety Troy Polamalu? He and his hair have graced the cover of many a magazine. He covers, blitzes, hits, plays the run, and has a serious nose for the ball. The third year former linebacker from USC seems to be everywhere on the field. Although almost everyone knows him, the person they should really know is free safety Chris Hope.
Hope is basically Polamalu's safety blanket, and his primary duty is to play centerfield and make a lot of the secondary calls. Slightly more than seven percent of Pittsburgh pass plays log Hope as the defender. Defending in the style of Brian Dawkins or Rodney Harrison, Hope is more of a hitter than a tackler. His style is not one of wrapping up but rather laying out the receiver (perhaps forcing him to "tap out"). Potentially as a result of this, Hope has one of the lowest average yards after catch on the team.
The completion rate of Hope's opponents was a low 48 percent, but more importantly, Hope has the lowest percentage of conversions for a first down or touchdown (29 percent). That is even more exaggerated on third down, when his conversion rate is only 17 percent. Despite being probably the most unheralded member of the secondary, Hope may be the best pure pass defender.
Meanwhile, Troy Polamalu is everywhere, even through his coverage skills are not the best on the field given his gambling style. His completion rate was 52 percent, and conversions to first downs were reasonable at 38 percent. However, on third down, 50 percent of passes against him were converted. Nevertheless, he is an all around package, and is integral to Pittsburgh's blitz schemes.
Here's how DVOA grades Pittsburgh against different types of receivers:
|vs. #1 WR||vs. #2 WR||vs. Other WR||vs. TE||vs. RB|
Seattle can test defenses early with its multifaceted attack. Hasselbeck can throw, and obviously league MVP Shaun Alexander can run. The Seahawks quite often run three- and four-wide receiver sets in order to spread the defense, and that's a good matchup for them because Pittsburgh ranks high against number one receivers and drops off against other wide receivers. The Steelers are also among the worst defenses against passes to running backs, as those pass-rushing outside linebackers leave big open holes in the flat. Fortunately for the Steelers and unfortunately for the Seahawks, Alexander is not much threat to catch the ball out of the backfield. Whether Seattle can exploit Pittsburgh's weakness with Maurice Morris or Mack Strong could be a key to the game.
Seattle's three main receivers are Darrell Jackson, Bobby Engram, and Joe Jurevicius -- all quality receivers. D.J. Hackett is also a respectable fourth receiver. Engram, who is a good possession receiver, will likely be covered by Taylor, and that should match up well for Taylor. It should be noted that Seattle's receivers move around. Engram is almost as likely to catch a pass on the right side of the field as he is the left. He could easily line up in the slot, and in such cases would likely match up against Townsend. That matchup is not as favorable to the Steelers. Unless the Steelers choose to follow coverage, Seattle may be able to create the matchups to its advantage.
The responsibility for Darrell Jackson will likely fall initially on Townsend, who is notably slower. Jackson is quick and strong. That may not help him in matchups against Taylor, but it should against Townsend. In three receiver sets, Jurevicius could pose a problem for McFadden, who is smaller and lighter; Jurevicius has six inches and 40 pounds on him. McFadden's ability to bat balls may not be as useful. In an attempt to neutralize Jurevicius, the Steelers may use Polamalu more in coverage and less in Seattle's backfield.
Although Pittsburgh's young defensive backs are tough, they do occasionally give up a big play. The Seahawks' real advantage, however, will be if they can spread the field and utilize multiple wide receiver sets to find the right matchups. That strategy will require Seattle's offensive line to pick up tricky blitzes and give Hasselbeck time to make progressions. It's a close one, but I say...
Advantage: Seahawks by a nose.
*Note: proverb also applies to Muhsin Muhammad
75 comments, Last at 05 Feb 2006, 4:30pm by DGL