You guys know Football Outsiders, at least most of you do. They produce all those funky, cool new stats that break down players into ways we didn't even think about a decade ago. Each year they produce a book called Football Outsiders Almanac 2009, which has chapters on all the NFL teams and even college stuff (available in PDF format). They sent me a copy and agreed to answer some questions about the Cowboys. If you're interested, you can get the book here.
Blogging The Boys: Last year, a lot of the Cowboys problems were blamed on injuries. According to your stats, we were in the middle of the pack in terms of team-health last year, so it must be a case of guys with no ready backup getting injured?
The rest below...
Football Outsiders: Well, not all injuries are created equal. Losing Terry Glenn for a season, for example, had far less of an impact than losing Tony Romo for four games, because quarterback injuries affect a team far more than wide receiver injuries, and the team had a far better replacement for Glenn (Patrick Crayton) than it did for Romo.
Because of the way the Cowboys are built, there are certain players that they simply just cannot lose and play the same way without. Romo is one of them. Flozell Adams, arguably, would be another. Jay Ratliff (who himself only got a chance to play after Jason Ferguson went down) and DeMarcus Ware are unquestionably two more. The team can handle, say, an injury to Marion Barber because they have the depth at running back and a good run-blocking offensive line up front.
BTB: The difference between the defensive performances in 2008 when Terence Newman was injured or out of games, compared to when he was healthy in the last part of the season, is remarkable. Is Newman one of the key components the Cowboys must have healthy in the secondary? More so than other players?
FO: It absolutely was; the Cowboys' pass defense was 24th in the league in DVOA while Newman was alternately hobbled and sidelined by a groin injury from Weeks 1-9. (For the uninitiated, DVOA is our core metric that measures performance against the league average for each play after accounting for the down, distance, situation, and the quality of the opposition; the Cowboys' pass defense DVOA over those weeks was 12.7%, meaning that they were 12.7% worse than a league-average team against pass plays during that timeframe.)
Once Newman came back, though, their pass defense immediately became the second-best in football, with a -12.6% DVOA. (Since you want to allow fewer yards or points than the league average, a negative DVOA is a good thing for the defense.) That was also borne out in Newman's individual metrics; he was astoundingly bad before the surgery, particularly in the first Redskins game, but was an elite corner after returning.
I'd probably add Newman to that list of "irreplaceable" guys above. There's talent in the backfield with Orlando Scandrick and Mike Jenkins (especially Scandrick, who was really fun to watch develop last year), but I think that even the most ardent of Cowboys fans would agree that neither of them are ready to be a #1 corner. Losing Newman -- or failing to have him at 100 percent -- would definitely have a dramatic impact on the team.
BTB: I've brought this up before, but without actual statistical evidence - Tony Romo is a slow-starter in games, but finishes strong. What did your stats say about this?
FO: This is exactly the sort of thing we can test for. Here's Tony Romo's quarter-by-quarter DVOA a year ago:
(He also had a -394.7% DVOA for the three times he dropped back in overtime.)
Most quarterbacks actually get worse as the game goes along; the league average DVOA for all other quarterbacks was 8.0% in the first quarter, 10.2% in the second quarter, 3.7% in the third quarter, and -1.7% in the fourth.
The difference was even more remarkable a year ago; Romo's first quarter DVOA in 2007 was -31.5%! His fourth quarter DVOA was only 14.3%, so he didn't show the huge late-game split that we saw in 2008, but his second quarter DVOA was 35.7%, and his third quarter DVOA was a very impressive 77.6%. He also had a -21.6% DVOA in the first quarter of games in 2006.
So, to answer your question, the data supports the idea that Romo's a slow starter. As to why that is? Hard to say. My educated guess is that Romo's a player who thrives when he's improvising; I think he's the best at the league at scrambling and making throws on broken plays or off of broken routes. As the game goes on, quarterbacks naturally get a feel for the sort of coverages they're seeing and the way that defenders are playing their receivers. The more snaps Romo gets in a game, the better he knows an opposing defense.
BTB: One big concern for the Cowboys has been the play of the offensive line which we thought was inconsistent and erratic last year. What do your stats say about the Dallas offensive line?
FO: I think that Cowboys fans might have been spoiled by a healthy, effective line in 2006 and 2007. Dallas had arguably the best offensive line in football in '07, and their five starting linemen managed to make it to 158 of 160 games; only Andre Gurode missed two games, and even then, those were the last two games of the 2007 season and had no bearing on the team's playoff status.
Injuries are a part of the game, though, and it's extremely hard for an offensive line to suit up year-after-year without seeing guys get healthy. The Giants have been lucky enough to see their five guys make it onto the field for every regular season snap in each of the last two seasons, and that level of continuity and reliability goes a long way in building their offense -- if you remember how poorly backup tackle Kevin Boothe played against the Cowboys last year when Kareem McKenzie went down, you know how just losing one player could have a huge impact on that team.
Obviously, last year, that changed. Losing Kyle Kosier for virtually the entire season had a huge impact on the team, even though he was arguably their least important lineman. Montrae Holland was no replacement, and Cory Proctor was disappointing.
By our metrics, the line wasn't that much worse. We use a stat called Adjusted Line Yards to try and separate the yardage gained by a running back on a given play from that of the yards created by the team's offensive linemen. By that metric, the Cowboys were 14th in both 2007 and 2008. Their Adjusted Sack Rate -- which considers a team's sacks allowed after adjusting for passes thrown and the quality of the defenses they faced -- went from seventh in the league to 13th, but some of that has to do with Brad Johnson and the eight sacks he took on only 78 attempts.
I don't know about the inconsistency, but the best way to ensure consistency is to have your five best guys lining up for every game. If the Cowboys can make that happen, I think the consistency issues will solve themselves.
BTB: Tell us what you can about the Cowboys special teams units and how they might improve in 2009?
FO: Well, one place they probably won't improve is on field goals and extra points. Nick Folk had a great year, adding 7.9 points above what an average kicker would have provided, the second-best total in the league. Unfortunately, we tend to find that performance on field goals from year-to-year is highly random; last year's number two kicker was Jeff Reed, who was league-average this year.
What is consistent, though, is a team's performance on kickoffs, and Dallas has been below-average each of the last two years. That's mostly on Folk; his kickoffs have been far below league average, while the kick coverage was slightly worse than league average. We estimate that the Cowboys lost 7.46 points of field position on kickoffs and kickoff returns against them.
Of course, they weren't good on punts or punt returns, either. Our estimates have Mat McBriar as -6.16 points below league average, with most of that coming due to below-average distances (relative to where he punted and the league averages) and the failed punt against the Cardinals that ended his season, while Sam Paulescu only cost the team 1.26 points, with most of that coming on return length. The less said about the team's returns, the better; no one came out looking good in those units.
Bringing in David Buehler should help; Buehler's no slouch on return units as a former linebacker, and he's got a stronger leg than Folk. He should be Folk's long-term replacement at kicker.
The team will also adjust from using a four-man wedge now that the league's banned wedges of more than two players. That should inspire them to re-think their special teams scheme, and replacing veterans like Keith Davis with draft picks like Victor Butler and Jason Williams should create opportunities for hungry players to earn spots on the roster. That should help their special teams improve some in 2009.
Thanks to Football Outsiders for taking the time to answer the questions. You can check the book out here.